Dr. Matevž Bošnak, univ. dipl. inž., dr. Andrej Zdešar, univ. dipl. inž., oba Univerza v Ljubljani, Fakulteta za elektrotehniko; doc. dr. Rok Vrabič, univ. dipl. inž., Univerza v Ljubljani, Fakulteta za strojništvo; Dr. Viktor Zaletelj, univ. dipl. inž., Epilog, Tehnološki park Ljubljana; Izr. prof. dr. Gregor Klančar, univ. dipl. inž., Univerza v Ljubljani, Fakulteta za elektrotehniko

Abstract:
This article presents new findings in the field of theoretical and practical aspects of the logistics of autonomous vehicles in the industrial sector, which are being investigated as part of an applied research project. The aim of the project is to use intelligent approaches and limited computing power to find optimal paths, temporal movement plans and task scheduling for a fleet of automated guided vehicles. The paper presents approaches to coordinated path planning for vehicles moving between stations connected with a network of intersections and roads. The environment map is written in the form of a graph and is used to design and test algorithms in the simulator. The basic task that a mobile agent has to perform is to travel between any two stations, and the agent can be at any location before performing this task. The implemented solutions are based on a route planning algorithm with safe time intervals and priorities, where conflicts can be solved by waiting in nodes or on roads. The basic solution is extended with the concept of lifelong planning with priorities, where new tasks appear during the execution of the simulation. We introduce the concept of planning with a safe location, which allows the compilation of partial plans into the final task and ensures the completeness of the algorithm.

Keywords:
path planning, automated guided vehicles, optimal path, SIPP algorithm, intralogistics