Autonomous RoboCar

Lead a team of Mechanical and Electrical Engineers through a multistage robotics competition. Devised hardware support for system integration, camera, GPS, and LiDAR. Worked on Linux/python code integration and improving vehicle performance by optimizing PID values. Given our efforts our car placed first place in a GPS navigation competition at UCSD.

For our final project we proposed to make a car that follows a laser pointer on the ground while simultaneously mapping its GPS coordinates to a local file. I developed a detection algorithm using OpenCV and basic image mask thresholding. This allowed the robot to isolate the laser pointer dot even on a busy or noisy background. The pixel location of the isolated laser dot was used to map the throttle and steering response accurately. The x-coordinate is mapped to left-right steering and the y-coordinate is mapped to throttle control. Both nodes are controlled pyVesc. GPS logging was accomplished with an open source GNSS library. We also incorporated an external speaker that plays cat sounds when the laser pointer is not visible in the camera frame.

Skills learned: fast prototyping, project management, traditional programming, and computer vision using Robot Operating System (ROS) and artificial intelligence deep learning