Accurate Driver Detection Exploiting Invariant Characteristics of Smartphone Sensors

Area of Science:

• Human-Computer Interaction
• Transportation Safety
• Sensor Technology

Background:

• Distracted driving remains a significant safety concern, with existing countermeasures proving insufficient.
• Current driver detection systems often rely on fragile designs and unreliable driving event data.
• There is a need for robust driver identification methods that are independent of specific vehicle actions or sensor placement.

Purpose of the Study:

• To propose and evaluate ADDICT (Accurate Driver Detection exploiting Invariant Characteristics of smarTphone sensors), a novel system for identifying drivers using smartphone sensors.
• To leverage invariant sensor dynamics and electromagnetic interplay for reliable driver detection.
• To assess the accuracy and efficiency of ADDICT across diverse driving scenarios.

Main Methods:

• ADDICT utilizes the inconsistency between gyroscope and magnetometer dynamics.
• The system analyzes the interplay between electromagnetic field emissions and engine startup vibrations.
• Experiments were conducted with participants across three vehicles, varying driving scenarios.

Main Results:

• ADDICT achieved an average accuracy of 89.1% in identifying the driver's smartphone across all tested scenarios.
• Accuracy remained high (>85%) even under extreme driving conditions.
• The system demonstrated marginal battery consumption.

Conclusions:

• ADDICT offers a robust and accurate method for driver detection, overcoming limitations of previous systems.
• The system's reliance on invariant characteristics ensures reliability regardless of smartphone position or driving actions.
• ADDICT presents a promising solution for enhancing driver safety and combating distracted driving.