Passive Monitoring of Dangerous Driving Behaviors Using mmWave Radar

Key Features


  • Privacy-preserving: mmDrive uses radar-based sensing instead of cameras, ensuring no visual data is captured, thus maintaining driver privacy while effectively identifying risky behaviors.
  • Unobtrusive Monitoring: Unlike wearable devices, mmDrive operates passively without requiring drivers to wear or interact with any hardware, making it a seamless addition to vehicles.
  • Comprehensive Detection: The system is capable of identifying a wide range of dangerous behaviors, such as yawning, steering anomalies, using a phone, and more, ensuring holistic monitoring of driver attentiveness.
  • Robust Noise Filtering: Integrated IMU sensors effectively filter out noise caused by road conditions like bumps and potholes, ensuring accurate detection in diverse driving scenarios.
  • Energy Efficiency: mmDrive employs a two-step classification approach, activating detailed behavior analysis only when risky driving is detected, conserving computational resources.
  • Real-world Validated: Tested across 7 drivers, 5 different car models, and multiple road conditions, mmDrive demonstrated consistent and reliable performance with a precision of 97%.

Contributors

prasenjit
Argha Sen

IIT Kharagpur, India

swadhin
Anirban Das

IIT Kharagpur, India

prasenjit
Prasenjit Karmakar

IIT Kharagpur, India

sandip
Sandip Chakraborty

IIT Kharagpur, India

Teaser Video

Publications


  1. Argha Sen, Avijit Mandal, Prasenjit Karmakar, Anirban Das, and Sandip Chakraborty: "Passive Monitoring of Dangerous Driving Behaviors Using mmWave Radar", PMC Journal, Volume 103, October 2024
  2. Argha Sen, Avijit Mandal, Prasenjit Karmakar, Anirban Das, and Sandip Chakraborty: "mmdrive: mmWave Sensing for Live Monitoring and On-Device Inference of Dangerous Driving", IEEE PerCom 2023

Funding and Support



For questions and general feedback, contact Argha Sen