Predictive Network Model for Real-Time Infrastructure Sensor Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network monitoring systems face challenges in providing real-time data due to the complexity and long processing times of simulation models, making it difficult to incorporate these models into real-time monitoring and decision-making processes.
Innovation Solution
A computer-implemented method that uses a predictive network model to generate predicted real-time data for a future time step, which is then input into a more complex network model, allowing for the generation of network state data in real-time, even for components or locations without direct sensor data, using graph neural networks for data fusion and normalization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If complex network simulation models are used to determine network state data, then measurement precision and insight quality are improved, but processing time increases significantly making real-time monitoring impossible
Solution Approach 1:
The system performs preliminary actions by training multiple machine learning models offline using historical data from complex simulation models. These pre-trained models capture the relationships between network parameters and states. During real-time operation, the pre-trained models rapidly process current sensor data to predict future network states, avoiding the need to run complex simulations in real-time while maintaining accuracy.
Solution Approach 2:
The system creates simplified copies of the complex simulation models in the form of machine learning models. These ML models are trained to replicate the behavior and output of the complex simulation models but with significantly reduced computational requirements. The copied models can then be executed in real-time on resource-constrained devices while maintaining the predictive accuracy of the original complex models.
2Loss of information
If real-time sensor data is collected from all network components, then data completeness is improved, but device complexity and data processing requirements increase
Solution Approach 1:
The system introduces machine learning models as intermediary components that process and interpret sensor data. These models act as mediators between raw sensor inputs and network state determination, automatically selecting relevant features and relationships. This intermediary layer reduces the burden on the overall system complexity while maintaining comprehensive data utilization through intelligent data processing.
Solution Approach 2:
The machine learning models serve multiple functions simultaneously: they process data from various sensor types, handle missing data through imputation, predict future states, and provide interpretability. This multi-functionality consolidates what would otherwise require multiple separate systems into a single unified framework, reducing overall device complexity while maintaining data completeness.
3Productivity
If complex models are executed in real-time, then productivity is improved, but device complexity and computational resources required increase
Solution Approach 1:
The system replaces expensive, complex simulation models with cheaper machine learning models that have significantly lower computational requirements. These ML models can be executed rapidly on resource-constrained devices such as edge computers or even embedded systems within the network. The trade-off accepts slightly simplified model architecture in exchange for enabling real-time execution on affordable hardware.
4Loss of time
If predictive models are used to generate future state data, then time advantage is improved, but measurement precision may be reduced compared to actual sensor data
Solution Approach 1:
The system implements feedback mechanisms where predicted future states are continuously compared with actual sensor measurements when they become available. This feedback is used to refine and retrain the predictive models, improving their accuracy over time. The system also uses feedback to identify when predictions diverge significantly from actual behavior, triggering model updates or alternative processing paths to maintain precision.
Data Source
AI summary
A computer-implemented method for determining data indicative of a state of a network. The method comprises determining a future time step based upon a current time step and a prediction time period, receiving first real-time data associated with the network, inputting the first real-time data to a predictive network model trained to generate predicted real-time data, to generate predicted real-time data for the future time step, executing a first network model based upon the predicted real-time data for the future time step to generate network model output data.