Motion Planning

Advisor: Dr. Yiqiang Han, Dr. Ardalan Vahid,

• Develop an iterative approach to optimize vehicle pose trajectory based on the balance between minimum distance and minimum curvature.

• Use dynamic programming technique to generate optimal velocity profile for the optimized vehicle pose trajectory with the constraint of maximum tire-force

' Reduce Global Hybrid A*'s searching time by controlling sampling distances and sampling steps. Enhance robot ability to navigate in partially-unknown with lower computation needed. '

• Perform detailed analysis on the effect of altering search distances and sample steps of global hybrid A* algorithm.

• Compare the new algorithm's result with D* Lite, A* and Hybrid A* in both holonomic and non-holonomic motion constraints scenario.

• Tested on Unmanned Aerial Drone and Autonomous RC car.

• Language, package and platform used in this project: ROS, Python, MATLAB

' This project is a partial fulfillment of CPSC-8810 Motion Planning Course in Clemson University. The main purpose of this project is to test two types of velocity-based collision avoidance algorithms (velocity-obstacle and power-law model). '

• Build velocity obstacle and power-law model in python.

• Test both 8-agents and multiple agents case with different start/goal states

• Language, package and platform used in this project: Python, Numpy, Tkinter