We have smart homes, smart highways, and even smart toothbrushes. Isn’t it about time we had smart warehouses too?
When the term “smart” is applied to various products, it usually has to do with internet connectivity. So, for example, a “smart” piece of field equipment would be able to send data on its condition and performance, allowing the manufacturer to remotely monitor maintenance needs and perhaps offer suggestions for how equipment users can improve performance.
While the smart warehouse also leverages the internet, it includes a lot more than just internet connectivity and analytics. It involves warehouse systems that are smarter, based on new levels of visibility and awareness, advanced optimization technologies, and increased system-based decision-making. It also leverages a number of supporting technologies, from the internet of things (IoT) to simulation and machine learning.
A framework for this smart warehouse is shown in Figure 1. It shows that although a warehouse management system (WMS) is necessary for achieving a smart warehouse status, it is not sufficient. Previously, thousands of companies depended on traditional warehouse management systems to drive high levels of efficiency. But there has, arguably, been only incremental progress in WMS functionality over the last 20 years. During that same time period, companies have needed to meet new throughput expectations, push back against rising costs, and enable shortened cycle times. These general business shifts are driving a new paradigm in warehouse operations and technology.
The components of a smart, automated warehouse
Enlarge this image
The smart, automated warehouse will be built on a number of capabilities and components beyond what can be achieved by a WMS alone. It will rely on technologies that can be flexibly deployed and combined to meet specific requirements. Critically, many of these new capabilities will be delivered by newer and complementary warehouse execution system (WES) software, which is related to but different from WMS. Below, we describe the key capability groupings and enabling technologies as shown in the smart warehouse graphic.
Core warehouse operations
The smart warehouse is built on top of core operations excellence, which will be delivered, in part, from an advanced warehouse management system. That operations excellence will also rely on pervasive use of mobile terminals and barcode scanning, system-directed activity, advanced task management, support for multiple picking and replenishment strategies, dynamic slotting, detailed labor reporting, and more.
In addition to guiding core warehouse operations, the smart warehouse is always “listening” to the environment. This awareness is generally provided by a WES and happens in a way that is fundamentally different than how a traditional WMS sees the world. A WMS is generally reactive in nature, processing the work as it sequentially arrives (physically or logically) at each next step in the fulfillment process. In comparison, a WES is “always on”—aware in real time of activity and constraints that can impact decision-making.
That awareness includes granular, real-time visibility of throughput and any bottlenecks set at user-definable levels. For example, the user could choose to have visibility of the case picking area as a whole or visibility of each level of a multilevel case pick module. The smart warehouse would know what is expected in terms of throughput in each area and will send alerts if throughput falls below expectations.
But there is a lot more here: That real-time visibility can be turned into powerful dashboards that give managers and supervisors a detailed look at where things stand across the distribution center (DC)—and what they should do next.
Here’s the cool part: The WES draws upon the same data being used to power the dashboards to make decisions about the flow of goods and work. For example, if a put wall area (an increasingly popular order-picking technology) is becoming congested, the smart warehouse will either slow down upstream pick activity or, for a period of time, send picks to an alternative path, such as to a manual cart pick, until the congestion dissipates. And it does this on its own.
Now that’s very smart.
This granular visibility of activity—current and planned—can then be used by simulation technology to provide the foundation for the intelligent and dynamic allocation of labor and resources, as discussed in the “Enabling technologies” section of this article.
Advanced software-based decision-making
Here is the reality: Even with advanced warehouse management systems, most warehouse operations are highly dependent on human decision-making about what work to release when, when to change order and replenishment priorities, and more. At the center of the smart warehouse is the ability of the WES to release orders and other work autonomously, without the need for human intervention, making the process more efficient. This automated release of work is based on a variety of attributes, including order priority, inventory and resource availability, optimization opportunities, carrier cut off times, and more.
The WES will also be able to reprioritize tasks as conditions in the DC change. While it’s true that warehouse management systems have had prioritization capabilities for many years, new smart warehouse capabilities will take prioritization to new levels.
Let’s take basic cart picking as an example. In a smart warehouse, when a picker scans the cart identification, the system will dynamically assign picks to that cart, based on the cart configuration and the goal of minimizing total travel time. But what if a very “hot,” urgent order comes in? In the smart warehouse, the system will scan the environment to see if any cart pickers have orders assigned to their carts that could be replaced with the hot, priority order—typically an order that hasn’t started any picks. But it will do so in a smart way, only assigning the new order if the pick locations are in front of the picker, so they do not have to reverse direction after having already completed picks along their path.
Instrumentation and user interface
The smart warehouse will increasingly automate the tracking and measurement/monitoring of inventory, equipment, and people by using technologies such as RFID, IoT, and real-time locating systems (RTLS). For example, in many cases, the smart warehouse will support RFID as an alternative to barcode scanning. RFID can eliminate many barcode scanning activities and automatically identify and prevent errors, such as “mispicks.”
