Sensor Fusion Weighting via Boundary Conditions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing sensor fusion methods in automated driving systems are susceptible to incorrect measurements and false positives/negatives due to external conditions, leading to degraded performance and safety concerns in safety-critical functions.
Innovation Solution
A method that assigns weighting factors to environment-related parameters from different sensor systems based on their boundary conditions, using an evaluation scheme that incorporates expert knowledge and neural networks to mitigate inaccuracies and leverage sensor capabilities, thereby improving fusion robustness and safety.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sensor data is fused into a secured environment model to represent the environment, then the scope and quality of the environment model are improved, but the system becomes susceptible to incorrect measurements and false positives/negatives due to external conditions
Solution Approach 1:
The patent applies preliminary action by evaluating boundary conditions before sensor data fusion occurs. The system determines whether each sensor system is operating within acceptable boundary conditions (e.g., weather, lighting, spatial range) before using its data, thereby preventing incorrect measurements from degrading the environment model quality in the first place
Solution Approach 2:
The patent changes the parameter of data availability by introducing a binary indicator (available/unavailable) based on boundary condition evaluation. When a sensor system operates outside acceptable boundary conditions, its data is marked as unavailable and excluded from fusion, effectively changing the state of the parameter from continuous data to binary availability status
2Productivity
If common algorithms for object detection and classification are used, then processing efficiency is improved, but the algorithms fail under degraded sensor performance due to false positive and false negative determinations
Solution Approach 1:
The system performs preliminary evaluation of boundary conditions before object detection algorithms process sensor data. By determining in advance whether a sensor system is operating within acceptable conditions, the system prevents algorithms from processing degraded data that would lead to false positives or negatives, thus maintaining both efficiency and accuracy
3Adaptability or versatility
If sensor systems operate under diverse boundary conditions (weather, spatial range, object properties), then system adaptability is improved, but the performance and capabilities of individual sensors degrade under certain conditions
Solution Approach 1:
The patent applies local quality by evaluating boundary conditions specific to each sensor system and its operating context. Instead of treating all sensor data uniformly, the system determines locally whether each sensor is operating within acceptable conditions for its specific type and spatial range, allowing the system to adapt to diverse conditions while maintaining reliability through selective data exclusion
Data Source
AI summary
A method for fusing environment-related parameters includes providing at least one environment-related parameter of a first sensor system; providing at least one environment-related parameter of a second sensor system; providing at least one respective boundary condition that existed when the respective at least one environment-related parameter of the respective sensor system was determined; assigning the at least one respective boundary condition to the respective at least one environment-related parameter of the respective sensor system; providing an evaluation scheme which assigns weighting factors to the respective environment-related parameter of the respective sensor system depending on the respective boundary conditions; and fusing, with a fusion device, the at least one environment-related parameter of the first sensor system with the at least one environment-related parameter of the second sensor system according to the assigned weighting factors to determine the representation of the environment of the sensor systems.