Manufacturing systems, such as the flexible manufacturing system (FMS), have recently changed to accommodate the production of various types and volumes of products. In the FMS, however, it is difficult to change the layout of the factory and solve problems that arise. Therefore, the importance of automated guided vehicles (AGVs) is increasing because they can flexibly respond to changes in facilities and factory layouts. However, there have not been any studies that take into account the many indefinite and accidental elements regarding AGV systems. Applications of knowledge in one domain to a different domain have been drawing much attention. Such activity is called a mimetic solution. We investigated applying the knowledge of traffic engineering regarding passenger transport to conveyance of AGVs. We propose an autonomous conveyance system for AGVs based on taxi transportation strategies to solve indefinite and accidental problems. The system focuses on applying traffic engineering knowledge regarding a flexible taxi system. A taxi is a transport unit in a traffic system involving high flexibility in traveling routes and arrival/departure points.
We also applied the waiting mode of taxis at stations where AGVs pick up and drop off products (P/Ds) as AGV rules to our system and investigated the system’s effectiveness and adaptability to schedule changes in the factories. To adopt a waiting mode as AGV rules, we determine the arrival/departure points that AGVs wait and changing the number of AGVs and product intervals. In addition, collision between AGVs must be considered. If a collision occurs, we have to change the factory schedule. Therefore, we took into consideration AGV collisions. We estimated the matching time, conveyance efficiency, and number of approaches as assessment functions. The matching time is the period between when a load is generated and received, conveyance efficiency is the ratio of total distance to distance traveled while empty, and approaching AGVs denotes the risk of collision. Specifically, we discuss the effect of the number of AGVs on these parameters by considering the ratio of the number of AGVs to the number of P/Ds. We demonstrated that our system is effective in terms of high conveyance efficiency and reducing the number of approaches without decreasing matching time and with a suitable number of AGVs.