Virtual Vehicle Environment Generation From Multi-Sensor Road Data

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for creating virtual vehicle environments for testing highly automated driving functions require significant personnel and cost expenditures due to their complexity, lacking efficient and cost-effective integration of pre-captured video image data, radar data, and lidar point clouds.

Innovation Solution

A computer-implemented method using machine learning algorithms to generate a virtual vehicle environment by comparing pre-captured data with stored synthetic objects, selecting or procedurally generating objects based on similarity measures, and integrating them into a virtual environment, reducing the need for manual object recognition and direct assignment.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of manufacture

If manual configuration and import of objects stored in an object library is used to create the virtual vehicle environment, then the scene setup can be performed, but high personnel and cost expenditure is required

Engineering Contradiction:
Improveease of creating virtual vehicle environmentVSAvoidefficiency of creating virtual vehicle environment
Core Design Contradiction:
Ease of manufactureVSProductivity

Solution Approach 1:

The patent uses pre-captured video image data, radar data, and lidar point clouds as templates to automatically generate virtual vehicle environments. Instead of manually configuring objects, the system copies and reconstructs real-world scenes using sensor data, thereby eliminating manual effort while maintaining scene accuracy.

Inventive Principle:
Principle #26Copying

Solution Approach 2:

The patent replaces the manual mechanical process of object configuration with an automated computer vision system. Machine learning algorithms automatically detect, classify, and place objects in the virtual environment based on pre-captured sensor data, substituting human labor with computational processing.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Ease of manufacture

If manual configuration and import of objects is used for scene setup, then the virtual vehicle environment can be created, but high cost expenditure is required

Engineering Contradiction:
Improveease of creating virtual vehicle environmentVSAvoidcost expenditure
Core Design Contradiction:
Ease of manufactureVSLoss of energy

Solution Approach 1:

The system creates virtual environments by copying real-world sensor data rather than manually creating each object. This approach reduces costs by using automatically captured data from video cameras, radar, and lidar systems already present in the vehicle, eliminating the need for expensive manual scene construction.

Inventive Principle:
Principle #26Copying

Solution Approach 2:

The system performs self-service by automatically processing pre-captured sensor data to generate virtual environments without requiring external manual intervention. The machine learning algorithms autonomously complete the entire pipeline from data capture to virtual scene generation, reducing operational costs.

Inventive Principle:
Principle #25Self-service

3Productivity

If pre-captured video image data, radar data and lidar point cloud are used with machine learning algorithms, then the efficiency of virtual environment creation is enhanced, but complex processing is required

Engineering Contradiction:
Improveefficiency of virtual environment creationVSAvoidcomplexity of processing system
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent segments the complex processing task into distinct modules: video image data processing, radar data processing, and lidar point cloud processing. Each sensor type is handled by specialized machine learning algorithms, dividing the overall complexity into manageable independent components that can be processed separately and then integrated.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces feature vectors as an intermediary representation between raw sensor data and the final virtual environment. Machine learning algorithms extract essential features from multi-source sensor data and represent them as compact feature vectors, which simplifies the subsequent object detection and scene reconstruction processes.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentEP3869390B1Computer-implemented method and system for creating a virtual vehicle environment
Publication Date: 2025.07.23 DSPACE SE & CO KG
  • EP3869390B1 patent drawingFigure 1
  • EP3869390B1 patent drawingFigure 2
  • EP3869390B1 patent drawingFigure 3

AI summary

The invention relates to a computer-implemented method for generating a virtual vehicle environment (U2) for testing highly automated driving functions of a motor vehicle (1) using pre-acquired video image data (10a), radar data, and/or a lidar point cloud of a real vehicle environment (U1). The invention further relates to a system (2) for generating a virtual vehicle environment (U2) for testing highly automated driving functions of a motor vehicle. The invention also relates to a computer program and a computer-readable data carrier.