Hardware application of human-machine interface in smart air conditioners using hand tracking

The aim of this paper is to present a real time application for air condition systems using hand recognition and tracking. The hardware requirements of the system, limited by only a simple USB webcam and a Beagle Board, which is a compact and low cost hardware. Our approach uses a simple motion detection trying to eliminate unnecessary regions of the image and a recognition algorithm based on Haar classifiers. An intelligent method and a practical application for detecting four hand movement instructions have been proposed for air condition systems in smart homes using Beagle Board. The system translates the detected gestures into different functional inputs and interfaces to air conditioner. The proposed method is also intelligent because it has an optimized multi layered neural network which is also optimized by a swarm-intelligence-based global optimization algorithm, Artificial Bee Colony Algorithm.