Ensemble of RNN Classifiers for Activity Detection Using a Smartphone and Supporting Nodes

Abstract

Nowadays, sensor-equipped mobile devices allow us to detect basic daily activities accurately. However, the accuracy of the existing activity recognition methods decreases rapidly if the set of activities is extended and includes training routines, such as squats, jumps, or arm swings. Thus, this paper proposes a model of a personal area network with a smartphone (as a main node) and supporting sensor nodes that deliver additional data to increase activity-recognition accuracy. The introduced personal area sensor network takes advantage of the information from multiple sensor nodes attached to different parts of the human body. In this scheme, nodes process their sensor readings locally with the use of recurrent neural networks (RNNs) to categorize the activities. Then, the main node collects results from supporting sensor nodes and performs a final activity recognition run based on a weighted voting procedure. In order to save energy and extend the network’s lifetime, sensor nodes report their local results only for specific types of recognized activity. The presented method was evaluated during experiments with sensor nodes attached to the waist, chest, leg, and arm. The results obtained for a set of eight activities show that the proposed approach achieves higher recognition accuracy when compared with the existing methods. Based on the experimental results, the optimal configuration of the sensor nodes was determined to maximize the activity-recognition accuracy and reduce the number of transmissions from supporting sensor nodes.

Model Klasyfikacji Aktywności Na Bazie Czujników Dla Systemów Vr

“…Ta kwestia ma szczególne znaczenie, ponieważ zmniejszenie częstotliwości próbkowania o osiem razy może zwiększyć żywotność urządzenia trzykrotnie [7]. Jednak autorzy w pracy [8] wykazali, że próbkowanie równe 20 Hz umożliwia rozpoznawanie bardziej skomplikowanych ruchów. Dlatego też, ta wartość została przyjęta jako bazowa do dalszych badań.…”

“…Model jest implementacją rekurencyjnej sieci neuronowej (rysunek 3) gdzie podstawowa struktura została dobrana na podstawie publikacji [8,13]. Struktura ta została dostosowana do ciągu wejściowego oraz liczby klas.…”

Ambient intelligent framework for modelling critical medical events based on context awareness

“…Apart from this, different ambient sensors may offer ambiguous or noisy data, which makes it quite computationally challenging to perform interpretation [10]. From the viewpoint of contextual awareness towards activity detection, there is a higher degree of variations associated with user activity with a lack of personalization and adaptation, thereby making it quite challenging to construct a universal model [11]- [15]. Another significant challenge is constructing a comprehensive contextual model towards an activity determination system by integrating information from multiple sources [16]- [20].…”

Monitorowanie I Klasyfikacja Ruchów Użytkownika W Środowisku Vr

“…Model RNN (rys. 3.1) został zaadaptowany na bazie publikacji[8,9], przy czym proces dostosowywania i liczba operacji wstępnego przetwarzania danych zostały dostosowane do danych. Liczbę epok treningu dostosowano do analizowanej warstwa odrzutu (ang.…”