Networked Autonomous Vehicle Simulator for Modular Algorithm Testing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing autonomous vehicle simulators require long development times for full algorithms due to the need for integrating subsets of algorithms within predetermined virtual environments, limiting diverse evaluations and collaborative development.
Innovation Solution
An autonomous vehicle simulator system utilizing a network platform that divides autonomous vehicle data into multiple data sets, allowing multiple users to share and operate these data sets, facilitating more diverse algorithm evaluations and collaborative development.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If one integrates subsets of algorithms within a predetermined virtual environment to build a full algorithm interface for simulator application, then the algorithm can be verified in the simulator, but it takes a long time for the entire algorithm to be applied and limits diverse evaluations
Solution Approach 1:
The patent segments the autonomous vehicle algorithm into three independent modules: perception module, planning module, and control module. Each module can be developed, tested, and verified separately in the simulator without requiring integration of the entire algorithm. This segmentation allows parallel development and reduces the time needed to apply the complete algorithm to the simulator.
Solution Approach 2:
The patent creates a universal simulator platform that can execute and verify any of the segmented algorithm modules independently. The simulator serves multiple functions: it can verify perception algorithms separately, planning algorithms separately, and control algorithms separately, enabling diverse evaluations without requiring full algorithm integration for each verification.
2Ease of operation
If a developer inputs data into the simulator with a subset algorithm, then the simulation can be performed, but it can only be combined with other algorithms allocated within the predetermined virtual environment, making varied evaluations difficult
Solution Approach 1:
By segmenting the algorithm into independent perception, planning, and control modules, the patent enables developers to input and test any single module or any combination of modules in the simulator. This modular approach allows varied evaluations of different algorithm components without being constrained by predetermined virtual environment allocations.
Solution Approach 2:
The patent implements a dynamic configuration system where the simulator can adaptively combine different segmented algorithm modules based on user needs. Developers can dynamically select and combine modules (e.g., perception + planning, or all three modules) to perform varied evaluations, making the simulation system flexible and adaptable to different verification scenarios.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The present invention relates to an autonomous driving vehicle simulator system that operates a simulator, which is operated on the basis of data in a database, wherein one piece of autonomous driving vehicle data in the database is segmented into a plurality of data sets and each data set includes at least one data packet, so that one data packet is selected in each data set. Therefore, the autonomous driving vehicle simulator system enables multiple users to share their own algorithms with each other and allows users' own algorithms to be interchangeable with other algorithms, so as to enable simulation in a more diverse environment.