April 19, 2017
by Vinod Baya
At the same time that robots’ capabilities are expanding, they are becoming easier to use and more affordable.
Robots have been long seen as the stuff of a tech-centric faraway future thanks to their portrayal in pop culture, and they have been considered exclusive to the largest and most technologically advanced enterprises. But robots have long been making a huge impact in manufacturing-intensive businesses such as automotive, electronics, and defense. However, they remain largely ignored by a mainstream business world that is hard-pressed to see how they could be beneficial.
That is finally changing. At the same time that robots’ capabilities are expanding, they are becoming easier to use and more affordable. Robotics technology seems to be on the threshold of going mainstream.
The global robotics market is already growing at a healthy clip. Unit sales for 2016 were expected to reach 6.69 million units, and for the next four years, a cumulative average growth rate of 26.5 percent, according to research from Technavio. It also forecasts that by the end of 2019, nearly 14 million robotic units will be sold every year. This uptick is echoed in PwC’s Digital IQ research: Today, 15 percent of business leaders say they are making significant investments in robotics technology; in three years, nearly one-third (31 percent) will be.
As their capabilities improve, robots are slowly but surely finding their way into every corner of every industry. In the future, people can expect robots to help out in the operating room, for example, because they can mimic human motions—but with a far greater level of precision and without getting tired. They might plant and harvest crops, able to pick the most delicate fruit without damaging it. Even your next T-shirt might be sewn by a robot rather than a human. Robots might also come in soft versions that emulate human muscles and can take on delicate tasks, handle fragile objects, and change their shape to work in different environments.
While capabilities are on the rise, costs are dropping. An industrial robot has a price tag of several hundred thousand dollars, but newer, more modular, flexible, and collaborative robots are starting to hit prices under $30,000—about what you’d pay for a company car. These robots are increasingly doing more than working the assembly line. They are assisting workers in business operations and interacting directly with customers.
How are robots becoming part of the mainstream? Here are five big trends that will help bring them into just about every aspect of modern business.
1. From a structured to a dynamic environment
Think about the traditional domains of robots, such as automobile manufacturing plants, highly computerized warehouses, and steel mills—all environments where robots work well because their surroundings are tightly controlled, highly structured, and pre-engineered with great precision. A robot needn’t make a decision about which part goes in which slot on a car coming down the assembly line because it places the same part into the same slot every time.
Robots are now starting to move beyond such structured environments and into more dynamic and less predictable ones, where people, objects, and even other robots work together to create a more fluid atmosphere. Where will you find robots in the coming years? You’ll see them in hospitals, where devices like the Aethon TUG robot roams the halls and delivers medications without fear of drugs being spilled or stolen. Savioke’s Relay is built for the hospitality market and delivers towels, toothbrushes, and minibar items directly to customers in their rooms.
The Blue River Technology See & Spray machine can perform weed treatment on commercial farms by detecting which plant is a weed and which isn’t. Why use robots in on the farm? “Every industry, including agriculture, has some jobs that are non-differentiating and onerous. And in some cases, such tasks [like removing weeds] go undone because either no one wants to do them or the business cannot justify the cost of doing them,” says Blue River vice president of business development Ben Chostner, “Skipping these tasks limits the performance the business is able to achieve.”
2. From working in isolation to working alongside humans
Highly automated factory floors are so isolated that lights-out manufacturing—where operations literally can be performed in the dark—became popular in the early 2000s. Humans simply aren’t useful or necessary in certain environments, not just for productivity reasons but also due to safety concerns. To keep human workers safe, robots work in isolation behind metal cages, and any interaction with humans—for maintenance, service, or material transfer—could pose a disruption and reduce overall productivity.
Tomorrow’s robots are getting downright cozy with humanity, to the point where a new term has been coined to describe them: Cobots, short for collaborative robots, are smaller, safer, simpler, and friendlier versions of their industrial counterparts. They are plug-and-play by design and built to be affordable even for small businesses. Some analysts anticipate cobot pricing will hit $15,000 to $20,000 per unit by 2020.
Lighter weight than typical robots, cobots have built-in sensors and technology to keep safe the people who work with and around them. Many of these cobots also can monitor their speed and proximity to humans and can modulate the power or force exerted in case of contact. For instance, variable impedance actuators help some cobots stiffen (to transfer energy to carry out a task, for example) or soften (to absorb force when coming in contact with a human or object).
What might cobots do for businesses? Cobots made by Rethink Robotics can perform high-end product testing (such as circuit board verification), complex assembly, or other machine operations. Cobots from Universal Robots can paint, weld, drive screws—you name it. In general, the goal is to offload the repetitive, boring parts of a production process while letting humans focus on adding value. But rather than sequestering robots in a dark factory of their own, people are designing them to be safe and flexible enough to work alongside human operators.
3. From programmed to learned behavior
Historically, robots have been prized for their ability to follow instructions precisely. These instructions were typically delivered via explicit programming of the robots’ actions. This approach was acceptable when their environment was constant. In dynamic environments, however, it is virtually impossible to anticipate every situation and to program every necessary robot action.
Today, robots learn from human trainers or through practice to complement their programmed instructions. This learning is key, because many actions a person might want a robot to do simply can’t be anticipated during programming. For instance, there is no way to program how a robot will grip the wide range of objects in a warehouse or a home. But that behavior can be learned.
