Adaptive Predictive Coordination of MDR Conveyor Systems for Real-Time Size-Based Sorting of Cylindrical Products

Motor-Driven Roller (MDR) conveyor systems are increasingly being used in modern warehouse and intralogistics systems due to their modularity, energy efficiency and the ability to implement the Zero-Pressure Accumulation (ZPA) principle. Despite their wide industrial application, most existing ZPA solutions rely mainly on reactive control based on the current occupancy status of the conveyor zones, without using prior information about the incoming loads. This limitation is especially significant when transporting cylindrical loads with different sizes and dynamic characteristics. In this work, a framework for adaptive predictive coordination of MDR conveyor networks (Adaptive Predictive Coordination for MDR Conveyors—APC-MDR) is proposed, designed for sorting and coordinating the transportation of cylindrical loads in real time. The proposed approach uses a photoelectric sensor and light beam interruption time measurement to estimate the diameter of each incoming load. The obtained information is used for both load classification and sorting-branch assignment, as well as for adaptive speed control based on the product size and the occupancy of the selected downstream branch. A mathematical model of the coordination mechanism has been developed, integrating load estimation, predictive sorting, adaptive speed coordination and preparation of downstream conveyor sections in a single PLC-based architecture. The system is implemented on an experimental MDR conveyor platform built with a Siemens SIMATIC S7-1212C PLC, ConveyLinx Ai2 modules and PROFINET communication. The conducted experimental studies show reliable classification and sorting of cylindrical loads, with a sorting accuracy of approximately 98.33%. The obtained results also show a reduction in the average transportation time and a limitation of the number of stop–start events compared to conventional reactive control strategies. The study confirms that using information about the geometric characteristics of the load as a predictive control variable can improve the coordination and efficiency of MDR conveyor systems while maintaining contactless ZPA operation.