A Stochastic Framework of Millimeter Wave Signal for Mobile Users: Experiment, Modeling and Application in Beam Tracking

In this paper, we propose to employ the Markov jump linear system to model the characteristics of millimeter wave (mmWave) signals, such as signal strength and angle-of-arrival (AoA), for mobile users. We integrate the spatial correlation between line-of-sight (LoS) and none-line-of-sight (NLoS) beams to the stochastic framework for better modeling mmWave. To handle the channel dynamics in mmWave communications, we derive a Markov gain-state channel model based on channel measurements in an indoor environment from our field measurement. The proposed stochastic model is used to estimate the transition probability of each channel. Then it is integrated into the Markov jump linear system framework to construct the linear estimator for channel variables, such as AoA and channel gain. The simulation result demonstrates the outperformance of the proposed framework against the previous Kalman-filtering-based tracking algorithm.