MIT Autonomous Vehicle Project

Overview

4 week program at MIT's Beaver Works Summer Institute where I worked hands-on with autonomous vehicles
Six-month online course prerequisite to qualify for the program, covering the basics of control algorithms, Gazebo Simulations, motion planning, OpenCV in Python, and the Robotic Operating System (ROS)
During the 6 week summer course, specific control algorithms like proportional-integral-derivative controllers, ways to use Gazebo to test the car offsite, global motion planning systems, HSV filtering and Canny Edge Detection, and how to send and receive data using ROS were also covered
Worked on algorithms that:
Acted as project manager and facilitated the organization of tasks in our group and verified that everyone understood what they had to complete, keeping everyone on task
Final two weeks involved a course with solid and mesh walls and obstacles, colored paths that had to be followed, and AR tags that had to be read
Worked with another teammate to create a yellow road and AR follower while the other half of the team worked on a potential field controller and AR logic controller
End result was that vehicle that could go through the obstacle course and race against other cars

RPReplay_Final1577412093.mov

Oversaw the design and programming of visualization processing programs using HSV, shape comparison, and a proportional–integral–derivative (PID) controller to control the vehicle
Algorithm determined center of track relative to the car and what adjustment needed to be made to keep the vehicle in the center of the track
Achieved 85% accuracy after multiple rounds of fine-tuning