Research On Topology Optimization Algorithm For Indoor Multi-Node Cooperative Localization

With the rapid development of science and technology,indoor and outdoor positioning technology has become a hot research topic.In outdoor environments,Global Navigation Satellite System(GNSS)can provide high-precision positioning services.However,in the complex and changeable indoor environment,accurate location information cannot be determined due to the occlusion of obstructions such as walls,tables and chairs.These obstacles cause the propagation environment of the signal to be the None Line of Sight(NLOS)environment,resulting in a large error in the measurement distance.Most of the current research focuses on the improvement of localization algorithms.It cannot be ignored that the topological position of the anchor node also has an important impact on the localization performance.This paper mainly studies the positioning topology optimization design problem in various indoor scenarios to determine the optimal deployment scheme as much as possible.The main work is summarized as follows.Firstly,in scenarios where the requirements for positioning accuracy are not high and the number of anchor nodes is small,approximate positioning can meet the positioning requirements.Based on this scenario,the node probability perception model is used to calculate the regional localization rate in this paper.At the same time,a new anchor node deployment algorithm RPSO-DV is designed.The algorithm combines the advantages of the three algorithms,including the Resampling Particle Swarm Optimization(RPSO)algorithm,the PSO-D algorithm and the improved Virtual Force(VF)algorithm,so as to determine the optimal deployment scheme.The RPSO algorithm improves the overall quality of the particles and the diversity of the population through the resampling strategy;the PSO-D algorithm improves the particle velocity and position update formula,which improves the global search ability of the population in the early iteration and the local search ability in the later iteration;Based on the VF algorithm,a virtual force threshold is introduced in this paper.The position of the anchor node is adjusted by the virtual force between nodes and the virtual force of the boundary to the node,so that the node distribution is more uniform.The simulation results show that the RPSO-DV algorithm has better performance in multiple groups of scenarios.Secondly,considering that obstacles in indoor NLOS environments have a great impact on positioning performance,an anchor node deployment algorithm for multilateral positioning scenarios is designed.For the convenience of research,this paper simplifies the positioning model.That is,the target node is not located by the anchor node in the NLOS environment.At the same time,a new evaluation function is proposed.In addition to the positioning rate and positioning accuracy,the evaluation function also includes a penalty term.Penalize the situation that the target node has too many or too few locatable anchor nodes,thereby achieving the purpose of increasing the positioning area.This problem is solved using an improved Gravitational Search Algorithm(GSA).The distributed structure is introduced to group the population,the sub-population is repaired according to the distance threshold,and the sub-population is evolved by using three levels of interaction.The poor particles are eliminated through the information exchange between the sub-populations,so as to determine the optimal deployment scheme.The simulation results show that the evaluation function and the GSA algorithm proposed in this paper have good performance.Finally,in the indoor NLOS multilateral positioning scenario based on obstacles,the differences in the deployment positions of anchor nodes and human tendency factors are considered.Aiming at the scenario where some anchor nodes have been deployed in advance,the deployment scheme of the remaining anchor nodes is designed.In order to improve the GSA algorithm proposed above,an IGSA algorithm is proposed to solve it.Use an improved uniformity index to evaluate particle quality.At the beginning of the iteration,the particles are screened,and the better particles are selected for iteration.The optimal solution is determined through population grouping,sub-population restoration and evolution.The simulation results show that the IGSA algorithm proposed in this paper has better performance under different proportions of mobile nodes.