Multi-target direct localisation via approximate l0 norm minimisation

Proposed is a novel sparsity-based direct position determination appoach to locate multiple targets in a region of interest. Since targets lie at a few points in the discrete spatial domain, this inherent sparsity can be exploited to convert the localisation problem into a sparse recovery problem. An iterative reweighted sparse recovery algorithm is also proposed to enhance the sparsity, especially under low signal-to-nosie ratio conditions. Furthermore, the locations of targets can be obtained from noisy signals by solving an approximate norm l₀ minimisation problem, even if the number of targets is not known a priori. Simulation results demonstrate the performance of the proposed algorithm on the location estimation accuracy.