Distributionally Robust Kalman Filter Fusion for Multi-sensor System

The aim of this study is to develop a robust module through the integration of distributionally robust Kalman filter with multi-sensor fusion. Leveraging the established reliability of the robust Kalman filter operating at a single station, we propose its extension to a multi-sensor distributed system, observing a reduction in error resulting from the integrated data fusion across multiple sensors. By employing a target tracking example, we demonstrate the superior performance of our bi-causal optimal transport fusion framework.