Sparse Inversion Localization of Multiple Sources With a Wireless Sensor Network

To address the challenge of achieving simultaneous co-frequency multi-target localization based on a network of collaborating sensors, a block-sparsity based model is proposed by considering energy attenuation and signal delay. Due to lack of accurate sparsity information in practical scenarios, a Block Residual Ratio Detection (BRRD) scheme is developed, leveraging the statistical characteristics of block residual ratios under noise-only conditions. Furthermore, a blind sparsity multi-source localization algorithm is developed, building upon traditional block-sparse algorithms. Through theoretical analysis and simulations, its capability for simultaneous co-frequency multi-source localization under low signal-to-noise ratio (SNR) conditions with blind sparsity is demonstrated. Simulation results indicate that the proposed algorithm outperforms traditional ones employing received signal strength (RSS) based compressed sensing method. Additionally, a better performance is achieved than the conventional sparse localization algorithm in case of unknown sparsity, indicating robustness of the proposed method.