The Research Of Human-robot Interaction Technology In Dual-robots Based On Dual-hands Gesture Control

Research on human-robot interaction technology is long-standing.The human-robot technology based on gesture control has a prominent advantage in the aspects of expression richness,universal environment and interactive comfort,which is expected to become one of new human-robot interaction in next generation.Based on the study of hand gesture interaction in the early stage of the lab,this paper presents a human-robot interaction in dual-robots based on dual-hands gestural control.The main contents are as follows:(1)The hybrid sensors of Leap Motion and Kinect are placed vertically to detect the position and pose of the hands in real time,which can effectively solve the hand gesture detection in the presence of occlusion,which can reduce the number of resetting the gestures during the human-robot interaction to improve the efficiency of interaction.(2)An iterative algorithm based on orthogonal matrix is used to calibrate and transform the local coordinate system of the hybrid sensor.The adaptive low-pass filter algorithm based on velocity is used to filter the data acquired by each sensor.And then unify the data into the global coordinate system.Finally,the data is fused by adaptive weight to get the final dual-hands gesture data.(3)Robot motion decision control is performed on the obtained dual-hands gesture data.Firstly,Gaussian mixed regression model is used to predict the lost gesture data frame.Then,the adaptive multi-space transform based on the distance of dual-robots is used to convert the gesture space data into robot workspace data.Finally,the collision detection algorithm based on capsule bounding box is applied to dual-robots collision avoidance detection during the control of the dual-robots cooperative task.At the end of the paper,the experiment of dual-robots screw based on dual-hands gesture control is carried out to validate the feasibility and validity of the research.