“Our approach is to leverage the state of the art in computer vision and to combine it with smart gripping hardware and big data. These technologies, when integrated tightly by our team, enable robots to grasp a wide range of objects reliably,” explains Yaro Tenzer, founder of RightHand Robotics. For example, researchers at University of California, Berkeley, are working on algorithms that allow robots to learn new skills through trial and error, mimicking the way a human would learn a task.
“Mostly it’s collecting images and building up the training database,” says Ben Chostner of Blue River Technology, describing how his company trains robots used in agriculture to pick a different type of produce. “The robot uses the training database to learn the behaviors applicable to a new type of plant.”
The concept of autonomy—think self-driving cars and automated drones—is heavily wrapped up in learning new behaviors. In fact, learning is starting to take center stage in robotics. Companies like Brain Corp. and Neurala are building learning capability into the operating systems of robots, so the companies can deliver devices that are learning-ready. While programming and machine learning will continue to be important, over time the balance will shift away from programming to learning. A focus on learning will make it possible for robots to take on a much wider range of tasks than ever.
4. From custom-built products to versatile platforms
Previously, if anyone wanted to build a robot, they needed to build all the functionality—in both hardware and software—from the ground up. As a result, most robots today are proprietary solutions, custom-built to work in specific settings.
Robotics technology is advancing through modular platforms that include hardware, operating systems, packaged libraries, and cloud-resident services. For example, Toyota’s Human Support Robot and SoftBank’s Pepper are the chosen standard platforms for the RoboCup@Home competition. Both support the developer community to build new capabilities on their platforms. The Open Source Robotics Foundation supports the Linux-based Robot Operating System widely used in academia, government labs, enterprises, and startups. There are also open source libraries for motion planning, perception, simulation, and other capabilities.
Robots of the future likely will be designed with modularity in mind, starting with a platform upon which a customized solution can be built. The beginning of this trend is already evident in offerings from companies like Modbot, which has developed a flexible set of components that can be constructed à la Lego bricks to create the robot you need—and reconfigured later when those needs change. BlackDog Robotics offers a modular, mobile base that can be variably configured for hazardous materials abatement, fire and rescue, or law enforcement use.
Such platforms lower barriers to entry for businesses that otherwise probably wouldn’t consider getting started with robotics. But these platforms also open the door for an ecosystem of innovation in which other companies can create modules that work with established robotics platforms.
5. From the back office to the front office
Perhaps the most profound shift will be robots moving from the factory floor, where they have been hidden from human view, to many other points along the enterprise value chain—including the front of the business, where they will interact directly with customers and employees.
In doing so, robots will go beyond simply making products to moving, manipulating objects, and serving customers (see the graphic below). Opportunities for automation will emerge across enterprise operations, increasing the potential to innovate through the use of robots. Companies such as Seegrid, Fetch Robotics, and inVia Robotics are automating many more operations in the warehouse setting. RightHand Robotics, Kinova, and Kinema Systems are developing sophisticated manipulation technology that may automate many more of the fulfillment tasks. Robots from providers such as Savioke, Aethon, and Toyota’s robotics division are on a path to work directly with humans as essential support providers—not just at the office, but also out in the field, on the road, and in the home.
Challenges to adoption
Although robots’ capabilities are increasing while costs are going down, certain businesses might not adopt robots today for many reasons, including:
- Maturity of the market: The industrial robot market is mature, but service robots, cobots, and others are fairly new. Many enterprises are reluctant to adopt emerging technology. Customers want assurance that the technologies they pick will last and will be supported for the foreseeable future.
- Fast rate of evolution: Because technological capabilities are evolving rapidly, many enterprises would rather wait to see which solutions catch on before committing to a particular solution or approach.
- Perceptions regarding labor displacement: The concern that robots will eliminate human jobs can create challenges with labor unions.
- Compliance: Regulations regarding safety, although necessary, can be a barrier for smaller businesses that do not have the scale and resources to monitor and report on compliance.
- High-cost system engineering and installation: While the cost of robots is coming down, the cost of system engineering, installation, and change management can sometimes be prohibitive.
Given the diverse tasks and environments for which robots are becoming suitable, businesses can find many opportunities to innovate and apply the emerging capabilities. And as a result, they could set themselves apart from competitors and create new efficiencies. Now is a great time to understand the technology and its potential for creating new business opportunities, to invest in some pilots, and to eventually integrate the technology into your operations.
To evaluate the potential and fit for emerging robotics solutions, businesses should ask themselves the following questions:
- Are there automation efforts that we once analyzed but shelved because of high cost?
- Have we evaluated robots in functions other than manufacturing?
- Are there tasks that we currently do not perform or perform infrequently because of the high cost associated with those tasks?
- Are there tasks that we do not perform because they are not core to our business, but we would like to perform if we could do so economically?
- Have we audited our processes and identified repetitive, mundane, or undifferentiated tasks done by humans that could be carried out by robots?
- Have we explored what gains in productivity, efficiency, and waste reduction we could achieve by deploying robots in different operations?
- Are there tasks that require high precision and dexterity for which we have difficulty securing human talent to perform?