Cooperative Navigation Filtering Algorithm Based on Improved Maximum Correntropy

Aiming at the problem that the thick-tailed non-Gaussian noise generated by the complex environment and complex flight dynamics leads to the poor navigation effect of the master-slave Unmanned Aerial Vehicle (UAV) conavigation, an Improved Maximum Correlation Entropy Kalman Filter (IMCKF) algorithm is proposed. To address the algorithm’s sensitivity to correlation entropy error and to outliers, the stability of the filtering is ensured by introducing residual weight coefficients and estimating the cholesky decomposition matrix of the noise by using expectations with robustness to weight the residuals; and it is used to optimize the kernel-width updating, which solves the problems of the algorithm relying on the kernelwidth’s local convergence criterion and the uneven convergence speed. Simulation results show that the algorithm can effectively improve the accuracy of UAV formation cooperative navigation.