A Computationally Efficient EK-PMBM Filter for Bistatic mmWave Radio SLAM

Millimeter wave (mmWave) signals are useful for simultaneous localization and mapping (SLAM), due to their inherent geometric connection the propagation environment channel. To solve SLAM problem, existing approaches rely on sigma-point or particle-based approximations, leading high computational complexity, precluding real-time execution. We propose a novel low-complexity filter, based Poisson multi-Bernoulli mixture (PMBM) filter. It utilizes extended Kalman (EK) first-order Taylor series Gaussian approximation of filtering distribution, applies track-oriented marginal multi-Bernoulli/Poisson (TOMB/P) algorithm approximate resulting PMBM as (PMB). The filter can account different landmark types in radio multiple data association hypotheses. Hence, it has an adjustable complexity/performance trade-off. Simulation results show that developed greatly reduce cost, while keeps good performance user state estimation.