Path Planning

2D Potential Guided Rapidly-Exploring Random Trees (P-RRT) Planner

RoboCup Standard Platform League Field. Image by RoboCup Federation.

Abstract

The goal of this project was to implement a 2D potential guided rapidly-exploring random trees (P-RRT) planning algorithm on NAO Robots from SoftBank Robotics in simulation. These robots are used in the RoboCup Standard Platform League, which is a robotic soccer league. The planner and environment were simulated in MATLAB. Five different experiments were created to test the planner’s performance with stationary and moving target positions and obstacles. The P-RRT planner’s performance was also compared to the performance of using only a potential field planner and a 2D RRT planner in appropriate situations. The path planner maneuvered an individual robot within the boundaries of a miniature soccer field, avoided collisions with dynamic teammates and opponents, and required low computational ability. The P-RRT planner was also able to successfully navigate a general path planning problem environment.

Videos showing the planner in four different environments are located below.

Full Report Found Here.

Note: Project was conducted with Zachary Fisher.

2D P-RRT Planner Result in RoboCup Style Static Environment

2D P-RRT Planner in a RoboCup Style Environment. The environment has 9 randomly generated dynamic obstacles, which represent the other 9 robots present on the field during a RoboCup soccer match.

2D P-RRT Planner in a Challenging Static Environment. The environment has 19 randomly generated static obstacles, which represent the 19 other players present on the field during a human soccer match, excluding goalies.

2D P-RRT Planner in a Challenging Dynamic Environment. The environment has 19 randomly generated dynamic obstacles, which represent the 19 other players present on the field during a human soccer match, excluding goalies.

2D P-RRT Planner in a Dynamic Environment with a Fixed Local Minimum. The environment has 9 randomly generated dynamic obstacles and 1 static block obstacle.

Page updated

Google Sites

Report abuse