Research On Mobile Robot Indoor Localization Technology Based On WLAN

High precision autonomous localization is an important subject in mobile robot research field,and is a significant foundation to carry out other researches and applications.So far,satellite positioning system such as GPS have been obtained high-precision localization in outdoor open ground scenes.However,the penetration of satellite signals is limited,and it is unreasonable to implement indoor localization,therefore the high precision indoor localization of mobile robot has gradually become a hot research point.Mobile robot indoor localization technology based on WLAN gives the possible to combine high precision positioning with low cost due to its convenience deployment,low cost,high reliability,convenient for measurement,it has a wide application prospect and great research value.To improve the accuracy of indoor localization for mobile robot in WLAN,the following works have been done in this paper:(1)In order to obtain high-quality signals,the WiFi signal features are analyzed from 6 expects in this paper including the feature extraction of signal intensity array,the probability distribution of signal strength,the effect of mobile robot orientation to signal reception,the interplay amount different AP,the relationship between distance and signal intensity and the influence of human body to WiFi signal;(2)This thesis on the basis of naive Bayes positing method presents a method called the probabilistic locating method based on improved particle filter.Its innovation focuses on 3 aspects:firstly,in the offline data acquisition stage the kernel density estimation is employed to fit the probability distribution of signal strength.Compared to Gauss distribution model,this method describes the regularity of WiFi signal distribution more well and truly,and enhance the accuracy of fingerprint matching.Secondly,in the online position estimation phase,this method introduces the maximum strength minimum variance strategy to select AP according real-time WiFi signals,it not only reduces the wastes time of algorithm,but also improve the anti-interference performance of collected WiFi signals effectively.Finally,a modified particle filter is presented to carry out matching refined location for mobile robot.This method considers the robot velocity,course angle information and the constraints of indoor environmental structure,correct the estimated value of robot position dynamically,then further improved indoor location precision for mobile robot.(3)The experimental platform of indoor localization for mobile robot in WLAN is built,and abundant contrast experiments amount the probabilistic locating method based on improved particle filter,Naive Bayes positing method and K-nearest algorithm are carried out.Results show that the probabilistic locating method based on improved particle filter respectively achieve 63.8%,72.8%,83.8%,92.4%,95.6%and 98%accuracy rate in 1m,2m,3m,4m,5m and 6m positing precision,and obviously increases the degree of positing accuracy.