A gallery of CORN-DRIVE growing system I designed and implemented. The CORN-DRIVE system is a dual-mode, autonomous indoor cultivation platform that transforms precision agriculture by transitioning from traditional timer-based irrigation to Vision-Defined Actuation (VDA). Operating as a decentralized Neural Rack, its primary Edge-AI tier utilizes a low power CPU and highly optimized computer vision to continuously monitor plant phenotypes, instantly translating early biological stress markers into targeted microcontroller interventions. To ensure maximum experimental rigor and provide a reliable control group, the system also features a robust, AI-independent baseline tier that relies on standardized soil moisture sensors and precise real-time clock scheduling. By successfully bridging the gap between high-throughput biological phenotyping and real-time robotic actuation, CORN-DRIVE provides a scalable, open-source blueprint for the next generation of intelligent, self-learning automated farms. Corn plants are grown to feed the larvae of Bicyclus anynana butterflies.
The first generation of CORN-DRIVE system that utlizes two microcontrollers to detect water levels, temperature, humidity, TDS, and pH and controls multiple pumps, light timing, and wind speed.
The second generation racks that optimizes on the findings from the first generation system and utilizes a modular controller platform that detect the levels of water, TDS, temperature, and humidity and prodives control to the pumps and light timing.
The CORN-DRIVE: Neural Rack (running corn_AI_v1.0) is an autonomous indoor cultivation platform that replaces traditional timer-based irrigation with real-time Vision-Defined Actuation (VDA). By fusing low-cost edge AI with embedded robotics, it continuously monitors visual plant phenotypes to deliver water and nutrients precisely when the crop's biological state demands it.
Corn plants growing in a microgreen tray for around 5 weeks.
Corn plants harvested after 5 weeks of growth.
An early photograph of the greenhouse showing the CORN-DRIVE under construction.
Corn rack_1.0 with the lights off state.
Corn harvest after 4 weeks of growth in microgreen trays.
An array of corn modules under contruction at NUS.
An advanced version of the controllers that intergrates information from multiple different sensors including light intensity, pH, TDS, water temperature, chamber temperature and humidity, and intergrates information from Corn-AI to provide highly specific response against water, nutrient, and light stress.
The controllers with the sensors.
A view of the internals of the corn module.
A total of 6 cameras montior and sends back signal to main microcontroller against water and nutrient stress.
Corn module rack 3 during the night time with lights off.
Corn plants after 5 weeks of growth, harvested and taken back to the insectary to feed Bicyclus anynana larvae.
A plant after 6.5 weeks of growth with corn tassels.
Corn plants after 5 weeks of growth ready for harvesting.