Localization of Multiple RF Sources via Priori Knowledge-Aided Bayesian Compressive Sensing in UAV-Based WSN

In the unmanned aerial vehicle-based wireless sensor network, the compressive sensing approach can simultaneously locate multiple ground radio frequency sources, while the existing algorithms’ localization accuracies would deteriorate confronting the heavy noise. Hence, we exploit the potential priori knowledge and propose a robust localization algorithm named priori knowledge-aided Bayesian compressive sensing. Firstly, the problem of multi-emitter localization is transformed to that of sparse recovery, whose sparse dictionary is constructed based on the lognormal signal propagation model. Besides, we replace the Gaussian prior in the conventional hierarchical model with the Laplace prior to enhance the localization accuracy. Finally, the sparse recovery problem is solved under the sparse Bayesian learning framework, where the priori knowledge is incorporated to increase localization accuracy further. The improved performance of the proposed algorithm is testified by the simulations where both the state-of-the-art methods and Cramer-Rao lower bound are compared.