Challenges - Sustainability

1. Challenges - Sustainability
Nature of Challenge
Product Related
Domain / Discipline of Challenge
Sustainability
Description (With Technical Specifications)
Background
Electronic waste, or e-waste, refers to discarded electronic devices such as
computers, smartphones, TVs, etc. These devices contain valuable materials such as
metals, plastics, and glass, but also harmful substances such as lead, mercury, and
cadmium. When e-waste is not properly recycled or disposed of, it can pose serious
environmental and health risks. For example, e-waste can contaminate soil and
water with toxic chemicals, or release greenhouse gases when burned. E-waste is
being generated at a rapid pace due to the increasing demand and short lifespan of
electronic products. According to the Global E-waste Monitor 2020, the world
generated 53.6 million metric tons of e-waste in 2019, and this amount is expected
to grow by 38% by 2030. This means that we will soon have e-waste piling up across
the globe, creating a huge challenge for waste management and resource recovery.
Challenge / Requirement
Reduce the environmental and social impact of electronic waste, or e-waste, which is
being generated at a rapid pace due to the increasing demand and short lifespan of
electronic products
Deliverables
1. A design proposal that describes your chosen solution in detail, including its
objectives, features, benefits, limitations, costs, risks, and implementation plan.
2. A prototype or simulation that demonstrates how your solution works in practice,
using appropriate tools and methods.
3. A presentation or report that showcases your project outcomes and explains how
your solution addresses the problem statement and meets the evaluation criteria.'
2.Recycling of Thermoset Materials
Nature of Challenge
Process Related
Domain / Discipline of Challenge
Sustainability
Description (With Technical Specifications)
Background
The applications and uses for thermosetting plastics have grown substantially over
the years. The success and benefits of this type of plastic is due in large part to its
favorable plastic properties.
Thermoset materials are a type of polymer that undergo a chemical reaction when
heated, cured or mixed with other substances. This reaction creates strong cross-
links between the polymer chains, making them rigid and durable. This also means
that thermoset materials cannot be melted or reshaped once they are formed and
hence difficult to recycled by conventional methods that involve heating and
remolding. Thermoset materials are often discarded as landfill after their use, which
poses environmental and economic challenges.
Challenge / Requirement
Design a sustainable and cost-effective solution for recycling thermoset materials.
Deliverables
- A technique and a prototype of the solution, considering the technical, economic,
environmental and social aspects; data pertaining to properties of recycled material
- Evaluate the performance and the feasibility of the solution, using appropriate
metrics and criteria
- Prepare a report and a presentation to showcase the solution and its benefits
Known Bottlenecks / critical areas / Key factors to be addressed
Recycling technology which can help to recycle any size and quantity of Thermoset
material without causing much effect to the environment.
3. Product and its Carbon foot print
Nature of Challenge
Product Related
Domain / Discipline of Challenge
Sustainability
Description (With Technical Specifications)
Background
A carbon footprint is a measure of how much greenhouse gases, like carbon dioxide,
are released into the atmosphere by making and using a product. There is a growing
need to know the carbon footprint of that a product and also including the method
of disposal for recycling & reuse.
To reduce the carbon footprint of a product, we need to consider not only how it is
made, but also how it is disposed of. Keeping track of the carbon footprint of
products throughout their supply chain, will involve capturing during manufacturing,
transportation, and selling them. Collecting and reporting this information can be
costly and time-consuming .
Challenge / Requirement
Develop and end to end solution that can measure and communicate the carbon
footprint of a product in a way that is accurate, efficient, and transparent for both
the companies and the customers, while taking into account the whole life cycle of
the product from production to disposal. The targeted user will be the manufacturer
or the intermediate agents in the supply chain or the end user
Deliverables
Framework/Solution that captures the carbon footprint for a given product that has
the required interfaces for various agents in the supply chain to interaact such as
partner, manufacturer, end user.
Known Bottlenecks / critical areas / Key factors to be addressed
Data privacy, Security, Ease of entry of data, Ease of access of data, ease of flow of
data, ease of reporting, Connectivity (Internal & external use), Legality, Design
Considerations, Content and user satisfaction etc..
4.Sustainable And High-Performing Wind Turbine Blade
Nature of Challenge
Product Related
Domain / Discipline of Challenge
Sustainability
Description (With Technical Specifications)
Background
Wind turbine blades are made of different materials that have different functions
and properties. The main material is reinforcement fibres such as glass fibres/fabric,
that make the blades strong and stiff. The reinforcement fibres are embedded in a
matrix material, that holds the fibres together and gives the blades shape. The
matrix material such as thermosets like epoxies, polyesters, vinyl esters,
polyurethane also protects the fibres from damage and corrosion. The blades also
have a sandwich core e.g. balsa wood, foams such as polyvinyl chloride, polyethylene
terephthalate, which makes the blades lighter and more stable. The blades also have
some other materials, such as metals and coatings, that are used for connecting the
blades to the turbine or for improving their performance and appearance. Currently,
nearly all utility-scale wind turbine blades are manufactured using traditional
thermoset resin systems such as epoxy, polyester, or vinyl ester. Decommissioned
thermoset blades cannot easily or economically be recycled at the end of life, with
the vast majority of them ending up in landfills. The wind turbine industry is aiming
at zero waste, recyclable turbine structures and is looking for ways to further reduce
its greenhouse gas footprint
Challenge / Requirement
Develop a sustainable solution for wind turbine blade with materials that are easily
recyclable and /or generate zero waste.
Deliverables
Material composition, manufacturing process, LCA analysis, DFM, CAE simulation,
Mold flow, Performance, Durability, Reliability data
Known Bottlenecks / critical areas / Key factors to be addressed
Performance & Cost same as as existing product, ease of processing, ease of
Recycling & reusability
5.Integration of CAD with LCA Tool
Nature of Challenge
Product Related
Domain / Discipline of Challenge
Sustainability
Description (With Technical Specifications)
Background
Life Cycle Assessment (LCA), which is a technique used to evaluate the
environmental impact of a product throughout its entire life cycle. This includes the
extraction of raw materials, manufacturing, transportation, use, and disposal. It helps
determine the sustainability level of a product.
Conducting an LCA manually can be time-consuming and labor-intensive for a
complex Bill of Materials (BOM) with 200+ items.
Challenge / Requirement
Develop an efficient and automated integration method to seamlessly connect to Life
Cycle Assessment (LCA) tool with commonly used CAD software or platforms.
The solution should enable seamless data exchange between CAD and LCA tools,
streamlining the sustainability validation process and promoting more
environmentally conscious product design and development practices.
Measuring of the LCA should be as simple as possible (like measuring the product's
weight in a single click).
Easy to visualize and measure LCA of the product (LCA - mainly Global Warming
Potential in terms of like CO2 eq. - due to the product's material and processing),
1. API – can be used to connect CAD or PLM with LCA softwares
a. OpenLCA and SimaPro can be connected as per our understanding
b. Other LCA software are also trying to connect with CAD or PLM for faster
calculation
2. OpenLCA software provides free data and SimaPro provides API for some cost
a. Different CAD can be tried
b. Search in open sources will help to find most suitable / compatible CAD and LCA
software
Deliverables
A solution that can be integrated with existing LCA tool with no/limited
customization and CAD software
Known Bottlenecks / critical areas / Key factors to be addressed
1.Measurement of LCA of the product, due to its material and processing (Global
Warming Potential - CO2 eq.)
2. As simple as possible measuring method
3. Location specific measurement of LCA is good to have
4. Aiming to reduce the time and effort required to conduct Life Cycle Assessments
for products with complex Bill of Materials (BOM) containing 200+ items.
6. A sustainable Kraft paper based cushoning system for packaging
Nature of Challenge
Product Related
Domain / Discipline of Challenge
Sustainability
Description (With Technical Specifications)
Background
Foam is a common material for packaging, but it is not good for the environment
because it is not reusable or recyclable. There are Kraft paper used in corrugated
boards for packaging but these corrugated boards are not made to give the desired
cushioning effect similar to foam. The requirement is to design a cushioning system
out of Kraft paper to match the cushioning properties of the Foam used in packaging
of consumer products so as to replace the single use foam in packaging
Challenge / Requirement
Design and demonstrate a cushioning system out of existing Kraft paper to match the
cushioning properties of the Foam used in packaging of consumer products so as to
replace the single use foam in packaging. The solution should make use of the
existing Kraft paper sheets and make layer keeping in the mind the effectiveness of
mass production.
Deliverables
Design of the new paper based cushioning systems which can be changed based on
the shape of the foam, FEA simulation to show the results, cost comparison of the
designs, and manufacturing feasibility.
A concise report depciting the quality and performance of your cushioning system,
measured by various tests such as compression, impact, vibration, etc., results and
conclusion
Known Bottlenecks / critical areas / Key factors to be addressed
Design of cushioning system using Kraft paper if required with various GSM and
layering
7. Plastic pollution is a global problem " How to minimize plastics part weight
in Automotive Exterior, Interior and Lighting parts like housing, bracket and
non- functional plastics.
Nature of Challenge
Product Related
Domain / Discipline of Challenge
Sustainability
Description (With Technical Specifications)
Background
The usage of plastic auto parts in the automotive industry today has proven to be a
success beyond expectation. Many factors are responsible for the popularity of
plastic, including its various chemical, Thermal and mechanical properties. However,
they also pose a serious environmental problem, as they are not easily biodegraded
and can harm ecosystems and human well-being. Since the plastics used in
automotive applications are made from more complex polymers, their
biodegradability is further reduced.
Challenge / Requirement
Create an alternative material that is biodegradable and can be used to can be used
for headlamp, bumper, interior plastics and other automotive parts. Test their
durability, safety and functionality and balance the environmental and economic
aspects of using biodegradable plastics in the automotive industry.
Deliverables
Alternative bio-degradable solution which can reduce plastic material weight without
compromising its performance.
Example: proposal of natural fiber reinforced plastic composites (plastic + Natural
fiber material)
Known Bottlenecks / critical areas / Key factors to be addressed
Alternative natural material shall be as similar as plastic physical properties,
aesthetics, and chemical properties also have lighter, quieter, and have a more
sustainable life cycle.
Weight reductions could be 20% from current part weight.
8. Alternative for SCR system to control NOx emission
Nature of Challenge
Product Related
Domain / Discipline of Challenge
Sustainability
Description (With Technical Specifications)
Background
A SCR (Selective Catalytic Reduction) system works to reduce the NOx emissions from
automobiles. A SCR system is a device that uses a chemical reaction to convert the
harmful NOx gases into harmless nitrogen and water. The SCR system consists of
several components, such as a urea tank, an injector, a slip catalyst, control unit et
all. A SCR system adds extra cost and weight to the vehicle, and it requires regular
refilling of the urea tank. To achieve the current emission regulation, a SCR system is
used as an additional attachment to control the NOx (this system required for all the
above 2L engine volume automobiles)
Challenge / Requirement
An alternative system/solution to control the NOx in after treatment system without
affecting the performance avoiding the Urea tank (AdBlue solution refilling).

