Optimizing Distribution and Fulfillment Center Operations with Computer Vision and Digital Twins
A fulfillment center is a critical node in providing optimal customer service in a supply chain network and for e-commerce. Therefore, improving order fulfillment time is critical to world-class operations. A key process in order fulfillment is decanting and picking — activities that consume the most time when operating a fulfillment center in general, and fulfilling an order in particular. We'll analyze an actual system that stores products within an automated system and releases orders to a picking station. We'll demonstrate the use of two key platforms from NVIDIA — (1) Omniverse, to create digital twin 3D assets and an architecture enabling variations in simulation models to address different scenarios and strategies aimed at improving system performance, and (2) Metropolis, to enable highly scalable intelligent video analytics applications to provide high-quality perception data and operational situation awareness. We'll discuss specific key performance indicators to compare different strategies and scenarios, such as order fulfillment lead-time, picks per man-hour, average picking time per order, and average time to pack an order. These proposed solutions will provide insights that lead to improvements in order processing time, order fulfillment rate, and increased operator efficiencies.