Driver Condition Detection Using Cabin Context and Adaptive CNNs

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing vehicle safety systems for autonomous vehicles do not adequately consider the impact of driver input on overall safety, particularly in scenarios where the driver retains control over the vehicle, and fail to integrate driver expression and interior environment data for enhanced safety features.

Innovation Solution

An integrated system that uses machine learning methods, including convolutional neural networks with adaptive filters, to analyze driver images and vehicle interior conditions, combining classification results to determine driver intent and provide safety warnings or interventions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If driver condition detection systems focus only on driver facial expressions, then the system complexity is reduced, but the measurement precision and reliability of driver intent assessment deteriorates

Engineering Contradiction:
Improvesystem complexityVSAvoiddriver intent assessment accuracy
Core Design Contradiction:
Device complexityVSMeasurement precision

Solution Approach 1:

The patent combines driver facial expression analysis with interior environment analysis into a unified detection system. The system integrates multiple data sources including driver facial images, interior environment images, and contextual information to comprehensively assess driver intent and condition, thereby improving measurement precision while managing system complexity through integrated architecture.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The detection system is designed to perform multiple functions: analyzing driver facial expressions, analyzing interior environment conditions, detecting driver intent, and assessing safety risks. This multi-functional approach allows the system to gather comprehensive information from various sources to improve driver intent assessment accuracy without requiring separate dedicated systems for each function.

Inventive Principle:
Principle #6Universality (Multi-functionality)

2Reliability

If the system integrates multiple data sources including driver expressions and interior environment, then the reliability of safety determination is improved, but the device complexity increases

Engineering Contradiction:
Improvesafety determination reliabilityVSAvoidsystem integration complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent segments the detection system into distinct functional modules: a driver condition detection module that analyzes facial expressions, an environment detection module that analyzes interior conditions, and a risk assessment module that integrates both data sources. This segmentation allows the system to process multiple data sources reliably while managing complexity through modular architecture, where each module handles specific tasks independently before integration.

Inventive Principle:
Principle #1Segmentation

3Measurement precision

If additional information is requested when classification results are inconclusive, then the measurement precision of driver condition assessment is improved, but the loss of time in decision-making increases

Engineering Contradiction:
Improvedriver condition classification accuracyVSAvoiddecision-making time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The system performs preliminary classification of driver condition using available data before determining whether additional information is needed. By pre-establishing classification thresholds and decision criteria, the system can quickly assess whether existing information is sufficient or if additional sensor data should be collected, thereby minimizing decision-making time while maintaining classification accuracy.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system implements a feedback mechanism where classification results are continuously evaluated against predefined thresholds. When classification confidence is insufficient, the system requests additional information from sensors and re-evaluates the classification. This feedback loop ensures high measurement precision while managing time loss through iterative refinement rather than requiring all data upfront.

Inventive Principle:
Principle #23Feedback

Data Source

PatentEP3652715B1Integrated system for detection of driver condition
Publication Date: 2025.09.03 HUAWEI TECH CO LTD
  • EP3652715B1 patent drawingFigure 1A
  • EP3652715B1 patent drawingFigure 1B
  • EP3652715B1 patent drawingFigure 2

AI summary

Methods, apparatus, and systems are provided for integrated driver expression recognition (174) and vehicle interior environment (102, 152) classification to detect driver (104) condition for safety. A method includes obtaining an image of a driver of a vehicle and an image of an interior environment of the vehicle. Using a machine learning method (112), the images are processed to classify a condition of the driver and of the interior environment (102, 152) of the vehicle. The machine learning method (112) includes general convolutional neural network (CNN) and CNN with adaptive filters. The adaptive filters are determined based on influence of filters. The classification results are combined and compared with predetermined thresholds to determine if a decision can be made based on existing information. Additional information is requested by self-motivated learning if a decision cannot be made, and safety is determined based on the combined classification results. A warning (150) is provided to the driver based on the safety determination.