A Novel Adaptive Radio-Map for RSS-Based Indoor Positioning

Wireless positioning in indoor environment is growing rapidly with the development of location based services. Among these technologies, fingerprinting positioning has been widely used, and its performance is satisfactory in precision requirement. The radio-map is the key to fingerprinting positioning. However, once the application scenarios changed, and the radio-map doesn't get updated in time, the performance will be seriously affected. In this work, we develop a novel adaptive radio-map for RSS-based indoor positioning. Firstly, we collected the signal characteristics of a plurality of reference points and checked nodes in the scene to create a static radio map. Then the static frequency was calculated based on the scene map for each reference point path-loss parameters. According to the reference points, path-loss parameters will be clustered and the location area was divided into a plurality of sub-regions. Finally, we checked node robust linear regression of RSSI fingerprints, including the reference points' in each sub-region and the check nodes' in the sub-region. Experimental results show that the algorithm can effectively improve the RF map change over time and the dynamic changes in the positioning accuracy of the indoor environment.