WasteShark Platform

An end-to-end system for autonomous waterway cleanup built by Theta Tau – combining a React frontend control center, a Node.js backend with MQTT + MongoDB, embedded navigation scripts for the robot, and computer-vision tooling for debris detection.

---

Repository Structure

Path	Description
frontEnd/	Vite + React 18 dashboard, authentication flow, robot telemetry UI, Tailwind styling, and integration layer ready for backend APIs.
backEnd/	Express server exposing REST APIs, MQTT bridges, MongoDB persistence, Docker Compose orchestration, and JWT authentication middleware.
microController/	Python scripts for Raspberry Pi / Pixhawk integration (navigation, LED tests, remote enable), along with deployment notes.
ML/	YOLOv5-based perception stack, training configs, and deployment scripts for detect-and-drive automation.
hardware/	Hardware-specific documentation and assembly notes (see folder README).
TailscaleSetup.md	Secure VPN setup instructions for remote robot connectivity.

---

Quick Start

Prerequisites

• Node.js 18+ and npm (frontend/backend development)
• Docker & Docker Compose (recommended for backend + infrastructure stack)
• Python 3.10+ (microcontroller scripts, ML tooling)
• MQTT client (optional, for manual robot telemetry testing)
• Git LFS if you plan to version large ML weights or CAD assets

---

Frontend (Vite + React)

The web dashboard lives in frontEnd/ and provides authentication, robot monitoring, and admin tooling.

cd frontEnd
npm install      # install dependencies
npm run dev      # start Vite on http://localhost:5173
npm run build    # production bundle
npm run lint     # check formatting / lint rules

Key files:

• src/App.jsx – router + provider composition
• src/context/AuthContext.jsx – JWT storage and auth helpers
• src/pages/Dashboard.jsx – telemetry polling, robot controls, and layout
• src/api/*.js – API facades (swap mocks for backend endpoints)

Set VITE_API_BASE_URL in .env once the backend URL is known:

cp .env.example .env
echo "VITE_API_BASE_URL=http://localhost:3000" >> .env

---

Backend (Express + MQTT + MongoDB)

The backend lives in backEnd/ and exposes REST endpoints, manages JWT auth, proxies robot commands over MQTT, and persists state in MongoDB.

Run with Docker (Recommended)

cd backEnd
cp .env.example .env         # populate Mongo + MQTT credentials
docker compose up -d         # starts backend, MongoDB, Mosquitto, Nginx (if configured)
docker compose logs -f       # follow logs

Local Node.js Development

cd backEnd
npm install
docker compose up mongo mosquitto-broker -d   # spin up dependencies
npm run dev                                   # or node server.js

HTTP endpoints auto-load from api/http/**, MQTT handlers from api/mqtt/**, and middleware (JWT, robot ownership) resides under middleware/. Schemas are defined with Mongoose in schemas/.

---

Microcontroller & Navigation

Scripts in microController/ target the onboard Raspberry Pi / flight controller:

• navigation_code.py, nav_barebones.py – autonomous routing and motor control routines.
• LED_Test_Code.py – hardware diagnostics.
• microcontroller_remote_enable.txt, raspberryIP – deployment notes.

Most scripts assume Python 3 with access to hardware GPIO / MAVLink libraries. Review each file’s header comments before running on the robot.

---

Machine Learning (YOLOv5)

The ML/ directory contains a customized YOLOv5 fork for debris detection and autonomous decision making:

• yolov5/detect*.py, detect_and_drive*.py – inference pipelines, some coupled with telemetry outputs.
• yolov5/train.py, data/, runs/ – training configs, dataset pointers, and experiment artifacts.
• yolov5/export.py, hubconf.py – model export for edge deployment.

Follow standard YOLOv5 setup (PyTorch, CUDA/cuDNN if available) and consult ML/readMe for environment details. Use virtual environments to avoid dependency conflicts with the frontend/backend stacks.

---

Hardware Docs

All mechanical/electrical assembly instructions live inside hardware/. Refer to that folder’s readMe (and any PDFs/CAD files tracked elsewhere) before modifying the physical system.

---

Development Workflow

1. Create a new branch for your feature/fix.
2. Update environment files (.env, docker-compose overrides) instead of hardcoding secrets.
3. Run relevant tests / linters:
   • frontEnd: npm run lint
   • backEnd: npm test
   • ML: python -m pytest or YOLO training dry-runs as applicable
4. Submit a pull request with a clear summary and testing evidence.

---

Support & Communication

• Open GitHub issues for bugs or feature requests.
• Use the team’s Tailscale mesh (see TailscaleSetup.md) for remote hardware debugging.
• Sync with the robotics, frontend, backend, and ML subteams during weekly standups to keep interfaces aligned.

---

License

Back-end services are currently distributed under the ISC license (see backEnd/README.md). Confirm licensing requirements with the project lead before publishing additional components or datasets.

---