GNSS/INS Tight-Coupled Robust Navigation Technology Research For Visible Satellite Shortage

In satellite navigation positioning,due to the defects of satellite navigation itself,the positioning accuracy and reliability of satellite navigation in urban canyons and other areas are reduced.With the development of intelligent transportation and unmanned driving technology,there is an increasing demand to solve the problem of satellite positioning in urban areas.Satellite positioning is characterized by high accuracy and error accumulation over time outdoors.However,GNSS-only positioning in urban areas with dense high buildings,satellite signals will be blocked by buildings,trees,overpasses,electromagnetic interference generated by high voltage,etc.However,inertial navigation system has the characteristics of being less affected by the environment and being able to provide high-precision carrier motion information in a short period of time,but its errors will accumulate rapidly with time,which shows complementarity with satellite navigation system.Therefore,integrating the advantages and disadvantages of satellite navigation system and inertial navigation system,and integrating the information of satellite navigation and inertial navigation are the research characteristics in relevant fields at present.This paper presents an improved GNSS/INS tight-coupled algorithm.In addition,a multi-source fusion attitude initialization algorithm when satellites are unavailable,a gross error detection suppression and a time-varying noise estimation method in a complex environment are proposed to finally realize stable navigation when the number of satellites is short and insufficient in an interference environment.In this paper,the GNSS/INS tight-coupled system model under the traditional EKF framework is firstly established,and its observability under different visible satellites is analyzed.Based on the traditional GNSS/INS tight-coupled model,a GNSS/INS tight-coupled model based on multi-epoch joint observation is designed.The receiver multi-epoch states are put into the state vector,and the pseudorange observation of multiple epochs is delayed updated uniformly,thus improving the observability of the system in a short time.Experiments show that the system can ensure the accuracy and stability of navigation in a short time visible satellites shortage(2 or 3 visible satellites).Aiming at the state initialization problem of GNSS/INS system,when GNSS is unavailable,an attitude initialization algorithm based on geomagnetic and gravity measurement is established and an improved attitude estimation algorithm based on magnetic field is proposed on this basis.Finally,a simulation experiment is carried out.The experiment shows that the attitude initialization method based on geomagnetic and gravity measurement is not applicable in dynamic environment,and the method based on geomagnetic alone can ensure good estimation accuracy.Finally,aiming at the problems of observation gross error and time-varying noise existing in the environment,an adaptive filtering method based on multi-epoch observation is proposed,and performance tests are carried out in GNSS/INS tightcoupled algorithm and geomagnetic/INS attitude estimation algorithm.The experiments show that the geomagnetic/INS attitude estimation algorithm has good adaptive estimation effect,but the estimation performance decreases in GNSS/INS tight-coupled algorithm.