IoT IoT-based garbage collection vehicle optimization path provision system using artificial intelligence

摘要

It collects traffic information from the ITS server through the ITS (Intelligent Transport System) network installed in connection with the closed network of the integrated urban operation center, and collects geographic information from the GIS (Geographic Information System) server, As a system for collecting video information from a VMS (Video Management System) server, the garbage collection vehicle optimization route providing system according to the present invention is a garbage collection vehicle through an external network connected to the closed network of the urban integrated operation center. A linked collection server that collects data on a garbage loading amount, a collection amount by unit area, and a time required for garbage collection by the garbage collection vehicle from from; A database server storing data collected by the cooperative collection server; And a garbage collection management server that derives optimal route data of the garbage collection vehicle using the data received by the linked collection server. The garbage collection management server. A first reinforcement learning model and a second reinforcement learning model constructed as independent learning models; A collection path providing means for providing an optimal path selected from among garbage collection optimal paths respectively calculated by the first reinforcement learning model and the second reinforcement learning model to garbage collection vehicles scheduled to proceed; And an estimated time calculated by evaluating the selected one of the selected first reinforcement learning model and the second reinforcement learning model based on the actual garbage collection time, and calculating the estimated time required for the remaining learning models. And a path evaluation means for selecting a learning model to provide an optimal path for collecting garbage at the next time by comparing the first reinforcement learning model and the second reinforcement learning model by evaluating on the basis of.