Online Learning-based Waveform Selection for Improved Vehicle Recognition in Automotive Radar

TL;DR

This paper introduces a reinforcement learning method for automotive radar waveform selection that enhances target recognition by adaptively choosing optimal waveforms based on real-time data, improving classification performance.

Contribution

It presents a novel satisficing Thompson sampling approach for online waveform selection in automotive radar, enabling rapid learning and adaptation from measurement-level simulations.

Findings

Effective waveform strategies learned quickly in large candidate sets

Adaptive bandwidth and code selection improve classification accuracy

Method demonstrates robustness in interference mitigation scenarios

Abstract

This paper describes important considerations and challenges associated with online reinforcement-learning based waveform selection for target identification in frequency modulated continuous wave (FMCW) automotive radar systems. We present a novel learning approach based on satisficing Thompson sampling, which quickly identifies a waveform expected to yield satisfactory classification performance. We demonstrate through measurement-level simulations that effective waveform selection strategies can be quickly learned, even in cases where the radar must select from a large catalog of candidate waveforms. The radar learns to adaptively select a bandwidth for appropriate resolution and a slow-time unimodular code for interference mitigation in the scene of interest by optimizing an expected classification metric.