Mobile Sensor Fusion for Driver-Passenger Identification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems for tracking driving behavior are limited in accurately identifying the driver and attributing data to a specific individual, lacking the ability to determine when someone is driving a vehicle, which is crucial for insurance and risk modeling purposes.
Innovation Solution
A method using a mobile device with an accelerometer and gyroscope to determine vehicle entry and exit times, orientation, and position by analyzing yaw measurements, acceleration events, and angular changes, allowing for precise identification of driving events and associating them with the correct user.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If external devices are integrated with vehicles to track driving behavior, then vehicle-related data can be measured, but the ability to accurately identify the driver and attribute data to a specific individual is limited
Solution Approach 1:
The mobile device serves multiple functions: it acts as both a general-purpose communication device and a specialized driving behavior tracking system. The device uses its existing sensors (accelerometer, gyroscope, magnetometer) for both standard mobile applications and driver identification, eliminating the need for separate dedicated hardware while achieving accurate driver attribution through multi-functional sensor data analysis
Solution Approach 2:
The system uses the mobile device's own built-in sensors to perform driver identification without requiring external add-on hardware. The device self-identifies the driver by analyzing data from its inherent accelerometer, gyroscope, and magnetometer, making the system self-sufficient and reducing overall system complexity while maintaining high measurement precision
2Reliability
If hardware devices are used to track driving behavior, then vehicle data can be read, but the system cannot determine when someone is driving versus sitting as a passenger
Solution Approach 1:
The system exploits the asymmetric movement patterns between drivers and passengers. Drivers exhibit characteristic yaw rotations when entering and exiting the vehicle, along with specific acceleration patterns during driving maneuvers. Passengers lack these asymmetric movements, particularly the rotational components. By analyzing the asymmetry in movement patterns across multiple sensors, the system reliably distinguishes drivers from passengers with high accuracy
Solution Approach 2:
The system transitions from analyzing only linear acceleration (single dimension) to incorporating rotational dimensions through gyroscope and magnetometer data. By adding angular velocity and orientation measurements as new dimensions of analysis, the system creates a multi-dimensional sensor fusion approach that enables reliable detection of driving events and driver identification, solving the difficulty of distinguishing drivers from passengers
3Measurement precision
If mobile device sensors are used to detect driving events, then driver identification can be achieved, but precise measurement of entry and exit times requires complex analysis
Solution Approach 1:
The system monitors changes in sensor parameters (acceleration magnitude, yaw angle, angular velocity) to detect entry and exit events. By establishing threshold values for these parameter changes and detecting when they are exceeded, the system precisely identifies transition moments without requiring complex algorithms. The use of parameter change detection simplifies the analysis while maintaining high temporal accuracy for entry and exit time measurement
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables accurate attribution of driving data to the correct user, providing insights for improving driving behavior, reducing insurance costs, and enhancing risk modeling by distinguishing between drivers and passengers, thereby incentivizing safe driving and reducing accidents.
Implementation Method 1
obtain data from the accelerometer of the mobile device, and detecting, one or more braking events and one or more forward acceleration events
Implementation Method 2
obtain angular measurement data from the gyroscope of the mobile device
Data Source
AI summary
A method of transforming measurements from an accelerometer of a mobile phone to a reference frame of a vehicle in which the mobile phone is disposed includes classifying the measurements into those caused by the vehicle. Using these measurements, the method detects positive or negative acceleration events, determines a direction of an event with respect to the mobile phone, and determines a direction of the event with respect to the vehicle. Based on the two directions, the measurements are transformed to the reference frame of the vehicle.