Sunday, June 19, 2011

augmentation: the other side of the robotics coin

Toward the end of the first installment of my response to Martin Ford's The Lights in the Tunnel I said the following:

just as technology enables automation, it also enables augmentation - strength amplification, protection against environmental hazards, heads-up displays providing just-in-time information, enhanced senses, precise manipulation, eye tracking, voice recognition and synthesis, etc. - making what the average human worker is able to perform a moving target

Under pressure from the growing complexity of the aircraft it produces, Boeing has been a pioneer in using virtual reality overlays to provide people possessing general skills with the specific technical information needed to perform the tasks at hand, making it unnecessary for workers to be experts on the systems they build or maintain. That expert knowledge is maintained in a database and served to the worker just when it's needed. (Such an approach could also enable farmers to maintain robotic machinery with which they had no prior experience.)

DARPA has demonstrated keen interest in providing soldiers with wearable equipment that can enhance their strength and stamina, allowing them to carry more weight further, faster, over terrain too difficult for wheeled vehicles. They have also funded fully robotic solutions, but so far that augmentative approach looks more promising.

These examples combine nicely. A factory or maintenance facility worker with a powersuit would be able to handle heavier parts without the need for hoists, and a soldier with a heads-up display would be less likely to get lost, or to waste time and effort on inefficient paths.

Telepresence and teleoperation make it possible for human workers to be on the scene, instantly, when needed. Using the example of an automated transportation system this could mean welcoming passengers and verifying that they and their belongings are entirely inside the vehicle before closing the door, ascertaining a destination, operating active components to secure assistive devices, checking whether problems develop en route, insuring passenger security at the destination before opening the door, checking whether the vehicle needs to be cleaned or repaired before being used again, and actually directing the vehicle anywhere it needed to go outside of the track/guideway system, with the aid of onboard sensors and intelligence. For each of these functions there might be an automatic mode, with a human operator monitoring in questionable circumstances and intervening whenever the automatic mode proved inadequate, when experience suggested that it would be likely to do so, or when a particular passenger had indicated a preference for dealing with a human operator and time permitted.

When a human is part of the solution, you get a highly evolved brain and basic senses in the bargain. Ford makes the point that, for many jobs, what a human brings to the table is more than is needed, and that providing technical analogs for just the portion that is needed is commonly either already possible or within reach. While I grant the truth of this, I also want to point out that within reach and affordable are far from being the same thing, and that just because you can replace a person with a machine in a particular circumstance doesn't mean that doing so constitutes a reasonable business decision. Moreoever, the 'excess' capacity of a human worker may be just what's needed to prevent an anomalous situation from turning into a disaster, saving the company far more than the difference between wages and benefits and the cost of ownership and operation of some replacement machine.

Because technical augmentation tends to move human workers from mind-numbing work into positions where they are both more stimulated and have a higher level view of the overall operation, it also pays off in terms of developing experience in those workers as individuals and in the workforce as a whole.

Aside from the simplest repetitive tasks, the return on the investment dollar for technology to enhance a human worker's capabilities is very likely to be both greater and more immediate than the return on investment for the more sophisticated technology needed to actually replace that worker.

No comments: