Research On SVR Indoor Positioning Model Based On Lightning Search Algorithm Optimization

The location-aware service finds the location of the user through a mobile device with location technology and provides services related to its location according to the location of the user.With the rapid development of the information age,the application of location-aware technology is becoming more and more important,from the precise positioning in the military,the guidance of the sea route of the sears,the route guidance of people driving,the accurate index of indoor destinations.With guidance.The Global Positioning System(GPS)has relatively accurate outdoor positioning,but often encounters no signal problems indoors,such as many shopping malls,playgrounds,etc.,due to obstacles,it is impossible to stably connect GPS satellites.Recently,indoor positioning schemes such as cellular positioning skill,RFID orientation skill,IBeacon orientation skill,ultrasonic positioning technology,and WLAN positioning technology have been proposed.Accurate positioning requires stable signal data and excellent position derivation.Accurate signals can cause excessive cost,and complicated calculations can lead to low efficiency.Therefore,how to develop a reasonable and efficient indoor positioning solution is an urgent problem to be solved..In view of the wide use of WLAN,the high penetration rate of WLAN connected to mobile phones,and the need to re-install signal transmitters to detect signal data,this paper uses signal data from different devices to integrate data and combine machine learning methods.An indoor location fingerprint localization method based on lightning search algorithm support vector regression(LSA-SVR).Using the signal data on the receiver,according to this method,high-precision position information can be quickly obtained,which provides a reference for people's position perception in the room.The main work of this paper is as follows:1.Analyze the principle and application method of WLAN technology in detail,master the characteristics of its signal data,and study the signal data integration scheme.2.Thoroughly study the principle of support vector regression machine and fully understand its theory.When using support vector regression machine,the choice of training parameters has an important impact on the prediction performance of the model.Therefore,in order to obtain a good performance model,it is necessary.Get the parameters that make it efficient.This paper introduces the newly proposed lightning search algorithm to optimize the support vector machine and parameters,penalty factor and insensitive loss function.Based on this,the LSA-SVR indoor positioning model is constructed.Through the simulation experiment,the error verification of the model based on LSA-SVR indoor position prediction is carried out,and the optimization model is compared with the traditional regression method,the regression method based on particle swarm or fish algorithm or genetic algorithm optimization,and the results show that LSA-SVR has better predictive power and robustness than support vector regression models based on fish algorithm,genetic algorithm or particle swarm optimization.Finally,chaotic optimization lightning search algorithm is used in support vector regression model based on fingerprint location.The combination makes the model more complete,with more accurate predictions and better robustness and generalization.