Mitigating the effects of residual biases with Schmidt-Kalman filtering

Abstract

Fusion of data from multiple sensors can be hindered by systematic errors known as biases, which generally include both deterministic and stochastic components. The deterministic errors can be estimated and then used to debias the sensor measurements prior to fusion. However, the remaining stochastic part typically referred to as a "residual" bias, can still severely degrade tracking performance. Specifically, residual biases can lead to degradation in covariance consistency, data mis-association, and spurious/redundant tracks, if left unaddressed. This paper presents an algorithm based on the Schmidt-Kalman filter for mitigating the effects of residual biases on sensor attitude and measurement generation. The algorithm incorporates the residual bias covariance into the track state update and "shapes" the state covariance. We also introduce a Schmidt-IMM filter implementation to address the problem of maneuvering targets. Simulation studies are presented to demonstrate the effectiveness of the Schmidt-Kalman and the Schmidt-IMM filters in the presence of residual biases. Significant improvements are shown for data association, covariance consistency, and position/velocity accuracy.