Combined Path Planning and Locomotion for Hexapod robots in cluttered environments using Deep RL
Avideh Zakhor, Professor
Electrical Engineering and Computer Science
Applications for Spring 2024 are closed for this project.
In this project, we will develop RL policies to teach a hexapod robot navigate in cluttered environments. We will initially develop separate policies for climbing joists, stairs, and squeezing under objects. Then we will distill and combine them using hierarchical RL. This work builds on top of existing work which demonstrates successful use of RL for joist climbing by the hexapod.
Qualifications: * Develop novel RL algorithms for locomotion using IsascGym simulation engine
* Test and demonstrate RL algorithms in actual hardware
* Develop path planning and navigation algorithms to enable the robot to choose the 'best' path/ strategy to get from one point to another point in cluttered environments.
Hours: 12 or more hours
Related website: http://
Engineering, Design & Technologies