Autonomous Driving Scenario Creation With Relative Actor Actions

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current autonomous driving simulation systems lack the ability to define the relative actions of non-ego vehicle actors in relation to ego vehicles, making it difficult to accurately simulate scenarios such as collisions and rotations, which hinders the testing and verification of autonomous vehicle software.

Innovation Solution

A method and apparatus that allow users to set the location and actions of non-ego vehicle actors relative to ego vehicles at specific time points, enabling precise definition of their behavior, including velocity, rotation, and trigger conditions, to create more accurate simulation scenarios.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If individual actor actions are defined separately without relative parameters, then the scenario creation process is simple, but the accuracy of relative timing and spatial relationships between actors cannot be achieved

Engineering Contradiction:
Improveaccuracy of relative timing and spatial relationshipsVSAvoidcomplexity of scenario definition system
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent introduces relative parameter definitions (relative position, relative velocity, relative acceleration, relative orientation) to transform the scenario definition system from absolute coordinate-based definitions to relative parameter-based definitions. This allows precise specification of spatial and temporal relationships between actors while maintaining a systematic approach to scenario creation.

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent adds a relative reference frame dimension to the scenario definition system. By defining actor parameters relative to a reference actor (typically the ego vehicle) rather than in absolute world coordinates, the system enables precise control over relative timing and spatial relationships without significantly increasing operational complexity.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Reliability

If relative parameter definitions are implemented between actors, then accurate simulation of relative actions can be achieved, but the scenario creation process becomes more complex

Engineering Contradiction:
Improveaccuracy of autonomous vehicle behavior testingVSAvoidease of scenario creation
Core Design Contradiction:
ReliabilityVSEase of operation

Solution Approach 1:

The patent establishes a reference actor (ego vehicle) first, then defines other actors' parameters relative to this reference. This preliminary action of setting the reference frame simplifies subsequent relative parameter definitions, as all other actors can be defined in relation to this established reference rather than requiring complex inter-actor coordinate transformations.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system transforms complex spatial and temporal relationships into manageable relative parameters (position, velocity, acceleration, orientation) that can be independently adjusted. This parameterization approach maintains reliability by preserving physical relationship accuracy while improving ease of operation through systematic parameter control.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP4239486A1Apparatus and method of creating scenario for autonomous driving simulation
Publication Date: 2023.09.06 WOVEN BY TOYOTA INC
  • EP4239486A1 patent drawingFigure 1
  • EP4239486A1 patent drawingFigure 2
  • EP4239486A1 patent drawingFigure 3

AI summary

A method and apparatus for creating a scenario for an autonomous driving simulation. The method includes setting, based on a first user input, a location of a first target object at each of a first plurality time points on a map for a simulation scenario, the first plurality of time points comprising a first time point and a third time point; and defining, based on a second user input, an action of a second target object that is different from the first target object in relation to at least one of the set locations of the first target object.