Particle filter based synthetic aperture reconstruction approach for real-time 3D wireless local positioning

In this paper we introduce a precise real-time 3D indoor localization concept using a combination of the MISAS (multilateral inverse synthetic aperture secondary radar) approach and particle filtering. The synthetic aperture approach is a promising locating technique with effective multipath suppression. However, the practical usage in real-time applications is limited by its huge computational load of conventional SAR reconstruction algorithms. Instead of evaluating the PDF (probability density function) of the target position in a grid-based approach, we utilize a set of weighted particles to represent the PDF. The particle number can then subsequently be reduced in the following iterations through a resampling algorithm. It is shown by simulation and experiments with a 5.8 GHz FMCW secondary radar with 140 MHz bandwidth, that 3D localization of a robot TCP (tool center point) in mm-range with a few hundred particles is possible even in a dense multipath environment. With our new approach the computational effort has been reduced to below 0.1% compared to a grid-based evaluation. Therefore, this novel extension of particle filtering gives an attractive solution for the real-time application of 3D MISAS localization.