How warehouse automation robotics transformed the supply chain

New kinds of robots

Enterprises have traditionally inquired about better warehouse management software. However, robots have become the No. 1 call topic since the pandemic, according to Dwight Klappich, vice president analyst at Gartner, who has 42 years' of industry experience with warehouse technology.

He is also seeing a significant shift in the reasons for adopting automation technology. Managers have historically focused on using automation to reduce headcount. Now, enterprises are struggling to keep enough people on their teams, while the cost of labor is going up simultaneously. "Even with the higher cost, they are still not able to find enough people, and that is why they are looking at robotics," Klappich said.

This interest is driven by a new breed of warehouse robotics designed to augment operations and collaborate with humans. Industrial robots have been around for decades but were often limited in their capabilities and cordoned off to protect people and equipment. Emerging warehouse robots use AI to navigate a facility, pick items and orchestrate collaboration with people.

Types of enterprise warehouse robotics systems include:

• collaborative picker robots that follow humans to collect and carry items for an order;
• goods-to-person robots that bring shelves closer to human pickers;
• heavy payload robots that are essentially autonomous forklifts;
• static robotic picking systems that pack orders or boxes;
• engineered automation that reimagines warehouse design to support high-efficiency robots;
• autonomous drones that fly around and constantly inventory tall shelves; and
• facilities management robots that mop floors and cut lawns.

Warehouse automation trends

Robotic warehouse automation is currently a small portion of overall warehouse automation spending, but it is growing rapidly. Advisory firm Interact Analysis predicts the mobile robotics market could grow from approximately $3.6 billion in 2021 to $18 billion in 2025. The number of warehouse facilities with mobile robots is also projected to expand from 9,000 in 2020 to 53,000 in 2025. In addition, Interact Analysis forecasts a total of 2.1 million robots in service by the end of 2025 with 860,000 of these shipped during that year alone.

Interact Analysis includes mobile robots as a component of the broader warehouse automation market, which it expects to grow from $49.6 billion in 2020 to $69 billion in 2025. Fixed automation systems such as automated storage and retrieval systems (AS/RS), conveyor belts and sorters will eclipse other forms of automation for the foreseeable future.

AS/RS is a more capital-intensive approach to optimizing a facility for space and efficiency. In some AS/RS examples, mobile robots might travel on dedicated tracks to retrieve goods from totes stored across multiple floors. This can be more efficient but costly to set up. Instead, a simpler robot could be used to operate in a specific, tightly controlled environment.

Warehouse robotics in the supply chain

Florian Pestoni, CEO of cloud-based robot management platform InOrbit, is seeing the greatest adoption in autonomous mobile robots (AMRs) that either follow a person as they collect items or move goods from one location to another.

Gabriel Aguiar Noury, robotics product manager at Canonical, the publisher of Ubuntu, observed robots are getting better at moving and manipulating goods -- two fundamental activities in a warehouse. Some of the simpler robots are essentially pallets on wheels. In more automated environments, these goods can be automatically moved within and between baskets.

Robots are also getting better at manipulating goods at scale. Health and safety regulations only allow a human to move a maximum of 50 pounds, and people become tired at the end of a long shift. "Robots can easily lift more [weight] and [do it] faster, keeping the flow of goods moving throughout the warehouse," Noury said. "Truck unloading is one of the most challenging jobs in a warehouse, and the use of robots can also limit injuries and health hazards."

Supply chain automation

Pestoni believes the top benefit of warehouse robotics for the supply chain lies in automating the manual movement of goods, especially when little critical thinking is required. He expects warehouses and retailers to start exploring ways to automate other aspects of the supply chain, such as autonomous trucking for long-haul trips. Companies are also exploring ways to automate the delivery leg of the supply chain. For example, instant logistics firm Zipline is developing drones that bring medical supplies to remote hospitals. And small AMRs and autonomous vehicles are starting to automate the trip from a fulfillment center to a customer's home.

Warehouse automation robotics could also improve supply chain visibility by automating inventory processes. Historically, most warehouses were not designed to be easily navigable, but inventory robots use computer vision, deep learning and standardized product identification to tackle this challenge. "Companies like Ware and Gather AI use computer vision and deep learning to scan warehouse shelves with speed and accuracy that human employees cannot rival," Nuce said.

Historically, most warehouses were not designed to be easily navigable, but inventory robots use computer vision, deep learning and standardized product identification to tackle this challenge.

Accurate stock and location data can make the difference between winning and losing a customer. The broader adoption of new inventory data standards could facilitate data exchange across trading partners. "The more information all trading partners can [have] about the products moving through the supply chains, the better they can anticipate demand, avoid disruptions and deliver the right products to the right place at the right time," Nuce explained.

The growth of warehouse robotics could also make it easier to deploy AI software for order slotting, said Zach Gomez, senior director of global logistics at Realtime Robotics. Slotting refers to the location where items are stored. In the traditional approach, similar products are stored near each other but not always shipped together.

"Recently, the software has used AI to analyze past order data and store items closer to items that are traditionally shipped together to reduce travel time and increase efficiency," Gomez said. For example, a grocery store fulfillment center might stage potato chips next to soft drinks as they are commonly sold together.

Navigating the warehouse

Warehouse robots are improving thanks to innovations in sensors, machine vision, AI, computing power and new hardware. These technologies have paved the way for increased awareness, enhanced problem-solving abilities and physics simulations that help robots learn faster. "With these innovations come the modern-day tools we're seeing in warehouses every day such as depalletizing robots, autonomous mobile robots and AS/RS," said Thomas Evans, CTO at Honeywell Robotics.

Older warehouse robots, called automated guided vehicles (AGV), followed fixed routes through a warehouse or facility by sensing wires buried in the ground or looking for unique markers. But these systems had limitations; they would stop and wait for any obstacle that blocked their designated paths to be cleared. AMRs, on the other hand, combine advanced sensors, AI coprocessors and machine vision algorithms to bring more flexibility to warehouse robotics.

With the addition of AI and continuous learning, AMRs can react in real time and adopt workflows to efficiently navigate diverse levels of demand and operational changes. "As such, there has been an explosion of AMR adoption in warehouses to tackle the challenges of ecommerce and scarce labor," said Matthew Cherewka, director of business development and strategy at Vecna Robotics.

AI can also improve the guidance systems for autonomous inventory drones. Skydio, which makes autonomous drones, has been working with Ware AI to help organizations incorporate autonomous inventory capabilities into their workflows. Skydio CEO Adam Bry said his team uses advanced computer vision and AI techniques to build real-time 3D maps of complex environments, plan collision-free paths around obstacles and automate complex tasks.

Planning for an autonomous supply chain

Some enterprises are even finding ways to teach the goods to move themselves. For example, Seoul Robotics worked with a vehicle manufacturer in Europe on the infrastructure to drive completed cars into a parking facility. The system uses Lidar sensors and centralized AI navigation systems embedded in the factory to control vehicles that may lack advanced sensors or self-driving capabilities.

This technology reduces the cost per vehicle and helped teams redesign the factory to be more efficient. "Navigating vehicles around a manufacturing facility is costly, challenging and prone to human error," said Jerone Floor, vice president of products and solutions at Seoul Robotics. AI-enabled warehouse orchestration engines can also improve coordination between robots and humans. Experts like Vecna Robotics' Cherewka believe this has become increasingly important when facing new global supply chain challenges and growing consumer demands.

Down the road, improvements in warehouse automation robotics will improve the creation and management of autonomous retail systems, such as automated replenishment and checkout-free experiences. "Warehouse automation will not only respond to current in-store conditions but also start to anticipate needs such as approaching expiration dates," Nuce said.