The Key Technology Of Perception Enhanced Robot ARM Teleoperation

With the development of robot imitation,robots can learn to imitate human behavior in on-line or off-line fashion based on the human action data,meanwhile the technologies and devices for collecting human action data are becoming more and more mature.The development of robot bionics increases the DOF of robots,which makes them flexible enough to imitate some complicated actions very well.However,only few works focus on the mechanism of robot perception feedback during the imitation.This article designs a perception enhanced robot arm teleoperation system which provides robot perception feedback.This system enables the remote robot imitating operator's action through teleoperation and gives the operator feedbacks on visual and tactile perception,which enhances robot arm perception.The main contributions of this paper are as followed:1)We implemented a hand action recognition method based on Kinect and color mark,which can collect hand action data when the hand is moving.2)We designed and implemented a mapping algorithm to convert hand action into robot operation instructions based on the collected hand action data,and the performance of this algorithm was also evaluated.3)We built a feedback system,which provides visual and tactile feedback during the runtime of the robot arm based on Kinect and pressure sensors.4)We implement an application which can teleoperate the robot arm to move objects based on the perception enhanced robot arm teleoperation system,and the performance of the system was evaluate through experiments.