Boolean Learning Machine for Electronic Nose Signal Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional sensor systems for detecting chemical substances and mixtures, such as those used in electronic noses, are complex and expensive, making them unsuitable for mass consumption applications, as they rely on neural networks that fail to interpret sensor responses correctly when implemented in simple microcontrollers or inexpensive electronics.
Innovation Solution
A sensor arrangement using a Boolean learning machine, specifically trained with the Hamming Clustering algorithm, generates binary classification rules to process and classify response signals from an array of sensors, enabling discrimination between predetermined outcomes, which can be implemented in inexpensive devices like microcontrollers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional neural network processing systems are used for sensor signal interpretation, then measurement precision and detection capability are improved, but device complexity and cost increase significantly
Solution Approach 1:
The patent replaces expensive, complex neural network processing systems with inexpensive, simple microcontrollers that can be mass-produced. The Boolean learning machine uses basic logic circuits and lookup tables instead of complex neural network hardware, making the system affordable for mass consumption applications while maintaining adequate detection capabilities for the intended use cases.
Solution Approach 2:
The patent substitutes the mechanical/computational complexity of neural networks with a Boolean logic-based system. Instead of using continuous mathematical operations and floating-point arithmetic required by neural networks, the invention uses discrete Boolean logic operations and pre-computed lookup tables that can be efficiently implemented in simple digital circuits and microcontrollers.
2Measurement precision
If neural networks are trained on computers with high calculation accuracy, then measurement precision is improved, but the system fails when implemented in simple microcontrollers with lower calculation accuracy
Solution Approach 1:
The patent performs the complex training and computation work in advance during an offline phase using high-precision computers. The results of this training are stored as pre-computed lookup tables and Boolean classification rules. During actual operation on simple microcontrollers, the system only needs to perform simple table lookups and Boolean logic operations, eliminating the need for high calculation accuracy during runtime and ensuring compatibility with low-cost hardware.
Solution Approach 2:
The patent introduces lookup tables and Boolean classification rules as intermediaries between the training phase and the execution phase. These intermediaries translate the complex neural network training results into a format that can be efficiently processed by simple microcontrollers, bridging the gap between high-precision training environments and low-precision deployment hardware.
3Measurement precision
If complex neural network systems are used for substance detection, then detection precision is improved, but ease of manufacture and mass production capability deteriorate
Solution Approach 1:
The patent adopts inexpensive microcontrollers and simple digital circuits that can be mass-produced using standard semiconductor manufacturing processes. These components are far easier to manufacture at scale compared to complex neural network hardware, enabling cost-effective mass production while maintaining sufficient detection precision for consumer applications.
Solution Approach 2:
The patent changes the computational parameters from continuous floating-point arithmetic to discrete Boolean logic operations. This parameter change enables the use of simple digital circuits and microcontrollers that are manufactured using standard CMOS processes, dramatically improving ease of manufacture and mass production capability compared to systems requiring complex neural network processors.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach allows for the development of an affordable artificial olfactory sensor system capable of qualitative and quantitative detection of chemical substances, suitable for mass consumption, by using a programmable sensor arrangement that can be specialized for different substances and adapted to varying conditions, while maintaining precision and reliability.
Implementation Method 1
the shifting of the resonant frequency of a quartz crystal
Implementation Method 2
the variations of electrical resistance or impedance of a film of chemically sensitive and intrinsically conductive material
Data Source
AI summary
What is described is a sensor arrangement of the electronic nose type and a method for the qualitative and quantitative detection of chemical substances and/or mixtures of substances in an environment. The sensor arrangement comprises an array of sensors (10), capable of emitting a set of response signals correlated with the presence and/or concentration of at least one chemical substance, and an electronic processing and recognition system including a Boolean learning machine (34), arranged to classify the response signals generated by the sensor array (10) by the application of at least one predetermined collection of binary classification rules, adapted to discriminate between a pair of predetermined complementary outcomes of the detection.