Recognising Activities in Real Time Using Body Worn Passive Sensors With Sparse Data Streams: To Interpolate or Not To Interpolate?

Abstract

Recent emergence of small, lightweight, batteryless (passive), and therefore maintenance free, wearable computing platforms such as sensor enabled RFID (Radio Frequency Identification) tags provide new opportunities for low cost and unobtrusive activity monitoring. Unfortunately, data streams from passive sensors are uniquely characterised by sparsity and noise. Consequently, readily extracting features that require a data stream with a regular sampling rate, such as those based on frequency domain transformations and auto regressive coefficients are challenging. In this study, we propose an approach that reduces online interpolation errors to facilitate interpolating sparse acceleration data streams from a passive RFID tag with an on-board accelerometer sensor. We demonstrate using two datasets from older people that incorporating features extracted from interpolated acceleration data depicts a significant performance improvement compared with incorporating features possible from only raw acceleration data. However, this is achieved at the expense of a real-time prediction delay (>400%) due to preprocessing involved. Furthermore, we demonstrate that for these types of sensors, features readily available from a typical RFID platform can be successfully used instead of features extracted from an interpolated data stream to achieve similar or better activity recognition performance without preprocessing and, whilst, using significantly less number of features.

Sequence Learning with Passive RFID Sensors for Real-Time Bed-Egress Recognition in Older People

“…Features from acceleration sensor: As the W 2 ISP data stream is a non-uniform time series, a faithful calculation of frequency-domain features, as in [8], [9], [14], without pre-processing is not possible [33]. Hence, we focus on time-domain features [8]- [11], [14] and features based on biomechanics [8], [12], [34] by analyzing bed-egress motion illustrated in Fig.…”

Effectiveness of a Batteryless and Wireless Wearable Sensor System for Identifying Bed and Chair Exits in Healthy Older People

“…Previous studies have established the importance and utility of RSSI based features. In [ 39 ], the combined use of RSSI with acceleration based features improved the classifier performance compared to using acceleration based features by themselves; this study [ 39 ] also demonstrated that the combination of features provided similar or better performance to using time and frequency domain features from acceleration readings alone. In [ 27 ], variations in RSSI data were useful to determine changes in postures that would otherwise be difficult to discriminate using only acceleration based data, e.g., a person sitting in the chair and standing has the participant’s trunk to be upright in both postures.…”

“…Therefore, we combine time domain information in RSSI, phase, frequency channel and acceleration sensor data and employ three categories of features based on studies in [ 16 , 28 , 39 , 40 , 41 ]; we describe them in detail below.…”

Enhancing the Accuracy of Health Care Internet of Medical Things in Real Time using CNNets

“…Wearable Wireless Identification and Sensing Platform (W2ISP) Wearable WISP (W2ISP) devices are designed by [21]. As shown in Figure 1, W2ISP is a small, easy-to-maintain device that is wearable over clothing [22][23][24]. Thus, older patients can wear them for experiments.…”