Research On Indoor Localization Method Based On VIO Algorithm And Wi-Fi Fingerprint

With the coming of the era of Intenet of Things(IoT)and Big Data,location services are becoming more and more common in our lives: mobile phone navigation,medical aid,sports and fitness,precision advertising and art guides...all these domains require the support of location services.The Beidou Positioning System based on satellite and the GPS technology can highly meet the user's needs in the outdoor environment.However,in the indoor environment,the satellite signal is attenuated and scattered which leads to the low accuracy of GPS.In order to satisfy the needs in indoor environment,various indoor positioning methods are high-lighted in recent researches.We first introduce those indoor positioning technology.Under the analyse of GPS technology,three main methods are presented: indoor localization based on wireless networks,on computer vision and based on inertial navigation.There are two types of algorithms based on wireless network: geometric localization and fingerprint localization.Since the geometric localization needs to obtain the coordinate values of each access points and it needs to solve the clock difference between access points and localization device,fingerprint localization is easier to use in practice.The inertial navigation method can directly use various types of motion sensors without any additional access point.It has good concealability and high accuracy in a short distance.However it is easy to generate drift which results in cooperation between computer vision and inertial navigation.We focus on the VIO algorithm which combines computer vision and inertial navigation.This algorithm can improve the stability and accuracy of the localization.Based on the VIO algorithm,we analyze its drift in the long-distance environment and the positioning error in similar scenarios.In order to solve these two problems,we design an indoor localization system based on VIO algorithm and Wi-Fi fingerprint.By dividing the large area into small area,the Wi-Fi fingerprint is used for coarse localization and VIO is used for precise localization.As a result,the system optimizes the drift problem at the system level.Experiments show that the proposed system has better localization accuracy and stability in large-scale localization.Experiments also prove the possibility that using the proposed system a nation-scale Wi-Fi Fingerprint Database can be constructed so as to realize a large-scale high-accuracy indoor localization.