Learning skills autonomously is a particularly important ability for an autonomous robot. A promising approach is reinforcement learning (RL) where agents learn policy through interaction with its environment. One problem of RL algorithm is how to tradeoff the exploration and exploitation. Moreover, multiple tasks also make a great challenge to robot learning. In this paper, to enhance the performance of RL, a novel learning framework integrating RL with knowledge transfer is proposed. Three basic components are included: 1) probability policy reuse; 2) dynamic model learning; and 3) model-based Q-learning. In this framework, the prelearned skills are leveraged for policy reuse and dynamic learning. In model-based Q-learning, the Gaussian process regression is used to approximate the Q-value function so as to suit for robot control. The prior knowledge retrieved from knowledge transfer is integrated into the model-based Q-learning to reduce the needed learning time. Finally, a human-robot handover experiment is performed to evaluate the learning performance of this learning framework. Experiment results show that fewer exploration is needed to obtain a high expected reward, due to the prior knowledge obtained from knowledge transfer.

Purpose: Lower-limb exoskeleton systems enable people with spinal cord injury to regain some degree of locomotion ability, as the expected motion curve needs to adapt with changing scenarios, i.e. stair heights, distance to the stairs. The authors’ approach enables exoskeleton systems to adapt to different scenarios in stair ascent task safely. Design/methodology/approach: In this paper, the authors learn the locomotion from predefined trajectories and walk upstairs by re-planning the trajectories according to external forces posed on exoskeleton systems. Moreover, instead of using complex sensors as inputs for re-planning in real-time, the approach can obtain forces acting on exoskeleton through dynamic model of human-exoskeleton system learned by an online machine learning approach without accurate parameters. Findings: The proposed approach is validated in both simulation environment and a real walking assistance exoskeleton system. Experimental results prove that the proposed approach achieves better performance than the traditional predefined gait approach. Originality/value: First, the approach obtain the external forces by a learned dynamic model of human-exoskeleton system, which reduces the cost of exoskeletons and avoids the heavy task of translating sensor input into actuator output. Second, the approach enables exoskeleton accomplish stair ascent task safely in different scenarios.