MIMO fingerprinting-based particle filter for mobile positioning systems

This paper presents a particle filter processor for mobile positioning systems. The proposed particle filter employs a selective sampling scheme to reduce the required iteration of particle filtering and improve positioning accuracy. Moreover, a threshold IMH resampling algorithm is also proposed to enhance the hardware utilization rate and processing latency. This study also designs and implements the particle filtering processor with a flexible particle number and sampling threshold according different channel conditions. The proposed threshold IMH resampling not only improves the hardware utilization to more than 95% but also reduces the processing latency of the conventional IMH resampling algorithm. The chip implementation results show that the particle filter, when deployed in base-station, can track the location of 100 mobile users at 180 km/hr speed with less than 10-meter root-mean-square (RMS) error.