Maximum Correntropy Divided Difference Filter with Student's T-Kernel for Systems Experiencing Measurement Packet Loss
Ao Xu, Xinmin Song · 2025
In response to the issue of degraded state estimation performance in nonlinear systems caused by non-Gaussian noise and measurement data packet loss, this paper proposes a Student's t-kernel (STK) measurement packet loss maximum correntropy criterion divided difference filter (SKMC-PD-DDF). The traditional divided difference filter (DDF) relies on the minimum mean square error (MMSE) criterion, which is limited by the assumption of Gaussian noise and struggles to handle complex non-Gaussian noise scenarios encountered in reality. Therefore, this research introduces the maximum correntropy criterion (MCC) based on the STK into the DDF, leveraging the STK function's ability to capture the higher-order statistical characteristics of non-Gaussian noise, the estimation performance of the filter is enhanced. Moreover, the issue of measurement packet loss is also considered in the algorithm. Simulation results demonstrate that under conditions of mixed Gaussian noise, outlier interference, and random packet loss, the SKMC-PD-DDF exhibits outstanding performance, confirming its enhanced robustness and estimation accuracy in non-ideal measurement environments.