Real-Time Bias Estimation and Alignment of Two Asynchronous Sensors for TrackAssociation and Fusion

摘要

An extensive simulation study of the problem of relatively aligning two sensorsthat measure range, bearing, and elevation is described in this report. Simple simulations are used to demonstrate the effects of alignment errors on multisensor tracking. The theory and algorithms for relatively aligning the sensors are briefly summarized. This work presents the extension and simulation of algorithms to permit the use of asynchronous data. This is accomplished by using Kalman-filter-based prediction algorithms to time-align the state estimates from the two sensors. One-step, fixed-lag smoothing is also employed to improve the accuracy of the state estimates. The effectiveness of using the Interactive Multiple Model algorithm versus single model filtering in the tracking filters prior to bias estimation is also studied. Multiple model versions for prediction and one-step, fixed-lag smoothing of the track estimates are also applied and compared with their single model counterparts with respect to bias estimation accuracy. (MM).