Summer Project for CATS Lab (UW-Madison)
https://catslab.engr.wisc.edu/
Developed closed-loop control system for a modular autonomous robot platform, enabling it to detect lanes and navigate within them autonomously.
Designed Real-Time Lane, Cornet Detection, and Depth Calculation Algorithms using OpenCV:
- Applied Hough transformation to identify left and right lane boundaries (green & blue line)
- Developed a custom algorithm to generate driving line (yellow line) and recognize corner types (sharp turn, smooth turn)
- Increased accuracy by pre-processing images with reverse-perspective and Gaussian blur
- Designed an average depth calculation algorithm to detect obstacles in the robot's path, enabling effective obstacle avoidance and safer navigation
Generated Control Signals Using ROS2 Platform:
- Designed and implemented a PID control system to maintain precise control over the robot’s movements
- Programmed adaptive driving modes based on corner types, utilizing Ackermann steering for smooth turns and spin mode for sharp turns
- Integrated CAN communication protocol for reliable communication with the robot