Multi-agent Trajectory Planning

Time

May 2017 – July 2017

Summary

During this project, I work with Prof. Dimitra Panagou and Ph.D Kunal Garg to develop a simulation environment for multi-agent trajectory planning. This simulation environment can be generalized to different maps including buildings, trees, cars, people, etc., with different aircraft like quadrotors, airplanes, etc.

I use ROS (robotics operating system), gazebo, Matlab, c++ to interact with each other and create a visualized UI in gazebo to simulate the trajectory of multi agents. In detail, I use c++ to receive position of the quadrotors from ROS and send it to Matlab. Since Matlab is efficient in calculation, I use Matlab to calculate the speed of quadrotors based on the received postion using the multi-agent trajectory planning algorithm. Then Matlab sends speed back to c++ and c++ eventually sends speed back to ROS. This working system can be applied to many Multi-agent Trajectory Planning algorithms as long as they calculate the speed based on positions.

What’s more, to make the simulation environment more realistic, I add many different obstacles in the gazebo simulation environment like buildings, trees, cars and so on. The position and dimension of these obstacles can be adjusted easily by only changing a few variables. Moreover, with the help of SketchUp, I import many online models to gazebo which simulates a much more realistic world. In the meantime, in addition to the quadrotors models, I also create a Boeing-747 airplane model for those algorithms designed for airplanes.

Architecture

屏幕快照 2017-11-10 21.19.17.png

Demo Video

Quadrotors

Airplanes

 

Document

Click here to see detailed project document.

Related Code

You can contact me if you want to see the actual implementation. I can add you to the private Github repo.