Passive Radar at the Roadside Unit to Configure Millimeter Wave Vehicle-to-Infrastructure Links

TL;DR

This paper introduces a passive radar approach at roadside units to efficiently configure millimeter wave vehicle-to-infrastructure links, reducing training overhead and improving link reliability in dynamic vehicular environments.

Contribution

It proposes a novel passive radar receiver that estimates spatial covariance from automotive radars to assist mmWave link configuration, with a covariance correction strategy and similarity metric.

Findings

Covariance correction enhances radar and communication APS similarity.

Radar-assisted method reduces training overhead significantly.

Effective in non-line-of-sight scenarios.

Abstract

Millimeter wave (mmWave) vehicular channels are highly dynamic, and the communication link needs to be reconfigured frequently. In this work, we propose to use a passive radar receiver at the roadside unit to reduce the training overhead of establishing an mmWave communication link. Specifically, the passive radar will tap the transmissions from the automotive radars of the vehicles on the road. The spatial covariance of the received radar signals will be estimated and used to establish the communication link. We propose a simplified radar receiver that does not require the transmitted waveform as a reference. To leverage the radar information for beamforming, the radar azimuth power spectrum (APS) and the communication APS should be similar. We outline a radar covariance correction strategy to increase the similarity between the radar and communication APS. We also propose a metric to compare the similarity of the radar and communication APS that has a connection with the achievable rate. We present simulation results based on ray-tracing data. The results show that: (i) covariance correction improves the similarity of radar and communication APS, and (ii) the radar-assisted strategy significantly reduces the training overhead, being particularly useful in non-line-of-sight scenarios.