Multi-Sensor For Integrated Navigation Technology Under Urban Roads

Easy access to accurate and reliable trajectory data and continuous trajectory data is the key to the intelligent development of transportation,tourism autonomous driving and other industries.Due to Global Positioning System(GPS)and the Inertial Navigation System(INS)complement each other,increasing the overall performance.Due to the excellent complementary features between the Global Positioning System(GPS)and the Inertial Navigation System(INS)in terms of principle,structure,error and etc.,the INS/GPS integrated navigation system is shown to have superior performance in information acquisition of vehicle position.However,with the increasing complexity of urban roads,the INS/GPS integrated navigation system will accumulate the errors of independent INS positioning over time due to satellite occlusion,signal interference and other reasons,which rapidly reducing the positioning accuracy.To tackle the above problems,the thesis focuses on the theoretical and technical research about satellite navigation instability,and concentrates on the hardware and software to explore the approach that increase robustness of the vehicle navigation and positioning system.Specific research contents are as follows:Firstly,the algorithm uses the improved Gradient Boosting Decision Tree(GBDT)to maintain the continuity of the loss function and eliminates damaged data on vehicle,and then improves GBDT to assist Kalman Filter(KF).Furthermore,modeling random errors in SINS,which compensate and predict positioning error of the SINS/GPS integrated navigation system.In addition,in the process of model training,Particle Swarm Optimization(PSO)is introduced into GBDT to find high-quality parameters,which improve global optimization ability and convergence speed.The test on the real road indicates that the KGP data fusion method based on integrated learning effectively suppresses the random error of SINS to some extent and effectively reduces the error rate of the integrated navigation system.The SINS/GPS/OBD integrated navigation strategy is designed to improve the positioning performance of the vehicle integrated navigation system during the disappearance of the GPS signal.Firstly,SINS,GPS,and On Board Diagnostics(OBD)are built on the same hardware platform.The organic combination of the external sensor and the in-vehicle sensor decreases the random error caused by the external auxiliary sensor.Then,the KGP method is integrated with the federated filter and applied to the vehicle device platform to construct a multi-sensor adaptive navigation software and hardware fusion system.Tests on different types of urban roads indicate that the strategy greatly improves the accuracy of positioning information while compensating for positioning errors,and provides a continuous and reliable solution for vehicle location services in complex traffic environments.