An Optimized Method for Wireless Sensor Localization Using Heuristic-Based Adaptive Particle Swarm Fusion

The fusion localization algorithm represents a cornerstone of data fusion localization technology. However, traditional algorithms often suffer from target loss in complex environments. We propose a heuristic adaptive particle set-based wireless sensor fusion localization method to address the challenges of inaccurate and inefficient multisensor fusion localization in complex environments. Initially, to enhance fusion localization efficiency, the sampling particle set size is dynamically estimated based on the variance of sensor observational distances. This allows for updating the number of sampling particles at each instance, alongside establishing a particle set threshold. Subsequently, to prevent the reduction of particle diversity and the occurrence of local optima or algorithmic divergence, the rate of change in particle number per unit time serves as a criterion for the update of superfluous particles. Then, the particle swarm optimization (PSO) algorithm is employed to augment particle diversity under this condition. The particle weights corresponding to positional information are computed, filtered, fused, and used to update the posterior probability distribution. A simulation experimental platform was subsequently constructed utilizing the RRESCAN environment and MATLAB/SIMULINK, with the experimental model of the algorithm being established therein. Conclusively, a comparative experiment involving ultra-wideband (UWB) and inertial measurement unit (IMU) fusion positioning was conducted in a real-world setting. Our proposed method demonstrates a significant improvement in tracking efficiency and positioning accuracy for vehicle fusion positioning within complex environments through comparative analysis with conventional fusion positioning algorithms.