An Algorithm for Cooperative Navigation using Neural Network to Identify Colored Noise Online

In practical applications of cooperative navigation algorithms based on Kalman filtering, noise is often colored noise, which does not meet the requirement of Gaussian white noise for Kalman filtering. Therefore, an algorithm is proposed to identify and compensate for colored noise online using a BP neural network. Firstly, a cooperative navigation model based on Kalman filtering is established. Then, by studying the update process of Kalman filtering, the neural network is incorporated into the prediction calculation of the observations, following the online learning and updating of the Kalman filter. Mathematical simulations show that under conditions of colored noise, the results of the Kalman filter diverge. However, after correction by the BP neural network, the filtering converges, providing high reference value for improving the accuracy of cooperative navigation.