Agent based air purification system using hybrid physics-machine learning model
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing air purification systems fail to adequately address dynamic flow variations in particle concentration caused by human respiratory events like coughs or sneezes, lacking adaptive and autonomous capabilities.
Innovation Solution
A hybrid physics-machine learning model, comprising a digital twin with a physics-based compartment model and a machine learning model, predicts aerosol dispersion and concentration using a robotic cough agent to simulate events and train the system, enabling autonomous control of mobile air purifiers to remediate aerosol concentrations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a physics-based compartment model is used to predict aerosol concentration, then the system provides interpretability and foundational physical understanding, but it fails to capture complex dynamic flow variations caused by human respiratory events
Solution Approach 1:
The patent combines a physics-based compartment model with a machine learning model into a hybrid system. The physics model provides structural framework and interpretability, while the ML model captures complex dynamic patterns from sensor data. The ML model's predictions are integrated with physics model outputs to produce final aerosol concentration estimates, resolving the contradiction between physical interpretability and dynamic adaptability.
Solution Approach 2:
The machine learning model acts as an intermediary between raw sensor measurements and the physics-based compartment model. It processes complex respiratory event patterns and translates them into corrections or supplements for the physics model, enabling the system to respond dynamically to coughs, sneezes, and other respiratory events while maintaining physical consistency.
2Adaptability or versatility
If traditional air purification systems are used, then the system structure is simple, but it lacks adaptive capabilities to respond to dynamic particle concentration changes
Solution Approach 1:
The system implements dynamic adaptation through machine learning models that continuously learn from sensor data and adjust predictions in real-time. The hybrid architecture allows the system to adapt its aerosol concentration estimates based on detected respiratory events, particle flow patterns, and environmental conditions, transforming a static purification system into a dynamic, responsive system.
Solution Approach 2:
The system incorporates feedback loops where sensor measurements of aerosol concentrations are continuously fed into the machine learning model, which adjusts its predictions and controls purification device operations accordingly. This closed-loop feedback enables adaptive response to changing conditions while maintaining system coherence.
3Measurement precision
If machine learning models are trained extensively to capture complex aerosol dynamics, then prediction accuracy improves, but computational resources and training time increase
Solution Approach 1:
The system performs preliminary actions by pre-training the machine learning model offline using extensive aerosol event data. This offline training phase captures complex respiratory event patterns and flow dynamics. Once trained, the model can make rapid predictions during real-time operation without requiring extensive computational resources, thus resolving the contradiction between training thoroughness and operational efficiency.
Data Source
AI summary
In some embodiments, there is provided a system configured to receive an indication of an aerosol event at a first compartment of a room, wherein the room is divided into a plurality of compartments; receive, from at least one particulate measurement sensor located in the room and during a machine learning training phase, at least one particulate measurement for at least one compartment of the plurality of compartments of the room; train, during the machine learning training phase, a digital twin using aerosol event parameters comprising the indication of the aerosol event at the first compartment of the room and the at least one particulate measurement for the at least one of the plurality of compartments of the room; and provide the predicted concentration. Related methods, articles of manufacture, and systems are also disclosed.