On the Particle-assisted Stochastic Search Mechanism in Wireless Cooperative Localization

The wireless cooperative localization plays a key role in location-aware service. However, its objective function, e.g., the posteriori probability function, is commonly nonconvex due to nonlinear measurement function and/or non-Gaussian system disturbance. Moreover, due to the unavoidable reference node location error, the associated objective function is commonly intractable, which further complicates the cooperative localization. In this paper, a novel particle-assisted stochastic search (PASS) algorithm is proposed to realize the cooperative localization. Given a nonconvex objective function, the proposed PASS method can find out the global optimum in probability, assisted with its search particles, detection particles, and proposal particles. In addition, the PASS algorithm can harness the reference node location uncertainties in cooperative localization, by employing its proposal particles. The associated Cramer-Rao lower bound (CRLB), localization error propagation, computational complexity, and convergence properties are also presented to assess the proposed PASS-based cooperative localization. Finally, received signal strength-based localization is simulated to validate the effectiveness of the proposed PASS approach.