Controls

Neuro-Adaptive Controller (NAC)

Uses Neural Networks (NNs) to estimate robot system dynamics.
Used for trajectory tracking.
Adapts to change in any robot dynamics.
Proved to be stable in the sense of Lyapunov.
It is model-free due to which it can be implemented on any robot without prior training.

Base-Sensor Assisted Physical Interaction (BAPI) Controller for Collision Detection

Uses a Force-Torque Sensor at the base of the arm to sense interaction forces.
NN used to identify wrench due to interaction forces and separate from wrench due to robot motion.
The algorithm was able to clearly identify collisions along the entire robot body.
Eliminates the need of wrapping sensor arrays around the robot body to sense forces.

Base-Sensor Assisted Physical Interaction (BAPI) Controller for Joint Torque Estimation

Uses BAPI controller to estimate joint torques of a robot arm.
Uses dynamic estimation from NAC to predict the joint torque during arm movement.
It can be used to estimate user interaction forces during pHRI.

Google Sites

Report abuse