Fuzzy Logic Inference-Based Automated Water Irrigation System

Introduction

Among the developing countries, India is one of the quickest developing economies in the world today, and a very large part of this economy is dependent on its agricultural sector. Our country also has the second-highest population on the planet Earth, which means we always face a problem of scarcity of resources. One of the most important and scarce resources is freshwater. Along with all that, we face unreliable monsoons and isotropic climatic conditions. For a supportable, adequate and solid farm yield, the vulnerabilities concerning temperature, humidity, environment and so forth should be confronted. Irrigation, be it rural or homegrown, is a tedious and work concentrated errand. Furthermore, if not done as expected, outcomes in an unfriendly impact on the plant and the soil, and might also lead to wastage of water. It is additionally vital that the watering of land is done at the perfect time, else, it might harm the harvest/plant. Under or Over-irrigation can decrease the nature of soil as far as its health benefit. By appropriate utilization of Science and Technology, we can foster apparatuses and technology that are naturally smart and can help in diminishing ecological corruption and human end eavours simultaneously (Oza and Sharma, 2014).

Agrarian water’s minimal use capability, inadequacy and waste are huge issues of momentum improvement of the irrigated cultivating. The dry season is one of the major ecological pressure components for plant development, which summarizes exceeding any other elements. A water system is a framework that straightforwardly supplies water and other compounds agents to soil with the moderate and standard inventory. A proficient water system is subsequently critical to determine the impacts of the dry season. Throughout the long term, a large portion of our water irrigation system was constrained by manual experience. Along these lines, analysis of automated water systems has incredible importance. Another drawback of environmental change can’t deliver a wide scope of things like foods grown from the ground. Along these lines, we have considered something which will bring strategy by presenting some controlled framework that will control the moisture in the soil wherein the plants are creating, and feed the plants in a dry environment to pass on yields. If we can use an automated system being developed cycle, we can convey a wide extent of harvests in each season which will lessen support and import costs too.

Weather plays a vital role in the agricultural domain. India is a country where most of the places have either tropical or subtropical climates. Weather forecasting is an area that analyses the current and past weather-related data and predicts the atmospheric conditions with the help of machine learning algorithms. Many techniques were proposed to forecast climate conditions such as satellites, radar systems, buoys and land stations etc (Kunjumon et al., 2018). The machine learning algorithms can be incorporated to get accurate forecasting data. Few of the machine learning algorithms used in forecasting includes supervised, unsupervised algorithms, support vector machines (SVM), artificial neural network (ANN) and clustering and classification algorithms. These algorithms provide better and accurate analytical weather-related data. The irrigation of the crops is contingent on the current climate conditions. Over irrigation of water in the rainy season can introduce plant root related diseases such as fungus, whereas under irrigation in the sunny or dry season can cause the crop to dry out. Hence, we need accurate climate forecast data to ease the irrigation process without damaging the field crops.

Fuzzy logic was introduced to handle unpredictability in circumstances. The meaning of fuzzy is unclear or vague. These unsolved real-world issues can be addressed by Fuzzy logic with the help of a fuzzy rule base (Khatri, 2018). The fuzzy logic-based system contains antecedents (inputs), consequent (outputs), a membership function and a set of the fuzzy rule base. A fuzzy rule base is used to map an input-output based relation. Before utilizing the fuzzy rules, initially, the crisp values of antecedents need to be converted to the fuzzy logic with the help of membership functions. This process is called fuzzification. These linguistic inputs can be passed to an inference mechanism, whose task is to calculate the required consequent based on the values of antecedents and the set of fuzzy rules in terms of ”If-Then” statements. We have incorporated fuzzy logic in our system with the help of a few python modules to address the issues that limit the efficiency of agrarian irrigation.