(1.School of Automation,Wuhan University of Technology,Wuhan 430070,China;2.Chongqing Research Institute, Wuhan University of Technology, Chongqing 401135, China;3.Department of Computer Science, University of Liverpool, Liverpool L693BX,UK)
Abstract:During collaborative operations between humans and robots, the robot’s inability to accurately predict human motion intentions hampers the achievement of safe and friendly collaboration, resulting in inefficient human-robot collaboration. To predict the operator’s motion intentions during human-robot collaborative co-carrying, an operator-based framework for the perception of human hand motion intention is proposed in this paper. This framework primarily includes two aspects: 1) using an autoregressive method to predict hand motion trajectories and proposing a robot motion trajectory planning method for human-robot collaborative operations, enabling the robot to anticipate human motion intentions; 2) introducing a directional compensation method in hand motion trajectory prediction to reduce interaction forces between humans and robots. Experimental results show that compared to scenarios without motion intention prediction, the interactive forces on the robotic arm decrease by at least 45.9% when motion intention prediction is implemented. Furthermore, when directional compensation is introduced into the motion prediction method, a 30.4% decrease in interactive forces on the robotic arm is observed compared to scenarios without directional compensation. The maximum trajectory prediction error decreases by 32.4%, and the average error decreases by 6.8%. Compared to Kalman filtering trajectory prediction, interaction forces decreases by 52.2%.