With the robot technology and application become more and more popular,the more natural and intelligent human robot interaction technology is urgently needed.This paper proposes a human robot interaction method based on intelligent perception,which realizes the natural interaction between the human and the multi-degree of freedom robot.The main contributions of this paper is as follows:(1)The data of human skeleton,gesture and voice are obtained by a three-dimensional sensor,and the Kalman Filter is adopted to optimize the data of human skeleton.With the human skeleton measurements,the motion of human could be predicted by Gaussian Mixture Regression Model.Then,the intention of the human is analyzed from the measurements.(2)The three-dimensional colored point cloud of the environment is measured by the three-dimensional sensor,which is used to help the robot perceive the environment with the human intelligence and the point cloud segmentation and recognition algorithms.(3)The paper introduces a bounding box based on the capsule to modeling the human and the robot for collision detection,and then a safety robot path planning method is proposed with it and the artificial potential field.(4)A robot arrest experiment is designed to verify the effectiveness and practicability of the human robot interaction method proposed by this paper.The human robot interaction method based on intelligent perception fully combines the intelligence of human and robot.With the advantages of the mixed intelligence,the complexity of the human robot interaction method is reduced and therefore the efficiency of it is improved. |