| 摘要 |
Invention herewith described refers to the procedure for processing the measured received signal strength (so called RSSI) during the process of forming a "radio map" that will be used in localization of devices in the wireless communication environment. The m measured RSSI vectors for each reference point in the radio map (511) will be grouped in k clusters and centres of each cluster, called prototypes, become the representative RSSI vectors of the given reference point in the radio map (512). In this way, the data compression in the radio map is achieved as well as detection of the most representative values of the received signal strength for each point of the space. With that any reference point of the radio map is defined by k vectors of prototypes of the received signal containing strengths of the signal (RSSI) received from the base stations. For clustering m vectors of the measured RSSI signals, the algorithm for unsupervised learning is used, which is characterized by its ability that the output vectors preserve the input data topology. The measurement procedure of RSSI signals (500), their processing (510) and storage (520) is repeated for all reference points of the previously defined indoor space. The forming process of a "radio map" is an integral part of the localization method based on the impression of the received signal strength and is applicable at all widespread standards for the wireless communication, IEEE.802.15, IEEE.802.11, IEEE 802.16, etc.
|
