Scalar weighting optimal and adaptive information fusion kalman filter with feedback

Abstract

Based on multi-sensor optimal fusion criterion weighted by scalars in the linear minimum variance sense, a scalar weighting optimal fusion Kalman filter with feedback is given for discrete linear stochastic system with multiple sensors, and it has a two-layer fusion structure. Further, an adaptive information fusion filter with feedback is also given when the process noise covariance is unknown. The fused filter and process noise covariance in the fusion center is spread to all local subsystems by feedback at each time step. And the fusion filters have better precision than any local filter does. Applying them to a radar tracking system shows their effectiveness.