The result of this evaluation can then be sent to a database via Wi-fi, which makes the data available for further processing. In the IoTrash application, this data is presen- ted as a heat map and thus made available to the user. Users can be institutions such as the building management office, city planners or the city utilities. This enables them to identify heavily littered areas and handle them in the short and long term.

The issue and our motivation

The pollution of the environment in general is a major problem for the future. Not only in the seas, but already in the cities we can observe the pollution of our environment. Many fellow citizens do not seem to be affected by this. This can be seen on the one hand in the permanently overfilled waste bins and on the other hand in the waste which lies freely in nature.

The trend to get food-to-go on the way also contributes drastically to these observations, because it is not uncommon for bulky packaging to overfill the trash cans and thus contribute to their overloading especially in larger cities. This is also confirmed by employees of the responsible authorities:

Not only because of general problems such as climate change, but also to make the city more attractive, many initiatives have already been taken to fight this. Educational events and posters, city clean-up cam- paigns for local citizens and schools, but also digital innovations such as the sensory detection of the filling level of waste bins aim to help here. These ideas help to encourage residents to think more sustainable. The waste problem is nevertheless improving only slowly. In large cities in particular, it is becoming increasingly difficult to keep track and find explanations for littering.

IoTrash User Interface

Through IoTrash's user interface, long-term planning and control of waste reduction is enabled. Other benefits, such as identifying causes of widespread waste in cities is possible, and it also serves as a support for city and event planning of short- and long-term measures. A problem that has been difficult to represent statistically becomes tangible through numbers. A comparison of waste amounts across countries and cities is possible.

Trash Tracking

To check whether our concept is realistic, we built a simple prototype ourselves, which is equipped with various functionalities. One of the most important factors, the recogni- tion of the type of waste, was tested using a camera. It was connected to the computer by cable. The images from the camera were recognized with the machine learning program „MobileNet“. In order to assess how accurately the location of a waste object can be recorded, we also experimented with a GPS module connected to the microcontroller, which can be read out on a computer.

The command for object recognition should only be given when an object is picked up. Therefore we attached a button to the handle of the pliers, which should be the trigger for the machine learning program in the next step. This prototype was in our case a simple and tangible representation of our concept. For a realistic use, all functionalities must be fully developed and the machine learning software must be provided with as many littering images as possible.