Research On Interactive Map Construction And Path Planning Of Service Robot

Map construction and path planning are the core technology of autonomous mobile service robots.The wide application of human-computer interaction technology facilities the use of service robots,and it is the only way for service robots to integrate into the common people's life.In this paper,we study the interactive map construction of service robot and the path planning technology under dynamic scenario.The purpose of this research is to enable the service robot with the ability of perceive pedestrians in the environment and achieve efficient human-computer interaction,while building maps of unknown environment and moveing autonomously under navigation instructions.The technology of this subject can be applied to robot for the disabled and the elderly,service robots and other robots,in order to improve the practicality of the robotIn the process of research,we first study and select the appropriate 2D grid map construction algorithm and robot self-localization algorithm for the existing hardware platform and sensor characteristics.The experiment is used to optimize the robot construction and localization parameters.In order to realize the interactive map construction,we realized and verified three kinds of robot follow people schemes with the help of Kinect.The best performance scheme is selected from the test of robot follow performance.The APP based on rosjava is constructed to realize the speech interaction between the robot and the user as well as remote monitoring of map construction and path planning.In order to solve the problem of path planning and navigation in the dynamic scene of the service robot,the employment of pedestrian trajectory prediction information obtained via pedestrian perception module in the Lattice state grid search space greatly enhances the path planing ability of mobile robot in the dynamic scene.The research contents of this paper include the interactive map construction algorithm and the mobile robot path planning system based on pedestrian perception technology.In order to improve the performance of the map construction,the optical encoder and the 9-axis IMU are equiped for the robot to provide accurate odometry information.The kinetic robot follows the pedestrian function based on the kernelized correlation filter algorithm.We build the human-computer interaction and the robot remote monitoring APP based on the rosjava environment;Using the laser radar and the depth camera(Kinect)to achieve a robust pedestrian perception and pedestrian trajectory prediction system;Finally we build up the robot path planning system in dynamic scene based on pedestrian perception and trajectory prediction information.The service robot constructed in the project qualified with the ability of map construction and path planning in dynamic scene,robot following as well as command the robot to reach the designated position through the human-computer speech visual interaction APP.Monitoring of the state of robot while building map as well as moving autonomously in the environment throuth the APP facilities the map construction.Pedestrian perception module is robust as well as rapid in pedestrian perception procedure and the trajectory prediction.The robot has the ability to move independently in dynamic scenes.Eventually,the expected research goals were achieved.