Research On Localization And Navigation Of Indoor Mobile Robot Based On Multi-sensor Fusion

In recent years,with the rapid development of robot technology,indoor mobile robots have gradually infiltrated into many aspects of production and life.However,there are still many technical problems to be solved due to the changeable and compact indoor scenes.In this paper,the positioning and navigation of mobile robot system is carried out for the mobile robot applied in hospital logistics distribution.As a result of the narrow corridors,many obstacles and a large flow of people in the hospital,the traditional differential wheeled mobile robot is not applicable.Hence,in this paper,an omnidirectional mobile robot based on Mecanum wheel has been designed.Firstly,the walking mechanism of the mobile robot is designed based on Mecanum wheel,and then the driving unit is completed so that the body configuration of the robot is designed in the light of the requirements of logistics distribution.Secondly,the inverse kinematics model and forward kinematics model of the mobile robot are performed by an analysis of the kinematics.Thirdly,in order to perform the autonomous positioning and navigation,the software of the lower computer control system and the upper computer operating system is designed and developed for the mobile robot.With regard to the problems of low positioning accuracy and even positioning failure of indoor mobile robots,a novel multi-sensor fusion positioning scheme is proposed in this paper.First of all,the sensor scheme is determined through a detailed analysis of the sensor and indoor scenes applied in mobile robots.Secondly,through the research of multi-sensor fusion method,a multi-sensor fusion positioning framework is put forward based on wheel odometer,IMU,UWB and 2D lidar.In order to implement the proposed multi-sensor fusion positioning framework,firstly,extended Kalman filter(EKF)algorithm and Adaptive Monte Carlo(AMCL)positioning method are studied.Secondly,the mathematical model of sensor is brought into the algorithm under the framework of multi-sensor fusion,thereby designing the fusion positioning algorithm of the mobile robot.In line with the problem that the Adaptive Monte Carlo algorithm cannot initialize the particles accurately,the initialization parameters of the algorithm are improved,and the position data provided by UWB is used to implement the Autonomous Initialization Positioning(AIP)algorithm for particle initialization of the AMCL algorithm.In order to improve the autonomous navigation function of mobile robots in practical applications,the autonomous navigation algorithm is improved under the navigation framework of the mobile robot,and the multi-point autonomous navigation algorithm is performed,which can use the autonomous navigation function of multiple target points in the environment.Hence,PC software is developed on the basis of ROS robot operating system so as to implement the application of multi-sensor fusion positioning algorithm and the improved navigation algorithm on mobile robots.Finally,in indoor scenes,the mobile robot is used to experiment with the positioning and navigation algorithm.Firstly,experiments have verified the feasibility of the positioning framework.Compared to a single sensor,the positioning accuracy of multi-sensor fusion positioning is greatly improved with robustness.Secondly,the experiments show that the self-positioning of the mobile robot is more accurate when it starts,which validates the feasibility of the Autonomous Initialization Positioning algorithm.In addition,the multi-point autonomous navigation of the indoor mobile robot is performed,verifying the feasibility of the improved navigation algorithm.