How to Mobilize Mmwave: A Joint Beam and Channel Tracking Approach

Jiahui Li, Yin Sun, Limin Xiao, Shidong Zhou, Ashutosh SabharwalICASSP 2018 · DOI

TL;DR

This paper introduces a novel algorithm, RBCT, for simultaneously tracking the beam direction and channel coefficients in millimeter-wave (mmWave) communication systems. This is crucial for maintaining reliable connections with fast-moving mobile devices, and RBCT achieves this with high accuracy, speed, and low pilot overhead, outperforming existing methods.

What Problem Does This Paper Address?

Millimeter-wave (mmWave) communication is a promising technology for future wireless systems due to its high bandwidth. However, its effectiveness is challenged by the narrow beams required to overcome signal attenuation, making it difficult to maintain connections with fast-moving users. Traditional methods for acquiring and tracking these beams and the associated channel characteristics require a significant number of pilot signals, which becomes impractical with high mobility. Existing tracking algorithms often assume known channel coefficients or don't optimize the training beamforming vectors, leading to suboptimal performance. This paper addresses the critical need for an efficient and accurate method to jointly track both the beam direction and channel coefficients in dynamic mmWave environments, even with limited pilot overhead.

Methods

The paper proposes a two-stage Recursive Beam and Channel Tracking (RBCT) algorithm.

  • Stage 1: Coarse Beam Sweeping: This stage uses a set of $M$ pilot signals to perform an initial, broad sweep to estimate the beam direction and channel. It employs an orthogonal matching pursuit method to obtain an initial estimate of the beam direction.
  • Stage 2: Beam and Channel Tracking: This stage refines the estimates from Stage 1. It utilizes two pilot signals per time-slot and a recursive update mechanism. The core of this stage is a joint optimization that simultaneously updates the beamforming vectors and the estimates of the channel coefficient and beam direction. The algorithm is designed to be recursive, using past estimates to inform current updates, and is motivated by a maximum likelihood estimation approach. The paper also derives the Cramér-Rao Lower Bound (CRLB) for beam tracking in static scenarios and proves that their RBCT algorithm can converge to this bound with high probability.

Datasets

Not discussed in the paper. The evaluation is based on simulations.

Key Insights and Ideas

  • Joint Tracking is Key: The paper highlights that effectively tracking mmWave links in dynamic scenarios requires simultaneously estimating both the beam direction and the channel coefficients. Decoupling these tasks leads to suboptimal performance.
  • Optimized Beamforming for Tracking: Simply using existing beamforming strategies for tracking is insufficient. The RBCT algorithm optimizes the training beamforming vectors specifically for the tracking task, leading to improved accuracy.
  • CRLB Convergence: Demonstrating that the proposed RBCT algorithm can converge to the theoretical minimum Cramér-Rao Lower Bound (CRLB) for beam tracking in static scenarios provides strong theoretical backing for its optimality.
  • Robustness to Mobility: The RBCT algorithm shows remarkable performance even at high angular velocities, a significant challenge for mmWave communications.

Key Results

  • High Accuracy: The RBCT algorithm achieves significantly lower Mean Squared Error (MSE) compared to existing algorithms, converging close to the minimum CRLB in static scenarios.
  • Fast Tracking Speed: The algorithm supports higher angular velocities than reference methods. For instance, it can track a mobile moving at 5.45 degrees per second.
  • Low Pilot Overhead: RBCT achieves its performance with a minimal number of pilots (e.g., 10 pilots per second).
  • High Channel Capacity: The algorithm enables achieving over 95% of channel capacity even under challenging dynamic conditions.

Limitations and Future Work

  • Static vs. Dynamic: While the paper derives CRLB for static scenarios, the RBCT algorithm is designed for dynamic environments. The theoretical analysis of convergence to the CRLB in dynamic scenarios is more complex.
  • Hardware Assumptions: The paper assumes ideal phase shifters and omnidirectional antennas. Real-world hardware imperfections might affect performance.
  • Multiple Paths: The paper focuses on tracking a limited number of distinct paths. Handling a very large number of scattering paths might require further extensions.
  • Future Work: The paper suggests exploring the algorithm's performance with more complex channel models and hardware impairments. Further optimization of the step-size parameters for different mobility scenarios could also be a direction.

Summary updated

BibTeX
@inproceedings{li2018how,
  author = {Jiahui Li and Yin Sun and Limin Xiao and Shidong Zhou and Ashutosh Sabharwal},
  title = {{How to Mobilize Mmwave: A Joint Beam and Channel Tracking Approach}},
  booktitle = {IEEE International Conference on Acoustics, Speech and Signal Processing},
  year = {2018},
  doi = {10.1109/icassp.2018.8461760},
}
Figure