Autonomous Driving Simulation for User-Created Edge Case Testing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current simulation methods for validating autonomous vehicle systems are limited by the creativity and experience of experts, making it difficult to cover all realistic driving scenarios, and existing simulations are often restricted to predefined conditions.
Innovation Solution
A computer-implemented simulation method and system that allows users to create and execute diverse driving scenarios, enabling the testing of control units and driving programs in a virtual environment, where users can interact with both manually controlled and autonomous vehicles, collecting data on their interactions and behavior, and storing relevant data for analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If simulations are limited to predefined scenarios created by experts, then the simulation system remains manageable and controllable, but the coverage of all realistic driving scenarios is insufficient
Solution Approach 1:
The system enables users to autonomously create and execute their own driving scenarios without requiring expert intervention. The simulation platform provides self-service capabilities where any user can define test scenarios, upload them to the server, and execute them automatically, eliminating the bottleneck of expert-dependent scenario creation while expanding scenario diversity
Solution Approach 2:
The simulation server is designed to handle multiple functions: it stores diverse driving scenarios from various users, manages execution of these scenarios, collects test data from autonomous vehicle simulations, and provides a universal platform for all users to access and contribute scenarios. This multi-functional design allows the system to scale without proportionally increasing operational complexity
2Reliability
If millions of kilometers of real-world testing are conducted, then comprehensive validation data is obtained, but time consumption and costs increase significantly
Solution Approach 1:
The system creates virtual copies of real-world driving scenarios in a simulated environment. Instead of physically driving vehicles for millions of kilometers, the simulation server executes digital replicas of diverse driving situations, including rare and extreme cases, allowing comprehensive validation data collection in a fraction of the time and cost of physical testing
Solution Approach 2:
The system performs preliminary simulation testing of driving scenarios before real-world deployment. By pre-executing diverse scenarios including edge cases and failure conditions in the virtual environment, the system identifies and resolves issues beforehand, reducing the need for extensive post-deployment validation and accelerating the overall validation process
3Measurement precision
If diverse and extreme driving scenarios are tested, then edge cases and functional deficiencies are identified, but the test system requires more sophisticated scenario creation capabilities
Solution Approach 1:
The system empowers regular users to create and submit their own extreme and edge case scenarios without requiring specialized expertise. The platform provides easy-to-use tools for scenario definition and automatic execution, allowing anyone to contribute valuable test cases that challenge autonomous vehicle systems under unusual conditions
Solution Approach 2:
The simulation server provides automated feedback on submitted scenarios, executing them and returning test results that indicate whether edge cases were successfully identified. This feedback mechanism guides users in creating more effective extreme scenarios while maintaining ease of operation, as the system automatically validates and refines scenario quality
Data Source
AI summary
A computer-implemented simulation method is described for testing a driving program for the at least partial autonomous guidance of a test vehicle. The test vehicle is moved by the driving program on a simulated road. A second vehicle in the simulation method is moved as a function of control commands of a user. The test vehicle and the second vehicle are displayed to the user including the road, data relating to the guidance of the test vehicle being collected and stored during the simulation method. A system is also described.