UC Berkeley

Nitrate is a critical nutrient for plant growth and a primary component of commercialfertilizers. Unfortunately, nitrate production is an energy-intensive process and, when applied in excess, generates greenhouse gases and pollutes the water supply. Despite this, tools for monitoring nitrate and other nutrients in the soil are inadequate. Measurements must be made at a high spatial and temporal resolution to optimize agricultural water and fertilizer inputs.

We will present the model-based design, fabrication, and implementation of wirelessnitrate sensor nodes in a precision farming system. First, I will discuss the design, fabrication, and characterization of fully printed potentiometric nitrate sensors, describing in detail key performance metrics and benchmarking our sensors against existing sensors in the literature. We will next dene and demonstrate the optimized placement of soil sensors in agricultural fields using machine learning optimization approaches to determine the scale of production required to meet the project goals and inform the design space of the wireless sensor nodes. The nitrate sensors were modied, integrated into a WiFi-enabled sensor node, and characterized in varying solution and soil conditions. The sensor node was adapted into a low-cost, naturallydegradable, passive RFID nitrate sensor node, characterized, and modeled into a wireless sensor network sampled by autonomous UAV drones. Finally, I describe the challenges of interference and present preliminary results of a multianalyte nitrogen sensor array.