Combining Fill-Level Sensing With Route Optimization for a More Efficient Waste Collection

We provide a solution to the problem of collecting the waste of modern cities within the scheduled time spans, which has become increasingly challenging as the waste collection team is not supplemented accordingly in many places,e.g. by more staff or vehicles. Separating waste has become natural for urban populations, as recycling is ecologically necessary. However, the separation also leads to a continuously increasing number of containers which are logistically unmanageable without computer-aided collection scheduling. Regensburg in Germany has recently introduced a new program for the collection of biological waste, which extends the private collection of such waste by a few hundred public biological waste containers. Computer support was pretty basic so far, i.e. the collection times were recorded and the routes were sporadically rescheduled manually in an Excel sheet. We show that much shorter waste collection routes can be found automatically using a dynamic version of the classic Ant Colony Optimization (ACO) algorithm for shortening the collection routes. Moreover, we sketch an IoT (Internet-of-Things) approach to ''Intelligent Waste Containers'' which we equip with ultrasonic sensors monitoring the fill levels, which allows us to skip unused containers during the collection and reposition them on demand. By means of a computer simulation with online visualization, we illustrate that our optimized and dynamically adapted collection routes lead to significant time savings. Indeed, more than twice the number of currently used containers can be collected within the same time. We also report on the first experiments that we conducted with real sensors. The presented work is the result of a cooperation between four contributing parties: 1) Technical University of Regensburg (OTH); 2) Ara Institute of Canterbury, New Zealand; 3) Regensburg Governmental Department of Waste Management; and 4) an industry partner (kpit.com). We compare the outcome of our project with related work and we design some future perspectives and ideas for transforming Regensburg into an even more automatized and environmentally friendly ''smart city''