Digital Agriculture/Internet of Things

Reusable CSA Box

We’re designing a reusable container system for use in ZWI’s Community Supported Agriculture (CSA) box distribution.

• Materials: Selecting lightweight, food-safe stainless steel for durability and longevity.
• Mechanical design: Developing a foldable/flattenable frame for compact storage, with a secure stackable locking mechanism that eliminates the need for plastic cling wrap.
• Thermal management: Implementing a fine-mesh design to ensure airflow, preventing vegetable spoilage while maintaining structural integrity.

Potential future projects — “Project Horizon”: We’re also prototyping a wearable-style sensor system with a 3-camera array (front- and side-facing) to capture a wide-angle view of a user’s surroundings. This is in the idea phase, but if implemented, it would be developed in partnership with our ML subteam, which will translate the visual feed into real-time audio feedback. A key design priority is focusing heavily on user acceptance to ensure the final product feels genuinely supportive and helpful, rather than perceived as just another piece of isolating technology that doesn’t improve their daily life.

Digital Infrastructure for ZWI

We’re rebuilding and modernizing ZWI’s web infrastructure so it can scale to support both community outreach and government partnerships.

• Migrating “BYO – US Reduces” off an aging WordPress/Bluehost setup that was difficult to maintain and update, including rebuilding the site’s event registration form.
• Building a Complaint Tracking System to be a low-maintenance, flexible tool designed to be usable not just by ZWI, but by city and state government sites as well.
• Tech stack: Django for the backend, with React.js powering the static BYO campaign outreach site.

Computer Vision for Agriculture

In parallel, our ML subteam has been focusing on our long-standing project, building an end-to-end computer vision pipeline to detect Northern Leaf Blight (NLB), a common corn disease.

• Implementing a two-step model architecture combining a YOLO object detector with a CNN classification step.
• Pipeline: Images stored in AWS S3, models built in TensorFlow (with LabelMe for annotation and custom augmentation), trained on AWS EC2/Colab, and containerized with Docker for deployment.
• Goals: Deploying the finished model on AWS SageMaker and hosting a public-facing demo, so the tool is freely accessible.

Future & Potential Projects

Beyond our current work with ZWI, we’re exploring a new potential collaboration with Professor Max Zhang on a project called EngagedIoT, which would expand our scope into energy, infrastructure, and public health applications across Tompkins County:

• Energy Metering: Designing current-transformer-based power meters to monitor energy use at CCE offices and municipal water pumps.
• Infrastructure & Health: Building mobile road-surface monitoring tools, microclimate sensors for “dual-use” solar farms, and remote telehealth devices for the county.
• Social Impact: Developing privacy-preserving hardware to monitor food cabinet stock levels, supporting local food insecurity efforts.

This would give our team the chance to apply IoT and embedded hardware skills to an even broader range of civic and environmental challenges, including energy systems, public infrastructure, healthcare access, and food security, rooted in direct partnerships with local organizations and researchers.