Research On Acoustic Indoor Localization System Based On Smartphone

Nowadays,there are many kinds of key technologies of indoor localization system,including UWB(Ultra-Wide Band),Wi-Fi,magnetic,acoustic signal and so on,the technologies mentioned above have different advantages and disadvantages.How to realize the indoor localization system which can be accepted by the majority of users while taking into account the high precision positioning has been a hot spot and difficult problem in the field of indoor localization.This paper presents a kind of indoor localization scheme of acoustic signal based on smart phones.The scheme adopts the acoustic signal in the unaudible frequency band of human ear and makes full use of the microphone part,calculation function and communication function of the smartphones.It realizes high precision and robust localization in the indoor environment.The main work of this paper is as follows:(1)An indoor localization scheme based on acoustic signal is designed without strict clock synchronization.While most of the traditional acoustic signal indoor localization systems need accurate clock synchronization and it often leads to localization errors and additional overhead.The acoustic signal localization system proposed in this paper uses the acoustic signal collected by the main anchor to correct the synchronization error caused by the uncertain time delay of wireless signal,so it does not need strict clock synchronization.(2)Design process and implementation of this localization system is introduced.In this paper,firstly,the system framework and function flow are designed in detail,then the appropriate localization signal is selected by practical experiments.Finally,the system is designed and implemented based on the smartphone platform.Experimental results in actual scenarios show that the probability of the system localization accuracy within 15 cm is 97%,within 10 cm is 84%,and different models of smartphones have little effect on the localization results.(3)NLOS(Non-Line-of-Sight)probability estimation method for acoustic signals is proposed.Firstly,the acoustic signals collected during the same location period are segmented.Then the 9Ds channel features are obtained according to the channel characteristics.Finally,a method is proposed to determine the NLOS probability by comparing the 9Ds features of the signals from different anchors in the same period.(4)Time delay estimation based on NLOS probability is proposed.The accuracy of traditional generalized cross-correlation time delay estimation methods will decrease sharply in strong reflection environment or NLOS environment.To solve this,an adaptive threshold delay estimation based on the NLOS probability of acoustic signals is presented.The experimental results show the accuracy is greatly improved with the proposed scheme in LOS/NLOS environment.(5)An adaptive weight estimation based on NLOS probability is proposed.The indoor environment is mostly LOS/NLOS mixed environment indeed and the positioning error of traditional localization method turns out to be large.Thus,an adaptive weight estimation method combining NLOS probability of acoustic signals and channel delay information is proposed in this paper.The experimental results show that in the LOS/NLOS mixed environment,combining the time delay estimation method and weight estimation method proposed in this paper,the probability of the localization accuracy of WLM?P algorithm within 0.2m is 85%.