A score to settle . . . .
Autonomous Rocket League Robot
The robot was designed and built for the second phase of the mechatronics and robotics course MCEN5115. Over the course of ten weeks, I and four other teammates implemented the mechanical, electrical, and software knowledge learned during the course to design a robot to compete in a competition against various other teams in the class.
As the use of robotics continues to grow in industry and research, so has the need for autonomous robotics that can complete a wide range of tasks. One area of growing interest is the ability of a robot to track objects autonomously while navigating in an environment and avoiding obstacles. This has a variety of applications such as performing rescue operations in hazardous environments, delivering materials to remote areas, and assisting in military services.
In this project, we explore this area of research by designing a robot tasked with competing against other robots in a real-life version of the video game Rocket League, in which players control cars in a game similar to soccer.
The key challenge is to detect the ball and maneuver it into the opponent’s goal net, either by pushing it or kicking it forward, while also obstructing the opponent robot.
The robot can detect its surrounding environment in a variety of ways, including image processing, radio signal, and/or sonic detection. This task allows for the team to test different potential approaches and compare them to each other through the competition.
Soccer Robot
Design Overview :
The Algorithm behind Soccer Robot :
Path Planning:
Using Camera To Chase Ball:
Arduino Mega Code #1:
Arduino Mega Code #2:
Then comes the Testing Phase . . . .
Calibrating the Pixy Camera
Testing the Mecanum Wheel motion
Conclusion:
Robots that focused on the camera detection for finding and chasing the ball performed the best overall. Our own robot focused on a path-finding algorithm using the coordinates of the objects on the field.
When it was close enough to the ball, it would switch to the camera and use image processing to locate the ball. Our robot performed well when the ball was directly between it and the opponent's goal, but would get lost otherwise. This may be because our gyroscope wouldn’t calibrate correctly, resulting in the robot being unable to detect its location accurately. Additionally, when it came to our image processing, our robot had difficulty fully distinguishing the ball from the surrounding. A more accurate image processing program, working simultaneously with the path-finding algorithm, would allow the robot to maneuver the ball and aim it towards the goal more precisely.
