Research On Method And Mapping Of INS/Lidar/Polarization Skylight/Optical Flow Integrated Navigation

Multi-sensor integrated navigation is an information processing process in which information and data from different sensors or different information sources are designed to complete navigation and other tasks under certain criteria.Multi-sensor integrated navigation technology has broad application prospects in aerospace,intelligent transportation,logistics transportation and other fields.At present,there are two main difficulties in the research of multi-sensor integrated navigation technology.One is how to choose sensors for navigation problems in a specific environment,and the other is how to realize a multi-sensor integrated navigation system with strong robustness,strong autonomy,high precision and high real-time.Aiming at this problem,this paper constructs a special environment scene with GNSS rejection and sparse features.Aiming at the second problem,it is proposed that the polarization sensor can obtain the absolute heading angle information from the atmospheric polarization mode,and the optical flow sensor can make up for the deficiency that the polarization sensor can not obtain the speed information.Combined with inertial measurement unit and lidar,loose-combined and tight-combined inertial/radar/polarization/optical flow integrated navigation systems are established respectively.Based on the factor graph optimization technology,the accuracy,autonomy and robustness of the multi-sensor integrated navigation system are improved.The main research contents of this paper are as follows:1.In view of the high cost of realizing high-precision navigation with a single sensor in the GNSS rejection environment,and the problems of the traditional multisensor integrated navigation system,such as easy interference,poor autonomy,error accumulation,etc.,the integrated navigation system is established by introducing a point source polarized light sensor and an optical flow sensor,combining with an inertial measurement unit and a laser radar,adopting the back-end fusion method of loose combination factor graphs,constructing the measurement functions of different sensors and the error functions of factor graphs,and optimizing the error functions by using the nonlinear least square method to obtain the optimal state.Finally,the effectiveness of the proposed method is verified by simulation and experiment.2.Aiming at the problems that the sensors in the loose combination factor graph are easily disturbed by noise and the data fusion is insufficient,we adopt the form of tight combination of multiple sensors,introduce the light intensity information of each channel of the point source polarization sensor,obtain the light intensity difference at the adjacent time,and calculate the heading angle increment at the adjacent time.We use the information of inertial sensor and polarization sensor to correct the motion distortion of lidar,and at the same time,the lidar provides the initial value for inertial sensor,and through the information correction between sensors,we add the optical flow correction factor to eliminate the outliers in the system.A factor map based on multisensor compact combination is constructed,and the error function of the factor map is optimized to obtain the position and attitude estimation value of the compact combination navigation system.A carrier platform is built to carry out outdoor experiments.The analysis of experimental results shows that the proposed multi-sensor compact integrated navigation method based on factor graph can improve the estimation accuracy of position and attitude of navigation system compared with other methods.3.Aiming at the problem that the polarization sensor has a large error in heading calculation when the sky is blocked,which leads to the poor accuracy and environmental adaptability of the integrated navigation system,firstly,the observability of the integrated navigation system is analyzed with the help of the concept of observability of the integrated navigation system,which proves the observability of the integrated navigation system based on polarization.Considering that the point source polarization sensor can not quantitatively analyze the sky occlusion.The image polarization sensor is introduced,and the percentage parameter of polarization image occlusion is defined.Whether the data collected by the polarization sensor is available or not is judged by the set threshold.The outdoor experiment was carried out under occlusion environment.By comparing the heading error with the occlusion percentage of the image,the critical occlusion percentage of polarization data failure was obtained.