Driver Fatigue Monitoring Using Video AI for Autonomous Vehicle Oversight
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The susceptibility of human test drivers to fatigue while monitoring autonomous vehicles increases due to prolonged periods of inattention, drowsiness, and the need for immediate intervention, posing a safety risk as the vehicles' performance improves and human interventions decrease.
Innovation Solution
A method and system for estimating the likelihood of fatigue in test drivers using video analysis and machine-learning models, which identify behaviors related to fatigue, provide reliability scores, and recommend interventions to prevent fatigue events, integrating human operator feedback and historical data to improve monitoring efficiency and safety.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If continuous human monitoring of test drivers is implemented, then driver fatigue detection capability is improved, but system complexity and operational costs increase
Solution Approach 1:
The system creates a digital model or representation of driver fatigue state by analyzing video feeds through machine learning algorithms. Instead of requiring multiple human observers, the system generates a computational copy of the driver's alertness level based on visual indicators, enabling accurate fatigue detection through data replication and analysis rather than human multiplication.
Solution Approach 2:
The patent transforms the monitoring approach by changing parameters from subjective human assessment to objective video-based metrics. The system analyzes specific visual parameters such as eye closure duration, head position, and facial muscle tension from video feeds, converting qualitative fatigue assessment into quantifiable measurable parameters that can be processed automatically.
2Reliability
If frequent video analysis and human operator review are conducted, then fatigue event detection reliability is improved, but time consumption and operational efficiency decrease
Solution Approach 1:
The system applies partial monitoring action by using the AI model to continuously analyze video feeds at a basic level, then selectively escalating to full human operator review only when the model detects indicators suggesting potential fatigue. This partial automation approach maintains high reliability by combining continuous automated monitoring with targeted human verification, reducing overall time loss compared to requiring constant human review of all footage.
Data Source
AI summary
Aspects of the disclosure relate to models for estimating the likelihood of fatigue in test drivers. In some instances, training data including videos of the test drivers while such test drivers are tasked with monitoring driving of a vehicle operating in an autonomous driving mode may be identified. The training data also includes driver drowsiness values generated from one or more human operators observing the videos. The training inputs and outputs may be used to train the model such that when a new video of a first test driver is input into the model, the model will output an estimate of a likelihood of fatigue for that test driver.