Autonomous driving testing method based on multi-coalition swarm confrontation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional autonomous driving simulation testing methods struggle to generate high-risk boundary scenarios that dynamically interact with tested vehicles, leading to low efficiency and poor scenario fit, making it difficult to effectively test autonomous driving systems.
Innovation Solution
An autonomous driving testing method based on multi-alliance cluster confrontation, utilizing reinforcement learning to divide background vehicles into clusters with varying confrontation intensities and implementing alliance game methods for cooperative behavior decision-making, followed by trajectory planning to generate highly confrontational scenarios.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If traditional parameter combination method is used to generate testing scenarios, then the testing process is simple to implement, but the scenario fit with tested vehicles is low and high-risk boundary scenarios are difficult to generate
Solution Approach 1:
The patent transforms the testing scenario generation from static parameter combination to dynamic parameter adjustment based on vehicle states. Background vehicles adjust their behavior parameters (speed, acceleration, trajectory) in real-time based on the tested vehicle's state, creating high-fit scenarios that adapt to different testing conditions while maintaining implementation feasibility through algorithmic control.
Solution Approach 2:
The patent introduces dynamic interaction between background vehicles and tested vehicles. Instead of fixed parameter combinations, background vehicles continuously adjust their behavior based on real-time state feedback, creating living scenarios that evolve during testing. This dynamic approach enables generation of high-risk boundary scenarios while maintaining systematic control.
2Ease of operation
If traditional digital mapping of trajectory data is used, then the method is straightforward, but the ability to dynamically interact with tested vehicles is lacking
Solution Approach 1:
The patent implements feedback mechanisms where background vehicles continuously monitor the tested vehicle's state and adjust their behavior accordingly. This feedback loop enables dynamic interaction while maintaining operational simplicity through automated decision-making algorithms that process state information and generate appropriate responses without complex manual intervention.
Solution Approach 2:
Background vehicles autonomously adjust their behavior based on the tested vehicle's state without external intervention. The system serves itself by automatically generating high-fit scenarios through embedded decision-making algorithms, combining operational simplicity with advanced dynamic interaction capabilities.
3Reliability
If environmental confrontation method is used to generate testing scenarios, then the confrontation intensity with tested vehicles is high, but the complexity of controlling background vehicle behavior increases
Solution Approach 1:
The patent segments background vehicles into different types (aggressive, defensive, cooperative) with predefined behavior patterns. This segmentation reduces control complexity by providing modular behavior templates while maintaining high confrontation effectiveness through strategic deployment of different vehicle types in appropriate scenarios.
Solution Approach 2:
The patent uses parameter changes to control background vehicle behavior, adjusting key parameters like speed, acceleration, and trajectory based on tested vehicle state. This approach maintains confrontation effectiveness through dynamic parameter adjustment while reducing control complexity by focusing on key behavioral parameters rather than full motion control.
Data Source
AI summary
The present invention proposes an autonomous driving testing method based on multi-alliance cluster confrontation, aiming to improve the efficiency and accuracy of autonomous driving simulation test, and comprising S1—initialization of testing environment of autonomous driving; S2—decision-making for dividing background vehicle clusters; S3—decision-making for confrontation behaviors of the background vehicles; S4—trajectory planning for the background vehicles; and S5—looping through the steps S2, S3, and S4 until cluster confrontation testing tasks are completed. The method of the present invention uses reinforcement learning and alliance games to dynamically generate testing scenarios that are highly confrontational to vehicles being tested, which can find dangerous boundary scenarios for autonomous driving more quickly and improve the efficiency of simulation testing.