Accident Pattern Classification for Autonomous Driving Risk Assessment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing traffic accident prediction apparatuses are unable to determine accident risks for autonomous driving scenarios, as they are based on patterns learned from manual driving data and do not account for the unique factors in autonomous driving.
Innovation Solution
An accident pattern determination apparatus that assigns attributes to predefined traffic situations, acquires accident patterns from vehicle-related accident cases, and determines whether these patterns represent high or low accident risks for specific vehicles or drivers based on the attributes and learned patterns.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If accident patterns are learned from manual driving data, then accident occurrence patterns can be identified, but the system cannot determine accident risks for autonomous driving scenarios
Solution Approach 1:
The system changes the parameters used for accident pattern determination by introducing autonomous driving-specific parameters (sensor malfunctions, system failures, environmental conditions affecting sensors) alongside traditional manual driving parameters. This allows the same accident pattern determination framework to be applied to both manual and autonomous driving scenarios by adjusting which parameters are considered relevant.
2Measurement precision
If comprehensive simulations are performed to determine accident risks, then risk assessment accuracy improves, but processing time and computational resources increase
Solution Approach 1:
The system performs preliminary classification of traffic situations into high-risk and low-risk categories using learned accident patterns before conducting detailed simulations. By pre-identifying which situations are likely to result in accidents based on historical data and pattern matching, the system可以避免 performing computationally intensive simulations on clearly low-risk scenarios, thereby reducing overall processing time while maintaining assessment accuracy for critical cases.
Data Source
AI summary
In an accident pattern determination apparatus including a storage storing attributes assigned to respective ones of a plurality of predefined traffic situations, an acquirer acquires, for each of vehicle-related accident cases, an accident pattern that is a combination of traffic situations in the accident case, from the plurality of predefined traffic situations. A determiner determines, for each accident pattern acquired by the acquirer, whether the accident pattern is an accident pattern of high accident risk or an accident pattern of low accident risk for specific vehicles, based on the accident patterns acquired for the respective accident cases and the attributes assigned to respective ones of the plurality of predefined traffic situations.