Adaptive Driver Stimuli Saliency for Hazard Attention Control

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Partially autonomous vehicles face challenges in managing driver attention and cognitive state, leading to issues with receptivity and engagement, particularly in scenarios where human drivers are not adequately responsive to hazards due to overconfidence or anxiety regarding machine control.

Innovation Solution

A system utilizing predictive models, such as Convolutional Neural Networks (CNN) and Recurrent Neural Networks (RNN), to determine a saliency level for stimuli that adjusts based on the driver's cognitive state, aiming to seamlessly direct spatial attention to hazards without disrupting the driver's non-driving activities, using visual and audio stimuli tailored to the driver's characteristics and environmental conditions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If alert systems increase sensitivity to reduce false negatives, then hazard detection reliability improves, but false positive rates increase causing driver annoyance and reduced receptivity

Engineering Contradiction:
Improvehazard detection reliabilityVSAvoidfalse positive alerts
Core Design Contradiction:
ReliabilityVSObject-generated harmful factors

Solution Approach 1:

The system dynamically adjusts alert parameters (saliency level, modality, timing) based on real-time driver cognitive state measurements. When driver attention is high, alert sensitivity is reduced to minimize false positives. When attention drops, sensitivity increases to ensure hazard detection, thus resolving the contradiction between reliability and false positives.

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The alert system transitions from static threshold-based alerts to dynamic adaptive alerts that continuously adjust based on driver cognitive state. The system monitors cognitive metrics (eye tracking, response time, pupil dilation) and modulates alert characteristics accordingly, enabling the system to maintain high reliability while minimizing false positives through continuous adaptation.

Inventive Principle:
Principle #15Dynamics

2Ease of operation

If the system provides frequent alerts to maintain driver engagement, then driver receptivity improves, but cognitive load increases causing driver stress and reduced wellbeing

Engineering Contradiction:
Improvedriver receptivityVSAvoidcognitive load
Core Design Contradiction:
Ease of operationVSDevice complexity

Solution Approach 1:

The system applies alert action selectively based on measured driver cognitive state. Instead of continuous alerts, the system provides partial action (alerts) only when cognitive metrics indicate reduced attention or engagement. This targeted approach maintains driver receptivity while avoiding excessive cognitive load from unnecessary alerts.

Inventive Principle:
Principle #16Partial or excessive action

Solution Approach 2:

The system implements closed-loop feedback by continuously monitoring driver cognitive state (through sensors tracking eye movement, response patterns, physiological signals) and adjusting alert frequency and intensity accordingly. This feedback mechanism ensures alerts are provided optimally to maintain engagement without overwhelming the driver, balancing receptivity and cognitive load.

Inventive Principle:
Principle #23Feedback

3Measurement precision

If the system uses salient stimuli to capture driver attention, then hazard awareness improves, but driver distraction from non-driving activities increases

Engineering Contradiction:
Improvehazard awarenessVSAvoiddriver attention from non-driving activities
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The system applies different stimulus qualities and intensities to different spatial locations and contexts. Instead of uniform salient stimuli, the system targets alerts to specific visual fields or auditory zones relevant to the hazard location, and adjusts saliency based on driver gaze direction and attention focus. This localized approach improves hazard awareness while minimizing disruption to ongoing non-driving activities.

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP3755596B1Cognitive state-based stimuli
Publication Date: 2024.06.12 ADAM COGTECH LTD
  • EP3755596B1 patent drawingFigure 1
  • EP3755596B1 patent drawingFigure 2
  • EP3755596B1 patent drawingFigure 3

AI summary

Method, apparatus and product for providing cognitive state-based seamless stimuli. The method comprising: obtaining a cognitive state of a human subject; determining a target cognitive state for the human subject; determining, based on the cognitive state and the target cognitive state, a saliency level for a stimuli, wherein the saliency level is configured to cause the human subject to direct spatial attention to the stimuli, wherein the saliency level is configured to cause the stimuli to be seamless for the human subject given the cognitive state; and outputting the stimuli at the saliency level to be perceived by the human subject.