Autonomous Tissue-Interfaced Plant-Machine Interface for Real-Time Physiological Control of Phosphate

The emergence of smart agriculture has motivated the development of smart fertigation management system (FMS) to replace conventional scheduled-meter-based fertigation. For smart agriculture, these new FMS should possess five key characteristics: robotic, in vivo microneedle-based electrical impedance spectroscopy (EIS) sensing; machine-learning (ML)-based instant data interpretation; automated precise nutrient formulation and delivery; real-time feedback-driven fertigation adjustment; and identified optimal growth phosphate status. However, sensors, ML, or automated fertigation injector reported to date typically fulfill one or two requirements at the same time, and rare are examples of FMSs that fulfill all five requirements simultaneously. Here, we present a bio-integrated computerized FMS that meet all five requirements simultaneously. TheMobile Sensing Platform provides stable, high-throughput measurements (16 plants/h) with minimal tissue disruption, with equipped MN sensors accessing into plant compartments to enable instant detection of physiological phosphate status; the lightweight Extra Trees classifier translated impedance spectra into plant phosphate states with 89% accuracy in <1s; These predictions triggered automated nutrient formulation and delivery within 15 min and 0.25 g/L precision, demonstrating continuous electronic-biological feedback control. Deployment in indoor strawberry cultivation identified an optimal phosphate regime that improved yield while reducing fertilizer input relative to conventional recommendations. By bridging the gap between tissue-interfaced sensing and closed-loop actuation, this work establishes a prototype of potentially scalable bioelectronic framework that couples living plant physiology with autonomous electronics, unlocking new avenues for sustainable, bio-integrated automation in smart agriculture