Tracking With Sequentially Fused Radar and Acoustic Sensor Data With Propagation Delay

In some special scenarios, the merits of radar tracking may be lost due to great interference, while the passive acoustic signal can be detected. Then, fusing active radar and passive acoustic sensors can improve the tracking performance. However, a problem associated with the existing fusion approaches is that the measurements of radar and acoustic sensor assumed obtained simultaneously are actually from different emission times due to the significant acoustic propagation delay. Then, a sequential cubature Kalman filter (CKF) can be used for fusing heterogeneous sensors with delay. However, in sequential fusion, it is difficult to perform an accurate state prediction or retrodiction due to the unknown delay. Although the emission time is unknown, it satisfies the delay constraint. In view of this, we augment the delay constraint onto the motion equation to estimate the target state and emission time simultaneously. Then, two sequential fusion algorithms, called first-radar-then-acoustic fusion (FRTAF) and first-acoustic-then-radar fusion (FATRF), are developed. A wide range of simulations have been carried, and the performance of the FRTAF, FATRF, and single radar tracking is compared. Illustrative examples demonstrate that both two fusion algorithms outperform the track-to-track fusion (T2TF) and information matrix fusion (IMF) algorithms.