Inductive Proximity Sensor Signal Correction for Installation Independence

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Inductive proximity sensors face challenges in achieving high switching distances, F1 behavior, and low installation dependency, with existing solutions either being costly or requiring extensive individual training for each sensor, especially when dealing with varying installation situations and materials.

Innovation Solution

The proposed solution involves an inductive proximity sensor that uses a combination of classic signal processing and machine learning methods to evaluate sensor signals, allowing for the determination of correction values for environmental interference, thereby achieving high switching distances and F1 behavior with reduced installation dependency, without the need for individual sensor training.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If hardware measures such as foils or additional compensation coils are used to minimize installation jump, then installation dependency is reduced, but manufacturing costs increase and switching distance is reduced

Engineering Contradiction:
Improveinstallation independenceVSAvoidsensor structure
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent replaces hardware-based compensation mechanisms (foils, additional coils) with a software-based machine learning evaluation method. The sensor system uses a trained evaluation algorithm that automatically compensates for installation effects by analyzing sensor signals and comparing them against training data that includes various installation scenarios, thereby achieving installation independence without adding physical components

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent changes the evaluation parameters from simple amplitude detection to comprehensive signal analysis including phase, frequency, and temporal characteristics. By using machine learning models trained on diverse installation parameters, the system adapts its evaluation criteria to compensate for installation variations, effectively changing how the sensor interprets its signals rather than changing its physical structure

Inventive Principle:
Principle #35Parameter changes

2Measurement precision

If machine learning methods are used for sensor signal evaluation, then F1 behavior and switching distance are improved, but training effort and complexity increase

Engineering Contradiction:
Improvematerial recognition accuracyVSAvoidtraining time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent performs preliminary training actions during the manufacturing process where sensors are trained offline using controlled training setups with known reference objects. This preliminary training establishes the machine learning model before the sensor is deployed, so that during actual operation, the sensor can immediately utilize the pre-trained evaluation algorithm without requiring real-time training, thus reducing operational training time to nearly zero

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent implements self-service mechanisms where the sensor system automatically performs evaluation and correction using its own sensor signals and the pre-trained machine learning model. The system serves itself by automatically compensating for installation effects and material variations without requiring external calibration or manual intervention during operation, thereby eliminating ongoing training time requirements

Inventive Principle:
Principle #25Self-service

3Measurement precision

If additional signals from vertical or coaxial coils are used to determine correction values, then installation jump compensation is improved, but device complexity increases

Engineering Contradiction:
Improvecorrection value accuracyVSAvoidcoil structure
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent makes the existing sensor coil multi-functional by using it for both primary detection and installation effect detection. The same coil that detects target objects also captures signals influenced by installation conditions, and the machine learning evaluation method analyzes these signals to extract both target information and installation characteristics, eliminating the need for separate vertical or coaxial coils while maintaining correction accuracy

Inventive Principle:
Principle #6Universality (Multi-functionality)

Solution Approach 2:

The patent extracts installation-related information from the existing sensor signals without requiring additional hardware. The machine learning evaluation method separates and extracts the installation effect components from the composite sensor signal, isolating the installation jump characteristics for compensation while using the remaining signal for target detection, thereby achieving precise correction without adding physical components

Inventive Principle:
Principle #2Taking out (Extraction)

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 enables sensors to maintain high switching distances and F1 behavior across different installation conditions, recognizing target materials effectively, while reducing the need for extensive training and minimizing installation-dependent variations, thus optimizing sensor performance and cost-effectiveness.

Implementation Method 1

inductive proximity sensor for detecting an object

Methodology Applied
Scientific EffectElectromagnetic induction: Electromagnetic Induction

Data Source

PatentEP3968520B1Sensor for detecting an object and method for evaluating a sensor signal
Publication Date: 2022.08.31 SICK AG
  • EP3968520B1 patent drawingFigure 1
  • EP3968520B1 patent drawingFigure 1~2
  • EP3968520B1 patent drawingFigure 3~4

AI summary

A sensor (10) for detecting an object is specified, which has a detection unit (12, 22) for detecting a sensor signal and a control and evaluation unit (24) which is designed to determine an object property by evaluating the sensor signal, to determine a correction value for disturbances of the sensor environment from the sensor signal using a machine learning method and to take the correction value into account when determining the object property.