Adaptive IMU error correction algorithm for dual-antenna GNSS/IMU integrated vehicle attitude determination

Abstract In land–based applications, the attitude of a vehicle can be accurately estimated by integrating data from global navigation satellite system (GNSS) and inertial measurement unit (IMU). However, traditional dual-antenna GNSS/IMU integrated systems are susceptible to signal reflection, diffraction, and interruption, leading to reduced accuracy and reliability in GNSS challenging environments. To address these issues, this study introduces an adaptive IMU error correction algorithm for dual-antenna GNSS/IMU integrated vehicle attitude determination. Initially, an IMU-aided baseline length constraint model is developed to enhance the ambiguity resolution performance of the integrated system in signal-degraded areas by incorporating precise prior information from the IMU. Furthermore, an adaptive IMU error correction mechanism based on long short-term memory and particle swarm optimization is proposed to predict the vehicle attitude during GNSS outages and to constrain the attitude error drift measured by lMU. Field test results demonstrate that, compared to two other candidate algorithms, the proposed algorithm improves accuracy in roll, pitch, and yaw by 19.23%, 30.56%, and 67.12%, respectively, and by 12.50%, 10.71%, and 38.39% respectively. Moreover, in two distinct scenarios where GNSS is blocked for 120 s, it was able to consistently deliver accurate and stable vehicle attitude information, maintaining roll, pitch, and yaw accuracy within 0.08 degrees.