Vehicle Following Control Using Driver Gap Preference Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems struggle to accurately and efficiently acquire driving preference values of a driver for follow-up traveling with respect to a preceding vehicle, particularly during transitions between manual and automated driving modes.
Innovation Solution
A control device for a vehicle that includes a memory device and processor to calculate driving preference values based on a driving dataset, using functions specified in advance to determine vehicle-to-vehicle preferences and control target values for acceleration, deceleration, and timing, leveraging machine learning for data analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple driving preference values are acquired for more automated driving parameters, then the driving preference reflection accuracy is improved, but the learning complexity and data acquisition difficulty increase
Solution Approach 1:
The patent extracts the core driving preference information by focusing on vehicle-to-vehicle preference values as the primary parameter. Other driving preferences are derived from this core parameter through the stored function, rather than independently learning each parameter. This extraction approach reduces learning complexity while maintaining comprehensive preference coverage.
Solution Approach 2:
The system performs preliminary action by storing the function relating vehicle-to-vehicle preference to other driving preferences in advance. This pre-established functional relationship allows the system to derive multiple driving preferences from a single learned parameter, avoiding the need to learn each parameter separately and thus reducing overall learning complexity.
2Ease of operation
If driving preference values are learned during manual driving, then the acquisition ease is improved, but the learning efficiency and speed decrease
Solution Approach 1:
The patent changes the parameter being learned from multiple driving preferences to a single vehicle-to-vehicle preference value. This parameter reduction maintains ease of acquisition during normal manual driving while significantly improving learning efficiency, as the system now needs to learn one parameter instead of multiple parameters simultaneously.
3Productivity
If a single vehicle-to-vehicle preference value is learned, then the learning speed is improved, but the coverage of driving preferences decreases
Solution Approach 1:
The patent makes the vehicle-to-vehicle preference value serve multiple functions by establishing it as a core parameter from which other driving preferences are derived. This single preference value universally represents multiple aspects of driving behavior through the stored functional relationships, maintaining both learning speed and preference coverage.
Solution Approach 2:
The system performs preliminary action by pre-storing the functional relationships between vehicle-to-vehicle preference and other driving preferences. This allows a single learned parameter to expand into multiple derived preferences, ensuring comprehensive coverage without sacrificing learning speed.
Data Source
AI summary
A control device includes a memory device configured to store a first function, and a processor. The first function is specified in advance based on a driving dataset for a plurality of drivers. The processor is configured to: calculate, in accordance with the first function, a vehicle-to-vehicle preference value of a driver according to vehicle-to-vehicle information and subject vehicle speed acquired during follow-up driving in manual driving; calculate, based on the calculated vehicle-to-vehicle preference value, one or more remaining driving preference values for at least one of acceleration, deceleration, deceleration timing, and acceleration timing; and execute, based on the calculated one or more remaining driving preference values, a target value calculation process of calculating one or more control target values for the at least one of the acceleration, the deceleration, the deceleration timing, and the acceleration timing in automated driving with the driver on board.