Overview
The Autonomous Litter Detector Mapper is an innovative solution aimed at tackling environmental pollution by combining robotics and AI to detect and map litter. Designed as a low-cost, practical tool, this project showcases the feasibility of using autonomous vehicles for environmental cleanup in real-world settings.
Problem We Solve
This robot traverses a space, autonomously detecting and mapping litter while avoiding obstacles. The solution aims to provide an efficient, scalable method for monitoring and addressing pollution in various environments.
Implementation Details
The core components of this project are powered by a Raspberry Pi platform, combined with a camera, ultrasonic sensors, and servo motors. The project includes the following technical features:
- Navigation: The robot uses two ultrasonic sensors to implement an advanced obstacle avoidance system. The traversal is based on a pseudo-randomized Depth-First Search (DFS) algorithm, enabling efficient movement through the space.
- Object Detection: A camera captures snapshots of objects detected in the robot's path. These images are analyzed using TensorFlow machine learning object detection models, trained on 10,000 images in the TACO dataset, to classify litter items such as plastic bottles and paper debris.
- Mapping: Detected litter locations are stored in a two-dimensional undirected graph, allowing the robot to create a complete map of the area highlighting littered locations.
- Optimized Resource Use: Instead of continuously analyzing live video, the robot captures snapshots only when an object is detected, enhancing responsiveness and conserving energy for prolonged operation.
Avoidance System
The robot's movement system is built on a simple coordinate system and cardinal direction movement. When an obstacle is detected, the robot identifies whether it is trash using its machine learning model. Only valid litter locations are saved in the graph for mapping purposes.
Future Potential
This project demonstrates the potential of using autonomous robotics for environmental monitoring. Future enhancements could include real-time litter cleanup capabilities and integration with larger waste management systems to promote cleaner, more sustainable communities.
Conclusion
By combining robotics, AI, and environmental science, the Autonomous Litter Detector Mapper represents a significant step forward in addressing pollution. Its low-cost, scalable design makes it a practical solution for creating cleaner spaces in urban and natural environments.