Article Preview
TopIntroduction
Water happens to be the most vital element for sustaining life on the Earth both for humans and animals. Studies show that humans can live without food for three weeks but when it comes to water, humans cannot sustain more than three days at a stretch. The same theory applies to the crops that are grown in the field; they do require water. The depleting levels of the water table, irregular monsoons, climate change (Aryal, J.P., et al. 2019), water contaminations are posing as serious hurdles for developing a sustainable agriculture system (Tripathy S, 2019). In agriculture, even with the irrigated lands with adequate irrigation sources (S. Latha, 2019), the utilization of water for irrigation is not strategic. The problem that needs immediate attention is, how water, being a limited and vital resource can be smartly used for irrigation purpose to produce good quality yields capable of fetching good returns to the farmers. Precision Agriculture (PA) system would definitely be a suitable solution for most of the issues arising in the agricultural domain including irrigation. In order to provide efficient water utilization, the PA system needs to be tailor-made for the farmers especially from developing countries. The implementation success of the PA system largely relies on the implementation costs, implementation complexity, deployment time, and ease of maintenance. To ensure cost-effectiveness, the field sensor-based approach is recommended. PA system using field-based sensor approach for data collection, IoT for providing remote monitoring of parameters and using intelligence in the form of ML-based predictive model to provide closed-loop control of field parameters would definitely transform the traditional agriculture into a sustainable one. The reach of the Internet in every nook and cranny of the world has created a huge demand for applications involving things rather than peoples. Currently, there are around 8.3 Billion IoT connected devices worldwide and it is predicted that by 2025, this number will be more than double what it is today (21.5 Billion). IoT has already started benefitting the users worldwide in all the sectors be it healthcare, industry, education or agriculture. Though the implemented PA systems have seen the progress and success in the agricultural sector in countries like Australia (Jochinke et al. 2007), Belgium, Canada, and the United States to name a few, still the major chunk of farmers are yet to harness the rewarding benefits of it. Going by the literature, there are many implementations of IoT in the agriculture domain. Agriculture can be thought of as a complex system, which consists of sub-systems like soil preparation, seed implanting, irrigation, fertilization, weeding, harvesting, sorting, storing and transportation.
IoT was the main driving force for the development of the agricultural systems providing some of the outstanding solutions to the problems being faced by the farmers in the agricultural and the related domains. Also, security remains the main concern when IoT is involved in providing enterprise and business solutions to the customers. Limited researches focused on the security part of IoT implementations in agriculture as in (L. Vidyashree and B. M. Suresha, 2019), where an encryption method was proposed for securing agriculture data.
The work carried out in this research attempts to design a highly secured agricultural field monitoring and irrigation scheduling system by using IoT and Artificial Neural Network (ANN) based predictive model. The IoT part is responsible for data collection, storage, and visualization. The feed-forward neural network (FFNN) model uses the locally generated datasets from the IoT cloud-server as model inputs. The performance measurement of the FFNN predictive model was done based on MSE and R. The model was able to accurately predict the moisture values in the field with low values of MSE and high values of R, apart from this, the other the salient features of the proposed system are that the developed system uses of low-cost and easily available sensors, the main center of attraction of the developed system is the ESP32 DevKit V1 which hosts an ESP32 SoC MCU (capable of providing high security, ultra-low power requirement with built-in Wi-Fi and dual Bluetooth modules), open-source hardware/software platforms for prototyping and programming, open-source ThingSpeak™ IoT platform and API for data storage and visualization which also provides MATLAB® analytics on the cloud. Finally, an Android App is developed by using the Blynk platform for providing user-friendly irrigation control and automation in the field along with user notification in the form of email and SMS.