AI-Based Laundry Sensing for Adaptive Washing Machine Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional washing machines often fail to accurately sense laundry weight and quality, leading to variations in washing performance and potential damage to laundry, as they primarily rely on weight-based settings without considering the material, water content, and composition of the laundry.
Innovation Solution
A washing machine utilizing an artificial neural network based on machine learning to determine laundry weight and quality by analyzing current patterns during accelerated rotation, allowing for the selection of optimal washing modes based on the state of the laundry, including rotational speed, water supply, and washing cycle time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If washing operation is set based on laundry weight alone, then device complexity is reduced, but washing performance deteriorates and laundry may be damaged
Solution Approach 1:
The patent replaces conventional mechanical weight-based sensing with an artificial neural network-based sensing system that analyzes motor current patterns. The neural network model processes current data during accelerated rotation to determine both laundry weight and quality characteristics, enabling more accurate and reliable washing performance without requiring additional mechanical sensors.
Solution Approach 2:
The patent changes the sensing parameters from simple weight measurement to multi-dimensional analysis of motor current characteristics. By analyzing current patterns during accelerated rotation through a neural network, the system extracts multiple features including weight, material type, water content, and composition, thereby improving washing performance through enhanced parameter detection.
2Measurement precision
If washing operation is set based on laundry weight alone, then measurement precision is reduced, but device complexity is lowered
Solution Approach 1:
The patent substitutes conventional weight sensors with an intelligent sensing system using artificial neural networks. The system measures motor current during accelerated rotation and uses the neural network to infer both weight and quality characteristics, achieving high measurement precision without additional mechanical sensing components.
Solution Approach 2:
The patent introduces motor current as an intermediary measurement parameter. Instead of directly measuring weight and quality, the system measures current consumption during accelerated rotation, which serves as an intermediary that contains information about both weight and material characteristics. The neural network then decodes this intermediary data into precise measurements.
3Adaptability or versatility
If conventional sensing methods are used, then device complexity is reduced, but adaptability to different laundry types deteriorates
Solution Approach 1:
The patent transforms the control system from weight-based operation to a multi-parameter analysis system. The neural network processes motor current patterns to determine not only weight but also laundry quality attributes such as material type, water content, and composition. This enables the system to adapt to different laundry types by recognizing their specific current signatures during accelerated rotation.
Solution Approach 2:
The patent implements self-service through the artificial neural network that automatically learns and identifies laundry characteristics without user input. The system autonomously analyzes motor current patterns, classifies laundry types, and determines optimal washing parameters, providing adaptability to various laundry kinds without requiring manual selection or additional sensors.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Disclosed is a method of controlling a washing machine, the method including determining the state of laundry received in a washing tub (4) from output of an output layer of an artificial neural network pre-trained based on machine learning using a current value supplied to a motor (9) configured to rotate the washing tub during accelerated rotation of the washing tub as input data of an input layer of the artificial neural network (a first sensing step), selecting (S80, S280) one of a plurality of washing modes classified in consideration of the wear degree of laundry or washing strength based on the state of the laundry, and performing washing according to the selected washing mode (a washing cycle step (S90, S290)).