An Improved Particle Swarm Optimization Indoor Positioning Method Based on the Weighted Adaptive KNN Algorithm

The particle swarm optimization algorithm is used extensively for global search in indoor positioning. However, owing to the large search area and insufficient accuracy of the particle search, it may converge too early or fail to find the global optimal solution, which can negatively impact the accuracy of indoor positioning. In this study, we propose an improved particle swarm indoor localization method based on the weighted adaptive K-nearest neighbor (WAKNN) algorithm to address this issue. Our method first employs the WAKNN algorithm for fingerprint matching and then selects suitable reference points (RPs) to limit the actual presence area of the user. Subsequently, we use the improved particle swarm optimization (PSO) algorithm for localization. To enhance the PSO algorithm, we introduce a mutation operator that enables particles to reinitialize with a certain probability and jump out of the previously searched optimal value position, thereby avoiding the trap of local optima. Experimental results demonstrate that the positioning error of our proposed algorithm is smaller than that of the WAKNN and PSO algorithm based on maximum likelihood estimation.