Predictive Analysis based Efficient Routing of Smart Garbage Bins for Effective Waste Management

Abstract

The crux of an effective waste management system is a well-organized routing algorithm that aids to collect the garbage from the bins efficiently and thereby eliminate the overflow of the garbage bins. To automate the garbage collection, the bins use a smart integrated sensing system. Three sensors (ultrasonic, load and gas sensors) are equipped in the proposed garbage bin to sense the level, weight and Carbon-di-oxide (CO2) concentration of the garbage in the bin. A dedicated webpage is assigned to monitor the collected garbage bin sensor data continuously. If the bin reaches the maximum waste level or weight or CO2 level (threshold) it will be marked as “Ready to be picked up”. Scheduling the route(s) and allocating pickup vehicle(s) to collect the filled bins is a significant concern for the efficiency of any waste management system. This scheduling problem boils down to a capacitated arc routing problem (CARP).Considering multiple trips for the available vehicles, capacity of the vehicle, capacity of the bins to be collected, etc., as factors a heuristic algorithm for an efficient routing is proposed in this paper. The primary aimof the proposed heuristic routing algorithm is to reduce the total usage cost of vehicles by minimizing total traversed distance. Predictive analysis algorithms will aid the pickup trucks to be used to the fullest capacity even though there are only few bins to collect so that better efficiency is achieved. In this paper, simple linear regression and multiple linear regression algorithms are applied to suggest the bins (that will be filled in near future) to be added to the route that will help achieve maximum usage of truck capacity. The waste management web application allows the admin to add these bins to the route and authorize the same for the drivers.