Class Project Overview
6.141 is a capstone engineering class exploring autonomous driving through comprehensive labs and a final project consisting of a race and obstacle avoidance challenge. The class used the Tesse driving simulator integrated with ROS and python to control the car. The class was team-based for almost all the labs and the final challenge; my team had five people.
My Work
The main content of the class comprised six comprehensive labs that explored essential ROS functions, a simple wall follower, a more complex wall follower in Tesse, line following and cone detection, Monte Carlo localization, and path planning algorithms. The final challenge combined all the content from the labs, and every team raced each other and a staff solution. Our team finished the race first and was 0.2 seconds faster than the staff benchmark using a finely tuned pure pursuit controller and pre-planned trajectory. We then used a modified pure pursuit approach combined with object detection to complete the obstacle avoidance course. For more information on the scope of the final project, here is our final presentation, and here is our final report.
What I Learned
6.141 taught me about ROS architecture, autonomy, path planning algorithms, localization methods, teamwork, and controls. Through the labs, we learned about computer vision, color segmentation, edge/line detection, MCL, particle filters, pure pursuit control, path planning algorithms like rrt, rrt*, A*, and obstacle avoidance. This class was yet another experience working in a team under high-stress conditions and not only completing the class objectives but excelling and winning the race.
6.141 is a capstone engineering class exploring autonomous driving through comprehensive labs and a final project consisting of a race and obstacle avoidance challenge. The class used the Tesse driving simulator integrated with ROS and python to control the car. The class was team-based for almost all the labs and the final challenge; my team had five people.
My Work
The main content of the class comprised six comprehensive labs that explored essential ROS functions, a simple wall follower, a more complex wall follower in Tesse, line following and cone detection, Monte Carlo localization, and path planning algorithms. The final challenge combined all the content from the labs, and every team raced each other and a staff solution. Our team finished the race first and was 0.2 seconds faster than the staff benchmark using a finely tuned pure pursuit controller and pre-planned trajectory. We then used a modified pure pursuit approach combined with object detection to complete the obstacle avoidance course. For more information on the scope of the final project, here is our final presentation, and here is our final report.
What I Learned
6.141 taught me about ROS architecture, autonomy, path planning algorithms, localization methods, teamwork, and controls. Through the labs, we learned about computer vision, color segmentation, edge/line detection, MCL, particle filters, pure pursuit control, path planning algorithms like rrt, rrt*, A*, and obstacle avoidance. This class was yet another experience working in a team under high-stress conditions and not only completing the class objectives but excelling and winning the race.
