Structural Action Recognition in Body Sensor Networks: Distributed Classification Based on String Matching

Abstract

Mobile sensor-based systems are emerging as promising platforms for healthcare monitoring. An important goal of these systems is to extract physiological information about the subject wearing the network. Such information can be used for life logging, quality of life measures, fall detection, extraction of contextual information, and many other applications. Data collected by these sensor nodes are overwhelming, and hence, an efficient data processing technique is essential. In this paper, we present a system using inexpensive, off-the-shelf inertial sensor nodes that constructs motion transcripts from biomedical signals and identifies movements by taking collaboration between the nodes into consideration. Transcripts are built of motion primitives and aim to reduce the complexity of the original data. We then label each primitive with a unique symbol and generate a sequence of symbols, known as motion template, representing a particular action. This model leads to a distributed algorithm for action recognition using edit distance with respect to motion templates. The algorithm reduces the number of active nodes during every classification decision. We present our results using data collected from five normal subjects performing transitional movements. The results clearly illustrate the effectiveness of our framework. In particular, we obtain a classification accuracy of 84.13% with only one sensor node involved in the classification process.

Action-based scheduling technique for 802.15.4/ZigBee wireless body area networks

“…Some works have been previously done on action recognition using on-body sensors [12]- [14]. In these work accelerometer sensors are used to collect acceleration of limbs in different directions and gyroscopes are applied to collect angular velocities in two directions of inclination angle θ and azimuth angle φ.…”

“…In these work accelerometer sensors are used to collect acceleration of limbs in different directions and gyroscopes are applied to collect angular velocities in two directions of inclination angle θ and azimuth angle φ. Features are then extracted from the collected parameters in different time segments, e.g., mean, standard deviation, root mean square, first and second derivatives [14]. They are subsequently transmitted to the base station so that the action can be decided based on hidden Markov model (HMM).…”

An Ultra Low Power Granular Decision Making Using Cross Correlation: Minimizing Signal Segments for Template Matching

“…However, a basic requirement of movement monitoring applications is to detect actions first, and perform additional processing next. This application is usually referred to as action recognition [26][27][28]. Typical chain of signal processing for action recognition include filtering, segmentation, feature extraction, feature conditioning, and classification.…”

Multi-sensor fusion in body sensor networks: State-of-the-art and research challenges

“…Consequently, the last decade has witnessed tremendous efforts in utilizing smart technologies such as BSNs for health monitoring and diagnosis through physical activity monitoring/assessment. Recent years have seen considerable research demonstrating the potential of BSNs in a variety of physical activity monitoring applications such as activity recognition [9,10,11,15,16,17], activity level estimation [18], caloric expenditure calculation [19,20], joint angle estimation [21], activity-based prompting [53,54,55,56,57,58], medication adherence assessment [59,60], crowd sensing [61,62,63,64,65,66], social networking [67,68,69,70], and sports training [22,23,24,25,26].…”

“…Each segment has a multidimensional (feature) vector extracted from it, which will be used for classification [93,11]. The most widely used classification and event detection algorithms include k-NN (k-Nearest-Neighbor), Support Vector Machines (SVM), Hidden Markov Models (HMM), Neural Network (NN), Decision Tree Classifiers, Logistic Regression, and the Naive Bayesian approach [94,95,96,97,98,99].…”