Tracking technologies such as RFID can, in turn, help empower new types of smart capabilities. For example, IoT can be used to trace a lift truck driver’s actual movements and share that information with analytic applications to identify if workers are taking the most efficient travel paths to complete their work. IoT can also be leveraged to enforce social distancing or to identify “dwell times” when product isn’t flowing as it should.
As warehouse technology becomes smarter, the user interface for that technology will become more intuitive. The smart warehouse will increasingly leverage voice technology not only to improve picking and other distribution processes but also to change how workers (especially managers and supervisors) interact with warehouse software. It will enable managers to ask questions or request data via voice, and trigger a dynamic system response, moving to a form of person-to-system dialogue. Analysts call this “conversational voice,” in contrast to the “transactional voice” that has been in place for decades for order picking and other tasks.
Already today, there are applications in which workers use voice to request information from a WMS. Examples include calling on a mobile device for an updated status on the current picking wave or requesting replenishment status for an empty location awaiting a pick.
Material handling system optimization
As noted above, there are a significant number of both traditional material handling systems (such as sortation and pick-to-light) and new generation material handling systems (such as put walls, mobile robots, and goods-to-person) available to distribution managers today. That includes technologies, such as mobile robots and put walls, that are relatively inexpensive and highly scalable, meaning companies can start small and add to them over time based on success.
Regardless of the type of automation, smart warehouse software will seamlessly integrate with and optimize the performance of these systems, both individually and as a whole. It will provide a single platform for integrating with DC automation and orchestrating the flow of goods across heterogenous materials handling systems. This integration layer can be thought of as an operating system for managing the integration and performance of any number of automation technologies. For example, this single platform could be used to direct different mobile robot types from different vendors.
This integration layer would also directly connect with systems such as voice, smart carts, pick-to-light, put walls, and mobile robots without the need for any other software. Utilizing a single platform has many advantages, including lower total costs and the ability to optimize the performance of these systems within the full context of WMS/WES information. As a result, the integration layer would eliminate the process and information siloes that occur when the WMS “throws the orders over the wall” to the picking subsystems.
This “plug and play” capability will not only ease initial integration efforts but also enable the automation systems to be included in the larger orchestration of workflows. Both automated and nonautomated processing areas could be considered as a holistic ecosystem, optimizing the flow of work and total throughput. This is very different than how warehouse software has worked in the past with automation—and it is very smart.
To achieve these capabilities, the smart warehouse will be built on the foundation of several enabling technologies. These include:
A dynamic rules engine: The smart warehouse will use a rules engine to define and dynamically execute conditional rules relative to process and flow. These rules will consider capacities and constraints and be easily adaptable over time without custom coding.
In-line analytics: The smart warehouse will be instrumented with a rich array of dashboards and analytics that are increasing “in-line”—or embedded into the warehouse technology and directly relevant to the job being done by the user. These dashboard analytics will support real-time decision-making.
Simulation: The smart warehouse will leverage simulation tools to improve resource planning, “what if” scenario analysis, system testing, and more. The WMS software, for example, could forecast expected order volumes and profiles based on history and other factors, then simulate how the default labor and resource plan for the day/shift matches up. The result would be a dynamic, time-phased plan that identifies where workers will be needed in what quantities for, say, every hour of a shift.
Artificial intelligence/machine learning: Naturally, artificial intelligence (AI) and machine learning will play a growing role over time in the smart warehouse. For example, companies may use artificial intelligence/machine learning together with simulation software to continuously improve labor and resource plans. Simulation software may create a work plan based on estimates of processing times, carrier schedules, and more. The timing of this automated order release will be continually improved based on machine learning.
Taken together, these new capabilities of the smart, automated warehouse will usher in a step change in warehouse technology capabilities.
Smart warehouse benefits
The smart warehouse will deliver a wide array of benefits to shippers. These include:
This is not small stuff. This is seismic change for warehouse operations and enabling software, representing a new era of nearly autonomous warehouse software. It will deliver competitive advantage to companies that embrace the vision before their competitors.
The smart, automated warehouse isn’t just some academic vision. While the smart warehouse paradigm should be thought of as a journey not a destination—both in terms of the overall market and at individual distribution centers—most of the capabilities described here are available today, some more complete, others more developing. But there is a lot more to come, especially through enhanced use of AI and machine learning.
With the growing availability of less expensive and more scalable technology, it seems clear that a much higher percentage of companies will embrace material handling systems than is the case today. But many of the capabilities described in this article can drive value for nonautomated or lightly automated operations as well. Whatever level of automation they adopt, it’s time for companies of all sorts to start envisioning a much smarter, automated future for distribution operations.