Evaluation of Radial Basis Function Neural Network minimizing L-GEM for Sensor-based Activity Recognition

Shuai Zhang, Wing Yin Ng, Jianjun Zhang, CD Nugent, Naomi Irvine, Ting Wang

Research output: Contribution to journalArticle

Abstract

Sensor-based activity recognition involves the automatic recognition of a user’s activity in a smart environment using computational methods. The use of wearable devices and video-based approaches have attracted considerable interest in ubiquitous computing. Nevertheless, these methods have limitations such as issues with privacy invasion, ethics, comfort and obtrusiveness. Environmental sensors are an increasingly promising consideration in the ubiquitous computing domain for long-term monitoring, as these devices are non-invasive to inhabitants, yet certain challenges remain with activity recognition in sensorised environments, for example, addressing the challenge of intraclass variation between activities and reasoning from low-level uncertain information. In an effort to address these challenges, this paper proposes and evaluates the performance of a Radial Basis Function Neural Network approach for activity recognition with environmental sensors. The model is trained using the Localized Generalization Error and focuses on the generalization ability by considering both the training error and stochastic sensitivity measure. This measures the network output fluctuation with respect to the minor perturbation of input, to address the tolerance of the low-level uncertain sensor data. This approach is compared with three benchmark Neural Network approaches, including a popular deep learning approach using an Autoencoder, and it is evaluated with a simulated dataset as well as a number of publicly available datasets. The proposed method has shown advantages over the other models for all four evaluated datasets. This paper provides insights into the importance of model generalization abilities and an initial analysis of the limitation of deep Neural Networks with respect to sensor-based activity recognition.

LanguageEnglish
Pages1-11
JournalJournal of Ambient Intelligence and Humanized Computing
Early online date6 Mar 2019
DOIs
Publication statusE-pub ahead of print - 6 Mar 2019

Fingerprint

Air cushion vehicles
Neural networks
Sensors
Ubiquitous computing
Computational methods
Monitoring

Keywords

  • activity recognition
  • Uncertainty
  • Neural Network
  • localized generation error
  • radial basis function neural network
  • Activity recognition
  • Localized generation error
  • Neural network
  • Radial basis function neural network

Cite this

@article{f92f390eb2a6431ab77e48747245ce34,
title = "Evaluation of Radial Basis Function Neural Network minimizing L-GEM for Sensor-based Activity Recognition",
abstract = "Sensor-based activity recognition involves the automatic recognition of a user’s activity in a smart environment using computational methods. The use of wearable devices and video-based approaches have attracted considerable interest in ubiquitous computing. Nevertheless, these methods have limitations such as issues with privacy invasion, ethics, comfort and obtrusiveness. Environmental sensors are an increasingly promising consideration in the ubiquitous computing domain for long-term monitoring, as these devices are non-invasive to inhabitants, yet certain challenges remain with activity recognition in sensorised environments, for example, addressing the challenge of intraclass variation between activities and reasoning from low-level uncertain information. In an effort to address these challenges, this paper proposes and evaluates the performance of a Radial Basis Function Neural Network approach for activity recognition with environmental sensors. The model is trained using the Localized Generalization Error and focuses on the generalization ability by considering both the training error and stochastic sensitivity measure. This measures the network output fluctuation with respect to the minor perturbation of input, to address the tolerance of the low-level uncertain sensor data. This approach is compared with three benchmark Neural Network approaches, including a popular deep learning approach using an Autoencoder, and it is evaluated with a simulated dataset as well as a number of publicly available datasets. The proposed method has shown advantages over the other models for all four evaluated datasets. This paper provides insights into the importance of model generalization abilities and an initial analysis of the limitation of deep Neural Networks with respect to sensor-based activity recognition.",
keywords = "activity recognition, Uncertainty, Neural Network, localized generation error, radial basis function neural network, Activity recognition, Localized generation error, Neural network, Radial basis function neural network",
author = "Shuai Zhang and Ng, {Wing Yin} and Jianjun Zhang and CD Nugent and Naomi Irvine and Ting Wang",
year = "2019",
month = "3",
day = "6",
doi = "10.1007/s12652-019-01246-w",
language = "English",
pages = "1--11",
journal = "Journal of Ambient Intelligence and Humanized Computing",
issn = "1868-5137",

