Environment Sensor Behavior Modeling for Precise Virtual Road Tests
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for simulating environment sensors in driver assistance systems face challenges due to the complexity of vehicle components and the lack of precise specifications, especially for sensors, which hinders accurate virtual road testing.
Innovation Solution
A method involving the training of an artificial neuronal network using traffic scenarios to generate and analyze sensor data, allowing for the simulation of environment sensors by importing and processing data streams, including image streams, to replicate the behavior of real sensors, even under varying conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a physical model is used to simulate sensor operation, then a high level of abstraction is required, but the simulation precision deteriorates due to lack of exact sensor specifications
Solution Approach 1:
The patent creates a virtual copy of the real sensor by training an artificial neural network to replicate the sensor's behavior. The neural network is trained on actual sensor data from road tests, enabling it to copy the sensor's response characteristics without requiring detailed physical specifications. This allows high-fidelity simulation while maintaining adaptability through the learned model.
Solution Approach 2:
The patent transforms the sensor simulation approach by changing from fixed physical parameters to learned parameters through neural network training. The neural network learns optimal parameters from training data, allowing the simulation to adapt to different sensor characteristics without requiring precise prior knowledge of sensor specifications.
2Ease of manufacture
If vendors do not provide necessary sensor data, then model parameterization becomes difficult, but using artificial neuronal networks enables training with available road test data
Solution Approach 1:
The system performs self-service by automatically training the neural network model using only the road test data that is already available. Instead of requiring external vendor specifications, the system learns sensor characteristics directly from operational data, making the parameterization process self-sufficient and eliminating dependency on vendor-provided specifications.
Solution Approach 2:
The patent performs preliminary training of the neural network with road test data before the actual simulation is needed. This preliminary action creates a pre-trained model that can be directly used for simulation without requiring additional vendor data or complex parameterization at the time of use.
3Productivity
If virtual road tests are to be conducted, then sensor simulation is required, but current methods lack the precision for suitable test data
Solution Approach 1:
The system uses feedback from actual sensor measurements during road tests to train and refine the neural network model. The neural network continuously learns from the discrepancy between predicted and actual sensor readings, improving the precision of virtual sensor data generation with each training iteration.
Solution Approach 2:
The patent replaces the traditional mechanical/physical sensor model with an artificial neural network model. This substitution allows the system to capture complex sensor behaviors that cannot be described by simple physical equations, significantly improving the accuracy of virtual sensor data while maintaining computational efficiency.
Data Source
AI summary
A computer-aided method for training an artificial neuronal network for the simulation of an environment sensor of a driver assistance system includes the following work steps: reading in traffic scenarios of a test journey; deriving test journey data and the sensor data to be output by the environment sensor from the traffic scenarios; generating a data stream which depicts the traffic scenarios from the perspective of the environment sensor; outputting the data stream such that the environment sensor can generate sensor data on the basis of the data stream and can provide same to a data interface at which the test journey data and the sensor data to be output are also provided; and reading the provided data into the artificial neuronal network.