Research On Mapping And Navigation Of Indoor Service Robot

With the rapid development of computer and sensor technology,the types of robots have also extended from the beginning to some industrial robots that complete simple tasks to agriculture,medical,education,aerospace and other fields.At the same time that robot technology is booming,how to achieve accurate real-time positioning and navigation of robots has also become one of the hot issues in robot research and development.Simultaneous Localization and Mapping(SLAM)method has special significance in solving the problems of robot positioning difficulty,poor accuracy and high sensor requirements.This paper makes a comparative analysis of sensor selection,sensor fusion technology,map construction and navigation algorithms,which improves the accuracy of robot positioning and navigation and reduces costs.It also designs and implements indoor service robot map construction and navigation obstacle avoidance.The research of the paper mainly has the following aspects.(1)The mobile robot system is modeled and analyzed under the ROS platform,including the system model composed of robot motion and coordinate transformation,the sensor model and map model composed of sensors such as odometer,inertial measurement unit(IMU)and laser radar,etc.Subsequent algorithm simulation and implementation provide the foundation.It provides the basis for subsequent algorithm simulation and implementation.(2)On the SLAM problem,research on sensor data fusion and map construction.In terms of data fusion,the Extended Kalman Filter(EKF)and Unscented Kalman Filter(UKF)algorithms are analyzed and studied in detail.Combined with the robot motion model,MATLAB is used to simulate and compare the two algorithms.Simulation results show that: Over time,the error of UKF in SLAM is much smaller than that of EKF.Based on the use of Unscented Kalman filter,the SLAM algorithm based on Rao-Blackwellized particle filter(RBPF)is studied.The difference between RBPF and PF algorithm are analyzed through simulation comparison to determine the effectiveness of RBPF.Finally,a filtering algorithm combining UKF and RBPF was selected to realize the Simultaneous Localization and Mapping of the robot.(3)Contrasting and analyzing several popular algorithms for robot path planning algorithms.In this paper,the ant colony algorithm and the improved artificial potential field method are selected as global and local path planning algorithms and the algorithm is simulated.Combining the two algorithms as a robot navigation algorithm can achieve dynamic obstacle avoidance while ensuring the optimal global path.(4)Under the ROS system,a self-designed wheeled robot experimental platform is built.The robot's hardware circuit design and component selection are analyzed in detail.Programs such as upper and lower computer drives and serial communication.Embed positioning,mapping and navigation algorithms into the hardware platform.Finally,experiments of positioning and mapping are carried out in indoor and promenade environments,and experiments of navigation and obstacle avoidance were conducted indoors.The experimental results verify the feasibility of the positioning,mapping and navigation obstacle avoidance algorithm adopted in this paper.