Overview
Our project is to create a software application complete with both a Graphical User Interface and backend functionality that helps to reduce false alarm rates in Intensive Care Units. We do this through a multi-factored approach where analyze several vital signals in numerous different ways.
Project Need
Our project is necessary due to:
- 80 - 99% of triggered alarms are false or clinically insignificant.
- Large numbers of false alarms result in alarm fatigue.
- Previous attempts to reduce false alarm rates have focused on expensive hardware, by contrast our project is nearly free since it is software.
Application Features
Our software application consists of the following features:
- Multiple signal processing techniques performed on each signal then comparison
- Multiple signal comparison for alarm detection
- Training set development and used to create algorithm for alarm classication.
- Testing individual signal against developed training set.
- Signal classification using algorith developed by the training set and single signal test.
Archetectural Overview
Our system achetecture is shown by the following diagrams:
Module and Interface Description
Application modules and interfaces consist of:
- Removing Artifacts
- Signal Transformations
- Heart Variability
- Extracting Features
- Feature Selection
- Alarm Classification
- Utility Functions