Deliverables
- A working cost effective prototype that demonstrates the alternative system
avoiding the Urea tank (AdBlue solution refilling) while being able to control the NOx
in after treatment system without affecting the performance.
- A schematic diagram and A technical specification of the alternate solution,
showing the components, dimensions, materials, and functions of the system.
- A performance analysis and A cost-benefit analysis of the SCR system, comparing
the NOx reduction efficiency, and the weight and cost of the system with other
alternatives or benchmarks.
Known Bottlenecks / critical areas / Key factors to be addressed
alteration solution for Nox reduction, reducing the urea refilling may not help to
reduce the cost
 Power Electronics
1. Design and Develop an Efficient Pre-charge relay circuit
Nature of Challenge

Product Related

Domain / Discipline of Challenge

Power Electronics

Description (With Technical Specifications)

Background
At power on a power electronics circuit, such as a motor drive or a converter, the
capacitors in the circuit need to be charged up. This can cause a very large current to
flow for a short time, called inrush current, which can damage the circuit or the
power source. A pre charge relay is used to limit the inrush current. This reduces
the voltage difference and the current flow. After the capacitors are charged, the
relay opens and disconnects the resistor or thermistor from the circuit. There is
considerable power loss as relay is always energized during normal operation.
Challenge / Requirement
Design and implement a power electronics circuit that uses a pre charge relay to limit
the inrush current. RT1 and RT2 are the PTC devices in the input path at start-up.
They perform the duty of reducing the inrush current. Once specific duration is
complete, and inrush is in control a circuit or controller signal is used to turn ON RL1
across RT1 and RT2 allowing them to cool down for next power cycle.
Deliverables
Circuit concept
The deliverables are:
- A schematic diagram of the circuit, showing the components and their values.
- A simulation of the circuit, showing the voltage and current waveforms at startup
and steady state.
- A physical prototype of the circuit, using a breadboard or a PCB.
- A report that explains the working principle of the circuit, the selection of the
components, the results of the simulation and the testing, and any challenges or
limitations you faced.
Known Bottlenecks / critical areas / Key factors to be addressed: Low cost, low
power, optimum solution
 NextGen Communication
1.5G orchestrator for network slicing
Nature of Challenge
Product Related
Domain / Discipline of Challenge
NextGen Communication
Description (With Technical Specifications)
Background
The fifth generation (5G) of cellular network promises to be major evolution of
wireless technology. It is planned to be used in broad set of application scenarios
based on the different vertical dimensioning.
Challenge / Requirement
The future challenge of 5G orchestrator is to be able to address usecase/application
requirement by allocating and controling wireless resource allocation, Network
slicing, Network Monitoring and NFVi management. Network slicing emerges as a
cost-efficient solution for implementation of diverse 5G requirements .
Deliverables
1. Development of 5G Network Slice instance as per 3GPP standards to cater the
below categories
• Enhanced Mobile Broadband (eMBB)
• Massive Machine Type Communications (mMTC)
• Ultra-Reliable Low Latency Communication (uRRLC)
2. Provision of the Network slice through the 5G Orchestrator
3. Development of Demo applications to demonstrate the Network slice allocation
and usage
Known Bottlenecks / critical areas / Key factors to be addressed
One critical factor to be considered is 5G network slicing where it demonstrates
different slices in action providing services with different requirements of
functionality, latency, reliability, security and capacity.
2.Failure Prediction of 5G RAN
Nature of Challenge
Product Related
Domain / Discipline of Challenge
NextGen Communication
Description (With Technical Specifications)
Background
A?5G Radio Access Network (RAN) relies on a fully coordinated, multi-layer network
for various bands to provide wireless connectivity to devices. The Radio Access
Network consists of antennas, radios, baseband (RAN Compute), and RAN software.
To ensure the guaranteed QoS is delivered to customer, it is important to ensure we
have continuity of 5G services with expected QoS hence it becomes important to
predict the failure of the RAN
Challenge / Requirement
Prediction of RAN failure in 5G network well ahead of time (can assume at least 24
hours prior)
Deliverables
Expecting a AIML model to be developed to address the failure prediction using the
FM/PM data from Network
Known Bottlenecks / critical areas / Key factors to be addressed
Accurate predictive mechanism system
 Challenges - Metaverse

