A Batch-Recursive Algorithm For Passive Ground Target T racking

We propose a batch-recursive algorithm for tracking ground moving targets on constrained paths using bearings- only measurements in clutter collected by a ground moving observer. The proposed algorithm, in its batch stage, uses a fast and low-complexity procedure based on the maximum-likelihood method to obtain a raw estimate of the line-of-sight direction. Once this is achieved, such an estimate is employed at the recursive stage to initialize a regularized particle filter, which provides track maintenance by recursively estimating a modified polar coordinate representation of the state vector. The scenario represents a target moving along a realistic road network with junctions, roads branching or crossing, where the probability of having measured the line-of-sight bearing, among the multiple observed ones, is less than the unity. Realistic simulations are presented to support our findings.