Rewards-driven control of robot arm by decoding EEG signals.

Decoding the user intention from non-invasive EEG signals is a challenging problem. In this paper, we study the feasibility of predicting the goal for controlling the robot arm in self-paced reaching movements, i.e., spontaneous movements that do not require an external cue. Our proposed system continuously estimates the goal throughout a trial starting before the movement onset by online classification and generates optimal trajectories for driving the robot arm to the estimated goal. Experiments using EEG signals of one healthy subject (right arm) yield smooth reaching movements of the simulated 7 degrees of freedom KUKA robot arm in planar center-out reaching task with approximately 80% accuracy of reaching the actual goal.

Autonomous reinforcement learning with experience replay.

This paper considers the issues of efficiency and autonomy that are required to make reinforcement learning suitable for real-life control tasks. A real-time reinforcement learning algorithm is presented that repeatedly adjusts the control policy with the use of previously collected samples, and autonomously estimates the appropriate step-sizes for the learning updates. The algorithm is based on the actor-critic with experience replay whose step-sizes are determined on-line by an enhanced fixed point algorithm for on-line neural network training. An experimental study with simulated octopus arm and half-cheetah demonstrates the feasibility of the proposed algorithm to solve difficult learning control problems in an autonomous way within reasonably short time.