Estimating travel and service times for automated route planning and service certification in municipal waste management

Nowadays, route planning algorithms are commonly used to generate detailed work schedules for solid waste collection vehicles. However, the reliability of such schedules relies heavily on the accuracy of a number of parameters, such as the actual service time at each collection location and the traversal times of the streets (which depend on the specific day of the week and the time of day). In this paper, we propose an automated classification and estimation algorithm that, based on Global Positioning System data collected by the fleet, estimates such parameters in a timely and accurate fashion. In particular, our approach is able to classify automatically events like stops due to traffic jams, stops at traffic lights and stops at collection sites. The system can also be used for automated fleet supervision and in order to notify on a web site whether certain services have been actually provided on a certain day, thus making waste management more accountable to citizens. An experimentation carried out in an Italian municipality shows the advantages of our approach