Given the challenge that cigarette butts pose to the environment, specially at the beach where the tides can easily swallow any litter left by humans, two Dutch engineers from the company TechTics have developed an autonomous robot with sensors that can identify and remove cigarette butts and other debris on the beach. Martijn Lukaart and Edwin Bos created BeachBot, a prototype with an artificial intelligence (AI) system that aims to combat this environmental problem.
Cigarette butts are full of microplastics and toxic chemicals that don’t easily decompose and when they come into contact with water several harmful substances are released. According to the World Economic Forum, cigarette butts are the most common type of litter, with an estimated 4.5 trillion discarded annually.
To train the BeachBot (also called BB), Lukaart and Bos used about 200 images of cigarette butts that were processed by Microsoft’s Trove AI system. The mobile beach-cleaning machine can detect cigarette butts, pluck them out and dispose of them in a safe container. Bos and Lukaart are co-founders of TechTics, a consultancy based in The Hague that works to solve social problems with technology.
The prototype uses AI to learn how to best find the strewn filters, even if they are partially buried in the sand. BeachBot has completed a demonstration at Scheveningen beach, in The Hague.
When the water comes into contact with discarded cigarette butts, more than 30 toxic chemicals are released.
Lukaart and Bos, Co-founders of TechTics
Visitors to Scheveningen beach are familiar with a variety of trash strewn along the shore: plastic caps, glass bottles, candy wrappers and these cottony cigarette filters. And this expands to other beaches around the world.
The TechTics team created the first AI-based detection algorithm that specifically detects cigarette butts. They worked with students at Delft University of Technology in the Netherlands to create BB. To teach the bot, they showed the the AI system thousands of photos of cigarette butts, all scattered in various states, as partially hidden, so it could recognize and remember them.
To help accumulate those photos, they turned to Microsoft Trove, an app that connects AI developers with photographers through a data marketplace. In this case, people can submit their photos and TechTics pays contributors 25 cents per accepted image. The company is looking to collect 2,000 photos through this platform.
By 2050, the seas will contain more plastic than fish.
Lukaart and Bos, Co-founders of TechTics
“The system learns to see images like a child recognizing an object for the first time,” said Christian Liensberger, Trove’s program manager. The platform is based on the idea that people should be paid for their data, such as published photos, rather than simply giving it away on social networks or communication platforms, said Liensberger.
The goal is to have control and transparency within that process, allowing people to see how their data is used. Trove users can choose when to participate. The platform can collect all kinds of data and is currently helping to support a wide range of AI projects.
According to the World Economic Forum, the designers are now working on a new project: the MAPP detection robot. These robots are designed to work in outdoor spaces like parks and beaches, mapping and collecting litter data. They are able to communicate with one another to collect trash as a team.
