Estimability of Sensor Alignment Biases Using the Static Cramer Rao Lower Bound
Morten P. Topland, Oddvar Hallingstad · AIAA Guidance, Navigation, and Control Conference and Exhibit · 2008
The estimation of sensor alignment biases is essential in order to perform successful multisensor measurement fusion in a target tracking scenario. If the alignment biases are not estimated, they cause large errors at long target ranges which degrade tracking performance. The biases can be estimated relatively or absolutely. In the relative case the biases are estimated relative to a master sensor, while in the absolute case the biases are estimated in reference to a coordinate system independent of the sensors. As emphasized in this paper, absolute bias estimation is essential if it is desirable to perform sensor fusion in such a coordinate system. For this purpose the sensor locations are assumed to be known. The observability and estimability of the alignment biases are studied using the Fisher Information Matrix (FIM) and the Cramer Rao Lower Bound (CRLB) for static targets, and it is concluded that absolute estimability is acheivable when using three sensors or more. The results of the analysis are verified in simulations, where the bias estimator is shown to be efficient as it converges close to the CRLB for moving targets.