POLYPROBLEM report: The Circularity Code

Technology that optimizes urban waste management SMART CLEAN-UP All too often, plastic waste still ends up where it doesn’t belong and remains lost forever. In Berlin, for example, 66% of residual waste is disposed of incorrectly.22 Separating trash by type pays off in two ways since many communities charge for residual waste by weight. Whether artificial intelligence, robotics, the Internet of Things (IoT), or blockchain technology, all of these high-tech approaches have a role to play when it comes to digitalizing urban waste management and making waste routes more efficient through better control, thereby saving time, costs, and CO₂ and creating more transparency by involving the public. From New York to Amsterdam, Stockholm23 and Songdo, South Korea, many cities have already set out to become smart waste management cities.24 At the beginning of 2020, Villach, a city of 65,000 inhabitants in Carinthia (Austria) launched a pilot project as part of a public-private partnership model together with Saubermacher Dienstleistung AG to optimize municipal waste disposal with the help of artificial intelligence.25 Sensors were installed in 1,100 glass recycling bins that regularly measure the fill levels so that the containers are only emptied when they are actually full. In Berlin26, Berliner Stadtreinigung (BSR) has also tested wireless sensors on underground containers since 2020, which collect fill level data 24/727. The data is then transferred to cloud servers, where it is analyzed using artificial intelligence and visualized on a dashboard in order to compute optimal routes which are then transferred to the navigation systems of the waste collection vehicles. The Berlin waste management company expects the system to increase efficiency by up to 30 percent in terms of punctual emptying and to reduce CO₂ emissions.28 Wirtschaftsbetriebe Duisburg published a case study in 2021 to evaluate the benefits of fill level sensors. They placed up to five types of sensors in 13 different garbage cans to evaluate the opportunities and limitations of using sensors. Some of the tests run as part of the study revealed large deviations between sensor readings and the container’s actual fill level, which the study attributed to outliers, such as cases where waste was placed directly underneath the sensor. The study also shows that present intervals for emptying residential garbage cans and picking up different types of garbage such as residual waste, lightweight packaging, and paper, cardboard and carton are sufficient. Rarely is garbage picked up too early or too late. As a result, the analysis shows greater potential for the use of sensors in public garbage containers. This really goes without saying since it is more difficult to predict how quickly a garbage can in a public space will fill up than residential garbage bins.29 In Bangladesh, for example, the country’s second-largest mobile network operator, Robi Axiata, has provided the city of Dhaka with five hundred sensors to monitor the fill levels of public garbage cans and inform the relevant authorities when they need to be emptied.30 Back to Villach in Austria. In order to analyze the material composition of waste with the help of AI and to determine residential separation rates, the city has tested so-called recyclables scanners (infrared cameras and sensors) in garbage trucks – as is being done in 22 Ott (2023), p. 5 23 Shahrokni et al. (2014) 24 Joshi (2022) 25 Goldschald (2021) 26 Hinweis Duisburg 27 BSR (2020); BSR (2021) 28 Ott (2023), S. 8 29 see Hoffmann et al. (2021) 32

RkJQdWJsaXNoZXIy ODI5MzU=