The value of information in automated sorting: leveraging look-ahead in multi-conveyor systems with articulated robotics for pick and place

Production lines utilize robotic arm conveyor systems for sorting and packing items. The continuously moving conveyor belt requires fast decisions based on the current system state. The interplay between information availability, system layout configuration, and performance has not been explored, and is thus studied for a multi-conveyor system with strong sequence dependency. Introducing a probabilistic control scheme utilizing varying degrees of information about future system states, and a continuous time problem formulation, this study evaluates the performance of the real-time scheduling problem solution under different levels of information. Results from extensive testing show that increased future state information asymptotically reduces give-away and improves performance for the studied type of system. The findings highlight a relationship between information access, arm coverage configurations, and system performance. This research provides insights into systems of similar type and configuration, while also offering a framework for estimating performance under different constraints and objectives.