IEEE Internet of Things Journal · 2019 · 462 citations · 40 references
ComputingComputer ScienceComputer SystemsInternet Of ThingsMachine LearningSensorsAction RecognitionDeep LearningIot SensorsSensor NetworksSmart SensorsWearable Health TechnologiesActivity MonitoringActivity RecognitionContinuous MonitoringConvolutional Neural NetworksDeep Learning For Health MonitoringHuman Activity RecognitionWearable DevicesWearable Sensors
Human activity recognition is a key component of continuous monitoring frameworks for ambient assisted living, well‑being management, medical diagnosis, elderly care, rehabilitation, entertainment, and smart‑home surveillance. This study presents an innovative HAR system that combines wearable devices with deep‑learning techniques to recognize common daily activities at home. The system employs a wearable sensor equipped with an IMU and Wi‑Fi that streams data to the cloud, and a lightweight CNN that performs inference on low‑cost or embedded devices. The system accurately identifies nine daily activities with a 97 % accuracy rate.
Human activity recognition (HAR) is currently recognized as a key element of a more general framework designed to perform continuous monitoring of human behaviors in the area of ambient assisted living (AAL), well-being management, medical diagnosis, elderly care, rehabilitation, entertainment, and surveillance in smart home environments. In this paper, an innovative HAR system, exploiting the potential of wearable devices integrated with the skills of deep learning techniques, is presented with the aim of recognizing the most common daily activities of a person at home. The designed wearable sensor embeds an inertial measurement unit (IMU) and a Wi-Fi section to send data on a cloud service and to allow direct connection to the Internet through a common home router so that the user themselves could manage the installation procedure. The sensor is coupled to a convolutional neural network (CNN) network designed to make inferences with the minimum possible resources to keep open the way of its implementation on low-cost or embedded devices. The system is conceived for daily activity monitor and nine different activities can be highlighted with an accuracy of 97%.
40
Algorithms for hyper-parameter optimization
James Bergstra, Yoshua Bengio · 2011 · 3.2K citations · Full text