Multiple target localization based on probability fusion by spatial association

In detecting multiple targets by means of opportunity transmitters, parameter fusion of multiple sensors is a difficulty, which makes it expensive for the system to detect large numbers of targets simultaneously. This paper proposes a new positioning algorithm in the context of multiple non-cooperative illuminators, multiple sensors, to observe multiple targets. Firstly, each sensor extracts the TDOA of multipath and reconstructs the probability density function (PDF) of real parameters. Secondly, the central processing station fuses the probability from all PDFs according to the spatial association. Finally, the centre localizes the targets by multipeak searching on the fused probability function. This algorithm needs no pretreatment of parameter pairing, since all the TDOA values are matched automatically in the course of fusion. The algorithm has the advantage of inhibiting bad parameters to some extent. Numerical simulations support the characters of the new algorithm.