Driver Awareness Monitoring for Safe ADAS Handover
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing driver attention tracking features in advanced driver assistance systems (ADAS) are inadequate for ensuring driver vigilance, as they can be misused or abused, leading to hazardous situations where the driver is labeled as fully aware but unable to intervene during critical events.
Innovation Solution
A human machine interaction monitor (HMIM) is integrated with an attention monitor to provide a driver awareness estimator (DAE) that tracks both short-term and long-term gaze patterns, generating warnings and controlling ADAS features to ensure driver engagement and safety, using a driver awareness level escalation regime and kinesthetic responses.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If driver attention tracking features use algorithms to track and categorize visual attention, then the system can monitor driver gaze patterns, but the system may be misused or abused leading to false labels of driver awareness
Solution Approach 1:
The patent introduces a Human Machine Interaction Monitor (HMIM) as an intermediary system that independently verifies driver attention status. The HMIM acts as a mediator between the driver attention tracking feature and the automation control system, cross-checking the driver's actual engagement level against the automation system's expectations. This intermediary layer prevents false positive labels by providing an additional verification mechanism that can detect when the driver is not truly attentive despite the attention tracking algorithm indicating otherwise.
Solution Approach 2:
The system implements a feedback mechanism where the HMIM continuously monitors driver behavior and provides feedback to the automation system about the driver's actual attention state. When the HMIM detects that the driver is not sufficiently attentive (through analysis of glance patterns, response times, and interaction quality), it sends feedback signals to adjust or deactivate automation features. This closed-loop feedback ensures that the system responds to the driver's actual state rather than relying solely on potentially misleading algorithmic categorizations.
2Productivity
If the system labels driver as fully aware to keep automation active, then automation functionality is maintained, but the driver may be unable to intervene in hazardous events
Solution Approach 1:
The HMIM performs preliminary verification of driver attention status before the automation system relies on the driver's ability to intervene. By continuously analyzing multiple indicators of driver engagement (glance frequency, duration, quality of interaction) in advance, the system can detect early signs of driver disengagement. This preliminary action allows the system to take preventive measures (such as issuing warnings or gradually reducing automation reliance) before a hazardous event occurs, ensuring that the driver is truly capable of intervention when needed.
Solution Approach 2:
The system implements a cushioning mechanism by creating a safety buffer through the HMIM's continuous monitoring. When the HMIM detects declining driver attention, it provides a transition period where the system can gradually adjust automation levels or issue progressive warnings, giving the driver time to regain attention. This beforehand cushioning prevents abrupt transitions and ensures that the driver remains capable of intervention even as automation levels are adjusted, bridging the gap between maintaining automation functionality and ensuring driver readiness.
3Reliability
If the system provides multiple warning levels through HMI, then driver awareness can be monitored progressively, but the system complexity increases
Solution Approach 1:
The patent segments the driver awareness monitoring function into distinct components: the attention tracking feature that categorizes visual attention, the HMIM that independently verifies attention status, and the escalation regime that manages warning levels. Each component has a specific role and operates semi-independently, allowing the system to maintain high reliability through functional segmentation while managing complexity through modular design. The segmentation enables clear separation of concerns where each module can be developed and validated independently.
Solution Approach 2:
The system implements a dynamic escalation regime where the number and intensity of warnings adapt based on the driver's attention level and the severity of the situation. Rather than using a fixed complex warning sequence, the system dynamically adjusts the warning strategy - starting with subtle cues and escalating to more prominent warnings only when necessary. This dynamic approach maintains simplicity by using a flexible, adaptive structure rather than a rigid complex sequence, allowing the system to provide progressive monitoring while keeping the warning mechanism relatively simple and context-appropriate.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An apparatus comprising an interface configured to receive a plurality of sensor signals from a vehicle platform of a vehicle and present one or more control signals to the vehicle platform; and a control circuit configured to (i) detect whether an attention state of a driver is in an attentive state or an inattentive state in response to one or more of the plurality of sensor signals from the vehicle platform during a first window having a first duration, (ii) assess whether the driver is sufficiently attentive by monitoring the one or more of the plurality of sensor signals from the vehicle platform and determining whether changes in the attention state of the driver during a second window having a second duration that is longer than the first duration exceeds a threshold, and (iii) when the threshold is exceeded, transition operation of the vehicle to the driver and safely discontinue an automation system function of the vehicle.