A Fuzzy-Logic-Based Smart Irrigation Controller for Precision Agriculture

Precision irrigation utilizing sensors and IoT devices was introduced for efficient utilization of natural water resources in the agriculture sector. Most existing precision irrigation techniques are computationally complex. To solve this issue, a precision irrigation controller using a fuzzy inference system (FIS) is proposed. The proposed FIS uses the deviation from the reference soil moisture and the crop coefficient as inputs. Then, the optimal incremental control of irrigation volume is computed using a 28-rules rule base. While discrete linear quadratic regulator (DLQR) is considered one of the most accurate techniques in the control of closed-loop systems, it is computationally expensive and not feasible for real-time control using the IoT. On the other hand, proportional-integral (PI) controllers are computationally lightweight but struggle to achieve higher accuracy. Simulations using actual data in MATLAB show that the proposed fuzzy-based model predictive control (MPC) controller not only closely follows the behavior of DLQR but it does so with significantly lower computational complexity of $(O(r^{k}))$ , approximately the same as a PI-based controller. The proposed fuzzy-based MPC controller is implemented on real hardware for validation and testing using an IoT device. The IoT device not only exhibits similar behavior as the simulations, an improvement of up to 37% in execution time as compared to the state-of-the-art existing techniques is observed. The proposed technique is a step forward to the development of a simple and fast irrigation controller that will enable better scalability and application to a wide range of agricultural environments.