Autonomous Driving Test Scenarios With Non-Tracked Object Recognition

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing scenario-based testing methods for autonomous driving systems focus primarily on determination and control functions, neglecting the importance of accurate recognition, which is crucial for correct operation, necessitating a solution that incorporates recognition testing based on real-world data.

Innovation Solution

A system and method for generating autonomous driving test scenarios that include tracked and non-tracked object data, analyzing non-tracking factors, and integrating these elements into the operational design domain to create comprehensive test scenarios that evaluate recognition, determination, and control functions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If scenario-based testing focuses only on determination and control functions, then testing efficiency is improved, but recognition function testing is neglected

Engineering Contradiction:

Improvetesting efficiencyVSAvoidrecognition function testing

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The testing system is segmented into distinct modules: a scenario generation unit that creates test scenarios from real driving data, a recognition testing unit that specifically evaluates recognition functions, and a determination-control testing unit. This segmentation allows recognition testing to be conducted independently and systematically without compromising overall testing efficiency.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Real driving data is collected and processed in advance to generate pre-defined test scenarios containing various non-tracking objects and edge cases. These pre-prepared scenarios are then used for systematic recognition testing, eliminating the need for time-consuming manual scenario creation during actual testing phases.

Inventive Principle:

Principle #10Preliminary action

2Ease of operation

If normal data is used for recognition testing, then testing simplicity is maintained, but accuracy of recognition testing is insufficient

Engineering Contradiction:

Improvetesting simplicityVSAvoidrecognition testing accuracy

Core Design Contradiction:

Ease of operationVSMeasurement precision

Solution Approach 1:

The system transforms normal driving data into specialized test scenarios by modifying key parameters: introducing non-tracking objects with specific attributes (color, shape, size, reflectivity), adjusting object positions and motion states, and creating edge case situations. This parameter transformation maintains data structure simplicity while dramatically improving recognition testing accuracy.

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

Real driving data is copied and replicated to generate multiple test scenarios with varying conditions. Instead of creating entirely synthetic data, the system copies real-world data structures and systematically modifies them to include challenging recognition cases, preserving the authenticity and complexity of real driving environments.

Inventive Principle:

Principle #26Copying

3Device complexity

If tracked object data only is used, then data processing simplicity is improved, but comprehensiveness of test scenarios is insufficient

Engineering Contradiction:

Improvedata processing complexityVSAvoidtest scenario comprehensiveness

Core Design Contradiction:

Device complexityVSAdaptability or versatility

Solution Approach 1:

The system merges tracked object data with non-tracked object data to create comprehensive test scenarios. By combining these data sources and integrating them into unified scenario structures, the system achieves comprehensive test coverage including both standard tracked objects and challenging non-tracked objects without proportionally increasing processing complexity.

Inventive Principle:

Principle #5Merging (Combining)

Solution Approach 2:

The scenario generation system is designed with multi-functionality to handle diverse data types: it can process tracked object data, non-tracked object data, environmental data, and road information simultaneously. This universal processing capability allows a single system to generate comprehensive test scenarios covering various testing requirements without requiring separate specialized systems.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Data Source

PatentEP4575804A1System and method for generating data-based autonomous driving test scenarios for testing and evaluation of autonomous driving systems

Publication Date: 2025.06.25 KOREA ELECTRONICS TECH INST

Figure 1~2
Figure 3~4
Figure 5

AI summary

There are provided a system and a method for generating data-based autonomous driving test scenarios for testing and evaluation of autonomous driving system. An autonomous driving test scenario generation method according to an embodiment generates tracked objects by using tracked object data that an autonomous driving vehicle collects while traveling, determines some non-tracked objects among the tracked objects and reflects a non-tracking factor on non-tracked object data, and generates an autonomous driving test scenario by using elements within an operational design domain of an autonomous test scenario and the non-tracked object data reflecting the non-tracking factor. Accordingly, scenarios for testing not only determination, control functions but also a recognition function are generated based on data acquired by autonomous driving, so that more inclusive, detailed testing and evaluation can be provided for autonomous driving systems performing dynamic driving tasks.