Automated Driving Parameter Control for Driver Style Adaptation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing automated driving systems struggle to dynamically adjust driving parameters based on individual driving styles and environments, relying on manual settings that fail to adapt to driver preferences and conditions.
Innovation Solution
A method for determining driving parameters that utilizes real-time driving data acquisition, extraction of driving features, and a logistic regression model to dynamically adjust parameters such as longitudinal acceleration, deceleration, and lane changing duration based on driver style probabilities, using vehicle sensors and machine learning algorithms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If manual parameter setting is used in automated driving systems, then the system structure remains simple, but the system cannot dynamically adjust driving parameters based on individual driving styles and environments
Solution Approach 1:
The system performs preliminary classification of driver behavior into distinct driving styles (e.g., aggressive, conservative, normal) based on historical driving data. This pre-classification enables the automated driving system to quickly adapt parameters without complex real-time analysis, resolving the contradiction by preparing adaptation patterns in advance.
Solution Approach 2:
The system adjusts multiple driving parameters (acceleration, deceleration, lane changing, steering) based on the classified driving style. By changing these parameters according to pre-determined patterns associated with each driving style, the system achieves dynamic adaptability without requiring complex real-time decision-making structures.
2Productivity
If automated driving parameters are manually set, then implementation costs remain low, but the system cannot meet real-time requirements and driver preferences
Solution Approach 1:
The system classifies driver behavior patterns in advance using machine learning algorithms trained on historical data. This preliminary classification creates a lookup table of appropriate parameter settings for each driving style, enabling real-time adaptation without expensive complex computation during actual driving.
Solution Approach 2:
The system creates simplified representations (copies) of complex driver behavior patterns through classification into discrete driving styles. Instead of replicating the full complexity of individual driver decision-making in real-time, the system uses these simplified models to quickly determine appropriate parameter adjustments.
3Measurement precision
If complex machine learning models are used to classify driving styles, then classification accuracy improves, but system complexity and computational requirements increase
Solution Approach 1:
The continuous spectrum of driver behavior is segmented into discrete driving style categories (e.g., aggressive, conservative, normal). This segmentation simplifies the classification task by dividing a complex continuous problem into manageable discrete classes, improving both accuracy and computational efficiency.
Solution Approach 2:
The system transforms complex multi-dimensional driver behavior data into simplified classification parameters that capture essential driving style characteristics. By changing the representation from raw continuous data to discrete style categories, the system achieves accurate classification with reduced computational complexity.
Data Source
Figure 1~2
AI summary
A method for determining driving parameters includes: acquiring driving data of a driver; extracting driving features of the driver based on the driving data, where the driving features include first operation frequency of a first component of a vehicle and second operation frequency of a second component of the vehicle; determining, based on the driving features, a first probability that the driver has a first driving style and a second probability that the driver has a second driving style; and determining the driving parameters based on the first probability and the second probability, where the driving parameters include at least longitudinal acceleration and longitudinal deceleration. The method can dynamically adjust automated driving parameters to meet the driving style of the driver, thereby effectively improving user experience.