1.Echo and background noise removal in Metaverse application

Nature of Challenge
Product Related
Domain / Discipline of Challenge
Metaverse
Description (With Technical Specifications)
Background
One of the challenges of building the metaverse is to enable voice communication
amongst users in different locations, using different devices, and speaking different
languages. Voice communication is essential for creating a sense of presence and
realism in the metaverse, as well as for facilitating collaboration and learning. Voice
communication brings in background noise and echo which adversely impacts user
experience.
Challenge / Requirement
Design and implement a solution to enable multiplayer interaction in a metaverse
application with clear spatial and non-spatial audio communication without echo and
background noise.
Deliverables
POC using sample application.
Known Bottlenecks / critical areas / Key factors to be addressed
multiplayer collaboration is required in such application.
2.Flickering effect in VR & User Experience
Nature of Challenge
Product Related
Domain / Discipline of Challenge
Metaverse
Description (With Technical Specifications)
Background
Immersive devices, such as virtual reality (VR) and augmented reality (AR) headsets,
are becoming more popular and accessible. However, they also face a common
problem that affects the quality of the user experience: visualizing. Visualizing is the
process of creating and displaying realistic images that match the user's perspective
and environment. For example, in VR, the images should be consistent with the
user's head movements and eye focus. In AR, the images should blend seamlessly
with the real world and respond to changes in lighting and occlusion. If visualizing is
not done well, the overall user may be impacted. Flickering is one such issue that
can impact user experience.
Challenge / Requirement
Provide a solution to nullify the flickering, experienced in stand-alone VR devices,
while viewing a VR application.
Deliverables
Demonstration of Methodology/Guidelines using a sample application.
Known Bottlenecks / critical areas / Key factors to be addressed
VR devices, Metaverse apps.
3. Navigation in a LIDAR or scanner point cloud temporal scene
using Virtual Reality (VR)
Nature of Challenge
Product Related
Domain / Discipline of Challenge
Metaverse
Description (With Technical Specifications)
Background
Scanning a scene of an industrial machinery set up at a site can be made employing
scanners such as LIDAR, stereo imaging scanners etc., thereby generating point cloud
of the scene. Such point clouds in temporal domain can be used to navigate in a
virtual environment to have virtual tour of the scene helpful in simulations or virtual
walkthroughs
Challenge / Requirement
To implement Virtual Reality software technology and a VR head set to walk through
3D point cloud data sets of a scene captured in the temporal or time domain
Refer to research papers where a VR headset was used, commercial information will
be on internet.
Deliverables
1)Demonstration of the implementation using a VR head set
2)Software architecture/code/tools of the implementation
Known Bottlenecks / critical areas / Key factors to be addressed
Validate the outcome using specific VR tools to determine the quality of the output.
 Challenges - MedTech
1.Intelligent Radiologist Assistant (Knee MRI)
Nature of Challenge
Product Related
Domain / Discipline of Challenge
MedTech
Description (With Technical Specifications)
Background
Magnetic resonance imaging (MRI) of the knee is the preferred method for
diagnosing knee injuries. However, interpretation of knee MRI is time-intensive and
subject to diagnostic error and variability.
Challenge / Requirement
An automated end to end system and a model for interpreting, classifying knee MRI
and be able to prioritize high-risk patients and assist clinicians in making diagnosis.
Deep learning methods, in being able to automatically learn layers of features, are
well suited for modelling the complex relationships between medical images and
their interpretations.
Deliverables
1. A well performing state of the art classification model .
2. Inference code for the model validation
3. API (Application Programming Interface) along with the UI for the demonstration
of the model.
4. Wrapper script and conversion of the inference script from Python to C++ is
required as per the deployment's need.
Known Bottlenecks / critical areas / Key factors to be addressed
The main objectives are as follows -
1. Model should classify the knee injury in an efficient way and should give a better
accuracy as much as possible. Since, this is related to a medical domain, the solution
should be good enough to avoid false-positives and wrong classifications.
2. Create a model to identify or classify the type of knee injury a person has.
3. Create a portal where doctors and patient can login.
4. This portal should use our model to generate a sample report based on the
uploaded
MRI scan of Knee.
5. Doctor should be able to login and upload the MRI scans in the portal and can
generate report using the model in a bulk.
6. Patient should be able to login and generate report too if they possess their knee
MRI
scan or they should see the report generated from the hospital or doctor using our
model.
2. An end to end Lungs Tumor Segmentation AI model through
Attention Mechanism
Nature of Challenge
Product Related
Domain / Discipline of Challenge
MedTech
Description (With Technical Specifications)
Background
Segmentation of the lungs tumor using Artificial Intelligence and Machine Learning is
a demand of this era. It will help doctors and clinicians to have a clear understanding
about the presence of the tumor in lungs. The purpose of this project is to remove
the human errors, which may occur while detecting, annotating and segmenting the
tumor.
Challenge / Requirement
The main objective of the challenge statement is to devlop a end to end solution and
a AI enabled model that can help identify and segment the tumor presented in the
lungs. The solution should be able to use attention based mechanism so as to
concentrate on relevant stimuli.
Deliverables
1. A novel performant state of the art model based on attention mechanism.
2. Inference code for the model validation
3. API (Application Programming Interface) along with the UI for the demonstration
of the model
4. Volumetric analysis of the detected tumor ie. to calculate the area and volume of
the segmented tumor.
5. Wrapper script and conversion of the inference script from Python to C++ is
required as per the deployment's need.
Known Bottlenecks / critical areas / Key factors to be addressed
The main key factor of the solution is to develop a Lungs tumor segmentation model
based on attention mechanism. It should detect and segment the tumor in an
efficient way and should give a better accuracy as much as possible. Since, this is
related to a medical domain, the solution should be good enough to avoid false-
positives and wrong detection/segmentation.
Validation/inference script, API with the Integration of UI, volumetric analysis and
area calculation, Python wrapper and the conversion of the inference code into C++
must be addressed.
3. Virtual Bronchial Tree Navigation and Registration
Nature of Challenge
Product Related
Domain / Discipline of Challenge
MedTech
Description (With Technical Specifications)
Background
Bronchoscopy is a procedure where the scope with camera is passed through a
bronchial tree to detect abnormalities. Before the procedure doctors/
pulmonologists do a path planning using the CT lung images. Using CT images
bronchial tree can be segmented out and centerline or path navigation through the
animated 3D bronchial tree can be seen.
Challenge / Requirement
The challenge is to do registration of real-time bronchoscopy video and showcase in
the 3D model generated from CT lung dataset of that person to understand the
location where the scope has reached in the bronchial tree.
A offline video of real-time bronchoscopy and the CT images of Lung of the same
patient will be shared and a algorithm has to be developed for registration of both.
3D segmented airway model in STL format (segmented from CT lung dataset of a
patient). Virtual Navigation video on the STL model video. Real-time video of
bronchoscopy navigation on same patient.
Deliverables
Demo video of Bronchoscopy registration where the 3D model Virtual navigation
frames has to be registered with the real-time bronchoscopy frames and the current
camera position in real-time bronchoscopy has to be highlighted in the 3D model.
Known Bottlenecks / critical areas / Key factors to be addressed
• Virtual Bronchoscopy
• Navigation
• Registration
 Electric Autonomous Connected Vehicles
1.Recyclable and reusable batteries
Nature of Challenge
Process Related
Domain / Discipline of Challenge
Electric Autonomous Connected Vehicles
Description (With Technical Specifications)
Background
The global EV battery value chain, like others within industrial manufacturing, faces
significant environmental, social, and governance (ESG) challenges. There is an
urgent need to focus on sustainable usage of any engineering materials so as to
attain zero wastage. As battery demand continues to grow, the supply chain of the
raw material will face availability challenges as the materials used for battery
moudules are rare and limited. Producers and purchasers could mitigate potential
shortages of raw materials by redefining their strategies and operations to be
economic, transparent, sustainable, and circular.
The is a huge demand of batteries for Electricle Vehicles across the globe with limited
technology. There are few companies claiming an estimated 95% of EV batteries can
be recycled. However technical details to prove recyclability are limited.
Challenge / Requirement
This problem statement expected to focus on generating ideas, methodologies to get
100 % recyclable batteries and all it components through economically viable
options.
Deliverables
Methodology and guideline to achieve 100 % recyclable batteries considering
particular country (India) for the framework demo
A solution/methodologies/fraemwork to attain 100 % recyclable. A Complete
framework from raw materials to finished products with financial calculation for cost
comparison, sustainable benefits.
Business growth with expected usage for next 10 Yrs, alternate application for Reuse
Known Bottlenecks / critical areas / Key factors to be addressed
• Zero Wastage with minimal usage of Natural Resources
• Least Environmental Effect
• Affordable cost escalation
• Safer Usage
2.Substitute Fluid For Rubber Vulcanizing
Nature of Challenge
Product Related
Domain / Discipline of Challenge
Electric Autonomous Connected Vehicles
Description (With Technical Specifications)
Background
To vulcanize tire rubber, you need to heat it up and keep it at a certain temperature
for some time. This process changes the properties of the rubber and makes it more
durable and elastic. Steam is used as a medium to transfer heat to the rubber.
However, usage of steam results in challenges of usage of precious resources like
water to produce steam; the transportation of the steam from a central boiler plant
to the curing machine far away results in energy losses. There is a also a need to
follow huge statutory requirements of maintaining the super heated steam to
comply with boiler and pressure vessel codes to make sure of personnel safety.
Challenge / Requirement
Develop an alternate eco-friendly, sustainable, and energy efficient
method/technique to heat and maintain the temperature of tire rubber for
vulcanization using an alternate fluid (other than water/steam). The alternate fluid
shoudl have the following properties:
1 Non Hazardous and eco friendly
2 Does not interact chemically with Cured rubber (the medium flows inside the
rubber bladder)
3 should be capable of being heated to 210 Deg C
4 Should have specific heat capacity greater or equal to super saturated steam.
5 Easy to handle
Deliverables
Suggest a fluid, demonstrate and prove its eco friendliness and capability to meet the
above 5 point in problem description., Discuss its various aspects of temperature and
specific heat capacity and compare with steam. The fluid should also be rubber
friendly chemically
A report and presentation that summarizes your findings, results, and
recommendations, using clear and concise language and visuals.
Known Bottlenecks / critical areas / Key factors to be addressed
Temperature,( 210 deg C) , heat transfer co-efficient > super saturated steam, and
easy handling, The fluid could also be a ER fluid for example, whose properties are
controllable through controllers
3. Basement parking monitoring system
Nature of Challenge
Process Related
Domain / Discipline of Challenge
Electric Autonomous Connected Vehicles
Description (With Technical Specifications)
Background
A basement parking monitoring system is a way to keep track of how many parking
spaces are available in the basement of a building. It can help people find a parking
spot faster and use the space more efficiently. The system can show the live status of
the parking space on a screen or a sign at the entrance of the basement.
The current parking management is done by people who have to count and update
the number of parking spaces manually. This is not very effective because sometimes
there are not enough people to do this job or they make mistakes. This can cause
confusion and frustration for the drivers who want to park in the basement.
Challenge / Requirement
Design a basement parking monitoring system that can automatically track and
