Odor detecting method for refrigerating appliance and refrigerating appliance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing refrigerating appliances lack an effective method to accurately detect and manage odors, which negatively impact user experience due to the deterioration of perishable foods stored within.
Innovation Solution
An odor detecting method utilizing machine learning models trained with data from multiple rounds of odor detecting experiments on various ingredients, combining sample data to improve prediction accuracy, and integrating a deodorization module to address unpleasant odors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional odor detection methods are used in refrigerating appliances, then the device complexity is low, but the measurement precision of odor properties is insufficient
Solution Approach 1:
The patent replaces traditional mechanical/chemical odor detection methods with machine learning models that process environmental data (temperature, humidity, gas composition) to predict odor properties. This substitution enables high-precision odor detection without requiring complex specialized sensing hardware, thus improving measurement precision while controlling device complexity.
Solution Approach 2:
The patent introduces environmental data (temperature, humidity, gas composition) as intermediary parameters that correlate with odor properties. Instead of directly detecting odor, the system measures these intermediary environmental factors and uses machine learning models to infer odor characteristics, achieving accurate odor detection through indirect measurement.
2Measurement precision
If multiple machine learning models are trained with extensive sample data to improve odor detection accuracy, then the measurement precision improves, but the loss of time for data collection and model training increases
Solution Approach 1:
The patent performs comprehensive data collection and machine learning model training in advance, before actual odor detection is needed. Extensive sample data from Q rounds of experiments on M groups of ingredients is used to pre-train N machine learning models. This preliminary action stores learned patterns in the models, enabling fast real-time odor prediction without requiring additional data collection time during actual use.
Solution Approach 2:
The patent collects and processes more sample data than minimally required (M groups of ingredients across Q rounds of experiments) to ensure the machine learning models are thoroughly trained. This excessive data collection during the preparation phase improves model robustness and generalization, allowing accurate odor prediction with minimal processing time during actual deployment.
Data Source
Figure 1
Figure 2
AI summary
An odor detecting method for a refrigerating appliance and a refrigerating appliance are provided. The method includes: acquiring environmental data of gas in a refrigerating appliance (S11); inputting the environmental data to N preset machine learning models (S12) to obtain N odor detecting results; and determining odor properties of the gas in the refrigerating appliance according to the odor properties in the N odor detecting results (S13). The above solution can improve accuracy of determining the odor properties of the gas in the refrigerating appliance.