Low-power, low-cost custom electronic system for in-vivo plant stress monitoring and prevention

Environmental sustainability is a growing global concern, since agriculture significantly contributes to eutrophication and greenhouse gas emissions. Smart agriculture addresses these challenges by enhancing crop productivity while optimizing water and nutrient use through innovative sensing technologies. This thesis presents a low-cost, low-power, custom electronic system for in-plant health monitoring, via electrical stem bio-impedance and a leaf-mounted capacitive sensor called Phylloclip. The goal is to support precision irrigation and prevent fungal disease, by directly assessing plant physiological conditions. Two measurement architectures were developed and compared. The first one employs the FDC2112 capacitance-to-digital converter, where the Phylloclip forms part of an LC tank. Variations in stomatal conductance and transpiration alter the dielectric constant and, consequently, the resonance frequency. The digital value is then stored in a register and made available to a microcontroller using the I²C protocol. The second version exploits an LMC555 CMOS timer as a relaxation oscillator, integrating the sensor in its feedback loop and converting its variations into frequency shifts. This design reduces power consumption and cost, while maintaining comparable accuracy. The same approach is applicable to both stem and leaf, enabling indirect impedance and capacitance analysis without I²C, relying only on a timer used in input capture mode. In addition, an environmental sensor for air temperature and humidity has been included, allowing vapor pressure deficit (VPD) estimation and detecting when air temperature reaches the dew point, increasing the condensation risk. Through selective circuit activation and ultra-low-power MCU modes, the system achieves long battery endurance for field deployment. Data are transmitted through the LoRa protocol, enabling long-range, low-power communication for distributed sensor networks. The prototype has been tested for several months on tobacco, apple, and kiwi plants under varying temperature and humidity conditions, in greenhouse, climatic chamber, and open-field setups, to validate its stability and reliability in realistic operating scenarios. The obtained results show that the stem frequency is inversely proportional to the electrical stem bio-impedance, allowing early detection of water stress, before any visible symptoms appear. The Phylloclip capacitance is directly correlated with leaf transpiration, let the identification of critical periods associated to a higher risk of fungal infections and enabling the estimation of leaf wetness duration, key to diseases such as Peronospora, which require water to proliferate and penetrate leaf tissue through open stomata. This information, unavailable from stem monitoring alone, also permits the identification of periods when the stomata are more open, providing valuable information on the optimal timing for pesticide application. Finally, two custom PCBs have been designed, fabricated, and assembled, but only the one based on the LMC555 has been selected as the final implementation. This board integrates dual capacitance and quad impedance measurement in a compact shield for the LoRaTO platform, based on the STM32WL5MOC microcontroller. Dedicated firmware has been implemented to ensure MCU–hardware interfacing, resulting in a sustainable solution for Agritech and Precision Agriculture applications. Future improvements will integrate energy harvesting for autonomous, battery-free operation.