display the availability of parking spaces in real time, and improve the parking
experience for the drivers. The cost effective solution should consider minimum #of
sensors to cover a large area and the same tiime guiding the driver to the available
spots.
Deliverables
An end to end working prototype solution conisdering required deployment, apps,
hardware, navigation cosidering both the parking operator and the driver.
Known Bottlenecks / critical areas / Key factors to be addressed
1) Fully Automatic system
2) Guide car driver to find empty space for parking, Navigation to space available for
parking so as to assist the driver to follow the path to park.
3) Display number of the car parking space available at the entrance of the Basement
parking area
4.Glare caused a headlight’s high beam having a dazzling effect for
oncoming traffic.
Nature of Challenge
Product Related
Domain / Discipline of Challenge
Electric Autonomous Connected Vehicles
Background
The primary function of headlamps of a vehcile is to provide forward visibility for
drivers. At the same time, there’s a risk that intense headlamp illumination may be
directed toward oncoming vehicles. Such illumination, referred to as glare, can
reduce the ability of other drivers to see and may cause discomfort.
Glare is related to safety because it can degrade important aspects of driving
performance. Headlamp illumination produces glare for drivers of oncoming or
preceding vehicles (via the rearview or side mirrors).
Challenge / Requirement
Design an alternative and cost effective solution to avoid glare in headlamp system.
The goal of the system is to improve long-range visibility for the driver without
causing discomfort, distraction, or glare to other road users. The automatic
adaptation of the beam pattern may not only serve as a convenience feature for
drivers, but could result in increased, safety-beneficial upper beam use.
Automatic adjustment of the headlamp beam pattern can be accomplished through
advanced front lighting systems. ADB is a “long-range forward visibility beam that
adapts to the presence of opposing and preceding vehicles by modifying portions of
its beam pattern to avoid glare above lower beam photometry levels to the drivers of
opposing and preceding vehicles”
Deliverables
A working prototype/POC with analsis report. Currently Adoptive driving beam with
help of Camera system available in high end cars.
Expecting Cost effective unique solution , which is feasible to adopt all vehicles .
Known Bottlenecks / critical areas / Key factors to be addressed
Driving Performance, Driver Comfort and less glare ,Cost effective solution all kind of
vehicle.
5. Design and build Smart Tyre to achieve minimal Tyre wear
Nature of Challenge
Product Related
Domain / Discipline of Challenge
Electric Autonomous Connected Vehicles
Description (With Technical Specifications)
Background
Tyre wear is a major source of particle pollution that affects the air, water and soil
quality. It is more harmful than other types of pollution. It is estimated that a
significant portion of all the small particle pollution from road comes from tyre and
brake wear. There is a need to develop smart tyres that have less wear and tear.
Challenge / Requirement
Design smart tyres that has minimal wear and tear. Potential solutions include using
alternative materials that are more durable and less abrasive, creating new tyre
tread patterns that optimize performance and efficiency, and reducing the overall
weight of the tyres. The smart tyre is expected to have lesser or equal cost, if not
slight cost escalation is acceptable.
Deliverables
Simulations with necessary assumptions to explain the approach if possible test
samples can be made to conduct practical test.
Known Bottlenecks / critical areas / Key factors to be addressed
• Less wear and tear
• Cost implications
• Raw materials Source and availability
• Reuse / Recycling methodologies
6.Vehicle and passenger safety from confined waterlogging
Nature of Challenge
Product Related
Domain / Discipline of Challenge
Electric Autonomous Connected Vehicles
Description (With Technical Specifications)
Background
Urban water logging has become very severe owing to alteration in land use and
climate change impact. Narrow spaces like an underpass pose a risk to small cars
where drivers often misjudge the depth of water. Overestimating their vehicle's
ability to wade through flooded waters causes drivers to override safety.
Challenge / Requirement
Develop a solution that can detect and alert the driver preventing them from taking
the risk of wading the vehicle through the water logs
Deliverables
An end to end prototype solution that can detect water logs and depth of
accumulated water along with warning to the driver. The product may also be
configured to send GPS signal to authorities (e.g. Fire Department) in case a vehicle
trying to cross is stuck in the waterlogged area.
Known Bottlenecks / critical areas / Key factors to be addressed
Human safety and uninterrupted power supply to the warning system are key
factors. In case a walled passage leads to the underpass, the warning lights must also
be installed much before the flooded portion starts.
7. Anti-theft Security System for vehicle tires, battery, accessories
or body parts
Nature of Challenge
Product Related
Domain / Discipline of Challenge
Electric Autonomous Connected Vehicles
Description (With Technical Specifications)
Background
In general, professional auto theft is done with the objective of removing and selling
major body and drivetrain components or reselling the vehicle. Most of the
conventional vehicle anti-theft system are designed to detect/sense illegal
entry/access to vehicle and provide alerts. Theft of wheels, battery, accessories, body
parts et all can go undetected by anti-theft system or when the anti-theft system is
compromised.
Challenge / Requirement
Build an end-to-end anti-theft solution that can prevent/warn/detect the theft of
parts of the car.
Deliverables
• An Advanced Burglar system that detects any theft or robbery from vehicle
• Prototype of a work model
Known Bottlenecks / critical areas / Key factors to be addressed
Robbery of any part in the vehicle needs to be identified, alarm should be provided,
and the information needs to be communicated to the concerned stakeholders like
Owner of vehicle, Nearby police station and so on.
8.Managing EMI/EMC in high power drives design for Electrical
Vehicle
Nature of Challenge
Product Related
Domain / Discipline of Challenge
Electric Autonomous Connected Vehicles
Description (With Technical Specifications)
Background
Electric vehicles (EVs) are becoming more popular and advanced, but they still rely
on some basic principles of electric drives and motors. One of the main challenges in
designing EVs is to ensure that the power electronics systems, such as converters,
inverters, and active rectifiers, can handle high voltages, currents, and power levels
without causing electromagnetic interference (EMI) or electromagnetic compatibility
(EMC) problems. High Power electronics modules systems such as converters and
inverters, Active rectifiers for ratings in the range of 800VDC and above, 60amps and
greater , 250KW and greater have EMI/EMC compliance challenges.
Challenge / Requirement
Develop a innovative demonstrable prototype/POC/method either Hardware (HW)
or Software (SW) approach that address EMI/EMC in motor control for Electric
Vehicle, and which is immune to EMI/EMC issues for handling designs that can
manage ratings as mentioned in the background of the challenge statement.
Deliverables
- A Hardware or Software approach , methodology, concepts and solutions to
address EMI/EMC issues in motor control for electric vehicle
- A prototype of the proposed power electronics system, along with the testing and
validation procedures and results.
- A report that summarizes your findings, analysis, and recommendations, along with
the references and appendices.
Known Bottlenecks / critical areas / Key factors to be addressed
A design of a power electronics system that can operate at 800V DC, 60A, and 250kW
without causing or suffering from EMI/EMC issues, including the specifications,
components, layout, and simulation results.
9.Realtime Tyre Tread Depth
Nature of Challenge
Product Related
Domain / Discipline of Challenge
Electric Autonomous Connected Vehicles
Description (With Technical Specifications)
Background
Tyre tread grips the road during drive. If Tyre tread depth is not deep enough, car
loses traction and suffers extended braking times. Shallow tread grooves make it
harder to control the vehicle in wet weather and the chance of aquaplaning
increases, resulting is accidents.
OEMs today are trying to introduce automatic tyre tread depth measurement and
display to consumer it’s live data without major success. Currently either manual
measurement or if automated it is established in the vehicle yard and system is
placed out of the product.
Challenge / Requirement
Design and prototype a method to measure tyre tread depth realtime embedded
measurement method established in tyre and displaying runtime to the driver in
various running conditions.
Deliverables
A POC product demonstrating tyre tread depth measurement online with any
embedded measurement method established in tyre itself.
Known Bottlenecks / critical areas / Key factors to be addressed
A POC product demonstrating tyre tread depth measurement online with any
embedded measurement method established in tyre itself.
10. Air Traffic Control with VTOLs
Nature of Challenge
Product Related
Domain / Discipline of Challenge
Electric Autonomous Connected Vehicles
Description (With Technical Specifications)
Background
Alongside Electric Vehicle EVs, there is increasing research and development
occurring on eVTOL (Electric Vertical Take-Off and Landing). VTOL are vehicles that
can fly in the air and can take off and land without a runway which brings in
challenges for air traffic control - a system that manages the movement of aircrafts in
the sky and prevents collisions. Air traffic control is already difficult for existing
aircrafts, such as planes and helicopters, because there are many of them and they
fly at different speeds and altitudes. For eVTOLs, it could be even more complicated,
because they might fly at lower altitudes, closer to buildings and other obstacles, and
in more crowded areas.
Challenge / Requirement
Develop safe and efficient air traffic management for eVTOLs
Deliverables
- Solution that will solve eVOTL traffic control challenge to make it a reality
- Minimal disruption to human activity
- Passenger safety
Demo of safe and efficient air traffic management for eVTOLS
Known Bottlenecks / critical areas / Key factors to be addressed
- Solution that will solve eVOTL traffic control challenge to make it a reality
- Minimal disruption to human activity
- Sustainability / carbon footprint consideration is a common denominator
11. Focus on Vehicle Interior design Access and Security products
Nature of Challenge
Product Related
Domain / Discipline of Challenge
Electric Autonomous Connected Vehicles
Background
The Vehicle Steering Wheel Switch Hand Capacitance sensors are designed to detect
the presence of the driver's hand on the steering wheel and provide feedback to the
vehicle's safety systems. The sensors need to perform reliably under various
environmental conditions, such as humidity and temperature.
Challenge / Requirement
How Might We achieve the Vehicle Steering Wheel Switch Hand Capacitance sensors
design performance PASS the Automotive Humidity and Tri Temperature testing’s @
( 24 deg C, - 40 deg C, +80 deg.)
Deliverables
# A detailed design specification that defines the sensor parameters, such as
capacitance range, sensitivity, response time, and accuracy.
# A test plan that outlines the test procedures, equipment, and criteria for the
Automotive Humidity and Tri Temperature testing's at ( 24 deg C, - 40 deg C, +80 deg
C).
# A test report that documents the test results, analysis, and recommendations for
improvement or optimization.
12.Detection of Porosity in Rubber Products
Nature of Challenge
Product Related
Domain / Discipline of Challenge
Electric Autonomous Connected Vehicles
Description (With Technical Specifications)
Background
Porosity detection is important step in tyre industry. Rubber product has porosity if
either it is not reached to the optimum cure stage or has some volatile material.
Optimum cure stage is not reached can happen for two situation :1. new
development of tyre size/compounds and thus not known the optimum cure cycle, 2.
Power failure or any other failure in process which results in less cure input to the
product. Major reason for porosity in tyre is under cure stage. Current practice is to
cut the tyre and visually see if there is any porosity exists.
Once we cut the tyre it has zero sale value so if we can detect this porosity without
cutting the tyre one can save the cost of the tyre.
Some work is available on new with topic "Millimeter Wave SAR Imaging for the
Detection of Porosity in Rubber Products"
Our intention is to develop any solution/device which can detect porosity in the level
of 0.02mm diameter without cutting the tyre.
Challenge / Requirement
How to detect porosity in the level of 0.02mm diameter without cutting the tyre.
Deliverables
Prototype system to detect porosity in the level of 0.02mm diameter in a rubber
block of 3-5 inch thickness.
Known Bottlenecks / critical areas / Key factors to be addressed
Prototype system to detect porosity in the level of 0.02mm diameter in a rubber
block of 3-5inch thickness.
 AI (EDUCATION):-
1. LEARNING ASSISTANTS:
Nature of Challenge

