Amity School of Engineering & Technology
Session 3:
Localization and Map Making
By
Dr. Ashok Kumar
Amity School of Engineering & Technology
Introduction to Localization and Map Making
•Definition of Localization: The process of determining the robot’s position
within an environment.
•Definition of Map Making (Mapping): The creation of a map that
represents the robot’s environment.
•Importance: Essential for autonomous navigation, allowing robots to
understand and interact with their surroundings.
Amity School of Engineering & Technology
Key Components of Localization and Map Making
1.Sensors: Devices like LIDAR, cameras, ultrasonic sensors, and IMUs
(Inertial Measurement Units).
2.Algorithms: Techniques like Simultaneous Localization and Mapping
(SLAM), Kalman Filters, and particle filters.
3.Mapping Process: Capturing environmental data and translating it into a
digital map.
4.Localization Process: Estimating the robot’s position using the map.
Amity School of Engineering & Technology
Types of Maps in Robotics
1.Grid Maps: Divide the environment into cells, where each cell contains
information about occupancy.
2.Feature-based Maps: Map key features (e.g., walls, objects) in the
environment.
3.Topological Maps: Represent environments as graphs, where nodes
represent locations and edges represent paths.
4.Occupancy Grid Maps: Each cell represents the probability of
occupancy.
Amity School of Engineering & Technology
Localization vs. Mapping
•Localization:
• Focuses on determining the robot’s position in a known or evolving
map.
• Can be done using GPS (outdoor) or sensors like LIDAR and cameras
(indoor).
•Mapping:
• Involves creating or updating a map of the environment while the robot
explores.
• Can be performed using a variety of sensors and algorithms like SLAM.
Amity School of Engineering & Technology
Simultaneous Localization and Mapping (SLAM)
•Definition: The process by which a robot builds a map of an unknown
environment while simultaneously keeping track of its location within that
environment.
Amity School of Engineering & Technology
SLAM Process Overview
1.Sensor Data Collection: Gathering data from sensors like LIDAR,
cameras, or sonars.
2.Feature Extraction: Identifying features in the environment such as
landmarks, walls, or obstacles.
3.Localization: Estimating the robot’s position in relation to the map using
sensor data.
4.Mapping: Updating the map with new data, while adjusting the robot’s
position estimates.
5.Optimization: Refining the map and localization estimate using techniques
like least squares or graph optimization.
Amity School of Engineering & Technology
Localization Techniques
1.Dead Reckoning:
1. A method of calculating position based on previous positions and
movements.
2.GPS-based Localization:
1. Used outdoors for global localization using satellite signals.
3.Visual Localization (VSLAM):
1. Uses camera data to track the robot’s position based on visual features.
4.LIDAR-based Localization:
1. Uses LIDAR to scan the environment and match observed data to a pre-
built map for localization.
Amity School of Engineering & Technology
Map Making Methods
1.Occupancy Grid Mapping:
1. Divides the environment into a grid, where each cell represents
whether the area is free, occupied, or unknown.
2.Feature-based Mapping:
1. Captures specific features in the environment (e.g., corners, edges,
landmarks).
3.3D Mapping:
1. Generates 3D models of environments using LIDAR, depth cameras,
or stereo cameras.
4.Semantic Mapping:
1. Involves labeling objects or regions in the map (e.g., “door,” “table”)
for higher-level reasoning.
Amity School of Engineering & Technology
Challenges in Localization and Map Making
1.Sensor Noise: Imperfect sensor data, requiring filtering and error
correction.
2.Environmental Changes: Dynamic environments can change, requiring
updates to the map and localization process.
3.Computational Load: Complex maps (e.g., 3D or large-scale) require
significant processing power.
4.Data Association: Identifying which features correspond to which parts of
the map or sensor readings.
Amity School of Engineering & Technology
Performance Metrics in Localization and Mapping
1.Accuracy:
1. How close the estimated position is to the true position.
2.Robustness:
1. The ability to maintain performance in noisy or dynamic environments.
3.Computational Efficiency:
1. The time and computational resources required to generate the map and
localize.
4.Scalability:
1. The ability to scale up to larger environments or more complex maps.
