A hybrid optimization method based on DBO-tuning BiGRU assisted AKF for seamless INS/GPS navigation

To enhance the navigation accuracy and continuity of the integrated navigation system (INS)/global positioning system (GPS) in satellite denied conditions, the study proposes a hybrid optimization seamless navigation strategy that utilizes dung beetle optimizer (DBO) to optimize bidirectional gated recurrent unit (BiGRU) assisted maximum versoria criterion (MVC)-based adaptive Kalman filter. Initially, for the information fusion challenge of INS/GPS integrated navigation system in complex environments, an adaptive Kalman filter based on MVC is designed, exhaustively considering the complexity of actual measurement noise and key parameters of each sensor in the INS/GPS system, thereby enhancing the accuracy and robustness of information fusion and providing satisfactory information samples for subsequent neural network training. Subsequently, the DBO is adopted to optimize the BiGRU, thus predicting the velocity and position observation information of the INS/GPS and addressing its performance deterioration during GPS outages. The BiGRU hyper-parameters are fine-tuned with the DBO to optimize the neural network's structure and enhance its prediction performance and robustness for time series information, which enables it to learn from the uncertainty of the system smoothly and accurately. Finally, a vehicle navigation platform and targeted ground vehicle experiments are designed to evaluate the proposed method. The experimental results demonstrate that the proposed strategy can effectively improve navigation accuracy and maintain navigation continuity in the actual complex environment.