}

TY - JOUR

T1 - Evaluation of Radial Basis Function Neural Network minimizing L-GEM for Sensor-based Activity Recognition

AU - Zhang, Shuai

AU - Ng, Wing Yin

AU - Zhang, Jianjun

AU - Nugent, CD

AU - Irvine, Naomi

AU - Wang, Ting

PY - 2019/3/6

Y1 - 2019/3/6

N2 - Sensor-based activity recognition involves the automatic recognition of a user’s activity in a smart environment using computational methods. The use of wearable devices and video-based approaches have attracted considerable interest in ubiquitous computing. Nevertheless, these methods have limitations such as issues with privacy invasion, ethics, comfort and obtrusiveness. Environmental sensors are an increasingly promising consideration in the ubiquitous computing domain for long-term monitoring, as these devices are non-invasive to inhabitants, yet certain challenges remain with activity recognition in sensorised environments, for example, addressing the challenge of intraclass variation between activities and reasoning from low-level uncertain information. In an effort to address these challenges, this paper proposes and evaluates the performance of a Radial Basis Function Neural Network approach for activity recognition with environmental sensors. The model is trained using the Localized Generalization Error and focuses on the generalization ability by considering both the training error and stochastic sensitivity measure. This measures the network output fluctuation with respect to the minor perturbation of input, to address the tolerance of the low-level uncertain sensor data. This approach is compared with three benchmark Neural Network approaches, including a popular deep learning approach using an Autoencoder, and it is evaluated with a simulated dataset as well as a number of publicly available datasets. The proposed method has shown advantages over the other models for all four evaluated datasets. This paper provides insights into the importance of model generalization abilities and an initial analysis of the limitation of deep Neural Networks with respect to sensor-based activity recognition.

AB - Sensor-based activity recognition involves the automatic recognition of a user’s activity in a smart environment using computational methods. The use of wearable devices and video-based approaches have attracted considerable interest in ubiquitous computing. Nevertheless, these methods have limitations such as issues with privacy invasion, ethics, comfort and obtrusiveness. Environmental sensors are an increasingly promising consideration in the ubiquitous computing domain for long-term monitoring, as these devices are non-invasive to inhabitants, yet certain challenges remain with activity recognition in sensorised environments, for example, addressing the challenge of intraclass variation between activities and reasoning from low-level uncertain information. In an effort to address these challenges, this paper proposes and evaluates the performance of a Radial Basis Function Neural Network approach for activity recognition with environmental sensors. The model is trained using the Localized Generalization Error and focuses on the generalization ability by considering both the training error and stochastic sensitivity measure. This measures the network output fluctuation with respect to the minor perturbation of input, to address the tolerance of the low-level uncertain sensor data. This approach is compared with three benchmark Neural Network approaches, including a popular deep learning approach using an Autoencoder, and it is evaluated with a simulated dataset as well as a number of publicly available datasets. The proposed method has shown advantages over the other models for all four evaluated datasets. This paper provides insights into the importance of model generalization abilities and an initial analysis of the limitation of deep Neural Networks with respect to sensor-based activity recognition.

KW - activity recognition

KW - Uncertainty

KW - Neural Network

KW - localized generation error

KW - radial basis function neural network

KW - Activity recognition

KW - Localized generation error

KW - Neural network

KW - Radial basis function neural network

U2 - 10.1007/s12652-019-01246-w

DO - 10.1007/s12652-019-01246-w

M3 - Article

SP - 1

EP - 11

JO - Journal of Ambient Intelligence and Humanized Computing

T2 - Journal of Ambient Intelligence and Humanized Computing

JF - Journal of Ambient Intelligence and Humanized Computing

SN - 1868-5137

ER -