User preference optimization for control of ankle exoskeletons using sample efficient active learning | Science Robotics

The proposed algorithm reliably identifies multiple users’ preferred settings of an exoskeleton controller in real time.