Self-tuning Measurement Fusion Kalman Predictor

摘要

For the multisensor system with unknown noise variances and with the different measurement matrices, the on-line estimators of the noise variances are obtained by a correlation method, and a self-tuning measurement fusion Kalman predictor is presented. Its basic principle is that the optimal fuser, accompanied by a recursive identifier of noise variances, will yield a selftuning fuser. It is strictly proved by using the dynamic error system analysis (DESA) method, that the selftuning Kalman fuser converges to the optimal Kalman fuser with probability one or in a realization. It can reduce the computational burden, and has asymptotic global optimality. A simulation example for a target tracking system with 3-sensor shows its effectiveness.