Research On Indoor Wireless Positioning Technology Based On Location Fingerprint

With the popularity of smartphones and the rapid development of the Internet, the demand of location-based services(LBS) grows rapidly. Recently, LBS are widely used in many social activities and have already won the acceptance and praise of the users. The key of LBS is how to obtain the position information, which can also be called location technology. The positioning precision affects the application’s effect of LBS directly. Although the satellite navigation and positioning technology is relatively mature and can meet the most demands of the outdoor location service, the application can not have a same performance when indoor, so the indoor location technology has attracted more and more attention in recent years. Because of the need for extra special hardware, it is difficult to apply and popularize widely for the indoor location technology due to the high cost of location and inflexible application. However, the WLAN location technology can use the existed WLAN infrastructures, these kinds of methods satisfy the localization as a mobile terminal using some special software instead of another dedicated devices, due to the fact that it can almost satisfy most of the accuracy demand of indoor position applications and has the advantage of lower cost, so recently the WLAN indoor position technology has become the most popular indoor position technology.The location fingerprint method has some advantages such as higher accuracy, stronger interference resistance and lower cost. All of those advantages make it becomes the most common used location method. Nevertheless, since the time-variation of RSS will decrease the position accuracy and lead to the fingerprint database outdated, this thesis proposes corresponding solutions, through experimental comparison and analysis, the validity and practicability of the proposed methods are verified. The main contents of the thesis are as follows:(1) In order to reflect the receiving signal distribution accurately, the thesis proposes a location fingerprint algorithm based on the blend double-peak Gaussian distribution. The algorithm uses different models to fit different RSS probability distributions, so the location fingerprint database is high reliable. Referencing the application of overlapping degree in map matching and skin color detection, the method measures the similarity of the two different location fingerprints by the overlapping degree of those two fingerprints. Then the K reference points which have the biggest similarity with the target point are give different weight respectively according to their similarities. Finally, we can estimate the position of the target point by the weighted average method.(2) Regarding the problem that the location fingerprint become outdated due to the time-variation of RSS, the thesis proposes a method to update and maintenance the fingerprint database by the users’ feedback information. The feedback information includes the location and the collected AP signal strength. We score the AP which is included in the location fingerprint according to the scoring mechanism, and update the fingerprint database by deleting the closed AP and adding the opened AP. For the situation that positioning results may be different from the actual location, we allow the users to revise the location results by themselves and judge the correction with the clustering detection method, the experimental results show that by using feedback information to update fingerprint database, the method is more feasible than artificial update or do nothing.