Research On Mobile Robot Localization Technology Based On WIFI And Vision

In recent years,with the increasing popularity of artificial intelligence,mobile robot technology has gradually become an important basis for measuring the advanced level of a country,and the effect in this research field in various countries has been remarkable.Most mobile robots must use indoor positioning technology to achieve indoor autonomous navigation.At present,WIFI positioning and visual simultaneous localization and mapping(visual SLAM)are widely used in indoor positioning technology.In a large indoor environment,although WIFI positioning has high real-time performance,simple deployment,and low cost,the positioning accuracy is low.Although visual SLAM has high positioning accuracy,it has the problems of low real-time performance and poor robustness.Therefore,in order to obtain an indoor positioning system with high real-time performance,strong robustness,and high positioning accuracy in a large indoor environment,this paper proposes an indoor positioning system that integrates WIFI and visual SLAM.Its main work includes the following aspects:1.WIFI signal preprocessing and positioning.This article expatiates the basic principles of WIFI positioning.In view of human activities,poor hardware performance and other factors that cause the low quality of the collected WIFI signal,this paper uses information gain as the evaluation standard to screen the access point(AP),and uses the Grubbs criterion to remove the abnormal value of the WIFI data.Finally,for the problem of low positioning accuracy of traditional K-nearest neighbor(KNN)fingerprint positioning algorithm,an adaptive KNN algorithm based on information gain is used to improve positioning reliability.2.Visual SLAM positioning and optimization.The article explains the basic principles of visual SLAM positioning.One of the most significant technologies in visual SLAM,the image matching algorithm,is described and optimized in detail.There is shortcoming that the mismatch rate of the traditional ORB matching algorithm is slightly higher.This article proposes to proposes a new feature descriptor of adjacent gray values to improve the matching rate of feature points.3.WIFI and visual SLAM fusion positioning technology.To establish the mapping relationship between the WIFI map and the visual SLAM map,this paper proposes the landmark method to realize the fusion of the two maps.Meanwhile,the single WIFI positioning has the problems of low positioning accuracy,and the single visual SLAM positioning has the problems of long calculation time and higher positioning failure rate.Therefore,this paper adopts the fusion of WIFI and visual Bag-of-words technology in visual SLAM to achieve rough positioning.Using rough positioning as a constraint,accurate positioning is achieved through visual SLAM.Experiments show that the fusion indoor positioning system proposed in this article has significant improvements in percent correct,accuracy,time and other performances.There is a lot of value that people research positioning in large indoor environments.