Our FestiCleaner robot autonomously identifies, locates, and cleans trash in real-time, and presents data in an intuitive dashboard. Built using an Arduino Uno Q and recycled cardboard, jar lids, etc. The robot's primary use-case is to clean festivals after events.
- Connect the robot and a computer to the same network (with appropriate device discovery settings). Firewalls, protected WiFi networks, etc. may cause issues. Self-hosting a hotspot on a phone should work.
- Configure the Arduino Uno Q to automatically connect to the WiFi network. The following commands may be useful:
Get the list of saved networks:
nmcli -t -f TYPE,UUID,NAME con
Delete a specific network. Use this ONLY on the main external network, like HackUPC or phone's hotspot:
sudo nmcli c delete <UUID>
Connect to a new network:
sudo nmcli dev wifi connect <WiFi-SSID> password <WiFi-password> - To start the robot, connect first the camera to the Arduino and then the battery.
- With our hotspot, the dashboard can be accessed at http://192.168.106.126:7000/.
Barcelona is known for its "festes majors", vibrant street festivals that turn neighborhoods into living art galleries. However, the aftermath often leaves plazas covered in plastic waste. FestiClean is a smart, autonomous rover designed to navigate the morning-after a street concert, identifying and collecting abandoned plastic cups and bottles to keep our city’s streets clean and sustainable.
FestiClean uses a multi-brain architecture to bridge the gap between low-level hardware and high-level AI:
- Computer Vision (The Eyes): Using the on-board camera and an optimized COCO-SSD object detection model, the robot scans the environment in real-time. It ignores static furniture and bystanders, focusing specifically on bottles and cups.
- Precision Navigation (The Brain): Once a target is spotted, the robot switches to a Pulse-Steering State Machine. It uses a Distance (ToF) sensor to approach the waste with millimeter precision, stopping exactly when the object is within reach of its custom-built gripper.
- Environmental Monitoring (The Safety): Designed for the hot Mediterranean summer, the robot features a Thermo sensor. It monitors local temperature and humidity, sending live telemetry to the cloud. If the temperature gets too hot, the system alerts operators via the dashboard.
We built a full-stack robotics ecosystem from the ground up:
- Hardware: Arduino Uno Q + Modulino Distance & Thermo Sensors + DC Motors + Servos.
- Embedded Logic: C++ State Machine for high-speed motor control and sensor polling.
- Edge AI & Bridge: A Python-based "Robot Brain" that runs Computer Vision and forwards data via HTTP POST.
- Backend: A FastAPI server (Python) that manages a SQLite database and broadcasts live data via WebSockets.
- Frontend: A high-performance React + Vite dashboard featuring a live video HUD, real-time telemetry gauges, and a remote "kill switch" for safety.
- Autonomous Pick-and-Place: From "Search" mode to "Drop-off," the robot handles the entire waste lifecycle.
- Zone-Based Tracking: Advanced motor logic that overcomes physical friction through discrete time-pulse steering.
- Live Dashboard: Operators can monitor the FestiCleaner from a single interface, seeing exactly what the robot sees.
- Sustainability-First: Built to tackle the specific plastic-waste profile of urban street festivals.