Home / Accueil


Why University Robotics Researchers
Need to Reach Out to Industry

by Andrew Goldenberg, Ph.D.

- Chief Technology Officer:
SuperRobotics Ltd., Hong Kong, China
ANZER Intelligent Systems Co. Ltd., Shenzhen, Guangdong, China
Engineering Services Inc., Toronto, Ontario, Canada
- Professor Emeritus
University of Toronto
 

One issue that has been pre-occupying me for a long time has been the relationship between academia and industry. I have read many reports in various media addressing the matter, but I have never found sufficient insight nor adherence to my views, until I read an article published in the IEEE Robotics & Automation Magazine, Sept. 2016 issue, entitled Robotics Academia and Industry: We Need to Talk!. The article is authored by Professor Erwin Prassler from Hochschule Bonn-Rhein-Sieg, Germany. He addresses, in part, the long-standing need for greater industry participation in university research. I found the article very interesting, and his assessment of the divide between the two communities by-and-large in agreement with my own.

In my experience, the issues raised in the article apply broadly to most industrialized countries. In this piece, I include many of the points made by Prof. Prassler, adding to them my own observations and beliefs. Whilst he wrote as an academic, I am largely claiming to present the industry perspective.

Why should university research community attract industry? Because: 
  • it is the major reason for its existence in terms of its role in education, training, and use of research results
  • focusing on research without solving operational problems of interest to industry leads to practical irrelevancy
  • by-and-large, industry needs the research community -- especially in the context of modern robotics: logistics, autonomous mobility, human-machine interaction, reasoning, and machine learning
  • ongoing dialogue between the two communities builds understanding of their divergent scopes and objectives: Research -- publish or perish; Industry -- economic relevancy
What kind of industry the research community should aim to attract? 
  • any industry that uses robotics
  • start-ups in robotics
What makes robotics research attractive to industry? 
  • industry does not perform basic research
  • industry needs to address advanced topics impacting future product developments
  • training and advanced education
  • robotics challenges
  • publishing survey articles
Origins of academia-industry divide  
  • the control systems research community was the first to address robotics research in late '70s
  • electromechanical design of robotics was not addressed by the early university research community concerned with robotics – it became the realm of industry therefore. In the '90s the research community became interested, however.
  • Computer Numerical Control was the first to address robotics as an industry
  • the computer research community was attracted to knowledge-base and early AI as a bridge to robotics in the '80s
  • while current university research meets Technology Readiness Level (TRL) 3-4; industry finds it useful only at TRL 7-8
  • industry does not approach the research community because it sees it as irrelevant to immediate issues
  • no strong evidence of operational links between university research and industry on issues of immediate concern
  • industry prefers to perform critical research internally (an IP matter)
  • embodiment of modern robotics technology requires intensive applied research
  • university research is primarily dependent on public or institutional funding
Suggested paths to rapprochement between university robotics research and industry 
  • industry addresses development of products
  • research community addresses basic research
  • both should aim to generate core technology to be embedded in products
  • product development and research to be performed simultaneously

 
“The primary aim should be to make the robotics industry create jobs for the masses, and not just for those with a Ph.D. Too often unsustainable businesses are created, conceived as high-end ‘advanced’ projects but lacking a fundamental business basis of growth that could lead to job creation.” 

Industry is best served by creating products that have a market value, as opposed to laboratory exploration that is the undertaking of basic research. Where academia undertakes applied-oriented research, it could be channeled towards ‘core technology’ that is the foundation of any product. Thus, product development and research would both contribute to the development of the core technology embedded in the products. Core technology, expressed through patents, trademarks, technical secrets and know-how would be valued by the market through its perceived efficacy in creating new products. This market value is further raised by the perceived market impact and penetration of the new products. In this manner, the research becomes relevant to creation of products that in turn may provide sustainment to the research efforts and cover their costs.

Academic and research-oriented institutions focus almost unilaterally on the development of basic technology. They are guided by perceived future market needs, competition between research institutions expressed by the publications and citations of each, and shear curiosity of research staff. These related undertakings are usually not linked directly to product development as done by industry. This leads to excessive generation of basic technology that may -- or may not -- be useful. Granted though, it may be ahead of the state-of-the-art, sometime by a decade or more; therefore, one cannot fully assess the impact of generating new basic and core technology.

The fact is that core technology that is directly related to market-driven products is rarely addressed outside those businesses whose main undertaking is to develop the products in the first place. The university-based research could be a significant contributor to the product development if it would allow targeted research to dominate, as opposed to total disengagement from immediate use that loads heavily on the taxpayer money.

Robotics as an Industry 

If we look historically at the computer and robotics industries we would note that the latter is older. This may be a surprise if one considers the stage of existence of the computer industry. It is mature and already in its 4th or 5th global business life cycle. The names of Microsoft, Dell, HP, and many others are well known. Why is the robotics industry still in diapers? The main reason is that it evolved as a novel academic exercise instead of being undertaken fully by the business community. Nonetheless, nowadays there are well known names in the robotics industry such as Fanuc, Kuka, Adept, and many more suppliers of subsystems and components. However, as an industry, it still represents a very small number of employees globally, when compared with the computer industry, and is still repeatedly referenced and viewed as an ‘emerging technology,’ although the field is more than 40 years old.

It is time to leave out the ‘wow’ of futuristic applications featured in YouTube videos, and get serious with the maturity of the robotics field as a current employment opportunity.

The primary aim should be to make the robotics industry create jobs for the masses, and not just for those with a Ph.D. Robotics should benefit the society at large, instead of the few and far between often supported by ill-informed funding decisions. This applies to both private-sector investments that do not reflect market needs and its capacity to adopt the new technology, as well as government funding that seeks to demonstrate forward-looking economic policies to a public largely ignorant on matters of technology and its impact on the welfare of the society. In either case, too often unsustainable businesses are created, conceived as ‘advanced’ projects but lacking the fundamentals of business growth that could lead to job creation.

Robotics start-ups 

To grow a technology business there is a need to develop a business plan based on a unique business model, as opposed to relying solely on the uniqueness or advanced nature of the proposed technology. Good business models succeed in so far as competing at an advantage, whereas good technologies without a suitable business model do not. This issue relates primarily to start-ups; governments and media focus on the rate of start-up formation as a measure of success. They ignore growth and sustainability in relation to the investment made.

Capturing a novel and unique business model has been my primary pre-occupation throughout my entire business journey. It has proven to be successful. I recommend to not adopt a known or existing business model simply because it is used by others. Devise the unique model that suits you and your technology.

N.Ed.
Dr. Goldenberg is the recipient of IEEE Canada's 2016 A.G.L. McNaughton Gold Medal, the highest award given out by our organization. His citation was "for outstanding contributions in the field of robotics and automation including research and development and founding of high-technology companies." You can read his biography and that of all the other 2016 Award recipients by perusing the Awards section of the Fall 2016 issue of the IEEE Canadian Review.


Sign-in is not required to post a comment (this option becomes visible after clicking on "Name"). However, all comments are vetted first by the IEEE Canadian Review editorial team before posting. You can receive notification of postings (Subscribe button at bottom left of page) if you set up a Disqus account, and then enable e-mail receipt (https://disqus.com/home/settings/email ). However, at this time, notification of Guest postings is not a feature of the Disqus platform.


Last update / 2017-10-11/ la dernière mise à jour

IEEE Canadian Foundation

The Personal

IEEE Member Discounts

© Copyright 2017 IEEE – All rights reserved. Use of this website signifies your agreement to the IEEE Terms and Conditions.

A non-profit organization, IEEE is the world's largest professional association for the advancement of technology.