SmartWall: Novel RFID-Enabled Ambient Human Activity Recognition Using Machine Learning for Unobtrusive Health Monitoring

Oguntala, George; Abd‐Alhameed, Raed A.; Ali, Nazar; Hu, Yim Fun; Noras, James M.; Eya, Nnabuike N.; Elfergani, Issa; Rodrı́guez, Jonathan

doi:10.1109/access.2019.2917125

Cited by 72 publications

(28 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In most of these works, the line-ofsight (LOS) signal that is directly back-scattered by RFID tags is the dominant contributor to the received signal, as the reflection is usually much weaker due to the reflection loss at the reflecting surface and larger path loss from a longer traveling distance, which agrees with the observation that if the human activity occurs far from the the line between the tag and the reader, the associated change in RSSI or phase is significantly smaller than LOS signal, which limits the detection range and the sensitivity to sense subtle body movement. The idea to deploy multiple closely-spaced tags has been proposed to address this challenge [11], [13]. However, this approach inevitably increases the cost on tags and computational load for the system.…”

Section: Theorymentioning

confidence: 99%

“…This involves transmitting a continuous wave (CW) from the reader to the tag and the tag retransmitting a modulated signal back to the reader. RFIDs have been introduced for human activity tracking [9]- [13], and has been shown to be a promising technology in tracking a wide range of human activities. These include using wearable and on-object RFIDs [14]- [16] and also device free RFID activity recognition [11], [17].…”

Section: Introductionmentioning

confidence: 99%

“…This is not sufficient as it results in a limited detection area, making it impractical in most real use cases. Previous works have tried to mitigate this by adding more tags [11], [13], allowing for a broader coverage area, however this becomes a hassle as the user has to manually space out multiple tags to avoid mutually coupling to each other [18], and adds general complexity. By reducing the number of tags and have each tag cover a wider area, the practicality of the system is drastically increased.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

RFID Based Non-Contact Human Activity Detection Exploiting Cross Polarization

Zhu

et al. 2020

IEEE Access

View full text Add to dashboard Cite

This paper explores antenna's polarization for Radio-Frequency Identification (RFID) based non-contact human activity detection. For the first time, a cross circular polarization configuration between reader antenna and tag antenna is proposed to increase the sensing range and spatial sensitivity. Compared to conventional RFID configurations-linearly polarized (LP) tag and circularly polarized (CP) reader, a cross CP configuration can improve the signal-to-noise ratio (SNR) which leads to 230% increase in the detection area over traditional approach in our experiment. As the result of the improvement of spatial sensitivity, the proposed approach can detect subtle and small body movements at almost 4.5m from the reader, such as head movements or even respiration, which enables a low-cost, easy-to-use, non-contact and non-intrusive monitoring of the elderly and disabled peoples in places like hospitals and assisted living homes.

show abstract

Section: Theorymentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

RFID Based Non-Contact Human Activity Detection Exploiting Cross Polarization

Zhu

et al. 2020

IEEE Access

View full text Add to dashboard Cite

show abstract

“…However, in ambient data sensing, sensors are integrated into the target's environment. The readings from the sensors are collected, interpreted for possible patient tracking using various data analytics algorithms including machine learning, neural network and deep learning [8,9]. The key aim of patient tracking is change detection by identifying changes in metric which represents a change point in time-series data within an indoor environment [10].…”

Section: Introductionmentioning

confidence: 99%

RFID RSS Fingerprinting System for Wearable Human Activity Recognition

et al. 2020

Self Cite

View full text Add to dashboard Cite

Alternative healthcare solutions have been identified as a viable approach to ameliorate the increasing demand for telehealth and prompt healthcare delivery. Moreover, indoor ocalization using different technologies and approaches have greatly contributed to alternative healthcare solutions. In this paper, a cost-effective, radio frequency identification (RFID)-based indoor location system that employs received signal strength (RSS) information of passive RFID tags is presented. The proposed system uses RFID tags placed at different positions on the target body. The mapping of the analysed data against a set of reference position datasets is used to accurately track the vertical and horizontal positioning of a patient within a confined space in real-time. The Euclidean distance model achieves an accuracy of 98% for all sampled activities. However, the accuracy of the activity recognition algorithm performs below the threshold performance for walking and standing, which is due to similarities in the target height, weight and body density for both activities. The obtained results from the proposed system indicate significant potentials to provide reliable health measurement tool for patients at risk.

show abstract

“…For example, human activity recognition models can be used to classify data and alert carers about falls, unordinary behaviour, or about certain activities. Additionally, the models can analyse different behaviours for detecting strokes, eating behaviour, and tracking whether people are taking prescribed medication [2]. An approach to HAR is based on a sliding window procedure, where a fixed length analysis window is shifted along the signal sequence for frame extraction [3].…”

Section: Introductionmentioning

confidence: 99%

Classifying Imbalanced Multi-modal Sensor Data for Human Activity Recognition in a Smart Home using Deep Learning

Alani

Cosma

Taherkhani

2020

2020 International Joint Conference on Neural Networks (IJCNN)

View full text Add to dashboard Cite

In smart homes, data generated from real-time sensors for human activity recognition is complex, noisy and imbalanced. It is a significant challenge to create machine learning models that can classify activities which are not as commonly occurring as other activities. Machine learning models designed to classify imbalanced data are biased towards learning the more commonly occurring classes. Such learning bias occurs naturally, since the models better learn classes which contain more records. This paper examines whether fusing real-world imbalanced multi-modal sensor data improves classification results as opposed to using unimodal data; and compares deep learning approaches to dealing with imbalanced multi-modal sensor data when using various resampling methods and deep learning models. Experiments were carried out using a large multi-modal sensor dataset generated from the Sensor Platform for HEalthcare in a Residential Environment (SPHERE). The data comprises 16104 samples, where each sample comprises 5608 features and belongs to one of 20 activities (classes). Experimental results using SPHERE demonstrate the challenges of dealing with imbalanced multi-modal data and highlight the importance of having a suitable number of samples within each class for sufficiently training and testing deep learning models. Furthermore, the results revealed that when fusing the data and using the Synthetic Minority Oversampling Technique (SMOTE) to correct class imbalance, CNN-LSTM achieved the highest classification accuracy of 93.67% followed by CNN, 93.55%, and LSTM, i.e. 92.98%.

show abstract

SmartWall: Novel RFID-Enabled Ambient Human Activity Recognition Using Machine Learning for Unobtrusive Health Monitoring

Cited by 72 publications

References 28 publications

RFID Based Non-Contact Human Activity Detection Exploiting Cross Polarization

RFID Based Non-Contact Human Activity Detection Exploiting Cross Polarization

RFID RSS Fingerprinting System for Wearable Human Activity Recognition

Classifying Imbalanced Multi-modal Sensor Data for Human Activity Recognition in a Smart Home using Deep Learning

Contact Info

Product

Resources

About