Motion Planning for a Humanoid Robot Arm

In this project, I implemented multiple sampling based motion planning algorithms (RRT, RRT*, Bidirectional RTT) for a humanoid robot to perform grasping tasks, and compared the performances of different methods. The red path is the raw path from the planer, and the blue path is the optimized and smoothed path. The main function is implemented in Python, while the planning algorithms are implemented in C++. The python main function calls the C++ functions through the Openrave C++ plugins.

Required packages:

• Openrave
• Python 2.7
• Cmake
• Openrave C++ plugins

Instruction

To use the RRT method, run

 python rrt.py

To use the RRT* method, run

 python rrtstar.py

To use the Bidirectional RRT method, run

 python birrt.py