Research On PDR Positioning Technology Of Indoor Spacius Scene Based On Single Base Station Correction

Now there are various indoor positioning technologies,including ultra-wideband,Wi-Fi,visible light,etc.Each technology has its advantages and disadvantages.How to design a high accuracy positioning system as well as low cost and low computation is the main task in the field of indoor positioning research.Considering Pedestrian Dead Reckoning(PDR)keep high accuracy in short period,this paper proposed a smartphone-based indoor positioning system,in which PDR is intermittently corrected by single BLE beacon.In this way,sparse BLE beacon deployment is enough.The specific research contents and innovations are as follows:(1)Indoor positioning system for corridor scene: Due to integration,PDR's positioning error will grow up with the time.While traditional fingerprint and trilateration require multiple beacons.Therefore,this paper proposed that PDR be corrected by Bluetooth RSS peak so that the accumulated error can be eliminated.Besides,the traditional RSS peak detection method is based on sliding window,which may miss or mistake the real RSS peak.Hence this paper detects the RSS peak from the global perspective in order to improve the robustness of RSS peak detection.What's more,taking into account that there is a certain delay in peak detection,PDR is used to estimate the pedestrian's walking displacement,and then the displacement is utilized to compensate above mentioned positioning result.Finally,traditional step length estimation method requires offline data collection and train progress to determine the step length model parameters,thus this paper proposed an online adaptive step length estimation method which online determine the suitable parameters for current user.(2)Indoor positioning system for spacious scene: PDR is a positioning technology that doesn't depend on external information;Hence it is not easily affected by the environment.Compared with the corridor scene,there is no wall constraint in the open scene,so on the basis of the RSS peak positioning in the previous section,further integrate the heading information to locate.Besides,this paper further combines Manhattan distance with orientation information to estimate the step model parameter,so that the curve trajectory has no effects.