Saturday, June 29, 2013

FRE tasks becoming more interesting

A sure sign that the state of some art is improving is when the difficulty of the tasks that must be performed in a competition is increased. That was the case with this year's Field Robot Event in Prague, which ended earlier today.

There was the usual navigating through rows of plants, but this year's tasks included detection of weeds (simulated by maize and oil seed rape) among the sunflowers, and detection of diseased or otherwise damaged sunflower plants.

In addition to this, team members following the robots were not allowed to carry any electronics that might be used to provide them with instructions. They could manhandle the machines to help them reorient at the ends of rows, at the cost of a penalty, but aside from that the machines were required to complete each task autonomously.

In my humble opinion, these changes are a hopeful sign that the intention is to move the competition towards real world applicability.

Saturday, June 22, 2013

sentry duty for ‘cultibots’

Earlier this month, at TED Global, science fiction author Daniel Suarez made a powerful argument for an international legal framework prohibiting the development of lethal autonomous systems, otherwise known as killer robots. He went on to suggest that machines designed to recognize menacing rogue behavior in other machines and raise an alarm could be used to protect against any killer robots that might be set loose despite the ban.

Perhaps that task, detecting the presence of a killer robot and raising an effective alarm, could be distributed across a wide variety of machines, particularly those working outside. ‘Cultibots’ would be ideal for this purpose, as they would also be geographically distributed, providing better coverage. I wouldn't suggest requiring such functionality, at least not yet, as that would simply delay deployment of robots for other purposes while this relatively sophisticated capability was developed, but once it becomes available it should be incorporated into any machine, with sufficient capacity to support it, that might be in a position to make early contact with a rogue.

Another, even more sophisticated capability that should be developed and incorporated into all types of robots is the ability to recognize and block any attempt to coopt or destructively repurpose themselves, whether on the part of a nearby machine or through a remote connection.

Suarez went on to suggest that machines dedicated to dealing with rogue robots might snare them and haul them away, but he says that, despite that they're dealing with other machines rather than people, they shouldn't be allowed to autonomously decide to destroy those other machines; that decision should always be made by a human being. Under most circumstances, other types of robots should avoid engaging a rogue, if possible.

A situation in which it might not be possible to avoid engagement would be if a human being were at risk, as expressed by Isaac Asimov's First Law of Robotics: “A robot may not injure a human being or, through inaction, allow a human being to come to harm.” Considering that an attack on a human might be the first clue that a robot was a rogue, and that a second's hesitation might make the difference between life and death, any robot in the vicinity should be ready to do what it can to block such an attack, perhaps even suspending other duties so as to be able to pay close attention whenever both a human and a robot of unknown trustworthiness are nearby – devoting any spare processor cycles to running scenarios and developing contingency action plans.

Of course, governments will want the ability to remotely repurpose whatever robots are available, not only to deal with rogues but to deal with all kinds of emergencies. However, there should be no way for them to turn ostensibly innocuous machines into combatants (beyond the requirements of the First Law), much less into autonomous killers. To ensure this, remote repurposing of ordinary robots should be constrained to a predetermined, short, fixed list of alternative modes. Any changes to that list should require either a ‘brain’ replacement or the physical presence of a factory rep to perform the reprogramming, and this too should be made part of international law.

Monday, June 17, 2013

higher fruit

FourCell

The low fruit in robotics is what we see around us now – most automation is either very limited in scope or else occurs in a controlled context, because even the most sophisticated machines are barely capable of operating in a truly uncontrolled environment, populated with with unknowns. The higher fruit remains largely hypothetical, for the moment.

But, at least as importantly, the low-hanging fruit involves fitting robots into what’s is already happening, as by substituting machines for people doing more or less the same work. Conversely, the higher fruit involves taking advantage of robotic technology to transform what's happening, substituting better practices for those that are currently convenient, or at least conventional. But transformation is, by its nature, disruptive.

I’d like to begin by laying out a conceptual, two-dimensional graph of possibilities, expanding upward and to the right from a point of origin. Let’s call the horizontal axis time, with what’s technically possible now near the left edge and what will sooner or later become technically possible arrayed to the right, according to how far in the future it can be expected to come within our reach. The vertical axis is a little trickier. At time zero (the present), you could call the bottom of the graph ‘current practice’ and points higher along the left edge represent increasing departures from current practice, all possible within the constraints of current technology, but which could be expected to involve greater investment and/or to encounter increasing resistance from vested interests, or other forms of economic or social inertia. Let’s label the vertical axis ‘disruptiveness’.

As you move to the right (applications of technology not yet within the state of the art), data points above the horizontal axis represent applications which, even if they were to become possible tomorrow, would still involve varying degrees of disruption of current arrangements. Current arrangements being financial, legal, social, habitual, or any combination of these, but specifically including current practice as what would be most directly disrupted.

Now, for the sake of simplicity, let’s reduce this graph to a four-celled, two-by-two grid (see graphic), with the lower-left cell representing current practice, the lower-right cell representing predictable advances in technology that, once available, can be applied without disrupting current arrangements, the upper-left cell representing transformations that are technically possible now, but which can’t be applied without disruption, and the upper-right cell representing what will become possible at some point in the future, given predictable advances in technology, but which can't be applied without disruption, unless current arrangements change in the meantime.

That upper-right cell is a kind of Pandora’s Box. It contains many wondrous things that, were they imminent, would nevertheless be characterized as impractical, because they somehow run afoul of current arrangements. Nevertheless, this is where you’ll find much of the potential of robotics.

Take the notion of Personal Rapid Transit (PRT) as an example of something overlapping between the upper-left and upper-right cells. PRT is an application of robotics to transportation which has been knocked about for several decades. Technically, it’s not such a great challenge that a determined effort couldn’t get it working acceptably on a large scale in fairly short order, a few years at most, but that determined effort can’t happen, because deploying such a system would require a large investment in new infrastructure and would work to the disadvantage of those who are already invested in some aspect of automobile transportation, including car owners since that infrastructure would inevitably mean taking some space from existing streets and some funding from their maintenance. Instead we get cruise control and automated lane following in otherwise conventional cars running on conventional roadways, and PRT research lives perpetually on the back burner, making slow progress.

To bring this back around to the application of robotics to agriculture, from a certain point of view it's easy to overlook the disruption involved – displacing monoculture, heavy equipment, and petroleum-based fuel, fertilizer, pesticides, and herbicides – as a net gain in itself. But not everyone would agree, certainly not most of the companies whose primary business is marketing such products to farmers. And most farmers would surely be pushed far outside of their comfort zones if the whole vision of replacing conventional agriculture with what is essentially robotic gardening were to be ready today and imposed by fiat tomorrow. Wherever you find yourself on that spectrum, it must be obvious that the application of robotics to plant care in a very detailed manner, on a plant-by-plant basis, if scaled up to a significant percentage of the land area currently under cultivation, holds the potential to be hugely disruptive. Because of this, I’m not particularly surprised to see research of this sort proceeding on a shoestring, privately, or outside of the US.

On the other hand, its potential for disruption is no measure of the intrinsic value of a potential technology, and some technologies are needed, even desperately needed, despite that they are likely to run afoul of current arrangements. I would argue that the application of robotics to making the best practices of horticulture scalable is one of these, perhaps even the poster-child example.

This is the fourth post in a series: Part 1, Part 2, Part3.