Maximum-Correntropy-Criterion-Based Extended Kalman Filter for Ultrawideband Distance Data Fusion under Colored Measurement Noise

This paper proposes a maximum-correntropy-criterion (MCC)-based extended Kalman filter (EKF) assisted by a finite impulse response (FIR) filter to mitigate the impact of colored measurement noise (CMN) on the accuracy of fused ultrawideband (UWB) distance measurements. We first design the EKF and FIR filter for handling CMN (denoted as cEKF and cFIR, respectively) based on the UWB localization scheme, which has high positioning accuracy compared with the received signal strength indication (RSSI)-based method. Subsequently, the MCC is applied to the cEKF, resulting in the proposed cMCCEKF. The cFIR is then utilized to assist the cMCCEKF when it encounters issues owing to singular matrix values. Finally, UWB-based indoor robot and human navigation experiments are conducted to evaluate the performance of the proposed cMCCEKF. The results demonstrate that the proposed filter significantly improves localization accuracy compared to the EKF and MCCEKF filter.