Warehouse automation solution through a goods-to-person AMR

The use of goods-to-person robotic systems in a warehouse enhances storage density by devoting specific areas for Automated Mobile Robots (AMRs). Nevertheless, operation in narrow aisles and confined spaces presents notable challenges. This study outlines a comprehensive approach for AMRs to autonomously navigate within narrow aisles of a warehouse, ensuring safety and streamlining order completion. The goal is to independently collect and transport bins with products referred to the order, to the workstation area, where workers finalize order packing. Precise positioning of fiducial markers on the floor, initiates a fusion-based localization system that merges information from wheel odometry and fiducial detection. The markers are organized as an undirected graph to facilitate navigation purposes and Dijkstra algorithm is employed to successfully generate both safe and optimal paths. A pipeline has been established for the retrieval and positioning of bins from shelves or workstations to optimize the order picking process. The proposed method has been tested on a simulation environment with realistic constraints stem from the development of an autonomous warehouse environment.