A Proposal of Path Planning for Robots in Warehouses by Model Predictive Control without Using Global Paths. Issue 37 (2022)

• Record Type: Journal Article
• Authors: Ishihara, Shinji; Kanai, Masaki; Narikawa, Ryu; Ohtsuka, Toshiyuki
• Is Part Of: IFAC-PapersOnLine. Volume 55:Issue 37(2022)
• Page Start: 573
• Page End: 578
• Publication Date: 2022
• Subjects: Model predictive control -- Optimization -- AMRs -- AGVs -- Path planning; Automatic control -- Periodicals; 629.805
• DOI: 10.1016/j.ifacol.2022.11.244
• Languages: English
• ISSNs: 2405-8963

Abstract

Abstract: With the development of the e-commerce business, there are high expectations for the use of automated transport robots to improve the efficiency of warehouses. In this study, we propose a path planning method based on Model Predictive Control (MPC) that optimizes the performance of the entire warehouse for multiple robots. In general, when the MPC is used to execute path planning, global path planning such as Dijkstra algorithm are used together. This is because, if the MPC is used alone, the existence of a local optimal solution will cause a situation where the robot cannot move. However, in order to use the global route, it is necessary to prepare the precise map information of the warehouse in advance. In this study, we propose a path planning method that avoids deadlock without using global path by improving an objective function for the MPC. The effectiveness of the proposed method is verified in a numerical simulation.