Industry Adapting to Collaborative Robots

In this exclusive interview with Yaskawa’s Chetan Kapoor, he explains how industry’s evolving mindset toward production, metrics and business models are paving the way for greater use of robots working alongside humans on the production line.

One of the first media properties I worked for back in the early 1990s was “Robotics World.” As a staff editor on that publication at the time, I often heard how I had missed the glory days of the robotics industry back in the early to mid 1980s when U.S. industrial adoption of robots was at its peak. By the early 1990s, most experts thought U.S. industry was satiated with robots and no more large-scale implementations would be happening any time soon.

Though things were quiet on the industrial robotics front for many years after that, robotics technology kept advancing. Now, with the emergence of collaborative robots, we seem to be on the cusp of a new age of robotic deployment that could well surpass the heady days of robot deployment in the 1980s.

To get more insight into this, I spoke with Chetan Kapoor, senior director of technology innovation at Yaskawa Innovation Inc. Kapoor has led robotics research projects for DARPA and NASA and served as chief scientist at the University of Texas at Austin. In his role at Yaskawa, he is focused on collaborative robots; development of a common software platform to integrate robots, motion and drives; and creation of an interface that makes it easier to program robots.

Kapoor will participate in the Collaborative Robotics panel at The Automation Conference and Expo 2016 to be held May 24-25 in Chicago. For more information and to register, visit:

Following is a transcript of my interview with Kapoor.

Automation World: In the past few years there has been a significant amount of new collaborative robot introductions. In your view, is this due more to technology advances or market demand?

Chetan Kapoor: The technology for collaborative robots has been there for quite a while. In fact, robots that had embedded torque and force sensing have been used in R&D environments for over 30 years. The same is true for the collaborative robot market. It has existed for some time, but was not targeted by the major robot vendors. Until recently, engineers developed new robots primarily because of the needs of automotive customers, with some innovation specific to the packaging market. The requirements from these customers are primarily “production rate” centric with significant upfront investment required to deploy robotics. On the other hand, there have been customers in general industry that always wanted a simpler robot that was easy to deploy and maintain using their existing staff and automation infrastructure, instead of requiring all the bells-and-whistles of industrial robots. Introducing collaborative robots that meet the needs of general industry has spurred the collaborative market. The initial success of collaborative robots in general industry is now driving interest in the automotive sector, which is leading to a response by major industrial robot suppliers.

AW: Based on your experience, where does the market stand on collaborative robotics today? Has there been much uptake yet, or are manufacturers still more reliant on the industrial robot types with which they are more familiar?

CK: Various market research companies have pegged the collaborative market at $1 billion in 2020 with an annual growth rate of 50 percent. As with any new technology, that is potentially disruptive. The three to four-year predictions for the collaborative robot market may not be met and the 10-year prediction may be surpassed. Early adopters who are using highly skilled labor to deploy such robots are driving the current growth rate in this market. As the deployment moves downstream to plant personnel, you will see a slowing down as the deployment life cycle elongates. As these deployments mature, you will see new metrics and business models for collaborative robots emerge. I believe the success of these business models will lead to exponential growth.

AW: I have seen a number of case studies about collaborative robotics that involve work alongside humans in the medical device and electronic equipment assembly industries. Based on your experience, are these industries leading the way in collaborative robot use or are others equally or more advanced?

CK: Both medical device and electronic assembly operations in the U.S. have not deployed significant numbers of robots like other industries. The reason is that their manufacturing processes have been designed for dedicated machines combined with human labor. Deploying an industrial robot in this environment involves changes to the process, staffing, and infrastructure. However, collaborative robots—due to their ease of use and reduced integration costs—have proven effective in these industries because their deployment is not disruptive to existing manufacturing processes. Given the success of collaborative robots in these industries, you will see other industries also accepting such robots. Eventually, you will see increasing deployment in automotive, as a significant part of automotive processes are still labor intensive and face the same constraints as the medical device industry and electronics industry.

AW: What do you wish industry knew better about robotics—both the technology and application?

CK: In terms of robotics applications, the tasks a robot performs in a collaborative application are not much different from standard industrial robots. The majority of these tasks will fall under the handling and assembly categories. However, the environments in which a collaborative and industrial robot will function will continue to be dramatically different. Instead of automation being designed around robots, collaborative robots will exist in unstructured environments. You will see flexible plant layouts with humans in the mix. Design engineers and the companies they work for will drive the value proposition by improving human productivity, increasing product mix and saving space versus simply trying to boost the production rate.

With regard to growth of the robotics market, the general public opinion is that smarter, artificial intelligence (AI)-driven robots will increase the market for industrial robots. This will happen, but the pace at which this will happen will be slower than expected. A larger market growth will come from simpler robots that are easy to deploy and maintain, thereby making them accessible to a larger customer base that currently needs automation. You will start seeing a lower density, but larger area spread of robot deployment as compared to the high-density deployment seen in a typical automotive installation.

The cost and non-standardization of end-effector tooling continues to be a bottleneck for robot deployment. Flexible and general purpose grasping systems that offer repeatable performance can significantly reduce the burden of integrating robots and profoundly change industry.

A common misconception about the industrial robot is that it is a stagnant technology, largely unchanged since the 1980s. In contrast, people point to computing technology, which has exponentially improved in the same time frame. However, a robot’s lifespan has increased from 5,000 hours to 80,000 hours; its smoothness of movement rivals that of a ballerina, and 2D vision applications have become commonplace. This dramatic improvement in industrial robots is partly responsible for the increase in quality-to-price ratio that consumers see in today’s automobile.

What today's industrial robots lack is the intelligent software and sensing required to operate in unstructured environments. However, this inability is directly related to the nominal developments in AI, which scientists and engineers conceived of in the mid-1950s. New approaches to machine learning that harness widespread computing and easy access to large data sets will ultimately meet the hopes that AI raised.

To read more of Kapoor’s insights on robotics, see his column on LinkedIn.

Companies in this article
More in Robotics