Abstract:

With the rapid development of wireless local area network(WLAN) technology, an important target of indoor positioningsystems is to improve the positioning accuracy while reducing theonline calibration effort to overcome signal time-varying. A novelfingerprint positioning algorithm, known as the adaptive radio mapwith updated method based on hidden Markov model (HMM), isproposed. It is shown that by using a collection of user tracesthat can be cheaply obtained, the proposed algorithm can takeadvantage of these data to update the labeled calibration data tofurther improve the position estimation accuracy. This algorithm isa combination of machine learning, information gain theory andfingerprinting. By collecting data and testing the algorithm in a realisticindoor WLAN environment, the experiment results indicatethat, compared with the widely used K nearest neighbor algorithm,the proposed algorithm can improve the positioning accuracy whilegreatly reduce the calibration effort.