Vehicle Digital Twin Validation Under Real Traffic Detachment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for autonomous driving testing are time-consuming, costly, and fail to adequately incorporate real-world traffic uncertainties, primarily focusing on ADAS perception tasks without considering vehicle dynamics and control algorithms, and lack comprehensive validation in real traffic environments.
Innovation Solution
Implementing a digital twin of the vehicle (DriveTwin) that simulates vehicle operation, aligns with real-time sensor feedback, and detaches from real driving when predefined criteria are met to allow deviations, enabling comprehensive validation and early detection of critical failures in real traffic scenarios.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional XiL testing (MiL, SiL, HiL, ViL) and proving ground track testing are used, then validation coverage is improved, but time consumption and cost increase significantly
Solution Approach 1:
The patent creates a digital twin (virtual copy) of the vehicle that replicates the physical vehicle's behavior and characteristics. This virtual model allows comprehensive testing and validation without requiring physical prototypes or track testing, significantly reducing time and cost while maintaining validation coverage. The digital twin can be rapidly instantiated and modified without the constraints of physical hardware.
Solution Approach 2:
The patent introduces a shadow mode testing mechanism that acts as an intermediary between simulation and real-world deployment. The shadow mode runs parallel to the physical vehicle, collecting and analyzing data without interfering with normal operations, enabling gradual validation and reducing the need for extensive physical testing.
2Loss of information
If shadow mode testing is used to test AI pipeline and perception performance, then data collection for ADAS improvement is enabled, but vehicle dynamics and control algorithms cannot be tested
Solution Approach 1:
The digital twin is designed as a universal testing platform that can evaluate multiple aspects of autonomous driving systems simultaneously, including perception algorithms, vehicle dynamics, control strategies, and safety systems. This multi-functional approach eliminates the limitation of shadow mode testing that focuses only on perception tasks.
Solution Approach 2:
The digital twin creates a comprehensive virtual replica that preserves all vehicle characteristics and behaviors, enabling testing of vehicle dynamics and control algorithms in a virtual environment. This copy allows researchers to safely test edge cases and failure scenarios that would be dangerous or impractical to test physically.
3Productivity
If simulation is used to reduce testing time and cost, then efficiency is improved, but mismatch between simulation and real-life traffic remains significant
Solution Approach 1:
The patent implements continuous feedback loops where data from physical vehicles is used to update and refine the digital twin model. This feedback mechanism ensures the virtual model remains synchronized with real-world behavior, reducing the simulation-reality gap while maintaining testing efficiency. The shadow mode also provides feedback by comparing virtual and actual vehicle responses.
Solution Approach 2:
The digital twin is designed as a dynamic model that can adapt its parameters and behavior based on real-time data from physical vehicles. This dynamic adjustment allows the simulation to evolve and improve its realism over time, rather than remaining a static approximation.
Data Source
AI summary
A method of operating a vehicle includes: sensing the vehicle's surroundings and vehicle's operating parameters by sensors; supporting the vehicle's driving by a first autonomous driving control receiving the sensed surroundings and the sensed operating parameters and the first autonomous driving control generating vehicle driving commands for controlling the vehicle's driving; simulating the vehicle operation by a digital twin of the vehicle that includes a second autonomous driving control generating vehicle operation control commands for controlling the digital twin's driving; aligning the vehicle simulation with the vehicle's driving by feedback of the sensed vehicle's surroundings and the sensed vehicle's operating parameters; and evaluating a detachment criterium and changing the operating mode of the vehicle simulation from being aligned with the vehicle's driving to a detachment of the simulation allowing a deviation of the simulated values to at least one of the sensed vehicle's surroundings and/or sensed vehicle's operating parameters.