Research On The Key Technology And Application Of Indoor Location Based On WLAN

As the rapid development of the wireless network technology, location-based services arecurrently being introduced to the various commercial applications. WLAN-based positioningtechnology could take full advantages of the existing wireless local area networks, which results inthe benefit of low cost. However, due to the complex propagation characteristic of radio wave inthe indoor environment, there is a great challenge in obtaining location information.Nowadays, the researches in WLAN-based indoor positioning technology mainly concentrateon personnel/asset positioning, moving target tracking, location navigating and low powerpositioning-label designing. Based on the existing theory of personnel/asset positioning andmoving target tracking, the paper proposes the positioning and tracking algorithms, which areapplied to positioning system prototype. By adopting the proposed algorithms, the responseaccuracy of the positioning system can be highly improved to meet the requirements.First, focused on the fact that the inconsistency of received signal strength (RSS) samplepopulation distributions would produce large positioning errors under indoor wireless local areanetwork (WLAN), an indoor positioning algorithm based on skewness-kurtosis testing is proposed.By using the testing method of skewness and kurtosis, the algorithm tests whether the RSS samplescome from the normal population or not. The probability density functions of the samples refusingnormal hypothesis are approximated by kernel function, which accurately estimates the populationdistribution of the RSS sample at the offline training stage. And the position of the target iscomputed by Bayes rule at the online stage. Experimental results show that the proposed algorithmleads to15percent improvement compared with the previous methods. Moreover, the proposedalgorithm can significantly reduce the workload of the off-line phase at the same positioningaccuracy.Second, the correlation between the previous measurement and the current position of thetarget is taken into account in the process of tracking. A new indoor target tracking algorithm basedon Kalman filter is proposed. Focused on the problem that tracking smoothness and accuracy arechanged with the ratio between the observed variance and the state variance, this paper mends theproposed algorithm. Due to the mended algorithm, the Kalman parameter values are adjustedaccording to the walking direction. Experimental results demonstrate that the tracking accuracy ofthe mended algorithm has been improved the by11%compared with the original.Last, the proposed location and tracking algorithms are applied to the implementation of thepositioning system prototype. And three application graphical display interfaces, including personnel/asset location, regional guard and target tracking are demonstrated. Furthermore, thedesigned indoor positioning system in this paper needs not to change the existing WLAN hardwaredevices. Due to the easily implement, it has the sustained market value for the long term.