Trajectory optimization techniques for closed-chain motion planning
- If you are interested in the proposal, please contact with the supervisors.
One of the most fundamental tasks in robotics is that of path planning: to find a collision-free trajectory to move the robot from one point to another. This problem is particularly challenging if the robots involved in the problem form one or more closed chains. This happens, for instance, in service robots that use two arms to manipulate objects (see the Baxter robot, for instance). In many situations, the fastest way to solve the problem is to determine a feasible trajectory ignoring the obstacles and latter on optimize this trajectory to avoid the collisions. However, the existing optimization methods for trajectories can not deal with closed chains.
The objective of this project is to adapt existing trajectory optimization techniques to the closed-chain case.
This project entails:
- Adapt the CHOMP and/or STOMP strategies to closed-chain case.
- Implement the strategies and integrate them in the CuikSuite.
- Perform experiments in simulation and possibly with real robots.
This project requires
- A good background of mathematical methods.
- Good C/C++ programming skill.