Parameterizable Skeletons for Autonomous Vehicle Simulation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Training and testing autonomous vehicles (AVs) in the physical world is costly, time-consuming, and challenging, especially for rare events, making it difficult to achieve comprehensive testing coverage without relying heavily on real-world road tests.
Innovation Solution
A simulator that uses parameterizable skeletons, attributes, and appearance features to create a virtual environment for AVs, allowing for the generation of diverse scenarios and efficient tuning of vehicle configurations, reducing the need for extensive real-world testing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If AVs are trained and tested in the physical world, then the testing results are realistic and reliable, but the process is costly, time-consuming, and cannot achieve comprehensive testing coverage
Solution Approach 1:
The patent creates virtual copies of real-world environments, vehicles, and scenarios in a simulation system. These virtual replicas allow comprehensive testing of autonomous vehicle configurations without requiring physical road tests, thereby reducing time and cost while maintaining testing reliability through realistic virtual scenario reproduction
Solution Approach 2:
The system performs preliminary testing and validation of autonomous vehicle configurations in the virtual environment before deploying to the physical world. By pre-testing thousands of scenarios including rare events in simulation, the system prepares vehicle configurations in advance, reducing the need for extensive real-world testing and accelerating deployment time
2Adaptability or versatility
If AVs are trained and tested in the physical world, then the testing coverage can be comprehensive, but the cost and time required are excessive
Solution Approach 1:
The simulation system creates virtual copies of diverse real-world scenarios, environments, and vehicle configurations, enabling comprehensive testing coverage across multiple conditions simultaneously. This virtual replication allows the system to evaluate thousands of scenarios in parallel, dramatically improving testing efficiency compared to sequential physical testing
Solution Approach 2:
The patent adds a virtual dimension to the testing process by creating a parallel simulated environment. This additional dimension allows comprehensive testing coverage to be achieved without proportionally increasing physical resources, as virtual scenarios can be generated and tested at minimal marginal cost compared to physical road tests
3Productivity
If a simulator is used to train and test AVs, then the process becomes efficient and scalable, but the accuracy and relevance of vehicle configurations may be reduced
Solution Approach 1:
The simulation system incorporates feedback loops where virtual testing results are used to refine and improve vehicle configurations, which are then validated in subsequent simulation iterations. This iterative feedback process progressively enhances configuration accuracy while maintaining high testing efficiency, bridging the gap between virtual simulation and physical reality
Solution Approach 2:
The system dynamically adjusts simulation parameters, scenario complexity, and vehicle configurations based on testing progress and performance metrics. This dynamic adaptation allows the simulation to evolve from simplified initial models to increasingly realistic representations, improving measurement precision while maintaining computational efficiency throughout the testing process
Data Source
AI summary
Examples of the present disclosure provide a computer-implemented system, comprising instructions for performing operations including: retrieving real-world data comprising skeleton attributes of various skeletons; receiving instructions to generate a simulated skeleton for a scene in a simulation; generating the simulated skeleton according to the scene based on the skeleton attributes, the simulated skeleton comprising a generic skeleton modified by scaling factors according to the scene; building a simulation asset using the simulated skeleton; and determining a reaction of a vehicle to the simulation asset in the scene simulated, the reaction of the vehicle being a function of a configuration of the vehicle.