Smell Sensor Transfer Function Features for Accurate Gas Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing smell sensor analysis techniques face challenges in achieving high accuracy due to the complexity of high-order feature values in output data, making it difficult to extract dynamic characteristics effectively.
Innovation Solution
An information processing apparatus that generates an Auto-Regressive with eXogenous input (ARX) model and computes a first-order lag transfer function feature value by subjecting the model to Z-Transform and partial fraction decomposition, simplifying the analysis of smell sensors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If high-order feature values from sensor output data are used for analysis, then the analysis can utilize the full information content, but the analysis accuracy remains difficult to achieve due to complexity
Solution Approach 1:
The patent transforms the feature values from high-order to low-order by changing the mathematical representation parameters. Specifically, it converts raw sensor output waveforms into transfer function parameters through system identification, reducing the order of features while preserving essential dynamic characteristics for accurate gas type discrimination
Solution Approach 2:
The patent extracts the essential dynamic characteristics from the complex high-order output data by identifying and isolating the transfer function parameters. This extraction process separates the meaningful dynamic behavior from the noise and redundancy in the original high-order feature values
2Ease of operation
If dynamic characteristics of smell sensor are extracted as low-order feature values, then the analysis becomes easier to perform with high accuracy, but the complexity of extracting these characteristics increases
Solution Approach 1:
The patent performs preliminary system identification to establish the transfer function model before actual gas discrimination analysis. By pre-characterizing the sensor's dynamic response through ARX modeling and Z-transform, the complex extraction work is done in advance, simplifying subsequent analysis operations
Solution Approach 2:
The patent introduces transfer function parameters as an intermediary representation between the raw sensor output and the final gas discrimination result. This intermediary form simplifies the relationship between input gas concentration and sensor response, making the analysis more tractable while maintaining accuracy
Data Source
AI summary
An information processing apparatus (20) includes a model generating unit (210) and a feature value computation unit (220). The model generating unit (210) generates an Auto-Regressive with eXogenous input (ARX) model of a smell sensor by use of input data controlling an input operation of gas including a smell component being a measurement target, and output data acquired by inputting the gas to the smell sensor, based on the input data. The feature value computation unit (220) computes a transfer function of the smell sensor relating to the smell component by subjecting the ARX model to Z-Transform, and further computes a first-order lag transfer function feature value of the smell sensor relating to the smell component by subjecting the transfer function to partial fraction decomposition.