RSSI-BASED MODIFIED K-NEAREST NEIGHBORS ALGORITHM FOR INDOOR TARGET TRACKING
Since the global positioning systems have been faced with high error rate in indoor target tracking, indoor target tracking methods have been developed. The high error rate is due to multipath propagation, which is one of the main problems of indoor tracking systems.K-nearest neighbors algorithm is used in this paper to reduce the complexity and raise the accuracy of indoor tracking system in comparison to different indoor tracking methods with high complexity that makes the system too expensive. This algorithm works based on Radio Signal Strength Indication; also some changes have been applied to this algorithm so we named it ‘modified K-nearest neighbors algorithm’. The changes include a novel weight vector that depends on the strength of signal which increases this scheme’s accuracy. At the same time, to mitigate the effects of multipath propagation, multichannel method is used and its influence on accuracy is shown in the simulation. Formulation in this paper is much simpler and different calibration is done to increase the precision. The simulations prove this scheme as a high exact indoor tracking scheme which has about one meter tracking error in a 100 square meters environment.
fixed position transceivers, multipath propagation, indoor tracking, radio frequency signal.