SoftManiSim is an advanced, open-source simulation framework designed for multi-segment continuum manipulators, offering a significant leap forward in accuracy and flexibility. Unlike traditional simulators, which often rely on simplifying assumptions like constant curvature bending and omission of contact forces, SoftManiSim employs a robust and rapid mathematical model to achieve precise, real-time simulations. Its versatility allows seamless integration with various rigid-body robots, making it applicable across a broad range of robotic platforms. Additionally, SoftManiSim supports parallel operations, enabling the simultaneous simulation of multiple robots and generating essential synthetic data for training deep reinforcement learning models.
git clone [email protected]:MohammadKasaei/SoftManiSim.git
cd SoftManiSim
python3 -m venv env
source env/bin/activate # On Windows use `env\Scripts\activate`
Before running the script, make sure you have execute permissions. Run the following command:
chmod +x install_dependencies.sh
To install all the dependencies, simply run:
./install_dependencies.sh
Wait for the script to complete. Once done, all the required dependencies should be installed in your environment.
The scripts folder contains several example scripts to help you get started. You can run any of these scripts using the following command:
python3 -m scripts.SCRIPT_NAMEFor example, to run the BasicTest_manipulator_two_robot, use:
python3 -m scripts.BasicTest_manipulator_two_robotAdditionally, we have developed a set of Gym Environment, for more detail check the SoftManipulatorEnv/SoftManipulatorEnv folder.
Below is the API documentation for the SoftManiSim class, detailing its essential methods, their arguments, and functionalities:
| Method | Argument | Description |
|---|---|---|
__init__ |
bullet |
Optional physics engine instance, defaults to None, initializes PyBullet if not provided. |
body_color |
RGBA color for the robot's body. | |
head_color |
RGBA color for the robot's head. | |
body_sphere_radius |
Radius of spheres used to build the robot’s body. | |
number_of_sphere |
Number of spheres constructing the robot’s body. | |
number_of_segment |
Number of segments in the robot's body. | |
gui |
Boolean to toggle the graphical interface, defaults to True. |
|
create_robot |
- | No arguments, sets up the robot's physical structure within the simulation. This function is invoked at the end of the constructor. |
move_robot_ori |
action |
Array of actions defining movement commands for robot segments. |
base_pos |
The base position of the robot in the simulation space. | |
base_orin |
The base orientation of the robot, specified as Euler angles. | |
camera_marker |
Boolean to display camera markers, defaults to True. |
|
calc_tip_pos |
action |
Array of actions affecting the tip’s position and orientation. |
base_pos |
The base position from which the tip's calculations start. | |
base_orin |
Base orientation affecting the tip’s calculation. | |
capture_image |
removeBackground |
Boolean to decide whether to remove the background from the image, defaults to False. |
in_hand_camera_capture_image |
- | No arguments, captures an image from the robot’s in-hand camera. |
is_robot_in_contact |
obj_id |
Object ID to check for contact with the robot. |
is_gripper_in_contact |
obj_id |
Object ID to check for contact with the robot's gripper. |
suction_grasp |
enable |
Boolean to enable or disable the suction grasp mechanism. |
set_grasp_width |
grasp_width_percent |
Percentage of maximum grasp width to set for the gripper. |
add_a_cube |
pos |
Position to place the cube in the simulation. |
ori |
Orientation of the cube, given as a quaternion. | |
size |
Dimensions of the cube. | |
mass |
Mass of the cube. | |
color |
RGBA color of the cube. | |
textureUniqueId |
Optional texture ID for the cube's surface. | |
wait |
sec |
Duration in seconds to delay the simulation. |
This table summarizes the main methods and their functionalities for users who want to integrate or work with the SoftManiSim class in their projects.
If you find our paper or this repository helpful, please cite our work:
@inproceedings{kasaeisoftmanisim,
title={SoftManiSim: A Fast Simulation Framework for Multi-Segment Continuum Manipulators Tailored for Robot Learning},
author={Kasaei, Mohammadreza and Kasaei, Hamidreza and Khadem, Mohsen},
booktitle={8th Annual Conference on Robot Learning}
}
This project is licensed under the MIT License.
This work is supported by the Medical Research Council [MR/T023252/1].