Research On Key Technology Of Wireless Indoor Positioning

In recent years,with the rapid development of mobile communication technology,people's demand for location-based services is increasing.Wireless indoor positioning technology has increasingly become the hotspot of current research.In general,the research on wireless indoor positioning includes two main directions:one is in line-of-sight(LOS)environment,which is a strict assumption,to further improve the localization accuracy,the other one is in non-line-of-sight(NLOS)environment,which is relatively difficult to locate,to come up with a feasible positioning scheme with an accepted positioning accuracy.First of all,the research background of wireless indoor positioning technology and the current research status at home and abroad are summarized.The main technical bottlenecks which restrict the further improvement of positioning accuracy are discussed.Secondly,this paper reviews the existing wireless indoor positioning technology.The physical layer technology such as ultra wideband(UWB),Bluetooth,WiFi(WIreless-Fidelity),as well as the location algorithm such as Time-of-Arrive(TOA),Time-Difference-of-Arrival(TDOA)?Received Signal Strength Indicator(RSSI)are introduced.Thirdly,a high-precision multiple-antenna indoor positioning system based on chirp signal is proposed,aiming at LOS scene.The proposed scheme reduces the average positioning error to 5 centimeters,and does not require clock synchronization between base stations.Specifically,the CRLB(Cramer-Rao low bound)of proposed algorithm is discussed.Then,from the aspects of tag,base station and server,this paper introduces the implementation of the proposed algorithm in two-dimensional(2D)dual-antenna scene,and extends it to multi-antenna positioning system in three-dimensional(3D)scene.Finally,this paper studies the indoor location algorithm based on machine learning for NLOS environment,and proposes two improved schemes for traditional kNN algorithm:a progressive indoor localization algorithm based on fully connected neural network,a modified kNN location algorithm based on dynamically updating fingerprint library.Through the simulation,these two improved algorithms reduce the location error over traditional algorithm by a few percentage points.After that,a Chirp-TDOA location algorithm based on machine learning is envisaged,which integrates all the work in this paper into the "physical layer signal,ranging technology,location algorithm" framework.