Location based fingerprinting techniques for indoor positioning

Abstract

Indoor positioning is a popular and a novel researched area mainly in terms of its vast applications on providing location based services. The requirement for indoor positioning has mainly occurred due to the infeasibility of the Global Positioning System (GPS) in indoor environments which consisting of many obstacles and having non-line-of-sight. As there is no exact method applicable globally For indoor positioning, several techniques are employed considering the environments they are being used. Among these techniques, fingerprinting provides several propitious features for indoor environments having obstacles and contains severe multipath channels. Our focus of study is to identify the applicability of audible sound for location based fingerprinting which could be augmented in environments with non-line-of sight and multipath wave propagating conditions, with low cost and a simple implementable solution. The study was conducted to locate the position of a receiver, by an acoustic signal generated from a fixed source transmitter at a controlled environment. Initially, the environment is divided into several grid points in order to generate a database comprising the location data. A large number of realizations were obtained at each of the grid points in order to create a robust database. The location is estimated via comparison of the feature vector of signatures generated at the receiver, with the pre-generated database. More focus was given on selecting suitable parameters to be used as unique signatures of a given acoustic signal of very short duration and to explore model based and non-model based methods on determining the position with minimum error.