Integrated Internet of Things (IoT) Technology Device on Smart Home System with Human Action Recognition Using kNN Method Muhammad Idris Siddiq, Ig. Prasetya Dwi Wibawa, Meta Kallista
School of Electrical Engineering, Telkom University.
Jl. Telekomunikasi, Terusan Buah Batu, Bandung, Indonesia
Abstract
IoT device technology is currently developing rapidly, for example in smart home systems that have several features including lighting, surveillance security, temperature control, water sensors, and smart electricity. IoT device consists of smart electricity integrated with human action recognition using sensor vision are developed in this work. In smart electricity system, we build some relays controlled by smartphone applications and web-based platforms. We can control the relays and monitor the voltage, current, and power used from electricity appliances that are connected to our IoT device. In human action recognition, we use a single RGB camera to capture some human poses into spatiotemporal sequences to get data for training. There are six poses for testing scenario, these poses will be clustered using kNN (k-Nearest Neighbor) method. Each human action that is recognized will be connected to an IoT device for controlling the switching mode on the relays in smart electricity system. The result in this experiment shows that the system successfully reads every single movement with quite good accuracy performance using confusion matrix.
Keyword: human action recognition, k-Nearest Neighbor (kNN), smart electricity, internet of things (IoT)
Keywords: Human Action Recognition, K-Nearest Neighbor (KNN), Smart Electric Control
