Apparatus control device and apparatus control method
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional apparatus control devices fail to consider unobservable values, leading to inadequate control of control target apparatuses due to the inability to estimate environmental changes accurately.
Innovation Solution
The apparatus control device utilizes a first learning model to predict observable values and a second learning model to estimate unobservable values, integrating these with observed values to calculate control values using a state prediction unit and control value selecting unit.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If only observable values from sensors are used for control, then the device complexity is reduced, but the control accuracy deteriorates due to inability to estimate environmental changes
Solution Approach 1:
The patent introduces learning models as intermediary components that bridge the gap between observable sensor data and unobservable environmental factors. The first learning model processes observable values to predict future states, while the second learning model estimates unobservable values, together enabling accurate control without direct measurement of all parameters
Solution Approach 2:
The patent replaces physical measurement devices (sensors) with computational models (learning models) to estimate unobservable values. Instead of installing sensors for every parameter, the system uses machine learning algorithms to infer unobservable environmental factors from observable data, reducing hardware complexity while maintaining control accuracy
2Measurement precision
If unobservable values are estimated using learning models, then the control accuracy is improved, but the device complexity increases due to multiple learning models
Solution Approach 1:
The patent divides the control system into distinct functional modules: an observable value acquisition unit for sensor data, a first learning model for predicting observable future states, a second learning model for estimating unobservable values, and a control value calculation unit. This segmentation allows each component to specialize in specific tasks, improving overall accuracy while managing complexity through modular design
Solution Approach 2:
The learning models serve multiple functions within the system. The first learning model not only predicts future observable values but also provides inputs to the second learning model. The second learning model estimates unobservable values that complement sensor data. This multi-functionality reduces the need for separate specialized components for each function
Data Source
Figure 1
Figure 2~3
Figure 4
AI summary
An apparatus control device (3) is configured to include an observed value acquiring unit (11) that acquires, from a sensor (2-n) (n=1, ..., N) that observes an environment in which a control target apparatus is installed, an observed value of the environment, an observation predicted value acquiring unit (12) that gives an observed value acquired by the observed value acquiring unit (11) to a first learning model (12a) and acquires an observation predicted value that is a future observed value from the sensor (2-n) from the first learning model (12a), and an unobservable value acquiring unit (13) that gives an observed value acquired by the observed value acquiring unit (11) to a second learning model (13a) and acquires an unobservable value that is a value not directly observed by the sensor (2-n) from the second learning model (13a). In addition, the apparatus control device (3) includes a control value calculating unit (14) that calculates a control value of the control target apparatus using the observed value acquired by the observed value acquiring unit (11), the observation predicted value acquired by the observation predicted value acquiring unit (12), and the unobservable value acquired by the unobservable value acquiring unit (13).