Magnetic Detection System Waveform Pattern Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional magnetic detection systems cannot determine the traveling direction of a magnetic body relative to a magnetic sensor, despite being able to differentiate between magnetic signals from the body and noise based on waveform patterns.
Innovation Solution
A magnetic detection system employing a waveform pattern classification unit that utilizes machine-learning to classify magnetic signals by generating fully connected layers from acquired sensor data, correlating waveform patterns with relative position and traveling direction, allowing determination of the magnetic body's direction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional magnetic detection systems use waveform pattern matching to differentiate magnetic signals from noise, then signal detection accuracy is improved, but the ability to determine traveling direction is lost
Solution Approach 1:
The patent transitions from conventional 1D waveform pattern matching to 2D waveform pattern distribution analysis. By mapping magnetic signals onto a distribution space with multiple dimensions (representing different waveform characteristics), the system can simultaneously preserve signal detection accuracy and extract traveling direction information from the distribution patterns across different dimensions.
Solution Approach 2:
The patent segments the waveform analysis into multiple standardized patterns (first through fourth patterns) representing different traveling directions. Each segment corresponds to a specific directional range, allowing the system to classify signals not just as present/absent but as moving in specific directions, thereby recovering the traveling direction information that was previously lost.
2Adaptability or versatility
If the magnetic detection system classifies waveform patterns into multiple categories, then traveling direction determination is enabled, but system complexity increases
Solution Approach 1:
The patent performs preliminary classification by mapping incoming magnetic signals to a predefined waveform pattern distribution space before detailed analysis. This preliminary action organizes the complex classification task into manageable regions (first through fourth patterns), reducing the computational burden during real-time operation while maintaining comprehensive directional classification capability.
Solution Approach 2:
The patent changes the parameter space from raw waveform data to normalized distribution coordinates. By transforming the signal representation into a standardized distribution space with defined boundaries and characteristics, the system simplifies the classification process while enabling multi-category differentiation for various traveling directions.
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
Enables accurate classification and determination of the traveling direction of a magnetic body by generating a waveform pattern distribution based on weighted features, effectively addressing the limitations of existing systems.
Implementation Method 1
a magnetic sensor arranged in water to acquire magnetic signals
Data Source
AI summary
The magnetic detection system (100) is provided with a magnetic sensor (1) and a waveform pattern classification unit (33c). The waveform pattern classification unit (33c) is configured to classify waveform patterns of magnetic signals acquired by the magnetic sensor (1) based on a waveform pattern distribution (60) generated based on a plurality of fully connected layers (52c) generated by weighting and connecting respective features in waveform patterns for each waveform pattern by machine-learning, and features in the waveform patterns of the magnetic signals.