Signal Observers for Reliable ODD Coverage in Virtual Tests
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for determining Operational Design Domain (ODD) coverage in virtual tests for automated driving systems are computationally intensive and difficult to implement in large test spaces, making it challenging to achieve comprehensive testing.
Innovation Solution
A method and system that generate observers for each element of the ODD, allowing automated monitoring and recording of test signals during virtual tests, enabling efficient and reliable determination of ODD coverage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional methods are used to measure ODD coverage in large test spaces, then comprehensive testing can be achieved, but the computational complexity and implementation difficulty increase significantly
Solution Approach 1:
The patent introduces observers as intermediary components that automatically monitor and record signal values during virtual tests. These observers act as mediators between the test execution system and the coverage analysis system, capturing relevant data without requiring complex post-processing or real-time analysis of the entire test space, thus reducing overall system complexity while maintaining reliable coverage measurement
Solution Approach 2:
The patent replaces traditional mechanical/manual methods of tracking ODD coverage with an automated software-based observer system. Instead of manually monitoring and recording test parameters, the system uses programmable observers that automatically detect when signal conditions meet ODD element criteria and record the corresponding coverage information, eliminating the need for complex manual tracking mechanisms
2Loss of information
If traditional scene element extraction methods are used during virtual tests, then complete test data can be captured, but the computational intensity increases significantly
Solution Approach 1:
The patent extracts only the specific signal values and test parameters that are relevant to ODD coverage determination, rather than processing and storing all scene elements generated during virtual tests. The observers are configured to monitor only those signals that correspond to predefined ODD elements, filtering out unnecessary data and reducing computational energy requirements while maintaining complete coverage information
Solution Approach 2:
The patent applies different processing quality levels to different parts of the test data. Instead of uniformly processing all scene elements with high computational intensity, the system selectively applies detailed monitoring and recording only to those specific signals and parameters that are locally relevant to each ODD element, reducing overall computational energy consumption while ensuring complete capture of coverage-critical information
Data Source
Figure 1
Figure 2~3
AI summary
The invention relates to a computer-implemented method and system (1) for generating at least one observer (10a, 10b, 10c) of at least one signal (12a, 12b, 12c) of a virtual test (14) for validating a predetermined function (11) of a motor vehicle, comprising the steps of providing (S1) a data set (16) comprising a plurality of elements (18a, 18b, 18c) of an operational design domain (20) of the predetermined function (11) of the motor vehicle, wherein each element (18a, 18b, 18c) of the operational design domain (20) is defined by a condition (22) of the at least one signal (12a, 12b, 12c) of the virtual test (14). The invention further relates to a system (1) for generating at least one observer (10a, 10b, 10c) of at least one signal (12a, 12b, 12c) of a virtual test (14) for validating a predetermined function (11) of a motor vehicle.