Virtual Pedestrian Generation for Standardized ADAS Intersection Testing

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current methods for testing Advanced Driver Assistance Systems (ADAS) at intersections are inefficient due to the time-consuming manual creation of realistic virtual pedestrians, which often fail to conform to industry and manufacturer testing standards, rendering virtualized testing invalid.

Innovation Solution

A pedestrian generation module that automatically generates realistic virtual pedestrians based on path and behavior specifications, using digital data from accident reports, industry, and manufacturer standards, integrated into simulation tools like CarSim or Prescan, ensuring conformance and realism.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If manual programming is used to create virtual pedestrians, then customization and uniqueness can be achieved, but the time consumption and labor intensity increase significantly

Engineering Contradiction:

Improveease of pedestrian creationVSAvoidtime to create virtual pedestrians

Core Design Contradiction:

Ease of operationVSLoss of time

Solution Approach 1:

The system pre-defines multiple pedestrian templates with standardized characteristics, behaviors, and attributes that conform to testing standards. These templates are prepared in advance and can be automatically instantiated during simulation setup, eliminating the need for manual creation of each pedestrian from scratch.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system automatically generates virtual pedestrians by copying and instantiating predefined templates with varying parameters. This allows rapid multiplication of realistic pedestrian models without manual intervention, maintaining consistency with testing standards while enabling batch generation.

Inventive Principle:

Principle #26Copying

2Adaptability or versatility

If ad-hoc manual creation of virtual pedestrians is used, then flexibility is maintained, but reliability and conformance to testing standards cannot be guaranteed

Engineering Contradiction:

Improveflexibility in pedestrian creationVSAvoidconformance to testing standards

Core Design Contradiction:

Adaptability or versatilityVSReliability

Solution Approach 1:

Testing standard requirements are incorporated into the pedestrian templates during their preliminary definition. Each template includes pre-configured attributes, behaviors, and characteristics that satisfy specific testing standard criteria, ensuring automatic conformance when instantiated.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system provides feedback mechanisms to verify that generated pedestrians meet testing standard requirements. This includes validation checks on pedestrian attributes, behaviors, and scenarios to ensure compliance with applicable standards before simulations are executed.

Inventive Principle:

Principle #23Feedback

3Manufacturing precision

If more virtual pedestrians are created to improve testing comprehensiveness, then test coverage increases, but the time and resources required for creation increase

Engineering Contradiction:

Improvetest coverage qualityVSAvoidcreation efficiency

Core Design Contradiction:

Manufacturing precisionVSProductivity

Solution Approach 1:

The pedestrian templates are designed to be universal and multi-functional, capable of representing various pedestrian types, behaviors, and scenarios through parameter variation. A single template can serve multiple testing purposes by adjusting its parameters, reducing the need to create entirely separate models for different test cases.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Solution Approach 2:

The system enables rapid generation of diverse pedestrian populations by changing parameters of base templates rather than creating new models from scratch. This includes modifying attributes such as walking speed, trajectory, reaction time, and behavioral patterns to generate statistically significant variations for comprehensive testing.

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS11810460B2Automatic generation of pedestrians in virtual simulation of roadway intersections

Publication Date: 2023.11.07 TOYOTA JIDOSHA KK

AI summary

The disclosure includes embodiments for automatically generating a virtual pedestrian for inclusion in a digital simulation. Some embodiments of a method may include automatically generating a virtual pedestrian based on a path specification and a behavior specification. The method may include executing a digital simulation that includes the virtual pedestrian crossing an intersection and a virtual vehicle responding to the virtual pedestrian crossing the intersection, wherein the digital simulation is operable to measure a performance of a virtual Advanced Driver Assistance System (virtualized “ADAS system”) included in the virtual vehicle to protect the virtual pedestrian when responding to the virtual pedestrian crossing the intersection. The method may include displaying a set of visualizations on a display, wherein the set of visualizations on the display visually describe the performance of the virtual ADAS system to protect the virtual pedestrian when responding to the virtual pedestrian crossing the intersection.