Mobile Location in MIMO Communication Systems by Using Learning Machine

The traditional mobile location systems are mainly based on trilateration/multilateration techniques. In wireless MIMO communication systems which utilize antenna array at both transmit and receive sides, the redundancy of multipath signals can be exploited to extract more parameters such as angle-of-arrival, angle-of-departure and delay-of-arrival using advanced array signal processing techniques. In this paper, based on estimated multipath signal parameters in wireless MIMO communication systems, we propose a novel machine learning approach to determine the position of mobile targets using only one base station. This approach adopted the nearest neighbor regressor as the learning machine to estimation the highly nonlinear relationship between the multipath signal parameters and the position of mobile target. The simulation results have demonstrated the viability of the proposed methodology. This solution breaks the bottleneck of conventional mobile positioning systems which have to require multilateration of at least three base stations.