Research And Implementation Of Robot Information Fusion System For Human-computer Interaction

In recent years,artificial intelligence has made great progress,robot and human contact more and more closely,there are more and more kinds of interaction between human and robot,but no matter what way,the robot is through the sensor to perceive the human body and interact with human,these goals is difficult to complete by single sensor,multi-sensor information fusion technology is indispensable means of the robot human computer interaction process.The essence of multi-sensor information fusion technology is to optimize the processing of multi-sensor information,eliminate redundant information,synthesize the effective information of each sensor to form complementary,more comprehensive information,and better characterize the internal and external environment of the robot.The accuracy of the robot's visual tracking is improved by means of the combination of depth information and color information;during the follow-up process of the human body target,and the accuracy of robot's movement is improved by means of the fusion of the robot based on the conditions.The main work and innovation of this thesis are as follows:This thesis presents a design scheme of robot information fusion system for human-computer interaction.The system is divided into three modules: human body perception module,robot visual tracking module and robot movement follower module.The human perception module,we through the motion compensation based on the inter-frame difference method to detect the moving target,we use skeleton to verify whether the target of the human body.The robot vision tracking module,we studied the wavelet transform fusion algorithm,we use Depth image and reverse projection method of fusion in the application of visual tracking,to improve the robot on the human target center of mass tracking accuracy.The robot movement follower module,we use the robot pose and inertial navigation pose information fusion,to real-time correct of the pose of the robot,it can effectively improve the accuracy of motion follow.This thesis design and implements a robot information fusion system for human-computer interaction.And we achieve the system and the relevant test in the robot.