Vehicle Function Verification Under Simulated Perception Failures
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing perception algorithms in autonomous and assisted driving systems often fail to meet expected functional safety and performance requirements due to hardware and environmental changes, leading to unreliable vehicle functions.
Innovation Solution
A method and apparatus for verifying vehicle functions by simulating various perception failure scenarios, allowing for the determination of the tolerance of vehicle functions to perception algorithm performance and guiding the development and optimization of the perception algorithm.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If perception algorithms are used for autonomous driving functions, then the vehicle can provide assisted driving services (AEB, ACC, NOA, HWP), but the performance is affected by hardware and environmental changes causing the functions to fail to meet functional safety requirements
Solution Approach 1:
The patent applies preliminary action by performing perception failure simulation before actual deployment. The system pre-generates failed perception results through simulation and uses these to verify vehicle function reliability in advance, allowing developers to identify and fix potential safety issues before the system operates in real-world conditions.
Solution Approach 2:
The patent introduces an intermediary verification system that sits between the perception algorithm and the vehicle control functions. This verification system uses simulated failed perception results to test and validate whether the vehicle functions can handle perception failures safely, acting as a mediator to ensure functional safety requirements are met.
2Ease of operation
If perception algorithms are deployed in real-world conditions, then the system can operate with actual sensor data, but hardware and environmental changes cause the perception algorithm performance to deteriorate
Solution Approach 1:
The patent applies parameter changes by systematically varying perception algorithm parameters under different simulated failure conditions. By testing with modified perception results (simulating hardware and environmental changes), the system identifies parameter adjustments that maintain acceptable performance across diverse operating conditions.
Solution Approach 2:
The system performs preliminary testing with simulated perception failures before real-world deployment. By pre-identifying how the perception algorithm degrades under various hardware and environmental changes through simulation, developers can optimize the algorithm and select robust parameters before encountering actual real-world variations.
3Productivity
If the vehicle function is verified with normal perception results, then the basic functionality can be confirmed, but the tolerance to perception algorithm performance degradation cannot be assessed
Solution Approach 1:
The patent applies partial action by separating verification into two stages: first verifying basic functionality with normal perception results, then performing additional targeted verification with simulated failed perception results. This approach efficiently confirms basic operations while specifically addressing perception failure tolerance without requiring complete re-verification of all functions.
Solution Approach 2:
The verification process is segmented into multiple independent verification scenarios. Each scenario tests specific vehicle functions under particular perception failure conditions. This segmentation allows efficient verification of basic functionality separately from failure tolerance assessment, enabling targeted testing that improves overall verification efficiency.
Data Source
AI summary
Embodiments of the present disclosure disclose a method and apparatus for verifying a vehicle function, an electronic device, and a storage medium, wherein the method includes: acquiring operation scenario data respectively corresponding to each of one or operation scenarios; determining, based on the operation scenario data respectively corresponding to each of the operation scenarios, first perception result information corresponding to a preset vehicle function of an ego vehicle in each of the operation scenarios; performing perception failure simulation processing on the first perception result information based on a perception failure simulation rule to obtain second perception result information corresponding to each of the operation scenarios; and verifying the preset vehicle function based on the second perception result information to obtain a verification result.


