Review: Lab course: Robotik und visuelle Sensorik

It was a high intensity semester, but it was definitely worth it. We all had a lot of fun and the students developed really neat and working versions of their ideas.

Kinect controlled humanoid robot

By capturing depth-images from a kinect camera and accquiring skeleton data from the NITE framework, angle information and coordinates were sent via TCP to a robot. This way it was possible to have robot replay the movements of someone standing in front of the camera in realtime.

Humanoid robot kick based on bezier curves

Our “old” kick did not support aiming, dynamic balancing or cancellation in case the ball was moved after the kick routine was started. It was really static, so one team of our students implemented a kick which is split into several phases, from strike-out to kick, every foot-movement based on 3d bezier curves creating smooth movements.

Humanoid robot picking up and throwing a ball

In robot football it is not common for the goalkeepers to use their hands. This project measured stress and temperature on the servo motors, controlled the center of mass and managed to pick up and throw a ball with our humanoid robot.

Custom walk implementation for humanoid robot

We are using the standard walk which is builtin in the robot middleware right now. There are several teams that implemented their own custom walk, based on inverse pendulum equations. This project tried to implement a custom walk based on these formulas, too.