Virtual Sensor Training for Interior Environment AI Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current environmental control systems in horticultural, agricultural, and architectural buildings face challenges due to sparse sensor data, poor correlation of daylight measurements, and suboptimal performance from fixed luminaire control systems, leading to inefficient energy use and occupant discomfort.
Innovation Solution
A system and method that uses a three-dimensional parametric CAD model to simulate environmental conditions and sensor data, generating a dense array of virtual sensors to provide training data for AI-based controllers, allowing for optimized control of lighting and environmental parameters while minimizing energy consumption and visual glare.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single daylight photosensor is used to measure lighting conditions, then system complexity is reduced, but measurement precision deteriorates due to poor correlation with interior daylight distribution
Solution Approach 1:
The patent creates a virtual copy of the physical environment through a 3D parametric CAD model. This virtual model includes virtual sensors that replicate the measurement function of physical sensors but can be placed throughout the simulated space to capture complete daylight distribution data without adding physical complexity
Solution Approach 2:
The patent introduces a simulation environment as an intermediary between the single physical photosensor and the AI controller. The simulation translates the limited physical sensor data into comprehensive virtual sensor data that accurately represents interior daylight distribution, bridging the gap between simple hardware and intelligent control
2Quantity of substance
If a sparse sensor array is used for economic reasons, then cost is reduced, but loss of information increases due to insufficient spatial coverage
Solution Approach 1:
The patent populates the virtual 3D model with a dense array of virtual sensors at strategic locations throughout the interior space. These virtual sensors capture environmental parameters (daylight, temperature, humidity) that would require numerous physical sensors, but exist only in simulation, preserving all spatial information without additional hardware cost
Solution Approach 2:
The patent performs preliminary simulation and data generation during the design and training phase. By pre-computing the complete environmental dataset in the virtual model before deployment, the system eliminates the need for dense physical sensor arrays during actual operation, having already captured all necessary information in advance
3Device complexity
If fixed luminaire control groups are used, then device complexity is reduced, but productivity deteriorates due to suboptimal energy efficiency
Solution Approach 1:
The patent transitions from static, fixed luminaire control groups to dynamic, adaptive control. The AI controller continuously receives virtual sensor data from the simulation and adjusts individual luminaire settings in real-time based on actual daylight conditions and occupancy patterns, optimizing energy efficiency without requiring complex manual reconfiguration
Solution Approach 2:
The patent implements a closed-loop feedback system where the simulation continuously monitors virtual sensor readings and feeds this information back to the AI controller. The controller then adjusts luminaire output accordingly, creating an adaptive system that automatically optimizes energy efficiency based on real-time environmental conditions
Data Source
AI summary
A dense array of sensors positioned in a virtual environment is reduced to a sparse array of sensors in a physical environment, which provides sufficient information to a controller that responds to environmental conditions and parameters in the physical environment in substantially the same manner as it would to the same environmental conditions and parameters in the equivalent virtual environment. Data from a sparse array of virtual sensors is correlated with data from a dense array of virtual sensors and is used for generating control signals for hardware devices that influence a real or virtual interior environment. The correlated data and the control signals are used to train an artificial intelligence based controller that then controls the values of the parameters of the interior environment. A model of the interior environment is created using basic parameters in a computer-aided design application.