Vehicle-in-Virtual-Environment for Autonomous Driving Testing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for developing and testing autonomous vehicles (AVs) rely heavily on simulations with lower fidelity and accuracy, making it difficult to safely and cost-effectively test and evaluate AV systems, especially in replicating real-world scenarios and extreme events, which are crucial for their development and deployment.
Innovation Solution
The Vehicle-in-Virtual-Environment (VVE) approach uses a computational system and software to generate a highly realistic 3D environment and simulate sensor information, allowing an AV to operate as if it is in a real-world setting, enabling safe, repetitive, and scalable testing and evaluation of AV driving algorithms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If traditional simulation methods (MIL, HIL, VIL) are used for AV testing, then development cost and time are reduced, but the fidelity and accuracy of testing results deteriorate
Solution Approach 1:
The patent introduces a virtual environment as an intermediary between the physical AV and the real world. The AV remains physically present but operates within a simulated virtual environment that provides realistic sensor inputs and scenarios without requiring actual public road deployment, thus achieving high-fidelity testing without the complexities of real-world deployment
Solution Approach 2:
The patent creates a virtual copy of the real-world environment that replicates physical conditions, sensor behaviors, and driving scenarios. This virtual copy allows for repeated, controlled testing of extreme and rare events without needing to physically recreate them, achieving high testing fidelity with reduced complexity
2Reliability
If public road testing is conducted to capture rare extreme events, then testing comprehensiveness is improved, but safety and operational risk worsen
Solution Approach 1:
The patent enables preliminary testing of AV systems in a virtual environment before actual public road deployment. By pre-testing various scenarios including rare extreme events in the virtual environment, the system identifies and resolves issues beforehand, ensuring comprehensive testing without exposing the AV to real-world safety risks
Solution Approach 2:
The patent converts the harmful aspect of real-world testing (safety risks) into a benefit by creating a virtual environment where dangerous scenarios can be safely reproduced. The virtual environment allows testing of hazardous conditions that would be too risky in reality, turning the limitation into an advantage for comprehensive yet safe testing
3Manufacturing precision
If city block replicas are created for testing, then urban scenario testing is improved, but environment adaptability and testing versatility deteriorate
Solution Approach 1:
The patent implements a dynamic virtual environment that can be reconfigured and modified programmatically. Unlike static physical city block replicas, the virtual environment can dynamically change scenarios, weather conditions, traffic patterns, and urban layouts, providing both accurate replication and high adaptability for testing various environments and conditions
Data Source
AI summary
The Vehicle-in-Virtual-Environment (VVE) technology described herein is directed to the use of a computational system and software that generates a highly realistic real world or representative 3D environment, generates raw perception and localization and communication sensor information from that environment, and feeds these to the autonomous driving computer system in an autonomous vehicle (AV) such that the autonomous driving system behaves as if it is in that environment.