Research On In-Motion Initial Alignment Technology Of GNSS-Assisted Low-Cost INS Integrated Navigation

The Inertial Navigation System(INS)is a navigation system that uses inertial measurement devices such as gyroscopes and accelerometers.It does not rely on external reference information and is particularly suitable for situations where there is no Global Navigation Satellite System(GNSS)or other navigation reference signals available.However,before navigation can take place,the INS requires an initial alignment to determine the initial attitude,velocity,and position of the vehicle.The accuracy and speed of the initial alignment directly affect the subsequent navigation accuracy and system response time,making fast and accurate initial alignment crucial for INS.For high-precision INS,the initial attitude can be obtained through initial alignment with a stable base,while the initial position and velocity can be obtained through GNSS.However,in emergency situations or low-cost INS,the initial alignment needs to be achieved with the assistance of the vehicle’s motion and external sensors.This paper addresses the initial alignment problem of low-cost INS/GNSS in a moving base scenario and proposes a low-cost INS/GNSS-aided integrated navigation initial alignment scheme.The scheme is designed,optimized,and validated.Firstly,this paper analyzes the effect of accelerometer bias on the accuracy of the multi-vector construction and proposes a new vector construction method.In the process of constructing vectors,vector subtraction is used instead of sliding window integration.This not only suppresses the error caused by the accumulation of accelerometer bias over time but also overcomes the disadvantages of sliding window integration,such as the need to store and integrate data within the window,increased computational complexity,and higher memory usage.Secondly,this paper designs a new Kalman filter model that reconstructs the system’s state and measurement equations.It utilizes the velocity and position information from GNSS to more accurately estimate the gyro bias and compensate for the misalignment of the vehicle frame,thereby eliminating the influence of gyro bias on the accuracy of the multi-vector construction.This improves the accuracy of the initial alignment in a moving base scenario and enhances the robustness of the algorithm.Finally,the proposed algorithm is compared and analyzed through simulation and vehicle field experiments.The simulation experiments verify the feasibility of the proposed initial alignment method.The vehicle field experiments use both vector subtraction and sliding window integration to construct the multi-vector,and then estimate the gyro bias and compensate for the misalignment of the vehicle frame using the proposed Kalman filter model and other Kalman filter models.The experimental results show that,in the case of low-cost INS/GNSS integrated navigation with a moving base,compared to the existing Optimization-Based Alignment(OBA)algorithm and its improved versions,the proposed method achieves better alignment accuracy and algorithm robustness.The algorithm converges the heading angle within 2 degrees in less than 2 seconds and maintains stable heading angle accuracy afterward.