earticle

영어

In this study, we analyze the service discovery protocol problem in a smart home environment and propose a deep learning-based prediction model. Currently, the service discovery protocols used in smart homes are divided into IP-based and non-IP-based, and there are limitations in service discovery owing to the interoperability issues between them. Although smart gateways support protocol conversion between heterogeneous networks, there is a limitation that smooth service discovery between smart home devices is difficult due to the lack of advertising and broadcasting functions for service discovery. To solve this problem, we propose a deep learning-based prediction model utilizing Recurrent Neural Networks (RNN) and Long Short-Term Memory (LSTM) networks. The experimental results show that the proposed model can accurately predict future service demands by learning various service request patterns and enables faster and more efficient service discovery than existing protocols. The proposed model has been shown to improve the interoperability between devices in a smart home environment that changes in real-time and significantly reduces the service discovery time.