Get started
Work

Enhancing safe driving through real-time driver monitoring using triboelectric energy harvesting and deep transfer learning

Year: 2025
Type: article
Abstract: In contemporary society, personal vehicles are indispensable for enhancing everyday convenience, and numerous advanced technological solutions for safe driving are being proposed, including autonomous emergency braking systems and lane departure warning systems. However, traditional sensors aimed at preventing incapacitated driving due to drowsiness, alcohol consumption, or sudden health issues often require additional power and may infringe upon driver privacy. A self-powered, nonintrusive motion recording system utilizing a triboelectric nanogenerator with an arch and cantilever structure (AC-TENG) as a sensor to track driver motions is introduced. Motion patterns of users are captured in real time by imaging the waveforms generated by the AC-TENG. To discern the peaks of the AC-TENG waveforms with greater precision and clarity, the roughness of the contact surface is enhanced using the reactive ion etching (RIE) technique, thereby amplifying the voltage by a factor of 2.1. The resulting output power of 2.77 mW and a corresponding power density of 3.08 W/m2 are demonstrated. Energy harvesting is facilitated by the proposed AC-TENG sensor, coupled with a power management integrated circuit (PMIC). Additionally, continuous wireless data transmission is enabled by using it as an input source alongside a solar cell. Motion data captured by the sensor are classified via deep transfer learning, enabling real-time monitoring of safe driving. Upon detection of any abnormal signals, the driver is promptly alerted by this system, and its application can be extended to functionalities such as transitioning the vehicle to autonomous driving mode.
DOI: 10.26599/nr.2025.94907483
Language: English
Open Access status: gold
Estimated APC paid: $5,290
OpenAlex ID: W4409562098
FWCI: 0.472
Cites: 36
Cited by: 2
Related to: 10