With the development of online shopping and the demand for automated packaging systems, we propose an Internet of Things (IoT)-based automated e-fulfillment packaging system and a 3-D adaptive particle swarm optimization (PSO)-based packing algorithm. The proposed system leverages the IoT to connect the data collection and conversion layer, the packaging management layer, the decision-making layer, and the application layer. A cyber network connects each robot, sensor, and smart machine to achieve high velocity, flexibility of procedures, and real-time information exchange. When customers order merchandise online, the orders are received and rearranged, and the deployment of items in a box is planned by the system. The proposed packing algorithm controls the arrangement of items. It compares the size and volume of items and boxes to choose a box of suitable size, as well as deciding on the optimal arrangement of items. This algorithm solves the difficult 3-D Multiple Bin Size Bin Packing Problem (3-DMBSBPP) by integrating an adaptive PSO-based configuration algorithm. Our simulation results show that the packing algorithm can deploy items appropriately, with all items packed inside their box without overlap and with an overall center-of-gravity close to the bottom center of the box. When all the items cannot be packed into a single box, the proposed dividing strategies split the items into groups to pack into two or more boxes of similar size. Furthermore, comparing with the real packages we assessed, the proposed algorithm has a competitive performance. Lastly, our robotic experiments show that the proposed packing algorithm can be implemented and executed by a robot and a manipulator. It also demonstrates the efficiency of this system, in which all devices communicate well with each other and the robots accomplish the packaging task successfully and cooperatively.
All Science Journal Classification (ASJC) codes
- Computer Science(all)
- Materials Science(all)