AI Module Classifier for Dynamic Sensor Data Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current approaches to automated driving rely heavily on AI models that vary in functional quality, lacking a dynamic combination of AI modules to ensure safe and effective performance across diverse situations, leading to limitations in adapting to changing conditions.
Innovation Solution
A method and device for creating a classifier that evaluates AI modules based on contextual parameters, determining their functional quality, and selecting the most suitable combination of AI modules and weights for processing input data from sensor systems in motor vehicles, enabling dynamic adaptation to various driving conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a single AI module or ensemble is used with optimization, then the system structure remains simple, but the functional quality and adaptability to different driving situations are insufficient
Solution Approach 1:
The patent implements a dynamic selection mechanism that chooses different AI modules based on contextual parameters such as lighting conditions, weather, and driving situation. The system transitions from a static single-module approach to a dynamic multi-module ensemble that adapts its composition in real-time, resolving the contradiction between simplicity and adaptability.
Solution Approach 2:
The system changes the operational parameters by selecting different AI modules based on contextual parameters. Instead of modifying a single module's architecture or training data, the system changes which module is active based on environmental conditions, achieving adaptability without proportionally increasing system complexity.
2Reliability
If training data is added or architecture complexity is increased to compensate for insufficient functional quality, then the functional quality of a single AI module improves, but the system becomes less flexible in adapting to diverse situations
Solution Approach 1:
The patent divides the AI processing task into multiple specialized modules, each optimized for specific conditions rather than attempting to create a single universal module. This segmentation allows each module to achieve high functional quality for its specific domain while the ensemble provides overall adaptability across diverse situations.
Solution Approach 2:
The system creates a universal ensemble of AI modules that can handle multiple different driving situations. Rather than making a single module universally competent across all conditions, the system achieves universality through the combination of multiple specialized modules, each contributing to overall system versatility.
3Adaptability or versatility
If multiple AI modules are dynamically combined on the vehicle side, then the adaptability to different situations improves, but the complexity of selecting and combining modules increases
Solution Approach 1:
The patent introduces a selection module that acts as an intermediary between the multiple AI modules and the input data. This intermediary evaluates contextual parameters and automatically selects the appropriate module combination, reducing the complexity burden from the overall system by centralizing the selection logic in a dedicated component.
Solution Approach 2:
The system implements feedback mechanisms where the performance of AI modules is continuously evaluated based on ground truth data and contextual parameters. This feedback loop enables automatic adjustment and selection of module combinations, reducing manual configuration complexity while maintaining high adaptability to different driving situations.
Data Source
AI summary
The present invention relates to a method for processing input data provided by a sensor system of a motor vehicle, and also relates to a classifier provided using such a method. In a first step, an AI module to be classified is selected. In addition, a suitable test data set is selected. The AI module is then applied to data points of the test data set. Associated ground truths and contextual parameters are known for the data points. On the basis of the outputs of the AI module, a functional quality is then determined for each of the data points. Finally, a classifier for the AI module is created, which outputs a functional quality for given contextual parameters.