Automatic Adaptation of an Humanoid Robot Gait to Different Floor-Robot Friction Coefficients
This project intends to develop a system capable of real time measuring different floor-robot friction coefficients in a humanoid robot. This humanoid robot must be capable of walk through several types of floors with different friction coefficients, automatically adapting its gait.
The use of instrumented shoes will allow the measurament of the floor-feet friction coefficent that will determinate the robot gait, the center of pressure (CoP) trajectory reference and the stability controller.
An haptic system will be used to train the balance controller acquiring real time feedback of the robot's CoP error to the reference.
Reinforcement learning will also be considered for the automatic adaptation of the robot gait to changes in the floor-robot friction conditions.
The instrumented shoes will also be used to study human gait disorders, allowing gait pathologies identification and their severity quantification.
The A2RH project(PTDC/EEI-AUT/5141/2014) has started on July 2016 and is scheduled to end by June 2019.