Identifying Autonomous Driving Difficult Areas via Manual-Automatic Data Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Automatic vehicles face difficulties in navigating certain road situations and environments, leading to unnecessary deceleration or inability to travel, which can result in accidents, and existing technologies struggle to accurately identify these challenging locations due to incomplete map information or erroneous data detection.
Innovation Solution
An information processing method that acquires both manual and automatic driving data from vehicles, calculates driving parameters, and compares them to identify areas where automatic driving is difficult, creating information that highlights these challenging locations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If automatic driving is implemented in all areas, then automation coverage is improved, but safety deteriorates due to inability to handle difficult areas
Solution Approach 1:
The patent divides the driving area into ordinary areas and difficult areas based on driving parameter comparisons. By segmenting the map information into different types of areas, the system can apply appropriate driving strategies for each, allowing automatic driving to operate safely in ordinary areas while identifying difficult areas for manual intervention or improved algorithm development.
Solution Approach 2:
The system performs preliminary identification of difficult areas by comparing historical manual and automatic driving data before actual automatic driving occurs. This advance preparation allows the system to pre-map difficult areas and adjust automatic driving strategies accordingly, preventing safety issues rather than reacting to them after they occur.
2Extent of automation
If map information is expanded to cover more areas, then automation coverage is improved, but measurement precision deteriorates due to incomplete or erroneous data
Solution Approach 1:
The patent establishes a feedback mechanism where automatic driving results and driving parameters are continuously collected and compared with manual driving data. This feedback loop allows the system to identify areas where automatic driving performance deviates from manual driving, thereby refining the identification of difficult areas and improving data accuracy over time through iterative learning.
Solution Approach 2:
The system uses its own automatic driving data and compares it with manual driving data to self-identify difficult areas and improve its performance. By leveraging its operational data, the system can autonomously refine its understanding of difficult areas without requiring external annotation or manual input, thereby improving measurement precision through self-learning.
3Measurement precision
If driving parameter comparison is performed for all areas, then identification accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent applies local quality by focusing computational resources on comparing driving parameters only in areas where differences between manual and automatic driving are detected. Rather than uniformly processing all areas with equal computational intensity, the system concentrates analysis on regions showing performance gaps, thereby improving identification accuracy while reducing overall computational complexity.
Data Source
AI summary
An information processing method includes: acquiring, from one or more first vehicles, manual driving information for each of the first vehicles; acquiring, from one or more second vehicles, automatic driving information for each of the second vehicles; calculating, for each area, a first value of a driving parameter correlated with a degree of difficulty in driving according to the manual driving information; calculating, for the each area, a second value of the driving parameter according to the automatic driving information; comparing the first value with the second value for the each area to determine at least one difficult area in which travel of a vehicle by automatic driving is difficult; and creating difficult area information.