Research Of A Mixed Fingerprint Indoor Positioning Algorithm Based On RSSI And Geomagnetic

With the increasing popularity of wireless networks and smart phones, indoor positioning technology has been rapidly developed. Since location information services are attracting more and more attention, accurate location information plays an important role in practical applications, such as the tracking and monitoring of patients, the supervision of key prisoners, the security of museum exhibits, the location of the commercial shopping cart and consumer navigation. In indoor positioning technology, WIFI-based RSSI positioning technology has high band width, high-speed, high coverage, low cost, which attract more and more people’s attention. And smart phone also integrates a sensor associated with the Earth’s magnetic field. Therefore, the studies the geomagnetic field RSSI and hybrid positioning technology in this paper has great practical significance.In this paper, based on RSSI and the geomagnetic field of indoor positioning technology, I have studied the fingerprint proposed hybrid location algorithm based on RSSI and indoor geomagnetic field. The work has been done as follows:First, in the offline construction of fingerprint library, since RSSI is susceptible of environmental interference, a combined calculation of hybrid and geomagnetic field has been proposed. The main ideas are following. First, collect the signal strength value of AP and RSSI from sampling point, and the three-dimensional coordinates of the geomagnetic field from the same sampling point, mix them together to form a combined fingerprint Then, based on the properties of the geomagnetic field and RSSI, use cluster analysis to specify the mixed fingerprint groups, do PCA analysis to each group of fingerprint component, and establish the comprehensive evaluation model. Finally, filter the fingerprints which get higher score than average, re-build a new fingerprint database. Experiments show that the new algorithm can effectively improve the positioning accuracy.Second, in online positioning of fingerprint database, the realization of a prototype positioning system has improved the positioning algorithm. When matching the fingerprint, we use the strongest/second-strongest AP selection method to screen fingerprints. When estimating the position, according to the traditional weighted KNN algorithm, we use K nearest neighbors signal sampling points and the Euclidean distance of mobile terminals as reference weights. We can ignore the fact that in practice environment, signals are easy to shack, which make the Euclidean distance of the nearing sampling points similar. We use the strongest and second strongest AP signal strengths as reference weights, to improve positioning accuracy, achieving the positioning of mobile terminals.