Product Related

Domain / Discipline of Challenge

AI (Education)

Background

Currently, the help provided for students at any level is ad hoc, scattered, and sometimes
irrelevant. It is a challenge for students to know where to look for what information that
applies to the subject they are studying. There is also no proper method to keep a student
engaged and motivated during studies especially when they are trying to learn a complex
concept.

Challenge / Requirement

How can we create an AI-powered chatbot that helps students by constantly quizzing them,
explaining concepts as when they arise and summarize lectures with key points? How can
we create a tool that automatically takes relevant notes and flashcards. In addition, how can
the system provide timely reminders for key events in the academic journey

Deliverables

1. Virtual Study Buddy that plugs into the academic subject and generates valuable and
intermittent quizzes

2. Lecture summarizer - Listens to recorded lectures and provides concise text summaries

3. Automated note-taker - A note taking function that transcribes lectures and highlights
important information for later review

4. Intelligent flash cards - a feature that generates personalized flash cards using NLP

5. Personalized reminders - a feature that analyzes a student's calendar and habits to

provide timely reminders about assignments, project due dates, meetings etc.

Known Bottlenecks / critical areas / Key factors to be addressed

Highly secure system,Highly accurate in its function, Mobile and web enabled, Can read in
text and image documents like pdf, excel, Word, ppt etc
2. CAREER COUNSELLING
Nature of Challenge
Product Related
Domain / Discipline of Challenge
AI (Education)
Background
Today, career counseling is wholly based on the quality and experience of the
counselors. The heavy dependency on individuals often leads to subpar experience
for a student when attempting to enter the job market in terms of knowing what
kind of jobs to apply for and how to create a viable career path.
Challenge / Requirement
How can we enable a better and more consistent experience for college students
entering the job market to understand the potential jobs that they can apply for
based on their academic program and academic courses. In addition, how can a
system be created that allows for realistic job interview situations including feedback
on how a student did on a mock interview.
Deliverables
An AI-based career counselor chatbot that is able to extract a student's skills from
resumes, cover letters, academic course work and projects to match industry jobs
and skills. The system should also have a chatbot that conducts practice interviews
and gives feedback to improve the student's skills and communication styles
Known Bottlenecks / critical areas / Key factors to be addressed
1. Highly secure
2. Should be able to work with documents, images and audio
3. Web and mobile enabled
3.CAMPUS ONBOARDING ASSIST
Nature of Challenge
Product Related
Domain / Discipline of Challenge
AI (Education)
Description (With Technical Specifications)
Background
The first time a student enters a college can be an intimidating experience. The
student is often trying to figure out where to go and what to do during the early days
in the university. Also, when a student goes to a new country, there may be
additional challenges in learning a new spoken language and understanding all the
slang and cultural connotations.
Challenge / Requirement
How can we make the life of first time college student easier? How do we provide an
easy way for a student to navigate through a new campus and guide them through
the various processes including opening bank accounts, locating the dorms, getting
food coupons etc. Also, how do we help a student translate signs and notices in
languages that are not familiar to them?
Deliverables
An AI-based campus navigation assistant is a mobile app that uses computer vision
and deep learning to guide students across the campus using augmented reality and
help. The app should be able to translate between languages and help users settle
down. It would be highly beneficial for the app to possess the capability to facilitate
language translation and assist individuals in acclimating to new environments. In
addition, this app should be able to translate between languages so students can
understand notices and signs on campus.
Known Bottlenecks / critical areas / Key factors to be addressed
1. Mobile app that can work on Android or iOS
2. Easily navigable interfaces
3. Light-weight and secure app
 Challenges - AI (Generic)
1.Multilingual Interpreter/Translator
Nature of Challenge
Product Related
Domain / Discipline of Challenge
AI (Generic)
Description (With Technical Specifications)
Background
Globalization has brought different ethnicity, cultures much closer today. Language
continues to remain a barrier to convey our messages to people in the globalization
and communication era. Translation of language from source to intended audience
at run time can help bridge the language barrier.
Challenge / Requirement
Develop a Software application which is capable of understanding multiple
languages (English, Hindi, Chinese, Spanish, Urdu, etc.) and be able to convert Input
lang. (mic) to Output lang. (speaker) of choice on the fly. E.g. Think of an Amazon
Alexa like device which can listen and instantly convert to your choice of output
language.
Deliverables
Software module must be platform independent. It can be configured on any
multimedia capable system e.g., Smart speaker, Mobile device, Smart gadget.
Known Bottlenecks / critical areas / Key factors to be addressed
Performance - Instant translation.
Accuracy
2.Differentiation and tagging of Synthetic Vs Real data:
Nature of Challenge
Product Related
Domain / Discipline of Challenge
AI (Generic)
Description (With Technical Specifications)
Background
Data is now everywhere and exponentially getting larger in volume and type. There is
a serious problem in identifying data that is fake and authentic data. Due to the lack
of controls and lack of regulations of AI data, numerous problems arise, including
data whose sources are unknown, data that is used to make inaccurate conclusions
as well as lead to wasted resources on storage, maintenance, and processing of false
data. Fake data can also be used to train algorithms, leading to biased results.
Challenge / Requirement
The ask is to create a solution that can somehow tag every piece of original work,
whether images, video, audio or text) with a unique digital signature or watermark
that is easy to verify for the public but impossible for bad actors to crack as it does
not follow human identifiable patterns. The watermark should be permanent and
non-erasable and cannot be deleted even if the digital asset is copied or duplicated.
The solution should also give data owners complete authority on who can and who
cannot train AI models with their data.
Deliverables
1. A digital watermark algorithm that can embed digital watermarks into data sets.
2. The watermark should work on images, text, audio, and video.
3. A trigger dataset that contains the digital watermarks
4. A method for verifying the integrity of data streams by extracting the digital
watermarks
5. A method for controlling access to the data sets based on the presence of digital
watermarks
Known Bottlenecks / critical areas / Key factors to be addressed
1. Robustness: The watermark should not be easily removed or rendered useless
when altered.
2. Reversibility: The watermark should be reversible.
3. Quality: The watermark should not reduce the purpose or quality of the data.
4. Security: The watermark should be secure and not easily broken or removed
3. Model Optimization and Compression
Nature of Challenge
Product Related
Domain / Discipline of Challenge
AI (Generic)
Description (With Technical Specifications)
Background
Generative Artificial Intelligence(GenAI), has gained significant attention in recent
years. These models are known for their large size and require substantial
computational resources for both training and making predictions.
Challenge / Requirement
Create a framework for open-source GenAI models (LLM and Diffusion based Model),
compress and optimize them with various techniques and compare the performance
and accuracy aspects for both training and inference perspectives.
Deliverables
Deliverables: Source code, documentation, and demo video.
Optimization Techniques & Evaluation
Use Cases and Demonstrations
Known Bottlenecks / critical areas / Key factors to be addressed
The use of existing tools like Openvino and similar types of tools can be used and
compared.
4.Prompting Framework For LLM
Nature of Challenge
Product Related
Domain / Discipline of Challenge
AI (Generic)
Description (With Technical Specifications)
Background
Large Language Models (LLMs) are gaining popularity in various fields, including
engineering. A new subject called "Prompting" is emerging, which involves using
these LLMs effectively.
Challenge / Requirement
Design and develop an innovative framework for Prompting Large Language Models
(LLMs) in engineering applications. The framework should explore and propose
effective methods for utilizing LLMs to solve specific engineering tasks. It must
provide clear guidelines on how to prompt LLMs with appropriate instructions and
constraints to achieve accurate and relevant outputs. The challenge aims to leverage
the power of LLMs and their language understanding capabilities to enhance
engineering problem-solving. In the solution, the user should also be able to upload
the relevant documents in the form of pdf, word, ppt, excel etc. type of files and
write the appropriate prompts to get the relevant information from the uploaded
documents.
Deliverables
Deliverables will be source code, documentation and video demo.
Evaluation and Performance Analysis

Known Bottlenecks / critical areas / Key factors to be addressed

The solution should try to focus on using open-source LLM models and must have
comparative reports of the accuracy and performance of the open-source LLM
models with the benchmark ones like ChatGPT or BARD.
5.Auto generation/modification of prompts based on factual
content of answers
Nature of Challenge
Product Related
Domain / Discipline of Challenge
AI (Generic)
Description (With Technical Specifications)
Background
Prompt engineering is emerging are where writing a right prompt is an art that
minimizes the hallicination and increases the factual content and conciseness of
generated output from LLM.
There is need of a system that automatically generates the right prompt or correct
the earlier prompt to get the right answer.
Challenge / Requirement
build a ML model that should generate and re-generates prompt that induces zero
short learning, or few shot learning or teach LLM model with examples until large
language model generates the right answer that should have minimum hallucinations
and maximum relevant factual information.
Deliverables
light weight LLM model or alternate DL model that controls regenerative prompt.
Mechanism to measure the perplexity and accuracy output.
simple UI to demonstrate this idea.
Known Bottlenecks / critical areas / Key factors to be addressed
light weight LLM model or alternate DL model that controls regenerative prompt.
Mechanism to measure the perplexity and accuracy output.
simple UI to demonstrate this idea.
 Challenges - AI (Medical)
1.GENERATION OF NEW MOLECULES SIMILAR TO EXISTING DRUG
Nature of Challenge
Process Related
Domain / Discipline of Challenge
AI (Medical)
Description (With Technical Specifications)
Background
Drug discovery is a critical and challenging areas in medicine. Drug discovery involves
a very prolonged and costly process of identifying potential drug candidates and
testing their efficacy and safety.
Challenge / Requirement
Generative AI can significantly accelerate this process by using deep learning
algorithms to identify potential drug candidates based on a large dataset of
compounds and their properties.
Virtual compounds are generated by AI algorithms and tested in silico, which means
in a computer simulation rather than a laboratory. This decreases the time and cost
involved in the discovery of new drugs.
Deliverables
1. potential drug candidates
2. new molecules optimized for a specific target
Known Bottlenecks / critical areas / Key factors to be addressed
Reduce the time consuming and expensive drug discovery process.
2.Maximizing Efficiency: Reducing Experimental Costs in Mass
Spectrometry Research
Nature of Challenge
Process Related
Domain / Discipline of Challenge
AI (Medical)
Description (With Technical Specifications)
Background
Generative AI offers a revolutionary approach to experimentation. It enables
researchers to simulate mass spectrometry experiments virtually, reducing the need
for resource-intensive physical trials. This virtual experimentation not only saves
costs associated with materials, equipment, and labor but also accelerates research
timelines by swiftly exploring a wide array of experimental scenarios. This speed and
efficiency empower researchers to make informed decisions and optimize
experimental setups before any physical work is undertaken.
The predictive capabilities of generative AI play a crucial role in data interpretation
and decision-making. Researchers can rely on AI-generated insights to predict mass
spectra, propose potential chemical structures, and guide the allocation of resources.
This results in robust and reliable data analysis, reducing human error and increasing
the confidence in research outcomes.
Moreover, the economic advantages of generative AI extend beyond cost savings. By
minimizing the wastage of resources, it contributes to sustainable research practices,
aligning with the global push for environmental responsibility.
Generative AI democratizes research by making advanced experimentation
accessible to researchers with varying levels of expertise. It stimulates innovation by
encouraging exploration of unconventional hypotheses and streamlining iterative
optimization. Additionally, its role in virtual screening and compound identification
accelerates drug discovery, materials science, and other research domains.
Challenge / Requirement
Collect diverse, high-quality mass spectrometry data for specific defined goal. Choose
an appropriate generative AI model like VAEs, GANs, or Transformers. Train the
model using the dataset, optimizing parameters and loss functions. Employ the
model for virtual experimentation, simulating mass spectra for different scenarios,
and predicting the effects of experimental variables.
Utilize the model for optimal experimental design, identifying conditions likely to
yield informative results. Integrate the generative AI into data augmentation
pipelines to enhance dataset diversity. For compound identification, employ the
model to propose potential chemical structures matching observed spectra.

Deliverables
Optimized experimental protocols for cost-effective mass spectrometry, a trained
generative AI model for virtual experimentation, enhanced datasets through AI-
generated data augmentation, predictive tools for compound identification, and
insights on optimal conditions.
Additionally, collaboration with domain experts, validated predictions, and
documented methodologies has to be shared.
Known Bottlenecks / critical areas / Key factors to be addressed
The solution should address critical factors such as accurate virtual experimentation
for efficient optimization of experimental setups, precise compound identification
through AI-generated predictions, streamlined data augmentation to enhance
dataset diversity, and collaboration with domain experts to ensure validation. Key
factors include the reduction of resource-intensive physical experiments, cost-
effective utilization of specialized instruments, and the integration of responsible AI
practices to ethically drive research efficiency.
3.Medical Device Cyber Security Threat Solution.
Nature of Challenge
Product Related
Domain / Discipline of Challenge
AI (Medical)
Description (With Technical Specifications)
Background
There are lot of advancement in medical industry where SW are used for
connectivity, diagnosis, and Patient health data storage. Though medical devices and
software are tested for cyber security threats using various software and manual
processes, there are always security vulnerability that can remain or come in through
upgrades.
Challenge / Requirement
Design a software solution using opensource AI stacks to detect security vulnerability
of medical devices by considering
a)Applicable regulatory needs for medical devices
b)list tests to be performed
c)calculate threat perception high/medium/low.
d)perform test on device or sw
e)list the threats
The solution can consider usage of AI for cyber security threat detection. The tool
can learn and get trained continuously and automatically to help improve the threat
detection in the medical and life science devices.
Deliverables
1. create SW application which can run on any platform such as computer , mobile
etc.
2. Connects to internet and send some data to another application or device
3. Software developed using open AI or any other stack shall be able to detect the
threat in application mentioned in 1 and 2.
[note: SW applications 1 and 2 can be free SW as well and can be downloaded from
internet if available but it shall be medical device application.]
Known Bottlenecks / critical areas / Key factors to be addressed
SW tool developed shall be able to find 1. applicable standards for cyber security 2.
identify threat in any connected device or SW.

