AgroTech 2.0
Project scope: automate irrigation decisions using real-time soil and temperature data. Context: Bogotá water crisis, 2024. Core problem: bridging Arduino sensor output to actionable software for non-technical users. Status: v2 complete.
The Challenge
The primary constraint was cost. The prototype had to be functional and affordable enough to be a realistic model for small agricultural operations, not a lab experiment. For v2 the challenge shifted to data acquisition: reading sensor output reliably over USB introduces real limitations. Port detection, signal noise, connection stability. Each requiring deliberate handling before a single useful data point reaches the application.
The Approach
The focus was not on building something impressive. It was on building something useful. The desktop application needed to give enough context about irrigation, what data matters, what thresholds mean, what a reading actually tells you about a crop, so that the information it displayed was actionable, not just visible. Useful for someone who understands agriculture, not just someone who understands software.
What Was Built
The system has three layers. The Arduino board runs the irrigation program from its IDE, controlling the pump and reading sensors. A Java listener monitors the USB serial port, captures raw sensor output, and writes it to a data file. The desktop application then ingests that file, enriches and formats the data, and presents it in a way that is useful now and structured for the analytics platform planned in v3.
The Outcome
University recognition in the first semester of 2024. The evaluation considered not just the technical prototype but the broader scope: the social, economic, ecological, and cultural impact of making precision irrigation accessible. That framing mattered. The project was not assessed as a piece of software but as a proposal for what technology could mean for Colombian agriculture given the right resources and investment.
What I'd Do Differently
Wireless communication. The USB connection between the Arduino and the PC introduces too many constraints for a real agricultural deployment: cable length, port conflicts, physical positioning of the board relative to crops. The next version needs the Arduino transmitting data wirelessly, which also changes how the Java layer listens and how the data pipeline is structured. The entire acquisition architecture would be redesigned around that constraint from day one.