The following headlines highlight how robots might be able to replace human workers.
- RoboBees with Laser Eyes Could Locate Disaster Victims
- How Robots Are Building a 3D-Printed Metal Bridge in Amsterdam
- Your Next Garbageman Could Be a Robot
(These three examples are from the LiveScience web site.)
In the 1970s, the factory of the future was to be the “next big thing”, a $25 billion market of which GE would have a 20% share.
GE lost at least $120 million (1980 dollars) in its efforts to develop robots for its factory of the future thrust.
Now, 45 years later, the headlines are still promising that robots have a future, so what went wrong? And do they have a future?
The answer to the second question is yes, because engineers now have a better understanding of the software needed to control robots, and because engineers recognize applications where robots can work.
When GE first embarked on its factory of the future program, it had been developing robots for several years, and believed that a robot could be trained to self generate the program for controlling its motions. Under this scenario, a robot would be manually walked through each step of a process, where it would remember the motions so they could be repeated continuously when the robot was in operation. GE also had a leadership position in the controls and communication devices needed to interconnect equipment, which supported its efforts.
The vision was sound, but an understanding of the technical requirements was lacking.
There were successes, but in nearly every success the processes were highly repetitive: Spot welding robots manufacturing automobile frames, and robots painting identical parts are examples.
All businesses within GE were encouraged, better stated ordered, to develop applications using robots in their plants. In one of my plants, an attempt was made to use a robot to weld various components, but the variations encountered were too great to allow the robot to work without constant supervision. No savings could be identified, and costs actually increased.
Robots, to be successful, must be able to control precise, intricate motions that are very often understood by the workman, but not by the programmer creating the code that is to control the robot. Situations, such as in welding, where the human operator can automatically respond to unexpected variations in the job at hand, are extremely difficult to program.
Repetitive applications are ideal candidates for robots, so RoboBees with laser eyes, programed to fly a predetermined route is realistic. Using drones to map agricultural fields with LiDAR to ascertain crop yields is already being done successfully.
An important lesson to be learned from the history of robots is that new technologies can take decades to become economically viable.
A good example is battery powered vehicles (BEVs). While BEVs work, they are not economically viable because of the high cost of the battery used to power the vehicle.
While there are new methods for designing and producing batteries being developed in the laboratory, such as Sakti’s solid state manufacturing of battery cells with material deposition, it could be years before they will dramatically reduce the cost of batteries in BEVs. It’s very likely that BEVs will be a niche product that can’t compete economically with gasoline powered vehicles for decades.
Solar power is another example where the process works, but is uneconomic. Someday, it may be possible to use the sun to generate electricity economically, but there is nothing on the horizon to indicate it will be economically viable within decades.
Even robots that are now becoming viable will have limited applications. It will require advances in software development, with self learning and sophisticated feed back loops, often referred to as deep learning, before robots become truly flexible.
Watch for my new book, which will be available in January.
* * * * * *
It’s easy to subscribe to articles by Donn Dears.
Go to the photo on the right side of the article where it says email subscription. Click and enter your email address. You can unsubscribe at any time.
If you know people who would be interested in these articles please send them a link to the article and suggest they also subscribe.
© Power For USA, 2010 – 2015. Unauthorized use and/or duplication of this material without express and written permission from this blog’s author, Donn Dears LLC, is strictly prohibited. Excerpts and links may be used, provided that full and clear credit is given to Power For USA with appropriate and specific direction to the original content.