4.AI based skin cancer diagnosis

Nature of Challenge
Process Related
Domain / Discipline of Challenge
AI (Medical)
Description (With Technical Specifications)
Background
Diagnosing Skin cancer is a time consuming and expensive process. AI algorithms can
identify patterns indicative of skin cancer by analysing a large dataset of skin images.
This can help dermatologists to make more accurate diagnoses and improve patient
outcomes.
Challenge / Requirement
AI algorithms can learn from a large dataset of medical images and identify patterns
associated with specific diseases.The algorithm can learn from a large CT scan
dataset and identify patterns indicative of lung cancer which can then be used in the
diagnosis of skin cancer. The algorithm can learn from a large dataset of skin images
and identify patterns indicative of skin cancer.
Deliverables
A trained model that can indicate the probability of having skin cancer in the near
future.
Known Bottlenecks / critical areas / Key factors to be addressed
Should be able to identify patterns from images like CT scans, MRIs and X-Rays
5. Personalized medical advice and recommendation chatbot
Nature of Challenge
Product Related
Domain / Discipline of Challenge
AI (Medical)
Description (With Technical Specifications)
Background
Chatbots are revolutionizing various industries and healthcare is no exception. By
leveraging artificial intelligence and natural language processing, medical chatbots
have the ability to assist patients, provide accurate diagnoses, and enhance the
overall healthcare experience.
Challenge / Requirement
Develop an AI based chatbot that can interact with patients about their symptoms
and deliver personalized medical advice.
Deliverables
Should be able to ask patients questions and deliver personalized medical advice.
Known Bottlenecks / critical areas / Key factors to be addressed
Patient centric and individual specific personalized medical advice.
6.Personalized treatment plans for patients
Nature of Challenge
Product Related
Domain / Discipline of Challenge
AI (Medical)
Description (With Technical Specifications)
Background
Chatbots are revolutionizing various industries and healthcare is no exception. By
leveraging artificial intelligence and natural language processing, medical chatbots
have the ability to assist patients, provide accurate diagnoses, and enhance the
overall healthcare experience. Generative AI algorithm can analyze a patient's
medical history, genetic information, and lifestyle factors to create a personalized
treatment plan. This can help doctors and healthcare providers make more informed
patient care decisions.
Challenge / Requirement
The algorithm should analyze a patient's tumor DNA and identify the genetic
mutations driving the cancer. Based on this information, the algorithm can
recommend a personalized treatment plan that targets specific genetic mutations.
The model should also be able to predict patient outcomes. The algorithm can learn
from a large patient data dataset and identify patterns associated with specific
outcomes.
Deliverables
Patient centric personalized treatment plan
Known Bottlenecks / critical areas / Key factors to be addressed
Most of treatment plans recommended by Physicians are disease and symptom
specific. They do not take into account patient's medical history, genetic information,
and lifestyle factors
The treatment plan generated by this solution should be individual specific
7.Improving patient outcomes Through Medical Simulations Assist
For Healthcare Provider
Nature of Challenge
Product Related
Domain / Discipline of Challenge
AI (Medical)
Description (With Technical Specifications)
Background
Sepsis is a life threatening condition. Studies have shown that early identification of
sepsis following rapid initiation of antibiotic treatment improves patient outcomes. A
generative AI algorithm that can simulate different scenarios for treating sepsis.
Challenge / Requirement
Generative AI creates realistic models for medical simulations and training purposes.
With hyper-realistic training environments, healthcare professionals can engage in
complex medical procedures and improve their skills and decision-making abilities
without real-world consequences.
Deliverables
Simulated scenarios and hyper realistic training models
8.Synthetic Medical image creation
Nature of Challenge
Product Related
Domain / Discipline of Challenge
AI (Medical)
Description (With Technical Specifications)
Background
Generative AI can help improve the accuracy and efficiency of using machine learning
in combination with medical imaging techniques, such as CT and MRI scans. Machine
learning models can automatically identify abnormalities in images and alert doctors
to potential issues.

Generative AI improves machine learning effectiveness via these steps:

1. A generative AI algorithm like generative adversarial networks (GANs) is trained on

real patient data.
2. The generative AI model generates new images that are similar to the input data
but can result in better machine learning due to increased volume or diversity in the
data.
3. AI model is trained on the synthetic data.
Challenge / Requirement
Create new medical images, such as X-rays or MRIs
Deliverables
A POC that can demonstrate creation of medical images and can be used to train
other models that to automatically identify abnormalities in images and alert health
provider potential issues.
9.AI-based clinical documentation
Nature of Challenge
Product Related
Domain / Discipline of Challenge
AI (Medical)
Description (With Technical Specifications)
Background
The goal of medical scribes is to pare down documentation burden on clinicians.
Physicians can spend up to six hours per day logging notes into an EHR, cutting into
time with patients and contributing to burnout.
Challenge / Requirement
Clinical documentation service that allows providers to automatically create medical
notes using Generative AI. Providers to build clinical applications that use speech
recognition and generative AI to create transcripts of patient visits, identify key
details and create summaries that can be entered into an electronic health record.
Deliverables
The source of generated text in the summary from the original transcript, making it
easier for doctors to review clinical notes.
 Challenges - AI (Others)
1.Image Semantic Segmentation
Nature of Challenge
Product Related
Domain / Discipline of Challenge
AI (Others)
Description (With Technical Specifications)
Background
Semantic segmentation is the task of classifying each and very pixel in an image into
a class. These may be needed for various applications such as self driving cars,
robotic system, damage detection et all. Collection of thousands and millions of
images and annotation for Semantic segmentations and scenario labelling involves
considerable efforts and is time consuming.
Challenge / Requirement
Generation of images for autonmous car scenario sush as road marking, highways,
streets, road users et all.
Annotaion of the generated images for semantic segmentation and scenario labelling
Deliverables
Expected delivery is using generative AI we should be able to create the images for
these scenarios road users, road markings, high ways, streets etc.
Known Bottlenecks / critical areas / Key factors to be addressed
Generating required images and which are auto labelled/annotated.
2.Image Processing in a Harsh Environment (underground)
Nature of Challenge
Product Related
Domain / Discipline of Challenge
AI (Others)
Description (With Technical Specifications)
Background
Navigating through dark and dusty areas is a common challenge for explorers,
rescuers, or researchers who work in environments such as caves, mines, or disaster
zones. These areas often have limited visibility, complex terrain, and unknown
hazards that can hinder the progress and safety of the mission. To overcome these
challenges, there is a need for effective tools and techniques that can help the
navigators to map, locate, and communicate in these difficult settings.
Challenge / Requirement
Explore the feasibility and benefits of using a thermal imaging camera for object
detection in low-visibility environments.

The thermal imaging camera would capture the temperature differences of the
objects in the scene and generate an image that shows their shapes and locations.
Alternatively, another sensor that can detect obstacles or motion in the absence of
visible light could be used.
Deliverables
Develop a PoC which would test and demonstrate the performance and reliability of
the imaging system under harsh conditions, such as low light, dust, smoke, or fog.
Known Bottlenecks / critical areas / Key factors to be addressed
Should work above the ground and scan underground environment and provide data
of the machine and environment.
3.Video and audio analysis of Live video/audio streaming to help
enhance end user experience
Nature of Challenge
Product Related
Domain / Discipline of Challenge
AI (Others)
Description (With Technical Specifications)
Background
All major video streaming services or apps are competing to ensure the best quality
of experience (quick video start, no buffering, quality playback without Audio/Video
issues etc). Video & audio analysis for application in the area of Live video streaming
on various OTT devices forms the backbone of collecting metrics on the end user
Quality of experience in the Video streaming world. This is required to identify and
benchmark golden batch of KPI's which the engineering team should meet and
improve.
Because of the various device platforms and complexities of video streaming (device
hardware, network speed, concurrency etc) and the ask to validate KPI or defects
that only the human senses (eyes & ears) can detect, the audio & video analysis on
Video playback on devices has been a challenge.

Challenge / Requirement
Develop compute vision based generic audio/video analysis algorithms leveraging
AI/ML models to come up with a best in class solution for Video & audio analysis on
streaming devices (mobiles/STB/Smart TV/Roku/FireTV/AppleTV etc) in Hitech &
Media industry segment.

Deliverables
•Live Stream Video Analysis compute vision No Reference algorithm (Real time video
monitoring to identify anomalies like macro blocking, blur, motion blur, poor quality,
colour distortions etc)
•No Reference compute vision algorithm for Video Analysis able to detect objects,
patterns and motion (no source video)
•4K Video Capture & analysis (quality analysis)
•Audio Analysis in video stream playback ( language detection, Audio quality
metrics, Deliberate vs unexpected audio Loss)
•Audio/Video Sync Validation on a video playback (no source video)
•Closed Captioning Language & Correctness (audio vs CC text)
•Closed Captioning sync with Audio during video playback (no source video)
•video playback A/V Quality at Content-AD transition (video and audio anomalies
when switching between ad and content and vice versa) - (Ads are dynamic and not
know prior)
4.Identification Of Flood Prone Area in Urban Settlement
Nature of Challenge
Product Related
Domain / Discipline of Challenge
AI (Others)
Description (With Technical Specifications)
Background
India regularly witnesses severe floods during the monsoon season. Monsoons are
becoming more erratic because of climate change and global warming. Flood water
can deviate from its usual course due to various factors, such as obstacles, urban
development, landscape changes, etc. One of the challenge of Urban areas is to
prepare for eventuality such as floods in NEW areas and having safety measures in
place to protect human lives and at the same time restrict damages.

Challenge / Requirement
AI based predictive analysis solution to predict flood risk prone area based on
aspects such as change in landscape, obstacles et all. The solution should also be able
to visualize the flood water path possibilities by analyzing the landscape changes.

