Virtual Sensor Mapping for Realistic ECU Object-Detection Testing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Testing electronic control units (ECUs) for vehicle sensor data is complex, expensive, and many scenarios can only be tested in real-world environments, which is impractical and costly, especially for extreme cases like accidents.
Innovation Solution
A simulator is developed to generate synthetic sensor data using a visibility module, mapping module, and output module to replicate sensor data in a virtual environment, incorporating sensor characteristics without disclosing detailed technical design, allowing real-time computation of partially processed sensor data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If testing is performed in real-world environments with physical sensors, then measurement accuracy and reliability are improved, but testing cost and complexity increase significantly
Solution Approach 1:
The patent creates a virtual copy of the physical sensor and its environment. A virtual sensor model replicates the measurement characteristics of the physical sensor, while a virtual environment reproduces real-world conditions. This allows testing to be performed on virtual sensor data without requiring physical sensors or real-world test scenarios, thereby maintaining reliability while reducing complexity and cost.
Solution Approach 2:
The patent introduces synthetic sensor data as an intermediary between the physical sensor and the control unit under test. Instead of directly connecting physical sensors to ECUs for testing, the system generates synthetic data that mimics physical sensor output and feeds it to the ECU. This intermediary approach allows comprehensive testing without the complexity of physical sensor integration and real-world test setups.
2Adaptability or versatility
If testing is performed in real-world environments, then extreme cases and accident scenarios can be tested, but cost and practicality deteriorate
Solution Approach 1:
The patent enables preliminary testing of extreme and accident scenarios in a virtual environment before real-world deployment. By pre-configuring virtual environments that represent extreme conditions (accidents, edge cases, rare scenarios), the system can test ECU responses to these situations without actually recreating dangerous real-world conditions. This preliminary virtual testing covers all possible scenarios including extreme cases while being safe, cost-effective, and easily repeatable.
3Measurement precision
If detailed technical design of sensor is disclosed for accurate simulation, then measurement precision is improved, but intellectual property protection deteriorates
Solution Approach 1:
The patent extracts only the essential measurement characteristics and functional behavior of the physical sensor without requiring disclosure of detailed technical design or proprietary implementation specifics. The virtual sensor model captures the input-output relationships, measurement ranges, and functional response characteristics needed for accurate synthetic data generation, while deliberately excluding sensitive design details, manufacturing processes, and proprietary algorithms. This extraction approach maintains measurement precision for testing purposes while protecting intellectual property.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
The application relates to a simulator (10) for simulating a sensor for object detection, wherein the simulator (10) is configured to generate synthetic sensor data (32) of the sensor by means of the simulation. The simulator (10) comprises: a visibility module (12) configured to generate object data (22) for at least one virtual object (O1-O6) located in a virtual environment (30), depending on the sensor; a mapping module (14) configured to determine at least two parameters of the virtual object (O1-O6) from the object data (22) and to represent the at least one virtual object (O1-O6) in a first mapping (24) depending on the at least two parameters; and a linking module (16) configured to transform the first mapping (24) into a second mapping (26), taking into account a characteristic function (46) of the sensor.wherein the second mapping (26) represents at least one virtual object (O1-O6) as a function of at least two parameters, an output module (20) configured to convert the second mapping (26) into the synthetic sensor data (32) and to make the synthetic sensor data (32) available for output. The application further relates to a method for simulating a sensor for object detection and to a computer program product.