High-precision Indoor Visible Light Positioning And Tracking

With the rapid development of the Internet of Things and wireless sensor technology,indoor positioning services are getting more and more attention.The traditional outdoor positioning system,GPS,has poor indoor coverage and cannot achieve accurate positioning results.Due to the need to add additional infrastructure and susceptibility to electromagnetic interference,indoor positioning technologies based on infrared,wireless LAN,Bluetooth,radio frequency and ultra-wide band cannot be well promoted and applied.Visible light positioning,which has the advantages of anti-electromagnetic interference,capable of lighting,saving installation cost,no radiation pollution,and high positioning accuracy,has become one of the hot research areas in indoor positioning.This article focuses on achieving high-precision indoor positioning and tracking.The main research contents are as follows:(1)First,the light source characteristics of LED,visible light communication model,channel response and noise model are introduced.The comparison of channel gain distribution between line-of-sight transmission link and non-line-of-sight transmission link is given.Several commonly used positioning algorithms are introduced,focusing on the positioning accuracy of the Received Signal Strength(RSS)algorithm,fingerprint algorithm and Time Difference of Arrival(TDOA)algorithm.Simulation results show that the positioning accuracy of the traditional RSS algorithm is not high,after centroid weighted optimization,the positioning accuracy is improved by about 49%.Although the fingerprint algorithm has high positioning accuracy,it requires a large complexity.TDOA algorithm can achieve high-precision positioning,but requires extremely high time delay estimation.(2)Due to multipath effects and noise,the indoor visible light positioning which adopts RSS method is inaccurate in estimating the distance,resulting in low positioning accuracy.A distance estimation method based on BP neural network optimized by genetic algorithm was proposed,which improves the positioning accuracy.The simulation results show that the average positioning error can reach 0.6428cm in the 5mx5mx3m positioning area.Compared with the traditional RSS weighted centroid method,the average positioning accuracy improves by about 96.4%.Within different positioning ranges and locations,the average positioning error is stable at the millimeter level and will not expand with the expansion of the positioning area.Similarly,after the GA-BP optimization of the time delay difference of the TDOA algorithm,it is unnecessary to calculate the time delay difference multiple times,which improves the calculation speed and positioning accuracy.Compared with the traditional TDOA algorithm,the positioning accuracy is improved by about 30%.(3)The RSS positioning algorithm is combined with the Particle Filter(PF)algorithm to achieve trajectory tracking of moving targets.The observation value of the PF algorithm is the distance value optimized by GA-BP,a single LED is used to track the target with uniform linear motion.Due to the random initialization of the particles,the tracking performance of the algorithm is not stable.Multiple LED PF algorithm(Multiple LED PF,MLED-PF)is used to track the target.The initial state of the target is calculated by GA-BP optimized RSS algorithm,which is assigned to the particles to complete the initialization,The tracking results of LED are averaged,problem of unstable PF tracking result is solved.At the same time,the occlusion factor is introduced.When an LED is blocked,use the Pauta criterion to remove gross errors and eliminate the overall tracking performance effect caused by blocked LED,the anti-occlusion ability of PF algorithm is improved.In the 5mx5mx3m positioning scenario,the average tracking error is stable at the centimeter level.