Linde is demonstrating the use of cameras and AI software in a semi-automated warehouse. Scanning cameras observe the warehouse in real time, while smart software provides the necessary control.
Warehouse management systems that control employees in real time have existed for decades. They ensure optimal workload distribution among employees. Order pickers receive their orders in their (handheld or machine-mounted) terminal, carry them out, and then receive a new order, while the WMS continuously optimizes the process. This technology will be further developed in the future. Linde Intern Transport recently demonstrated how this works at the Linde Summit at LogiMAT in Stuttgart, with a demonstration of an AI-driven warehouse.
The demonstration was made possible by integrating Linde warehouse software with Nvidia’s Omniverse platform. Warehouse trucks are equipped with cameras that continuously scan the environment and map the warehouse. (Pictured: a warehouse operator with a scanning camera.) This provides the current position of all employees and even all goods in the warehouse. A computer system can therefore address (potential) bottlenecks much more precisely. Consider, for example: a pallet being deposited in the wrong place, incorrect item counts during picking, a bulky pallet taking up more than one pallet space, or a truck stalled due to a dead battery. Irregularities are instantly identified and resolved by sending the correct instructions to employees or unmanned trucks. The Omnivers system uses simulation to rapidly calculate various options. Moreover, the system is self-learning (in popular terms: artificial intelligence), meaning the results are continuously evaluated and lead to further improvements.
Linde sees particular potential for this AI system in warehouses using Linde L-Matic Core. These are unmanned stackers for storing and retrieving pallets. The delivery and retrieval of pallets is usually performed in combination with manned pallet trucks. Such a system is susceptible to disruptions, for example, if a pallet is placed crookedly, is in the wrong position, or the inventory level is incorrect. The real-time camera system, with its powerful underlying calculation software, should provide a solution for this.
