Technology trends in distribution centers

Automation has long been present in distribution center around the world: automated storage and retrieval systems, automated guided vehicles, robotic arms, and automated conveyor sortation systems have been on the market since the late 1950’s. However, recent developments in artificial intelligence, electronics, and control software have allowed robots to be smarter, more flexible, and more useful for logistics. Unlike traditional automation this new wave of mobile robotics adapts faster to variable order profiles and they are able to operate in a broader spectrum of environments — Robots are coming out of the cage. This kind of flexibility is what is causing a high demand for robotics in order fulfillment centers as an strategy to satisfy the rising customer expectations in the e-commerce business.


The next generation of automated guided vehicles (AGV) can freely navigate through the facility without guidance mechanisms on the floor. Like Kiva systems, other automation providers (Swisslog, GreyOrange, Scallog, Greenzebach) offer similar smart goods-to-man technologies to support e-commerce demands. Adept Lynx is able to navigate safely alongside people extending its spectrum of applications. Furthermore, Lynx Conveyor, by Adept, and Open shuttle, by Knapp, combine AGV capabilities with motorized conveyor platform as an alternative to static conveyors that hardly adapt to rapid changes. Fraunhofer IML is developing a LivingLab Cellular Transport System where a swarm of AGVs operate with out the need of a centralized control providing a more flexible and reliable system.

Smart, free mobile, flexible AGVs are minimizing transportation times for inbound logistics. However, the dexterity necessary to identify an item in any possible orientation and extract it from the shelves is still a difficult task (see Amazon picking challenge). Researchers at the Karlsruhe Institute of Technology in Germany developed a smart picking cart, FiFi, that can follow the picker, receive commands by natural gestures, and travel autonomously through the warehouse. Locus offers a more independent system that can autonomously locate items, but still relies on humans for the extraction.

Smart, free mobile, flexible AGVs are minimizing transportation times for inbound logistics. However, the dexterity necessary to identify an item in any possible orientation and extract it from the shelves is still a difficult task (see Amazon picking challenge). Researchers at the Karlsruhe Institute of Technology in Germany developed a smart picking cart, FiFi, that can follow the picker, receive commands by natural gestures, and travel autonomously through the warehouse. Locus offers a more independent system that can autonomously locate items, but still relies on humans for the extraction.

Other industrial solutions like Fetch and IAM robotics are taking a step forward by being able to perform the complete picking activity independently from humans in unstructured environments. Fetch is a two-robot set that can freely navigate through the aisles of the warehouse, locate items, extract them from the shelve, and deliver the picking batch to shipping stations.

GridFlow, developed by Dr. Kevin Gue at Auburn University by then, is a high density storage system consisting of a set of unit-sized conveyor segments. Each segment communicates with its neighbors and operates without the need of a centralized unit overseeing the operation. Researchers at the University of Duisburg-Essen are developing an AS/RS that instead of a shuttle, has a lightweight platform that is moved by a system of parallel prestressed cables in front of the rack. This makes the device easy to set up in pre-existing storage systems, scalable, cheaper, and energy efficient.

Sophisticated control software is allowing robots to collaborate with humans at different levels. Baxter, by Rethink Robotics, is a collaborative robot that can operate safely alongside humans in the fulfillment centers in tasks that were usually performed only by humans: kitting, packaging, sorting, loading and unloading of small parts.

Outside logistics, there are some technological advances that are worth mentioning. Asimo, by Honda, is a humanoid robot designed to interact with humans. It can talk, follow the face of a person, recognize human gestures, look towards a sound and respond accordingly. Furthermore, Atlas, by Boston Dynamics, is leading the way in agility and dexterity. It focuses on sensors and sophisticated control to allow a robot to keep the equilibrium, stand up for itself after it has fall off, and perform the most delicate task with great agility.

 

Finally, there are three research trends that can revolutionize robotics: wearable robots, cognitive computing, and cloud robotics.

  • The MIT d’Arbeloff Lab has developed a wearable robot that provides a human with two extra robotic arms — the Supernumerary Robotic Limbs. The innovation resides in the fact that the arms can autonomously decide when, where, and how to assist the user. It uses sensors on the wrist of the user to determine if the human is on a task and decides the best way to help.
  • Cognitive computing equips robots with self-learning capabilities. Autonomous cars are programed to operate without human intervention, while a cognitive robot will learn how to drive by itself — it is autonomous knowledge acquisition. IBM (USA) and SoftBank Robotics (Japan) are working together to create the world’s first cognitive robot in the near future.
  • Robots have traditionally performed all processing and control within the standalone system. Researchers at Berkeley Laboratory for Automation Science and Engineering are working with Google to develop cloud robotic systems. Robots integrated with the cloud will have all the information of the internet available in real time, access to massive parallel computing, the ability to learn from other robots experiences, and most surprisingly, the ability to ask the human assistance in the face of unexpected difficulties.

Analyisis

Order fulfillment centers are already familiar with high levels of automation, automation suppliers are continuously providing innovative solutions, and research keep advancing the technology. Today’s leading technologies are being developed faster than what the logistics industry is adopting them. Despite the several options of flexible AGVs available in the market, manual order picking is more common in fulfillment centers. Similarly, there are other examples of underutilized automation technologies in packing, shipping, palletizing, truck loading/unloading, and storage systems. According to Fortna Inc “For some companies, the distribution center of the future is available today”.

However, detailed, variable, and unstandardized processes are still hard to automate. That is why, researchers and leading technology companies are working to make robots smarter and more agile. Cloud robotics and cognitive computing are expanding the ability to perform complex tasks and to coordinate better groups of robots. The ability of robots to self-learn tasks will give companies the flexibility to easily transfer robots between operations. In addition, collaborative learning will increase exponentially the efficiency of groups of robots.

The advances in the dexterity, agility, and mobility of robots are complementing are allowing robots to overtake humans in the most detailed procedures. Soon, robots are going to be able to handle unit loads of amorphous shapes faster and cheaper than humans. Wearable robots can be used by warehouse workers as a middle ground for those activities too complex to fully automate. The “super pickers” will be equipped with extra robotic arms that can be assisting to extract and carry orders or it can be processing independent orders by itself. Advances in tracking technologies and drones are leading inventory management to continuous models where the information is available in real time and with high accuracy. Finally, sophisticated machine learning are allowing robots to understand their environment. For example, sharper computer vision processing will help robots to read the label of a carton and troubleshooting: multiple labels, damaged labels, inconsistent addresses, or no label at all.

Robotics and automation aim to replace humans in the performance of physical activities. It has traditionally been seen as a strategy to avoid labor cost and/or square footage. But, in recent years, e-commerce has showed how automation can also add value in the race to speed up fulfillment where the competition is ruthless and customers are just one click away of your competitor. We cataloged it as a disruptive technology — big investment and entirely new operation model. The integration of automation and big data has the potential to create a self-driving distribution center — zero human participation.

Challenges

“The key to automation is the existence of standardized product ” – British steamship. Robotics is still challenging for unstandardized product and unstructured environments, where they have to be aware of their surroundings. However, advances in artificial intelligence and robotic control software are easing this difficulty. Moreover, it is difficult to asses quantitatively the economic benefits of automation. Savings in labor costs and leasing costs are explicit, but the potential competitive advantage in the market of having faster logistics is hard to foretell. In consequence, several automation technologies are currently underutilized.