Artificial Intelligence Enabled Smart Monitoring and Controlling of IoT-Green House

• Published in: Arabian journal for science and engineering (2011) Vol. 49; no. 3; pp. 3043 - 3061
• Main Authors: Riskiawan, H. Y., Gupta, Nishu, Setyohadi, D. P. S., Anwar, S., Kurniasari, A. A., Hariono, B., Firmansyah, M. H., Yogiswara, Y., Mansur, A. B. F., Basori, A. H.
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01.03.2024; Springer Nature B.V
• Subjects: Actuators; Artificial intelligence; Automatic control; Central processing units; CPUs; Data collection; Embedded systems; Engineering; Environmental control; Humanities and Social Sciences; Indoor environments; Internet of Things; Microcontrollers; Monitoring; multidisciplinary; Parameters; Predictive control; Research Article-Computer Engineering and Computer Science; Science; Smart green house; LSTM; Monitoring; Actuator; IoT
• ISSN: 2193-567X; 1319-8025; 2191-4281
• DOI: 10.1007/s13369-023-07887-6

Green houses are being built and expanded at a breakneck pace. The green house climate directly affects plant’s development, and its continuous indoor environment monitoring is critical. Most green house systems rely on manual temperature and humidity monitoring which can be inconvenient for personnel required to visit the green house daily and manually control it. This research uses a modern approach by implementing automated environmental control technology to improve green house control technology effectively. The integration of the Internet of Things (IoT) and artificial intelligence (AI) can develop a capacity for independently predicting and controlling IoT devices. The microcontroller controls the system, which serves as the central processing unit for sensors and actuators. The sensor data utilizes input parameters for the microcontroller, which processes it using the long short-term memory (LSTM) approach to anticipate output parameters for controlling actuators, such as fan exhaust, misting, and motor control. Intelligent control is not placed directly on the embedded system but on a framework known as Laravel through the results of knowledge trained by the LSTM method. Using the traditional embedded system, no data can be learned using only simple conditioning. The test findings from the LSTM training data obtained that the learning rate was less than 0.002 with a total of 250 steps, where the results were processed from the accumulation of data every minute for one month, are near to a good value. Consequently, the proposal can accurately forecast the actuator control in three minutes based on data collected from the accumulated data in the preceding three minutes. The system will collect data for future use to anticipate and optimise the melon product. The primary purpose of using AI, IoT and LSTM is that the system can see the wide range of agricultural applications by taking into account several other complex parameters, each owned by LSTM to be applied using IoT devices, so that they can be will be applied directly using publish and subscribe method.