Almost all legged robots depend on sensors, such as cameras, to move. When these robots run, the motion of the robot introduces noise into the data, making it hard to interpret. This project focuses on designing and creating a stabilization system to reduce motion of the sensors on the robots, thereby returning more useful and helpful data.
The baseline of motion data for this project is from the Minitaur robot. Data collected from the Minitaur shows how much motion is acting on the sensors. Our project design involves a mechanical system that can attach to different platforms independently of the robot itself. It uses four servo motors to counteract the motion and steady the camera. Software and hardware actively respond to the robot motion and compensate with motors for sensor balance.
The project system’s frame is 3D printed with cutouts in the linkages to reduce weight and connect the motors and the camera. This results in a lightweight product that also improves data clarity without compromising the operation or effectiveness of the robot. A camera serves as the main sensor, so video from the camera shows how the project improves data. Overall, the project can be applied to different platforms and improves sensor feedback.