Cross-product algorithms for source tracking using an EM vector sensor
Arye Nehorai, Petr Tichavský · 1999
We present two adaptive cross-product algorithms for tracking the direction to a moving source using an electromagnetic vector sensor. The first is a cross-product algorithm with a forgetting factor, for which we analyze the performance and derive an asymptotic expression of the variance of angular estimation error. We find the optimal forgetting factor that minimizes this variance. The second is a Kalman filter combined with the cross-product algorithm, which is applicable when the angular acceleration of the source is approximately constant.