Switchable Autonomous Driving Simulation Models for Lower Test Load
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current autonomous driving simulation systems are limited by their inability to integrate variably-sourced simulation data models, leading to unnecessary costs, processing burdens, and extended testing times due to the inclusion of non-essential models and high accuracy sensors for all applications.
Innovation Solution
A customizable autonomous driving simulation architecture that allows for the integration of multiple simulation data models from different sources, enabling the selective use or replacement of models with lower accuracy 'dummy' models based on application requirements, thereby reducing processing load and costs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a common simulation architecture integrates all simulation data models (vehicle simulator, sensor simulator, etc.) to ensure comprehensive testing capability, then testing reliability is improved, but device complexity and processing load increase
Solution Approach 1:
The simulation architecture is segmented into independent, switchable model modules (vehicle simulator, sensor simulator, core simulator) that can be selectively activated. Each module operates independently and can be connected or disconnected based on testing requirements, reducing overall system complexity while maintaining reliability when needed.
Solution Approach 2:
The simulation architecture implements dynamic configurability where the connection status of each simulation model can be changed during operation. Switches allow the system to transition between different operational states (full integration, partial integration, or minimal integration) based on the specific testing needs of different applications.
2Measurement precision
If highly accurate sensor simulators are integrated into the simulation architecture to ensure data accuracy, then measurement precision is improved, but processing load and cost increase
Solution Approach 1:
Different levels of simulation accuracy are applied locally based on specific testing requirements. Instead of using high-accuracy models universally, the system allows selective activation of accurate sensor simulators only when and where precision is needed for particular applications, reducing overall processing load while maintaining necessary accuracy.
Solution Approach 2:
The system enables dynamic changing of simulation parameters and model fidelity levels. Simulation models can be switched between different accuracy modes or completely disconnected based on the specific testing scenario, allowing optimization of processing load while maintaining measurement precision when required.
3Adaptability or versatility
If all simulation data models are integrated into the core simulator to provide comprehensive testing capability, then adaptability is improved, but processing time increases
Solution Approach 1:
The system performs preliminary configuration by allowing users to select and pre-configure only the necessary simulation models before executing tests. By anticipating which models are needed for specific application types, the system avoids unnecessary processing of unrelated models, reducing processing time while maintaining adaptability.
Solution Approach 2:
The core simulator is designed with universal interfaces that can connect to different simulation models as needed. This multi-functional architecture allows the same core simulator to work with various combinations of models (vehicle only, sensor only, both, or neither) without requiring separate systems, thus maintaining adaptability while enabling selective activation to reduce processing time.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A device and a method for providing an autonomous driving simulation architecture for testing an application. The method includes integrating a plurality of simulation data models with a first core simulator that runs an autonomous vehicle simulation for testing the application; selectively connecting the first core simulator to the application; and selectively disconnecting a first simulation data model, among the plurality of simulation data models, from the first core simulator.