Abstract
This paper describes a comprehensive framework for the engineering of smart agricultural systems using the Internet of Things (IoT) and Artificial Intelligence (AI) technologies for sustainable urban farming. To address the increased issue of food scarcity, lack of arable land, and environmentally constrained space in urban areas, we aim to design and implement a vertical farming system with advanced IoT sensor networks composed of DHT11 temperature and humidity sensors as well as lux light intensity sensors, and manual measurement of plant height to monitor crucial environmental parameters in real time. Data was gathered constantly at one-minute intervals for two weeks with both cloud and local systems to ensure robust and scalable data acquisition for a variety of use cases. Afterwards, short-term temperature forecasting enabling proactive climate control to optimize resource use and crop yield was done with classical time series approaches (ARIMA) and deep learning techniques (LSTM) based models. Analysis showed that the baseline for ARIMA was strong, but the LSTM model performed somewhat better, demonstrating the advantages of AI enabled predictive control in indoor vertical farming. The model increases environmental stability and energy efficiency while laying the groundwork for smart urban agriculture systems that fosters sustainable development.