Deliverables
Solution approach (Historical data collection, Image processing, Predictive analysis)
Solution design and implementation strategy (Tools, Opensource solutions)
POC of the AI model which can identify obstacles, change in landscape etc from the
images
Irregular flood water prediction analysis - Good to have
Known Bottlenecks / critical areas / Key factors to be addressed
Obstacles, objects and landscape change detection in flood sensitive regions.
Predictive analysis of the flood water path and identification of existing and new
areas which could get affected by flooding in future.
5.GenAI based RFP(Request for proposal) assessment againt tendor
Nature of Challenge
Product Related
Domain / Discipline of Challenge
AI (Others)
Description (With Technical Specifications)
Background
In Contruction segment contracting and sub-contracting there are multiple RFP
submitted by multiple vendors again any tendor, assessing those RFP in terms of
technical, financial, legal and work scope perspective is very time-consuming task for
tendor issuing organization. We are looking for GenAI based solution that can
generate a summary report for all RFP submitted against tendor in terms of financial
efficiency, timeline, legal requirements, qualification criteria, technical parameters
and previous track records.
Challenge / Requirement
proposed system should do following
1. UI should have ability to upload multiple RFP documents against any given tendor
document
2. It should Tabulaize the requirements from tendor and categorize them into legal,
finance, complaince etc
3. It should read content from all RFP from multiple vendors highlighting the
vaguness of the contractual requirements from corresponding RFP
4. It should generate summary comparasion report prefabally in tabular format
5. It should recommend top 3 vendors with justification who has potential to qualify
for contract
Deliverables
working POC
6.AI - Powered Project Management Assistance
Nature of Challenge
Product Related
Domain / Discipline of Challenge
AI (Others)
Description (With Technical Specifications)
Background
Project management involves various tasks, such as planning, organizing, and
monitoring a project from initiation to completion. Project managers often face
challenges, such as managing project timelines, resources, and budgets.
Challenge / Requirement
Develop an AI-powered project management tool that analyzes project timelines,
resource allocation, and risks. It can suggest optimal schedules and identify potential
bottlenecks.
Deliverables
A working prototype that can suggest optimal schedules and identify potential
bottlenecks through chat based interfaces.
Known Bottlenecks / critical areas / Key factors to be addressed
Artificial Intelligence within the chatbot
7.Course recommendation Engine
Nature of Challenge
Product Related
Domain / Discipline of Challenge
AI (Others)
Description (With Technical Specifications)
Background
Currently there is no easy way for a student to figure out what specific courses to
take and the degree to choose that aligns with his/her strength and interests. The
process of picking academic courses is often laborious, time consuming and requires
a lot independent research by the student.
Challenge / Requirement
Create an AI system to analyze a student's profile, academic history, interests and
goals to provide clear recommendations of courses to take and their impact and
value on the student's goal progression.
Deliverables
A prompt-based AI system that will read in data either from text or from documents
to provide personalized course recommendations to each student
Known Bottlenecks / critical areas / Key factors to be addressed
The solution should be highly secure, explainable and accessible from the web or
mobile device.
 Challenges - AI (Software Engineering)
1.AI chatbot for Android Apps
Nature of Challenge
Product Related
Domain / Discipline of Challenge
AI (Software Engineering)
Description (With Technical Specifications)
Background
Users of apps (android or otherwise) may find it challenging to navigate through
different parts of the application to perform different actions. AI Chatbots can be
helpful for the users to navigate and give suggestions about the different aspects of
the application for which the user might not be aware of. The history of the chat can
also be helpful to design future apps.
Challenge / Requirement
Develop a AI based Chatbots solution that can help users to navigate through
different parts of the application and perform different actions, give suggestions
about the different aspects of the application for which the user might not be aware
of. The solution can also demonstrate analysis of the usage of past chat history to
help design future apps/upgrades. The solution should be attempted for any android
app.

Deliverables
A working POC of AI chatbot that can take any android app and engage end user as
well derive analsys based on chat history.

Chat bots must navigate to different parts of the application in a flick of a minute and
without any hassle and can perform different actions based on User requests.
Known Bottlenecks / critical areas / Key factors to be addressed
Artificial Intelligence within the chat bot
2.Testcase generation automation from requirement document
Nature of Challenge
Product Related
Domain / Discipline of Challenge
AI (Software Engineering)
Description (With Technical Specifications)
Background
Software testing is a way of checking if a desirable system works as expected and
meets the requirements. To do this, you need to create test cases, which are specific
scenarios or situations that you use to test the system. Test case generation is one of
the most challenging stages in software testing.In most of the cases, test case
generation involves considerable manual efforts prone to manual error.
Challenge / Requirement
Develop a prototype of a tool/framework that can take requirements document
(Requirement document, Functional, Quality Req Doc et all) and be able to generate
testcases either on its own or with minimum manual effort/intervention.
Deliverables
- A prototype or a demo of the proposed solution, which can be a tool, a technique,
or a framework that can generate test cases automatically or semi-automatically for
a given system.
- A set of test cases that are generated by the proposed solution and by the manual
method, along with their quality and coverage metrics, such as correctness,
completeness, consistency, traceability, etc.
- A presentation that summarizes the main points of the report and the prototype or
demo, and showcases the advantages and limitations of the proposed solution,
productivity improvement compared to the manual method.
Known Bottlenecks / critical areas / Key factors to be addressed
This challenge statement focuses on reducing testing life cycle, effort reduction,
testing time span reduction, cost reduction, reducing human errors etc.
3.API generation using generative AI
Nature of Challenge
Product Related
Domain / Discipline of Challenge
AI (Software Engineering)
Description (With Technical Specifications)
Background
The API ecosystem is currently burgeoning. Every application is now API-based, and
API-driven development is the norm. With the proliferation of APIs, there is a critical
need to develop industry-standard APIs that are designed for the long term and are
easy to access. While API standards and templates exist, no publicly available
material will ensure an API meets all the standards of a good API design. Developers
spend days writing and modifying APIs, and there is no standard way of developing
APIs. The need for more valuable resources to create APIs is exacerbated as
thousands of new developers enter the workforce from academic institutions.
Creating sustainable APIs that follow all standards is thus a critical need for the
software industry. Finally, the challenge is the need for API is fast-paced and APIs
need to be developed at a much higher rate than even a few years back.
Challenge / Requirement
Create a generic LLM that can be trained on various APIs and can provide clear
guidelines and code to create a new API based on the specific prompts. The LLM
should also have the ability to be trained on newer APIs. The outputs should be well-
defined APIs that can be used in any code. Ideally the output APIs should be in the
most common programming languages like Java, Python. Go and Node.js
Deliverables
1. An LLM for generating APIs
2. Well-defined APIs as output following API standards
3. Support for multiple programming languages
4. Ready for implementation of industry-standard APIs
Known Bottlenecks / critical areas / Key factors to be addressed
1. Error-free code and solution
2. All code and models should be developed from scratch and not copied from any
online/offline sources
3. LLM is of manageable size (in terms of parameters, weights and size)
4. The LLM should follow all ethical AI, trustable AI and explainable AI criteria.
5. The LLM should have the ability to learn from new data and inputs.
4.AI assisted software development using LLMs
Nature of Challenge
Product Related
Domain / Discipline of Challenge
AI (Software Engineering)
Description (With Technical Specifications)
Background
In the current process of software development, each stage (architecture, code
development, testing etc.) goes through a completely new start and hence takes
significant time in overall delivery schedule. Using LLM, the development cycle can
be enhanced, enriched and also validated within minimum time.
Challenge / Requirement
LLMs can have broader impact in terms of prompt engineering, close domain LLMs,
private LLMs for code development and testing to tackle the overall problem
statement of software development cycle. The solution referred here will be an AI
assistance to the software developer.
Deliverables
An overall framework architecture diagram along with details about each block of
the development cycle where LLMs can be plugged into. It may also contain the
developed microservices and whole of the code containers to execute all the blocks
of the software development cycle.
Known Bottlenecks / critical areas / Key factors to be addressed
1. Prompt engineering
2. Type of LLMs
3. Accuracy
4. validation
5.LLM based Product design improvement based on stakeholder
feedback
Nature of Challenge
Product Related
Domain / Discipline of Challenge
AI (Software Engineering)
Description (With Technical Specifications)
Background
Customer feedback from various sources, such as surveys, reviews, social media, or
direct communication are analyzed to extract the key insights that may help improve
product/services. Prioritizing of improvements based on impact, feasibility, cost,
product vision and goals needs to be justified with evidence and logic. The
improvised product also needs to be thoroghly tested to avoid any negative side
effects or unintended consequences before product launch.
Challenge / Requirement
1. develop a LLM based pipeline that should read and extract actionable insight
(strength, weakness, customer preferences etc) for any specific product/product
family preferably medical device. multiple open source product review can be used
specific to one product family
2. A generative AI model should generate new product specification considering
customer feedback and competing products to minimize the hallucination.
3. LLM should provide step by step reasoning/justification for any change
recommended by LLM
4. Using A/B testing or causal model to do predict impact on revenue or any other
factors
Deliverables
A LLM Model with interfaces to interact based on new logs that are input
Known Bottlenecks / critical areas / Key factors to be addressed
1. develop a LLM based pipeline that should read and extract actionable insight
(strength, weakness, customer preferences etc) for any specific product/product
family preferably medical device. multiple open source product review can be used
specific to one product family
2. A generative AI model should generate new product specification considering
customer feedback and competing products to minimize the hallucination.
3. LLM should provide step by step reasoning/justification for any change
recommended by LLM
4. using A/B testing or causal model to do predict impact on revenue or any other
factors
6.AI enabled root cause analysis using LLM and DL
Nature of Challenge
Product Related
Domain / Discipline of Challenge
AI (Software Engineering)
Description (With Technical Specifications)
Background
Root cause analysis is one of challenging task considering collecting data from
various sources in various format infer insight from those data and create mind map
for root cause analysis. This specific problem becomes more complex when human
has to read, infer and deduce root cause from machine
Challenge / Requirement
We are looking for a system that must include LLM as component that should
perform following actions.
1. Can read machine generated log like OS log/device log/fault logs
2. it can identify unique pattern and label them in semi-supervised fashion
3. Impact on overall performamce should not be dependent on log format
4. it can generate a mind map based on RCA techniques like 5 whys, Pareto analysis
etc.
5. system should assign probability score for all potential root causes
6. Must include LLM as one of the component and simile UI to demonstrate
Deliverables
Working Prototype.
7.GenAI based end to end regression testing life cycle automation
Nature of Challenge
Product Related
Domain / Discipline of Challenge
AI (Software Engineering)
Description (With Technical Specifications)
Background
Regression testing is one of major time taker in testing lifecycle eventhough with
advent of various automation testing tools, creation and maintenance of automation
script code are still a major challenge, in addition tracing back automation script to
manual test cases then to requirement is cumbersome. in nutshell if we have genAI
based solution if we provide requirement document then can it create all related
artifacts with minimal human validation.
Challenge / Requirement
1. System should learn from existing requirmenent tracability matrix.
2. System should accept requirement documents in various file formats
3. It should generate following artifacts from requirement documents in order Test
Scenario> test case description> test steps> Automation test script > Execution log
analysis report
4. system should have feature of human validation and modification at each steps
5. basic UI to track the progress on each stage
Deliverables
working prototype
8.Automated Code Refactoring
Nature of Challenge
Product Related
Domain / Discipline of Challenge
AI (Software Engineering)
Description (With Technical Specifications)
Background
Code refactoring is the process of restructuring existing computer code—changing
the factoring—without changing its external behavior. Code refactoring helps in
cleaning the code and simplifying it. It makes software easier to understand.
Challenge / Requirement
Develop an AI-driven code refactoring tool that suggests improvements to existing
code. The generative AI can analyze code quality, performance, and best practices to
propose refactoring solutions.
Deliverables
A working POC of the tool that can take a code as input and does code refactoring.
The generative AI can analyze code quality, performance, and best practices to
propose refactoring solutions.
Known Bottlenecks / critical areas / Key factors to be addressed
Artificial Intelligence within the chatbot
9. Library dependencies analyzer tool
Nature of Challenge
Product Related
Domain / Discipline of Challenge
AI (Software Engineering)
Description (With Technical Specifications)
Background
In the Python language ecosystem, there are over 1000s of libraries that available.
The intricate web of dependencies between them often gives rise to significant
challenges within a working environment. Additionally, the process of incorporating
new versions of these libraries invariably demands extra effort from developers.
There is a need for a practical solutions in the form of effective recommendations
centered around compatible library versions that can aid the developers
Challenge / Requirement
An AI based model designed to analyze various libraries listed in the requirements.txt
file. The model checks for compatibility, suggests appropriate library versions, scans
the code, and offers essential recommendations for identifying obsolete code or
adapting to changes in functionality within new versions.
Deliverables
A working prototype of the analyzer with a features to aid developers
Known Bottlenecks / critical areas / Key factors to be addressed
generative ai, code analyzer, library analyzer.
 Challenges - AI (Transportation)
1.Interactive digital twin of Automotive digital cockpit
Nature of Challenge
Product Related
Domain / Discipline of Challenge
AI (Transportation)
Description (With Technical Specifications)
Background
A interactive digital cockpit is the future of automotive interiors. A car can effcetively
be seen as a computer on wheel running millions of lines of software code. Software
informs everything from AD/ADAS to other configurable features. This is what leads
to a software defined vehicle in which the software informs almost everything. A
Software-Defined Vehicle (SDV) is any vehicle that manages its operations, adds
functionality, and enables new features primarily or entirely through software.
Digital twin of a SDV can help in testing and monitoring before and after
deployment.

