Driving Scene Perceptual Load Estimation Using Visual Features

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current methods fail to accurately measure and recognize perceptual load directly from the visual driving scene in a non-intrusive manner, which is essential for improving driving safety and predicting a driver's ability to detect safety-critical events, as they either require intrusive measures or amalgamate workload sources, making it difficult to isolate perceptual load.

Innovation Solution

A control device that extracts scene features from the visual driving scene and determines perceptual load using a data-driven approach with crowd-sourced load labels, trained on reference video scenes, allowing for real-time analysis without interfering with the driver's activities, employing optical sensors and machine learning models like regression or classification to map scene features to load values.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If intrusive measures are used to measure driver workload, then measurement precision may improve, but ease of operation deteriorates as it interferes with driver activities

Engineering Contradiction:
Improveperceptual load measurementVSAvoiddriver operation
Core Design Contradiction:
Measurement precisionVSEase of operation

Solution Approach 1:

The patent uses the visual driving scene as an intermediary to indirectly measure perceptual load. Instead of directly measuring driver state through intrusive means, the system analyzes environmental features (traffic density, scene complexity, object counts) that correlate with perceptual load, thereby avoiding interference with driver operations while still achieving measurement precision

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent replaces direct physiological or behavioral measurement mechanisms with an optical sensing and computational analysis system. By substituting mechanical/intrusive measurement methods with optical scene analysis and machine learning algorithms, the system achieves accurate perceptual load assessment without physical interference with the driver

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Ease of manufacture

If workload sources are amalgamated, then ease of manufacture improves, but measurement precision deteriorates as it becomes difficult to isolate perceptual load

Engineering Contradiction:
Improvesystem implementationVSAvoidperceptual load isolation
Core Design Contradiction:
Ease of manufactureVSMeasurement precision

Solution Approach 1:

The patent segments the workload measurement task by focusing exclusively on perceptual load components related to the visual driving scene. Rather than amalgamating all workload sources, the system isolates and measures only scene-related perceptual demands using specific features like traffic density, scene complexity, and object counts, thereby maintaining measurement precision while simplifying the implementation scope

Inventive Principle:
Principle #1Segmentation

3Productivity

If real-time analysis is implemented, then productivity improves, but device complexity increases due to computational requirements

Engineering Contradiction:
Improvereal-time load determinationVSAvoidcomputational system
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent applies partial action by selecting only the most relevant scene features for analysis (traffic density, scene complexity, object counts) rather than processing all possible visual information. This selective feature extraction enables real-time perceptual load determination with reduced computational complexity, achieving productivity improvement without excessive device complexity

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP3465532B1Control device, system and method for determining the perceptual load of a visual and dynamic driving scene
Publication Date: 2024.08.07 TOYOTA JIDOSHA KK
  • EP3465532B1 patent drawingFigure 1~3
  • EP3465532B1 patent drawingFigure 4~5
  • EP3465532B1 patent drawingFigure 6~8

AI summary

The invention relates to a control device (1) for a vehicle for determining the perceptual load of a visual and dynamic driving scene. The control device is configured to: • receive a sensor output (101) of a sensor (3), the sensor (3) sensing the visual driving scene, · extract a set of scene features (102) from the sensor output (101), the set of scene features (102) representing static and/or dynamic information of the visual driving scene, and • determine the perceptual load (104) of the set of extracted scene features (102) based on a predetermined load model (103), wherein the load model (103) is predetermined based on reference video scenes each being labelled with a load value The invention further relates to a system and a method.