Research On Indoor Location Technology Based On Crowdsourcing

The WiFi fingerprint location method is the most popular in indoor location technology.This method needs to establish fingerprint database of received signal strength(RSS)from different wireless access points in the area to be located.However,in the actual indoor environment,there will be some problems,such as furniture position moving,people moving,the displacement or missing of AP and so on.This requires a lot of manpower and material resources to maintain and update the fingerprint database regularly.Therefore,using crowdsourcing idea and pedestrian dead reckoning(PDR)technology to improve the maintenance and update efficiency of WiFi fingerprint database has become a research hotspot.At present,the PDR technology requires the mobile phone to maintain a fixed horizontal attitude in the measurement.However,in the crowdsourcing scenario,mobile phone posture is diverse,which leads to the current crowdsourcing fast database building method is not practical.To solve this problem,this paper proposes an SVM based mobile attitude classification method after analyzing the indoor mobile phone usage scenarios.Especially for the direction information of pedestrian which is difficult to judge,this paper proposes a method of turning motion detection based on principal component analysis.The experimental results show that the method is effective in pedestrian attitude and mobile phone attitude detection.After getting the mobile phone attitude and user attitude at different times,this paper proposes a landmark assisted algorithm for pedestrian heading updating,which effectively integrates the information of AP at the turning point and the built-in inertial sensor of mobile phone.This paper proposes a turning point correction algorithm based on particle filter,which can effectively calculate the indoor pedestrian trajectory.At the same time,this method does not require the mobile phone to maintain a fixed attitude relative to the human body,which can effectively solve the problem of the limitation of mobile phone attitude in the traditional PDR method,and is suitable for indoor crowdsourcing scene.Finally,aiming at the problem of fingerprint database maintenance,this paper proposes an RSS data optimization algorithm based on Kriging interpolation.In the process of pedestrian movement,the mobile phone automatically collects RSS.Through the time stamp,the RSS label is corresponding to the specific location point to complete the fingerprint database maintenance.This paper also explains the framework of fingerprint database maintenance based on crowdsourcing in detail,and proves that this method can not only ensure the accuracy of fingerprint database location,but also effectively reduce the human and material resources needed for its maintenance,which has strong practicability.