The digital twin of the HMI interface of Digital Cockpit encompasses the following
features :
 Instrument Cluster
 Driver Cabin Monitoring
 Heads up Display
 Infotainment Systems
Challenge / Requirement
Building a Digital Twin of a Digital Cockpit for a software defined vehicle. The digital
twin of this HMI interface to encompass the following features :

 Instrument Cluster with simulated functionalities

 Driver Cabin Monitoring with simulated functionalities
 Heads up Display with simulated functionalities
  Infotainment Systems with simulated functionalities

The digital twin should include 3D responsive model, exploded view of scene
composition, Simulations and scenarios, AR/VR enablement.
Deliverables
3D modelling & visualization, Simulator for data, Hosting, Graphics UI
2.Extraction of 3D shape key points of objects from a 3D LIDAR
point cloud
Nature of Challenge
Product Related
Domain / Discipline of Challenge
AI (Transportation)
Description (With Technical Specifications)
Background
3D LIDAR is a device mounted on autonomous vehicles to scan area surrounding the
car in a definite range to gather scene information. Obstacle or objects in such a
scanned scene are captured as 3D point cloud and are not obvious to reveal the
information of shape of such obstacles or objects. Extracting such shape information
as 3D key points will form unique signature of the obstacle or object.
Challenge / Requirement
Develop and demonstrate a software algorithms that extracts shape information as
3D key points of objects. These points can be stored in an output file with a suitable
label.

Kitti data set of LIDAR point cloud from Velodyne LIDAR mounted on a travelling
vehicle available in public domain on internet.
Deliverables
Software algorithm and various considerations.
Known Bottlenecks / critical areas / Key factors to be addressed
Algorithms method to be a software code to have steps to extract shape information
3.AI based Traffic sign Detection Improvement
Nature of Challenge
Product Related
Domain / Discipline of Challenge
AI (Transportation)
Description (With Technical Specifications)
Background
Deep learning has enabled traffic sign detection for Autonomous vehicles. In few
scenarios there are limitations with traffic sign detection solutions specially in low
light conditions, weather conditions and in some cases where the traffic sign is
distorted.
Challenge / Requirement
Design and develop the traffic sign detection solution to overcome the challenges
forced by low light conditions, weather conditions, distorted sign conditions etc.
Deliverables
A trained model which can detect traffic signs in various conditions.
Known Bottlenecks / critical areas / Key factors to be addressed
Identify the traffic sign correctly in different conditions few examples rain, low light,
distorted signs.
 Challenges - Digital Manufacturing
1.End to End Digital Twin for Process/manufacturing Plant
Nature of Challenge
Product Related
Domain / Discipline of Challenge
Digital Manufacturing
Description (With Technical Specifications)
Background
Currently Plant Engineering is trying to develop Digital Twin related to Engineering of
New or Existing Process Plant and Integrated Asset Management is similarly doing it
for Running Plant related to Engineering Technology. If both the Digital Twins are
brought together on the same platform where in customer can have both
Engineering technology & Operational Technology viz, Running Plant data from PLCs,
SCADA, Laboratory, Field etc. getting synced with Engineering Technology viz.
Engineering Data and a single output Mathematical Twin is prepared where in client
is able to visualize his running plant and can get health check of the running plant on
a single click on a real time basis to avoid any malfunction.
Challenge / Requirement
Develop a mathematic model that integrates both ET and OT data in a single
platform
Deliverables
The model must generate anomalies and trigger solutions to address anomalies.
Known Bottlenecks / critical areas / Key factors to be addressed
1) All Engineering Data whether in soft or hard to be brought on the single platform
2) All Operational Data of the running plant to be brought on the single platform
3) Mathematical Model run by combining both the data viz. ET & OT and an Analytics
is performed
4) The output of the Analytics Model is brought onto the Dashboard for Visualization
2.An Autonomous Stock Checking Machine
Nature of Challenge
Product Related
Domain / Discipline of Challenge
Digital Manufacturing
Description (With Technical Specifications)
Background
Inventory management is the process of keeping track of how much raw material,
work-in-progress, and finished goods a manufacturer has at any given time. This is
important because it affects the production efficiency, customer satisfaction, and
profitability of the manufacturer. However, inventory management can be very
complex and time-consuming, especially when done manually. That's why some
manufacturers use automated solutions, such as software or robots, to monitor and
control their inventories more accurately and quickly.
Challenge / Requirement
Design and implement an automated solution that can help a manufacturer keep
track of their inventory levels, optimize their production efficiency, and increase their
customer satisfaction and profitability
Deliverables
_A prototype of an automated inventory management system that can monitor and
control the inventory levels of A manufacturer using sensors, software, and/or
robots.
- A report that explains the design and implementation of the system, the benefits
and challenges of using it, and the testing and evaluation results.
- A presentation that showcases the system and its features, and demonstrates its
functionality and performance.

3.Large scale system control using Digital Twin

Nature of Challenge
Product Related
Domain / Discipline of Challenge
Digital Manufacturing
Description (With Technical Specifications)
Background
Remote operations in a high-risk environment takes good amount of control system
to be involved. Increasing the usability and scalability of such remote operation is
possible using Digital Twins.
Challenge / Requirement
Create a digital twin of an upstream oil well using 3D CAD modeling, IOT sensors, and
VR visualization. First, a 3D CAD model of the oil well is created or imported from an
existing source. Second, a set of IOT sensors are attached to the physical model to
measure and transmit relevant parameters, such as pressure, temperature, and flow
rate. Third, a digital twin pipeline is established to connect the sensor data to the 3D
CAD model and update it in real time. Fourth, the data is visualized in a 3D
environment through a VR device, allowing the user to interact with the digital twin
and monitor its performance. Fifth, a feedback data pipeline is created, which
enables the user to control the transmission of data from the IOT sensors by
triggering it from the 3D model visualization. This way, the user can switch on and off
the data flow as needed.

Deliverables
Demonstrate the methodology with a POC.
Known Bottlenecks / critical areas / Key factors to be addressed
Digital Twin, IOT, Could.
4.Automation of engineering process
Nature of Challenge
Product Related
Domain / Discipline of Challenge
Digital Manufacturing
Description (With Technical Specifications)
Background
Plant engineering involves design and construction of industrial plants, such as power
plants, chemical plants, or oil refineries. Some of the tasks that plant engineers are
involved includes
- Creating layouts, which are drawings that show the arrangement and dimensions of
the equipment and piping in a plant.
- Creating 3D models, which are digital representations of the plant that can be
viewed from different angles and perspectives.
- Creating material requisitions, which are requests for purchasing or procuring the
materials for a project.
- Creating datasheets, which are documents that provide technical specifications and
performance data of the equipment and components in a plant.
- Creating lists, schedules, bill of materials (BOM), or material take-offs (MTO), which
are documents that specify the quantity and type of materials needed for a project.
These tasks can be time-consuming and tedious, especially if there are many changes
or revisions in the design
Challenge / Requirement
Develop a prototype of an AI based automation tool and algorithms that can
generate plant engineering deliverables automatically from various inputs. The
inputs can be text-based documents, such as design criteria, process flow diagrams,
or piping and instrumentation diagrams. The inputs can also be drawings in native or
PDF formats, such as equipment drawings, plot plans, or isometric drawings. The
automation tool should be able to use all these inputs to generate the deliverables in
the required formats, such as Excel, Word, PDF, or CAD files.
Deliverables
Demonstrate the algorithm

5.Human Occupational Safety within a confined working area

Nature of Challenge
Product Related
Domain / Discipline of Challenge
Digital Manufacturing
Description (With Technical Specifications)
Background
Occupational Health and Safety (OHS) focuses primarily on protecting employees in
the workplace from accidents, injuries, and exposure to harmful substances. There
can be cases where human's knowingly or unknowingly may enter restricted zone
that can be violation of safety regulation and be hazardeous to health.
To address this challenge, safety engineers need to design and implement systems
that can sense human presence in restricted areas, notify the relevant authorities or
personnel, and take necessary actions to stop or reduce the operation of the
equipment that poses a threat. This way, safety can be achieved by preventing or
minimizing the potential harm to the human.
Challenge / Requirement
Develop an end to end system to detect human presence and detect the if any of the
body parts (hand, foot, arm, leg, head, crosses the defined boundary zone.
1)Are of coverage is 3m * 4m * 2.5m (L*W*H)
2)Detection of human part (typical size can be 3cm * 3cm)
3)Response time to notify the detection of human out of defined boundary is less
than 1 sec
4)Notify which part of the body is crossed the boundary
5)Accuracy of detection should be 100% and and classification of body part should be
>95%
6)It should work in any dark, daylight conditions.
Deliverables
Working prototype to meet the defined use cases.
Known Bottlenecks / critical areas / Key factors to be addressed
Notify the event occurrence, classify the detected human part which entered outside
the defined zone.
1)Are of coverage is 3m * 4m
2)Detection of human part (typical size can be 3cm * 3cm)
3)Response time to notify the detection of human out of defined boundary is less
than 1 sec
4)Notify which part of the body is crossed the boundary
5)Accuracy of detection should be 100% and and classification of body part should be
>95%
6)It should work in any dark, daylight conditions.