Automated small-scale hydroponics: a low-cost fuzzy-IoT system for non-specialist users

Objectives Small-scale urban hydroponic systems suffer from low volumetric buffering capacity, causing rapid fluctuations in pH and electrical conductivity (EC). This study aims to develop and validate a low-cost, decentralized Deep Flow Technique (DFT) system utilizing a embedded Fuzzy Logic Controller (FLC) and an IoT framework to autonomously manage these physicochemical parameters for non-expert users. Methods An ESP32-based FLC was deployed to calculate corrective dosing natively, avoiding the latency of cloud processing. To rigorously evaluate controller robustness, 12 severe perturbation experiments (extreme acidification, alkalinization, nutrient depletion, and saturation) were conducted on a plantless hydraulic bench. The FLC’s transient response and steady-state accuracy were then statistically compared against a conventional threshold-based (On-Off) controller. Results The FLC demonstrated superior performance for linear EC dynamics, reducing the average settling time by nearly 50 % (34.9 s vs. 68.4 s) and achieving a marginally significant lower steady-state error. Conversely, controlling the highly nonlinear, logarithmic dynamics of pH proved challenging; the FLC’s symmetrical linear membership functions yielded persistent micro-oscillations, allowing the On-Off controller to achieve a statistically better Root Mean Square Error (RMSE) ( p = 0.0260). Conclusions The proposed embedded architecture successfully insulates small-scale systems from linear perturbations (EC), significantly lowering the technical barrier for novice operators. While managing non-linear parameters like pH requires future adaptive or asymmetrical membership functions, this FLC-IoT integration ensures automation, transforming users from manual laborers into remote supervisors.