Evan Krell & Arun Prassanth R.B.
This project explores using particle swarm optimization (PSO) for mobile robot path planning. In this system, a simulated Turtlebot is able to generate a map of the environment which it can then use to generate a path from its current position to a user-specified target position. A very high-level overview of the system flow is summarized in the System Flow diagram, below.
The Gazebo simulator is used for the Turtlebot and its environment. A virtual machine with the environment used is available here. The Turtlebot is controlled through a MATLAB script on a Windows computer. Communication is handled by ROS, which is supported by MATLAB's Robotics System Toolbox. Path planning is done on a separate Linux server. The MATLAB script coordinates the Turtlebot's activity, sends generated occupancy maps to the path planning server, and uses the solution paths to direct waypoint following. This is summarized in the System Overview diagram, below.
Below are Gazebo configurations, and a map generated by the Turtlebot. On the map is a solution path generated by PSO.
Setup 1: "Doorways"
Solution Path in "Doorways"
Setup 2: "Open Room"
Solution Path in "Open Room"
Mapping
Mapping an unknown environment
Path Following
Following a path to reach blue ball