Research And Prototype System Implementation Of Urban Environmental Satellite/inertial Navigation/Wi-Fi Fusion Positioning Technology

In recent years,the complex environment full-time positioning navigation technology of intelligent terminal equipment has become a research hotspot in the field of civil-level commercial navigation.With the rise of unmanned driving and the rapid spread of smart terminals,perceptual positioning and intelligent navigation,as an important part of driverless driving,are also favored by more and more technology upstarts.Integrated navigation technology based on Global Positioning System(GPS)and Strap-down Inertial Navigation System(SINS)is one of the important research links of unmanned driving,and the traditional GPS/SINS integrated navigation technology can effectively make up for the impact of short-lived GPS.However,in traditional applications,the sensors used by Strap-down Inertial Navigation generally have embedded devices with high precision and sensitivity,and the accuracy and sensitivity of the existing low-cost Micro-Electro-Mechanical System(MEMS)sensors integrated by smart phones are relatively low,which is not enough to support the sensing precision required by the traditional method.That leads to the system's overall accuracy is not high and the robustness is poor.On the other hand,after the GPS failure for a long time,the output of the INS is also easy to diverge.In order to increase the positioning limit of INS and the overall accuracy of low-cost system when satellite is missing,this paper uses neural network method to simulate Kalman filter output prediction result in GPS failure,smart phone Wi-Fi signal and ranging-based indoor positioning algorithm is also considered,and it is expected to improve the positioning navigation accuracy and positioning robustness of the intelligent terminal in a complex environment.The paper proposes a Position-Velocity-Update Architecture(PVUA)Neural Network.During GPS failure,PVUA is used to discover the non-linearity mapping relationship between the INS navigation information and the Kalman filter output.The paper also researched and partially reproduced the positioning algorithm using outdoor public Wi-Fi positioning,and integrated it with GPS/INS integrated navigation system to further improve the positioning performance of GPS/INS integrated navigation system.Experiments show that the PVUA-assisted GPS/INS integrated navigation system has a probability of positioning accuracy of less than 10 meters with a 90%probability of low-speed straight-line driving pattern during GPS failure,and the overall accuracy error of the GPS/INS system is about 20 meters.After the same sports car data for error CDF analysis,the positioning performance based on the Wi-Fi positioning assisted fusion optimization algorithm(Wi-Fi/GPS/INS system)has a certain improvement compared to the GPS/INS system,with a probability of 98%.The positioning error is less than 15m.