Barometric Pressure Event Classification for Low False Alarms

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing systems for detecting intruder events in indoor environments using barometric pressure sensors face challenges in distinguishing between different types of events, such as window opening, tilting, or breaking, and are prone to false alarms due to their simplicity or complexity in state identification.

Innovation Solution

A two-stage approach combining a predictor and classification block, utilizing a recurrent neural network and error signal determination, with preprocessing steps like high-pass filtering and normalization, and employing machine learning algorithms like Linear Discriminant Analysis (LDA) for accurate event classification, accounting for environmental factors like weather conditions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of manufacture

If a simple threshold-based approach is used for event detection, then the system is easy to implement and low cost, but it cannot distinguish between different types of events and generates high false alarm rates

Engineering Contradiction:
Improveimplementation simplicityVSAvoidfalse alarm rate
Core Design Contradiction:
Ease of manufactureVSReliability

Solution Approach 1:

The patent transforms the pressure signal from raw form to error signal by comparing predicted and actual values, changing the parameter representation to enable better event discrimination while maintaining computational simplicity

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent replaces complex mechanical sensor systems (motion sensors, contact sensors, glass break sensors) with a barometric pressure sensor combined with signal processing algorithms, achieving simpler hardware with improved reliability through error signal analysis

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

2Reliability

If a Hidden Markov Model is applied to classify events, then event classification capability is improved, but the system complexity increases significantly due to state identification and transition probability requirements

Engineering Contradiction:
Improveevent classification accuracyVSAvoidmodel complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent extracts only the essential features needed for classification by computing error signals between predicted and actual pressure values, eliminating the need for complex state identification and transition probability calculations required by Hidden Markov Models

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent divides the classification task into two stages: first generating error signals from pressure data, then classifying events based on error signal patterns, simplifying the overall process compared to comprehensive state-based models

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP3618019B1Apparatus and method for event classification based on barometric pressure sensor data
Publication Date: 2021.11.10 INFINEON TECHNOLOGIES AG
  • EP3618019B1 patent drawingFigure 1
  • EP3618019B1 patent drawingFigure 2
  • EP3618019B1 patent drawingFigure 3

AI summary

An apparatus for event detection is provided. The apparatus comprises an error signal determiner (110) configured for determining an error signal portion depending on a pressure signal. The error signal determiner (110) is configured to determine a predicted signal portion depending on a first signal portion of the pressure signal. Moreover, the error signal determiner (110) is configured to determine the error signal portion depending on the predicted signal portion and depending on a second signal portion of the pressure signal, wherein the second signal portion of the pressure signal succeeds the first signal portion of the pressure signal in time. Moreover, the apparatus comprises a classifier (120) configured for determining, depending on the error signal portion, whether an event of a group of one or more events has occurred or whether no event of said group has occurred.