COBOL – COmmunity-Based Organized Littering

Littering is a major problem that threatens the environment (fire hazards, air pollution, street flooding), society (human health hazards) and the economy (significant expenses for administrations and municipalities). Its roots are in society and it cannot be solved by the authorities alone: addressing littering requires everyone’s commitment and the cooperation of different professions and agencies.

The project focuses on illegal dumping and addresses the automated localization and identification of abandoned litter, and the management of its proper disposal, in an integrated and flexible way. Previous approaches, including those based on citizen engagement, required either too much effort from the community or technical solutions not easily accessible to municipalities.

COBOL will deliver:

• Lightweight littering reporting that lets devices report littering transparently (for example, automatically detecting litter in a selfie), in addition to explicit reports by citizens.
• Engagement models that involve all the stakeholders in the disposal process.
• Federated learning across multiple communities, giving prediction and detection capabilities even to small rural communities that could not otherwise generate enough data.
• A model-driven, self-adaptive solution that adapts to different contexts and handles unexpected events to guide citizens and authorities in the disposal process.

Pilots with two municipalities will demonstrate the effectiveness of the solution.