ADAS Control Parameters Using Age-Specific Driver Prediction Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing ADAS systems in vehicles operate generically and do not adapt to the individual driving styles of different drivers, particularly considering age-related changes in perception and reflexes, leading to suboptimal performance.
Innovation Solution
A method and device that adapt ADAS system control parameters by using machine learning to predict and determine control parameters based on driver age, utilizing data from a set of vehicles to partition and learn age-specific models, and apply these models to individual vehicles for personalized control.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a generic ADAS system operation is used, then the system is simple to implement, but it does not adapt to different driver types and age-related changes
Solution Approach 1:
The patent segments the driver population into different age groups (e.g., young, middle-aged, senior drivers) and creates separate control parameter sets for each group. This allows the ADAS system to adapt to different driver characteristics without requiring a completely new system architecture, thus improving adaptability while controlling complexity.
Solution Approach 2:
The system dynamically adjusts control parameters based on the detected driver age group and driving style. Instead of using fixed generic parameters, the system selects and modifies parameters in real-time according to the current driver's characteristics, enabling adaptation without permanent system reconfiguration.
2Manufacturing precision
If age-specific prediction models are learned for each group, then the control parameters are optimized for each driver type, but the data processing and model selection complexity increases
Solution Approach 1:
The patent performs preliminary learning and training of age-specific prediction models during a data collection phase before actual deployment. Multiple prediction models are pre-computed for different age groups using historical driving data. During operation, the system only needs to select from these pre-trained models rather than learning in real-time, thus achieving high precision while keeping runtime complexity low.
Solution Approach 2:
The system changes the parameter set based on the detected driver age group. Each age group has its own optimized parameter ranges and weighting factors. This allows the system to achieve high precision for each group by using group-specific parameters without requiring a completely different system architecture.
3Reliability
If control parameters are customized for each driver age group, then safety and performance are improved, but the system requires more data collection and processing
Solution Approach 1:
The patent creates a universal data collection framework that serves multiple purposes: it collects information needed for age-group classification, driving style detection, and parameter optimization simultaneously. This multi-functional approach reduces redundant data collection while achieving the same reliability improvements.
Solution Approach 2:
The system uses the driver's existing driving behavior data to automatically determine their age group and preferred driving style without requiring explicit user input or additional surveys. The system self-adjusts parameters based on observed driving patterns, reducing the time and effort required for customization while maintaining high safety standards.
Data Source
Figure 1~2
Figure 3~4
Figure 5
AI summary
The present invention relates to a method and device for controlling an advanced driver-assistance system, abbreviated ADAS, of a first vehicle (10). To this end, a first piece of information representative of the age of a driver of the first vehicle is obtained. First data representative of an environment of the first vehicle (10) and second data representative of driving parameters of the first vehicle (10) are received. A model for predicting control parameters of the ADAS is selected from a plurality of prediction models depending on the first piece of information. The prediction models have been trained beforehand using data obtained from a set (11) of second vehicles. A set of control parameters of the ADAS is determined by feeding the selected prediction model with the first and second data, with a view to controlling the ADAS.