The students in SE423: Mechatronics were tasked with designing a robot to navigate through an obstacle course, while collecting and depositing golf balls. Our teams strategy was to iteratively improve our robot by making it faster, more efficient in navigating courses, and forcing it to avoid collisions.
Here is an example of the path of the robot, with obstacles, and blue and orange golf balls:
The first iteration developed for this task used LADAR sensors to detect obstacles and walls which allowed the robot itself through the maze. Additionally, the front-mounted Open MV camera was tuned and calibrated to sense purple and orange golf balls located inside of the course.
The second iteration combined LADAR sensors with the internal gyro of the robot, to map out the locations of obstacles and golf balls. The Kalman filter was applied to improve performance of the robots location detecting. This proved to be unstable, as the robot was frequently hitting walls. At this stage, some logic was implemented to force several RC Servos mounted under the robot to grab detected golf balls and place them into chutes.
The final iteration used the Astar algorithm to map the most efficient course through the path. Astar was used in conjunction with the opti-track camera system for precise location tracking. This final iteration allowed the robot to avoid collisions while accurately detecting golf balls.
Here is a demonstration of the final iteration:
The team:
Thank you!
Comments