Default Course Project of CS4278/CS5478 Intelligent Robots: Algorithms and Systems
In this project, the task is to:
- Navigate the Mobot to the drawer
- Pick up the mug
- Place it into the drawer
We provide an example control code for the robot in simulation/main.py.
You should implement the navigation and motion planning algorithms by yourself to accomplish the task.
When the total score of the project is 100 points:
You should navigate the robot to go from the starting point to any place after the narrow passage at the right room. When you arrive the goal region, the main.py function will print a successful signal. You can rewrite the success judgment function as long as the target region is kept the same.
You will get corresponding points based on your total navigation driving distance (in meter):
45% for total distance ≤ 18; 40% for 18 < total distance ≤ 22; 30% for 22 < total distance ≤ 25; 20% for 25 < total distance ≤ 28; 10% for 28 < total distance ≤ 32; 0% for total distance > 32.
you should provide at least three successful navigation videos to show the robustness of your algorithm.
You should implement a motion planning algorithm for the provided stretch robot arm end-effector to go to any given point in the task space, including:
path planning (10%); collision avoidance (10%); inverse kinematics calculation (10%).
You should provide a video to show that your robot arm (robot link 18) can reach these tree given positions in the task space. You can check the success with the motion_planning_test() function in utils\tools.py:
[0.27, -0.71, 0.92] [-1.70, -3.70, 0.46] [1.45, -1.68, 0.59]
please provide the success videos of the arm reaching these positions successfully.
Successfully pick and place the target mug into the drawer. When your algorithm achieves this, the main.py function will print a successful signal. You can rewrite the success judgment function as long as the target region is kept the same.
10%: if only pick up the mug; 15%: pick and place the mug.
You should provide at least three successful pick and place (or only pick) videos to show the robustness of your algorithms.
Clear and readable project report (4%) Runable codebase (4%) Clear video results (2%)
5%: if you implement trajectory planning for the robot arm (including inverse dynamics)
5% if you can pick and place a randomly initialized mug. You can start random initialization of the mug at here. You should provide at least three successful videos for this part.
2024/11/01: add rubic and helper functions
2024/10/25: add attach and detach functions for grasping in simulation/main.py. Now you can use this to grasp and release the mug. Try it!
2024/10/24: add useful tools in utils/tools.py.
2024/10/22: fix the gravity coefficient; add useful tools in utils/tools.py.
2024/10/21: add keyboard control example code in simulation/main.py.
-
Create a new Anaconda environment:
conda create -n mm python=3.8 -
conda activate mm -
pip3 install pybullet numpy -
git clone https://github.com/NUS-LinS-Lab/Mobile-Manipulation.git
python simulation/main.py
PyBullet Documentation: https://pybullet.org/wordpress/index.php/forum-2/.
You can also use ChatGPT or Cursor to get the API of PyBullet.