# Adaptive Beam Tracking with the Unscented Kalman Filter for Millimeter Wave Communication

arXiv: https://arxiv.org/abs/1804.08640 · HTML: https://mlsanity.com/paper/1804.08640/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 23 Apr 2018.

## Built on by (most cited first, 3 of 3)

- [Deep Learning-based Beam Tracking for Millimeter-wave Communications under Mobility](https://mlsanity.com/paper/2102.09785/) (2021)
- [Zero-Shot Adaptation for mmWave Beam-Tracking on Overhead Messenger Wires through Robust Adversarial Reinforcement Learning](https://mlsanity.com/paper/2102.08055/) (2021)
- [GPS-Aided Deep Learning for Beam Prediction and Tracking in UAV mmWave Communication](https://mlsanity.com/paper/2505.17530/) (2025)

## Abstract (arXiv)

Millimeter wave (mmWave) communication links for 5G cellular technology require high beamforming gain to overcome channel impairments and achieve high throughput. While much work has focused on estimating mmWave channels and designing beamforming schemes, the time dynamic nature of mmWave channels quickly renders estimates stale and increases sounding overhead. We model the underlying time dynamic state space of mmWave channels and design sounding beamformers suitable for tracking in a Kalman filtering framework. Given an initial channel estimate, filtering efficiently leads to refined estimates and allows forward prediction for higher sustained beamforming gain during data transmission. From tracked prior channel estimates, adaptively chosen optimal and constrained suboptimal beams reduce sounding overhead while minimizing estimation error.