Autonomous Vehicle Log Perturbation for Realistic Simulation Coverage
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in efficiently generating realistic simulated log data for testing and improving their control software, as existing methods often result in unrealistic datasets and require extensive real-world data points or manual simulation creation.
Innovation Solution
The system perturbs real-world log data to generate simulated log data by modifying parameters such as object trajectories, speeds, and types, creating a simulated driving environment that mimics real-world scenarios, allowing for extensive simulation testing without new real-world data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If real-world log data is used directly for simulation testing, then the data reflects actual driving conditions, but the quantity of available test scenarios is limited and real-world data collection is time-consuming
Solution Approach 1:
The patent applies parameter changes by systematically modifying real-world log data parameters including object trajectories, speeds, types, and environmental conditions to generate diverse simulated scenarios. This allows multiplication of test scenarios from limited real-world data while maintaining realism through controlled parameter variations.
Solution Approach 2:
The patent creates simulated log data as copies of real-world log data, then applies perturbations to these copies to generate variations. This copying approach enables extensive scenario generation without requiring additional real-world data collection, resolving the contradiction between data realism and scenario quantity.
2Adaptability or versatility
If simulated log data is generated by manually creating simulation scenarios, then diverse test cases can be created, but the process is time-consuming and requires extensive manual effort
Solution Approach 1:
The patent implements self-service by automatically generating simulated log data through computational processing of real-world data with algorithmic perturbations. This eliminates the need for manual simulation creation, enabling diverse scenario generation without proportional increases in time investment.
Solution Approach 2:
The patent performs preliminary action by collecting and storing real-world log data in advance, which then serves as the foundation for generating multiple simulated scenarios. This preliminary data collection, combined with automated processing, reduces the time required for subsequent scenario generation while maintaining diversity.
3Reliability
If extensive real-world data collection is performed to improve simulation quality, then more comprehensive test coverage is achieved, but the cost and time of data collection increase significantly
Solution Approach 1:
The patent performs preliminary action by collecting real-world log data in advance, which then serves as the foundation for generating multiple simulated scenarios. This preliminary data collection, combined with automated processing, reduces the time required for subsequent scenario generation while maintaining diversity.
Solution Approach 2:
The patent creates simulated log data as copies of real-world log data, then applies perturbations to these copies to generate variations. This copying approach enables extensive scenario generation without requiring additional real-world data collection, resolving the contradiction between data realism and scenario quantity.
Data Source
AI summary
A method includes receiving a set of real-world log data defining a real-world driving environment and generated by an autonomous vehicle (AV). The set of real-world log data includes at least one set of real-world parameters defining at least one real-world object observed by the AV operating within the real-world driving environment. The method further includes generating, based on the set of real-world log data, a set of simulated log data defining a simulated driving environment and causing a simulation to be performed using the set of simulated log data. The set of simulated log data comprises at least one set of simulated parameters defining at least one simulated object within the simulated driving environment, and generating the set of simulated log data includes perturbing at least one real-world parameter of the at least one set of real-world parameters to obtain the at least one set of simulated parameters.