Meta-Learning for Image-Guided Millimeter-Wave Beam Selection in Unseen Environments

The use of alternate modalities, like images, for fast beamforming in the millimeter wave (mmWave)-band is being proposed to ensure high bandwidth connectivity in vehicular scenarios typically seen in the context of autonomous cars. Considering the dynamic deployment conditions, a car may encounter new environments which were not explicitly included in an apriori training dataset. In this paper, we propose to use the Model-Agnostic Meta-Learning (MAML) framework on the image data of the mmWave vehicle-to-infrastructure beam selection FLASH dataset, to overcome the generalization issues of a pre-trained model in unseen non-line-of-sight (NLOS) connectivity environments. MAML has additional advantages over traditional deep-learning techniques: (i) it uses a fraction of the data which, in turn, simplifies data collection and storage, and (ii) it results in equal or higher accuracy in optimal beam selection compared to the case when the new environment dataset is fully available during initial training. We show that our MAML implementation improves test accuracy of beam selection by up to 86% with fine-tuning when encountering an unseen NLOS environment compared to conventional supervised learning.

1. INTRODUCTION

Autonomous vehicles are poised to revolutionize transportation in the future. One direction towards realizing this vision involves relaying massive volumes of locally collected sensor data from the vehicle to a central cloud for optimizing routes and obstacle avoidance, in the order of 10 Gbps. Transmission in the millimeter-wave (mmWave) band makes such high data rates possible [1], though beamforming is required to overcome the high path loss in this band by channeling the radio frequency (RF) energy in narrow spatial lobes. We have previously used machine learning (ML) over contextual information from the environment, captured via images from a vehicle-mounted camera to enable beamforming faster than the standards-defined brute force approach. In this paper, we tackle the key problem of enabling our image-guided beamforming approach to work in environments not seen during training, while minimizing the costly overhead of new data collection.