Multi-Domain Magnetic Detector with Neural Network Position Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing sensors struggle to accurately determine the positional domain of a magnetic field source due to the complexity of magnetic field variations, which can lead to inaccuracies in safety-critical applications.
Innovation Solution
A sensor system utilizing a magnetic field sensing module and processing circuitry with a neural network to analyze multiple components of the magnetic field, generating probabilities for different positional domains, and outputting an identifier for the current domain based on these probabilities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional sensors are used to detect magnetic field components, then the device complexity is low, but the measurement precision of positional domain is insufficient
Solution Approach 1:
The patent combines multiple magnetic field sensing elements (at least two different types of sensing elements) into a single sensor device, merging their outputs with neural network processing to achieve accurate positional domain determination. This resolves the contradiction by integrating sensing capabilities that individually would be too simple but collectively provide the necessary measurement precision.
Solution Approach 2:
The patent introduces neural network processing circuitry as an intermediary between the magnetic field sensing elements and the positional domain determination. This intermediary processes the raw sensor data through trained neural networks to accurately classify positional domains, resolving the measurement precision issue without requiring the sensing elements themselves to be overly complex.
2Measurement precision
If multiple magnetic field sensing elements are used to improve measurement precision, then the measurement precision improves, but the device complexity increases
Solution Approach 1:
The patent employs different types of magnetic field sensing elements (e.g., Hall effect sensors, magnetoresistive sensors, fluxgate sensors) each optimized for detecting specific magnetic field components or characteristics. This local quality approach allows each sensing element to be specialized for its optimal detection task, improving overall measurement precision while managing device complexity through functional differentiation.
3Measurement precision
If neural network processing is applied to analyze magnetic field data, then the measurement precision of positional domain improves, but the use of energy increases
Solution Approach 1:
The patent performs preliminary action by pre-training neural networks with training data that includes magnetic field signatures and corresponding positional domain labels before actual operation. This preliminary training allows the neural networks to make accurate positional domain determinations during operation without requiring continuous complex computation, thereby reducing real-time energy consumption while maintaining high measurement precision.
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
Enhances the accuracy of determining the positional domain of a magnetic field source, improving reliability in safety-critical applications by providing precise positional information.
Implementation Method 1
a magnetic field sensing module that is configured to generate a plurality of signals, each signal indicating a magnetic flux density of a different component of a magnetic field that is produced by a magnetic field source
Data Source
AI summary
A sensor, comprising: a magnetic field sensing module that is configured to generate a plurality of signals, each signal indicating a magnetic flux density of a different component of a magnetic field that is produced by a magnetic field source; a processing circuitry that is configured to: receive the plurality of signals from the magnetic field sensing module; evaluate a neural network based on the plurality of signals to obtain a plurality of probabilities, each of the plurality of probabilities indicating a likelihood of the magnetic field source being positioned in a different one of a plurality of positional domains; generate an output signal based on the plurality of probabilities, the output signal encoding an identifier of a current positional domain of the magnetic field source.