Driver Inattentiveness Detection Using Head-Mounted Imaging

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current driver inattentiveness detection systems are inadequate as they rely solely on vehicle motion and may generate false alerts or miss inattentiveness, as they cannot accurately differentiate between driver distraction and regular driving maneuvers.

Innovation Solution

A system that uses a wearable device with image sensors to capture the driver's perspective and compare it with the vehicle's direction, processing the images to determine angular differences and generate alerts when the driver becomes inattentive, incorporating head movement and GPS data for more accurate detection.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If driver inattentiveness detection is based on vehicle motion monitoring, then the system can detect driver state, but it generates false alerts due to inability to differentiate between driver distraction and regular driving maneuvers

Engineering Contradiction:
Improvedetection accuracyVSAvoidfalse alert rate
Core Design Contradiction:
Measurement precisionVSReliability

Solution Approach 1:

The system segments the detection task into multiple independent measurement components: vehicle motion parameters (acceleration, steering angle, speed) are measured separately from driver behavior parameters (head position, eye gaze direction, facial expressions). By dividing the detection system into vehicle-based sensors and driver-based sensors, each component can be optimized independently, and the combined data provides more reliable detection with reduced false alerts.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system introduces an intermediary analysis layer that correlates vehicle motion data with driver behavior data. This intermediary processing layer compares the expected driver response to vehicle maneuvers against actual driver behavior, allowing the system to distinguish between intentional driving actions and genuine inattentiveness. The intermediary layer acts as a mediator that reconciles conflicting signals from different sensors.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Ease of operation

If driver inattentiveness is detected using steering angle tracking alone, then the system can monitor driver behavior, but it confuses regular lane swerving with driver distraction

Engineering Contradiction:
Improvemonitoring simplicityVSAvoiddetection accuracy
Core Design Contradiction:
Ease of operationVSMeasurement precision

Solution Approach 1:

The system merges multiple monitoring functions into a unified detection framework. Instead of relying on steering angle tracking alone, the system combines vehicle motion monitoring with direct driver behavior monitoring (head position, eye gaze, facial expressions). This merging of multiple independent monitoring functions creates a more accurate detection system that can distinguish between lane swerving and genuine distraction by looking at the relationship between vehicle motion and actual driver attention.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The system adds new dimensions to the detection problem by moving from one-dimensional steering angle monitoring to multi-dimensional driver state assessment. By incorporating head position, eye gaze direction, facial expression analysis, and vehicle motion parameters, the system creates a multi-dimensional detection space where genuine inattentiveness can be identified through pattern recognition across multiple dimensions, rather than relying on a single steering angle metric.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Data Source

PatentEP3316231B1Alert generation correlating between head mounted imaging data and external device
Publication Date: 2021.11.10 IONROAD TECH
  • EP3316231B1 patent drawingFigure 1
  • EP3316231B1 patent drawingFigure 2
  • EP3316231B1 patent drawingFigure 2A

AI summary

A method of alerting a vehicle's driver. The method comprises receiving a first set of image data imaging a driver's perspective, the first set of image data captured by at least one image sensor coupled to a primary device placed on the driver and directed outwardly from the driver, receiving a second set of image data, the second set of image data captured by a second device mounted on a vehicle driven by the driver, processing the first and second sets of image data to determine a comparison result, the comparison result determined according to image similarity between the first and second sets of image data, generating an alert according to the comparison result, and automatically presenting the alert to the driver