HRNN4F: hybrid deep random neural network for multi-channel fall activity detection

Research output: Contribution to journalArticle

Abstract

Falls are a major health concern in older adults. Falls lead to mortality, immobility and high costs to social and health care services. Early detection and classification of falls is imperative for timely and appropriate medical aid response. Traditional machine learning models have been explored for fall classification. While newly developed deep learning techniques have the ability to potentially extract high-level features from raw sensor data providing high accuracy and robustness to variations in sensor position, orientation and diversity of work environments that may skew traditional classification models. However, frequently used deep learning models like Convolutional Neural Networks (CNN) are computationally intensive. To the best of our knowledge, we present the first instance of a Hybrid Multichannel Random Neural Network (HMCRNN) architecture for fall detection and classification. The proposed architecture provides the highest accuracy of 92.23% with dropout regularization, compared to other deep learning implementations. The performance of the proposed technique is approximately comparable to a CNN yet requires only half the computation cost of the CNN-based implementation. Furthermore, the proposed HMCRNN architecture provides 34.12% improvement in accuracy on average than a Multilayer Perceptron.
Original languageEnglish
JournalProbability in the Engineering and Informational Sciences
Early online date23 Aug 2019
DOIs
Publication statusE-pub ahead of print - 23 Aug 2019

Fingerprint

Random Networks
Neural Networks
Neural networks
Network Architecture
Network architecture
High Accuracy
Sensor
Drop out
Sensors
Costs
Multilayer neural networks
Perceptron
Health care
Mortality
Healthcare
Skew
Learning systems
Multilayer
Regularization
Machine Learning

Keywords

  • falls
  • older adults
  • HMCRNN architecture
  • probabilistic networks
  • queuing theory
  • simulation

Cite this

@article{bcf76c180519403a8f93cb3b857b7147,
title = "HRNN4F: hybrid deep random neural network for multi-channel fall activity detection",
abstract = "Falls are a major health concern in older adults. Falls lead to mortality, immobility and high costs to social and health care services. Early detection and classification of falls is imperative for timely and appropriate medical aid response. Traditional machine learning models have been explored for fall classification. While newly developed deep learning techniques have the ability to potentially extract high-level features from raw sensor data providing high accuracy and robustness to variations in sensor position, orientation and diversity of work environments that may skew traditional classification models. However, frequently used deep learning models like Convolutional Neural Networks (CNN) are computationally intensive. To the best of our knowledge, we present the first instance of a Hybrid Multichannel Random Neural Network (HMCRNN) architecture for fall detection and classification. The proposed architecture provides the highest accuracy of 92.23{\%} with dropout regularization, compared to other deep learning implementations. The performance of the proposed technique is approximately comparable to a CNN yet requires only half the computation cost of the CNN-based implementation. Furthermore, the proposed HMCRNN architecture provides 34.12{\%} improvement in accuracy on average than a Multilayer Perceptron.",
keywords = "falls, older adults, HMCRNN architecture, probabilistic networks, queuing theory, simulation",
author = "Ahsen Tahir and Jawad Ahmad and Gordon Morison and Hadi Larijani and Gibson, {Ryan M.} and Skelton, {Dawn A.}",
note = "Acceptance in SAN Record recreated by library for author. Original record removed in error. AAM: 6m embargo",
year = "2019",
month = "8",
day = "23",
doi = "10.1017/S0269964819000317",
language = "English",
journal = "Probability in the Engineering and Informational Sciences",
issn = "0269-9648",
publisher = "Cambridge University Press",

}

TY - JOUR

T1 - HRNN4F: hybrid deep random neural network for multi-channel fall activity detection

AU - Tahir, Ahsen

AU - Ahmad, Jawad

AU - Morison, Gordon

AU - Larijani, Hadi

AU - Gibson, Ryan M.

AU - Skelton, Dawn A.

N1 - Acceptance in SAN Record recreated by library for author. Original record removed in error. AAM: 6m embargo

PY - 2019/8/23

Y1 - 2019/8/23

N2 - Falls are a major health concern in older adults. Falls lead to mortality, immobility and high costs to social and health care services. Early detection and classification of falls is imperative for timely and appropriate medical aid response. Traditional machine learning models have been explored for fall classification. While newly developed deep learning techniques have the ability to potentially extract high-level features from raw sensor data providing high accuracy and robustness to variations in sensor position, orientation and diversity of work environments that may skew traditional classification models. However, frequently used deep learning models like Convolutional Neural Networks (CNN) are computationally intensive. To the best of our knowledge, we present the first instance of a Hybrid Multichannel Random Neural Network (HMCRNN) architecture for fall detection and classification. The proposed architecture provides the highest accuracy of 92.23% with dropout regularization, compared to other deep learning implementations. The performance of the proposed technique is approximately comparable to a CNN yet requires only half the computation cost of the CNN-based implementation. Furthermore, the proposed HMCRNN architecture provides 34.12% improvement in accuracy on average than a Multilayer Perceptron.

AB - Falls are a major health concern in older adults. Falls lead to mortality, immobility and high costs to social and health care services. Early detection and classification of falls is imperative for timely and appropriate medical aid response. Traditional machine learning models have been explored for fall classification. While newly developed deep learning techniques have the ability to potentially extract high-level features from raw sensor data providing high accuracy and robustness to variations in sensor position, orientation and diversity of work environments that may skew traditional classification models. However, frequently used deep learning models like Convolutional Neural Networks (CNN) are computationally intensive. To the best of our knowledge, we present the first instance of a Hybrid Multichannel Random Neural Network (HMCRNN) architecture for fall detection and classification. The proposed architecture provides the highest accuracy of 92.23% with dropout regularization, compared to other deep learning implementations. The performance of the proposed technique is approximately comparable to a CNN yet requires only half the computation cost of the CNN-based implementation. Furthermore, the proposed HMCRNN architecture provides 34.12% improvement in accuracy on average than a Multilayer Perceptron.

KW - falls

KW - older adults

KW - HMCRNN architecture

KW - probabilistic networks

KW - queuing theory

KW - simulation

U2 - 10.1017/S0269964819000317

DO - 10.1017/S0269964819000317

M3 - Article

JO - Probability in the Engineering and Informational Sciences

JF - Probability in the Engineering and Informational Sciences

SN - 0269-9648

ER -