Research On The Key Technologies Of Indoor Location Based On WiFi

With the development of wireless location technology, indoor location has become a hot spot. Indoor location technologies at present mainly include optical tracking, A-GPS, ultrasonic and WiFi, etc. Indoor location technology based on WiFi has many advantages such as wide coverage area, high-speed information transmission, low cost, etc.WiFi is paid more attention by people.This subject comes from the joint research project of Dalian University of Technology and OKI Corporation-WiFi-based indoor positioning,the aim is to achieve 3 meter positioning accuracy under the indoor environment.Because under indoor environment signal fluctuation is very big, causes the positioning system's accuracy of the preliminary project is not high, cannot satisfy the 3 meter positioning error range to meet 75% requirements.In order to increase the pointing accuracy, can do reseach in following aspects:RSSI (Received Signal Strength Indicator) measurement, RSSI signal processing, indoor propagation models,positioning algorithm.This paper foucuses on three aspects:RSSI signal processing, indoor propagation model and positioning algorithmFirst, According to the state of positioning terminals, we divided the signal processing methods into static signal processing and dynamic signal processing. We manage to discover the distributed characteristics of signal strength in the indoor environment by analyzing the statistic characters of indoor signal strength, and propose one kind of static signal processingthe logarithm model. Then, given the disadvantages of existing dynamic data processing method, we propose a method of corrected outliers and weighted filtering.Second, we make simulations of some typical propagation model and compare their simulation results in current experimental conditions. The Motley-Keenan model is the best. We further improve Motley-Keenan model. Then,in the analysis of exiting location algorithm based on propagation model, we propose two methods of maximum likelihood estimate. We make simulation to compare locating accuracy and performance of these two algorithms.Last, the above signal processing, model and location algorithm are applied to the locating system in a real WiFi environment. The locating results are the distribution of cumulative error is 75% within 3.02-meter locating error. It fulfilled the project requirement.