mmDrive: mmWave Sensing for Live Monitoring and On-Device Inference of Dangerous Driving

TL;DR

This paper presents mmDrive, a mmWave radar-based system that accurately detects and classifies dangerous driving behaviors in real-time without relying on visual data, achieving over 95% accuracy.

Contribution

The paper introduces a novel Fused-CNN model utilizing mmWave radar data to detect and classify dangerous driving actions, demonstrating high accuracy in real-world tests.

Findings

Achieves over 95% accuracy in dangerous driving detection

Identifies unique range-doppler patterns for 9 dangerous actions

Validates effectiveness through real-world experiments with volunteers

Abstract

Detecting dangerous driving has been of critical interest for the past few years. However, a practical yet minimally intrusive solution remains challenging as existing technologies heavily rely on visual features or physical proximity. With this motivation, we explore the feasibility of purely using mmWave radars to detect dangerous driving behaviors. We first study characteristics of dangerous driving and find some unique patterns of range-doppler caused by 9 typical dangerous driving actions. We then develop a novel Fused-CNN model to detect dangerous driving instances from regular driving and classify 9 different dangerous driving actions. Through extensive experiments with 5 volunteer drivers in real driving environments, we observe that our system can distinguish dangerous driving actions with an average accuracy of > 95%. We also compare our models with existing state-of-the-art baselines to establish their significance.