Design Of Indoor Unmanned Vehicle MINS/LiDAR/OD Combined Positioning System

A single positioning method cannot meet the needs of the rapidly developing unmanned vehicles in terms of accuracy,stability,and safety.Multi-sensor information fusion technology has become a research hotspot for indoor unmanned vehicle positioning systems.Based on the research background of the indoor unmanned vehicle positioning system,this paper designs and builds the MINS/LiDAR/OD combined positioning system,conducts in-depth research on the sensor model,pose calculation,and combined positioning algorithm,and finally verifies it through simulation analysis and experiments.The designed combined positioning system can meet the application requirements of indoor unmanned vehicles.The main research contents of this paper are as follows:(1)Design and implementation of combined positioning system.According to the overall plan and design index of the unmanned vehicle positioning system,complete the sensor selection and software and hardware plan design.Complete the corresponding hardware circuit design and driver programming,based on the four-wheel unmanned vehicle chassis design and build modules such as sensors,data acquisition and processing,chassis motion control,remote control and display,and realize the MINS/LiDAR/OD combined positioning system.At the same time,according to the actual parameters of the sensors and the unmanned vehicle,an unmanned vehicle simulation environment was built in the Gazebo software under the ROS environment.(2)Sensor model and data processing method.Firstly,the initial alignment of MINS assisted by the magnetometer is studied,and some error terms are ignored due to the large error of MEMS devices,which simplifies the navigation solution equation;secondly,the LiDAR motion distortion correction and occupation grid assisted by MINS are studied.The map is updated and the LiDAR data is solved based on the Hector SLAM algorithm,and the position and heading changes of the unmanned vehicle are obtained.Finally,based on the four MGT010 Hall encoders installed on the wheels,the OD-based kinematics model and pose of the vehicle body are studied Extrapolate.(3)Research on multi-sensor combined positioning algorithm.The MINS/LiDAR/OD combined positioning algorithm based on EKF filter is designed,the state error equation and measurement equation of the combined positioning system are analyzed and established,and the multi-sensor data processing flow under different update frequencies is studied.(4)Test and analysis of the combined positioning system.In order to verify the reliability and accuracy of the designed combined positioning system,a single-sensor positioning test,a LiDAR motion distortion correction test and a MINSLiDAR/OD combined positioning test were carried out.The results show that: when MINS works alone,the error diverges quickly and does not have a separate positioning function;when LiDAR and OD are positioned separately,the position root mean square error is 0.17 m and 0.4m,respectively,which does not meet the positioning accuracy requirements of unmanned vehicles;under the assistance of MINS The LiDAR motion distortion correction effect is good;the accuracy of the MINSLiDAR/OD combined positioning system is better than that of a single sensor,and the position root mean square error under linear motion,rectangular motion,polygonal motion and indoor complex environment motion are 0.096 m and 0.084 respectively m,0.060 m,0.079 m.The position accuracy can meet the positioning needs of indoor unmanned vehicles and has certain application value.