MBSE Scenario Modeling for Complex Autonomous Driving Episodes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing machine learning techniques for autonomous vehicle control, such as self-driving cars, are ineffective in complex interactions and struggle with parsing large datasets to detect and extract episodes of interest, leading to safety concerns and inefficiencies in vehicle control actions.
Innovation Solution
A model-based systems engineering (MBSE) approach is applied to analyze high-level events from driving log data, identifying datasets of interest and extracting episodes of interest using state transition diagrams, which are then used to generate driving scenarios and parameterize agent and scenario models for autonomous vehicle operation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If machine learning techniques are used for vehicle control, then autonomous operation capability is improved, but effectiveness in complex interactions deteriorates
Solution Approach 1:
The patent introduces MBSE models as an intermediary between raw driving log data and machine learning training. The MBSE model analyzes high-level events and extracts episodes of interest, serving as a mediator that transforms unstructured data into structured training scenarios that improve effectiveness in complex interactions while maintaining autonomous operation capability
Solution Approach 2:
The patent extracts specific episodes of interest from large driving log datasets using MBSE state transition diagrams. This extraction process isolates critical complex interaction scenarios from the broader dataset, creating focused training data that improves reliability in complex situations without sacrificing overall automation capability
2Measurement precision
If large datasets are parsed to detect episodes of interest, then detection accuracy is improved, but time consumption increases
Solution Approach 1:
The patent applies MBSE models preliminarily to analyze and structure driving log data before the actual episode detection process. By pre-processing data through MBSE state transition diagrams and identifying high-level events in advance, the system improves detection accuracy while reducing the time required for subsequent episode extraction
Solution Approach 2:
The patent segments the large driving log dataset into manageable high-level events using MBSE analysis. This segmentation divides the complex parsing task into structured event categories, enabling more efficient detection of episodes of interest with both high accuracy and reduced time consumption
3Productivity
If MBSE model is used to analyze high-level events, then episode extraction efficiency is improved, but system complexity increases
Solution Approach 1:
The patent employs MBSE models that serve multiple functions: analyzing high-level events, extracting episodes of interest, and generating structured training data. This multi-functionality improves episode extraction efficiency while avoiding the need for separate specialized systems, thereby managing overall system complexity
Data Source
AI summary
A method for agent and scenario modeling is described. The method includes analyzing, through a model-based systems engineering (MBSE) model, high-level events extracted from driving log data to identify a dataset of interest from the driving log data. The method also includes extracting an episode of interest from the dataset of interest according to an MBSE state transition diagram. The method further includes generating a driving scenario of interest based on the episode of interest. The method also includes utilizing the driving scenario of interest to parameterize agent and/or scenario models used for autonomous operation of an ego vehicle.