A hybrid system integrating signal analysis and probabilistic neural network for user motion detection in wireless networks

This paper proposes a novel user motion detection (UMD) model called PNN-MSMD and its application on seamless handoff. The PNN-MSMD uses a novel signal analysis model to identify the movement of a mobile device without using a location system. The PNN-MSMD integrates multiple signal processing sensors with probabilistic neural network (PNN) to analyze the received signal strength (RSS) of a mobile device. The PNN uses the variation of RSS derived from the sensors as input training data to learn the moving patterns and generate the classification rules for detecting the motion state of a mobile device. The detection results can help mobile devices to enhance handoff processes. Computer simulations show that the proposed PNN-MSMD based handoff algorithm performs better than four traditional handoff algorithms and two motion detection based handoff algorithms. The PNN-MSMD method can save up to 82.56% power consumption and reduce up to 42.51% number of handoffs.