Research on Indoor Robot Localization Method Based on Clustering Optimizes Genetic Algorithm

To reduce the operation time of the visible light positioning system, a visible light 3D positioning system based on clustering optimization genetic algorithm (COGA) is proposed. This method uses genetic algorithm (GA) to assess the fitness of each individual of the population, then applies density based noise application spatial clustering algorithm (DBSCAN) to analyze the population structure and then adjust the population structure. By combining these two algorithms, we can search for the optimal solution on a global scale and adjust it on a local scale, integrated use of global and local information. Simulation results show that the location speed of single point is 43ms, which improves the real-time performance of the positioning system. In a simulation space of
$3\times 3\times 3.6\mathrm{m}^{3}$
, the average positioning error of this method is about 3.8cm, which achieves the precision of centimeter positioning and satisfies the need of indoor robot positioning.