K-NEAREST NEIGHBOR AND POSSIBILISTIC C-MEANS GROUPING BASED MIXED WIRELESS INDOOR POSITIONING METHOD CAPABLE OF ACCURACY OF A DATABASE SEARCHING PROCESS

摘要

PURPOSE: A KNN(K-Nearest Neighbor) and PCM(Possibilistic C-Means) grouping based mixed wireless indoor positioning method is provided to generate a database of a wireless fingerprint type by measuring a signal noise ration received from access points.;CONSTITUTION: A plurality of standard points collects first propagation characteristic value data. A database is built with the first propagation characteristic value data in a wireless fingerprint mode(S110). A plurality of access points receives a second propagation characteristic value at a test point. The standard point which has the first propagation characteristic value data corresponding to the second propagation characteristic value is searched(S120). The standard point is searched using a KNN/PCM(K-Nearest Neighbor/Possibilistic C-Means) mixed method.;COPYRIGHT KIPO 2012