Research On Fingerprint-based Wireless Indoor Location Algorithm

In recent years, more and more scenes need indoor location service. GPS can provide nice location server outside, but it can not be used indoor and the weather could affect it. So it needs other technology to get indoor location service. The development of wireless sensor network provides an indoor location method, which is much accurate and much convenient, wireless indoor location service. The accuracy of the method, which is based on received signal strength, would be influenced by the propagation of the signal, which is due to refection, transmission, diffraction and multipath fading. In order to solve the problem mentioned above, the paper puts forward a new indoor location algorithm (AFILA, Advance Fingerprint-based Indoor Location Algorithm), which is based on the fingerprint-based indoor location algorithm. Compared with other algorithm, it gets much better result without increasing load. The contribution of the paper is as follows:(1) To process the source data:because the source data is affected inside, it needs to get reliable source data to improve the accuracy. The algorithm smoothes the source data with Kalman. With the collision of the wireless link, the source data could be lost. The algorithm fixes it with the method of compensation. To the equipment with high frequency, the algorithm uses the average smoothing method to reduce the wave of source data.(2) To improve the location method:the algorithm is based on fingerprint. The paper explains it from selecting APs, solving neighbors, determining result. The paper also explains the algorithm with ANN and the problem of it.(3) To optimize track:the location track is not matching with the real track because of the wave of the source data during moving. The algorithm solves it with speech filter and barrier filter.At last, the experiments show the improvements advance the indoor location service and also can get the location track which is much closer to walking track.