Vehicle Braking Distance Learning via Segmentation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing vehicle systems struggle to accurately learn the braking distance based on a driver's sensation, especially at intersections, due to significant variations in braking behavior even under the same environmental conditions.

Innovation Solution

A driving control system that prioritizes learning the braking distance during manual driving when there is no preceding vehicle, using a multiple regression model to analyze deceleration starting speed and environmental factors, thereby isolating the driver's braking behavior from external influences.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If braking distance learning includes all manual driving cases (with and without preceding vehicles), then more learning data is available, but the variation in braking behavior increases and learning accuracy decreases

Engineering Contradiction:

Improveamount of learning dataVSAvoidbraking distance learning accuracy

Core Design Contradiction:

Quantity of substanceVSMeasurement precision

Solution Approach 1:

The patent segments the learning data by separating cases with preceding vehicles from cases without preceding vehicles. This segmentation allows the system to selectively use only the data from cases without preceding vehicles for learning the driver's braking distance, thereby eliminating the variation caused by different driving situations and improving learning accuracy.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent extracts and excludes data from cases where preceding vehicles are present, keeping only the data from cases without preceding vehicles for the learning process. This extraction of relevant data removes the harmful variation introduced by different traffic conditions and enables accurate learning of the driver's true braking behavior.

Inventive Principle:

Principle #2Taking out (Extraction)

2Adaptability or versatility

If learning includes cases with preceding vehicles, then diverse driving scenarios are covered, but the driver's true braking sensation cannot be accurately learned due to external influences

Engineering Contradiction:

Improvecoverage of driving scenariosVSAvoidbraking distance learning reliability

Core Design Contradiction:

Adaptability or versatilityVSReliability

Solution Approach 1:

The patent extracts only the essential data needed for learning the driver's braking sensation by excluding cases with preceding vehicles. This extraction ensures that the learning process is based solely on the driver's own braking behavior without external influences, improving the reliability of the learned braking distance.

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent applies different data selection criteria for different learning purposes. Specifically, it uses only data from cases without preceding vehicles for learning the driver's braking distance, while potentially using other data for other aspects of driving behavior. This local quality approach ensures that each learning parameter is trained on the most appropriate data.

Inventive Principle:

Principle #3Local quality

Data Source

PatentUS10787173B2Traveling assistance method and driving control device

Publication Date: 2020.09.29 NISSAN MOTOR CO LTD

AI summary

A traveling assistance method of the present invention for a vehicle capable of switching between manual driving by the driver and automated driving includes learning a braking distance of a case of stopping at an intersection during the manual driving by the driver, in which a braking distance of a case of no preceding vehicle in front of the vehicle is preferentially learned.