An Initiative to Keep America's Robotics Roadmap Relevant

Did you know the United States has a roadmap for robotics? It does! In 2006, a one-day workshop titled Science and Technology Challenges for Robotics was organized by George Bekey of USC, Vijay Kumar of UPenn, and Matthew Mason of CMU. A summary report of that workshop states There was an enthusiastic response to the workshop with over 85 participants. Discussions had to be cut short because of time constraints. This could clearly have been a two-day workshop. There were many volunteers who were ready to take on more responsibilities to promote the discipline. (Vijay Kumar has recently been interviewed on Robots Podcast and was mentioned on Robots.net even more recently.)

During the process which followed that workshop, Matthew Mason and Henrik Christensen of Georgia Tech collaborated on an essay which summarized the state of robotics and previewed the findings of the effort to produce a roadmap for robotics. (Before occupying the KUKA Chair of Robotics at Georgia Tech's College of Computing, Henrik Christensen was the founding Chairman of EURON, the European Robotics Research Network.)

The final roadmap report was presented in May, 2009, before the Congressional Robotics Caucus, however, in the effort to produce that report, the call for the formation of an American Robotics Network (9th slide) appears to have fallen by the wayside.

On January 22nd, Professor Christensen posed the question Are we ready for an American Robotics Network on his blog, saying that he had started a discussion regarding the organization of an American Robotics Network. He has also discussed the formation of such a network in a brief essay on his website. In the recent blog post, he says I would like to get this underway as soon as possible to make sure that we can leverage the momentum from a National Robotics Initiative. It will also be an important mechanism to make sure that we can maintain a push forward.

a minimal-hardware approach to weeding

The idea presented here applies only to weed seedlings. Weeds growing from tubers or invasive roots will need to be handled more aggressively, but seedlings, being poorly rooted, are vulnerable to methods that destroy their single meristem. Moreover, after a few years of careful weeding, they are the only type of weed that would persist, except for those growing from runners invading from adjacent land, around the perimeter of the plot, so this method would become gradually more sufficient.

In a nutshell, the idea is to use video imagery to locate seedlings, an expert system (the hard part) to distinguish between desirable seedlings and weeds, and a pulse laser to first make sure it has a clear path to the weed seedling (nothing in the way), focus on the portion of the seedling containing the meristem and then deliver one or more relatively high-energy pulses to heat it sufficiently to render the meristem inert, so that the cells are no longer capable of growth and division. It isn't actually necessary to kill the meristematic tissue outright, just inactivate it, so the higher energy pulses used to accomplish this should not need to be so powerful that they present any danger of fire.

Of course, if the machine carrying out this task maintains or has access to a very detailed map of the plot, which precisely locates and keeps an image archive of every seedling, the next time it passes nearby it can simply check whether the plant appears to have withered, or whether it has recovered and continued growth, in which case it may be time to call in heavier equipment. In this way it can build experience with just how much energy is required to stop the growth of a weed seedling of a particular type at a particular stage in its development. Weeds that survive the surgical approach of the laser can be dealt with by more conventional mechanical methods.

The video system should at least combine a wide-angle view with a telescopic view (needed to distinguish between weeds and desirable seedlings). Either or both might be binocular (stereo), for 3D capability, and the telescopic view in particular would benefit from the use of a sensor that could deliver partial frames very rapidly, to help assess the effectiveness of the laser pulses (how much does the meristem swell within the first tenth of a second?).

I call this a minimal-hardware approach because it involves little more than a pair of cameras, one wide-angle and the other telescopic (two pair for stereo video at both focal lengths) and a laser, on a mount with two degrees of freedom, both rotational, and some means of moving that mount around a plot or field. The real complexity would be in the software that deciphered the video input, deciding which seedlings to zap and which to let live. A high-pressure water jet could be substituted for a laser, but such an arrangement would be more challenging mechanically, because the nozzle would need to either come within a few inches of the seedling or use a significant amount of water to be effective. Too much water applied at high pressure might create other problems, for example encouraging the growth of fungi.

The knowledge necessary to distinguish between seedlings of various species would be an appropriate addition to the RoboEarth project.

a compromise between rails and walking directly on the ground

If the area to be covered by a farmbot is known, and limited, it might be tempting to outfit the land with rails and the machine with wheels to match, to keep the weight of the machine off the soil and improve its mobility, but in areas where production is constrained by low precipitation or short growing seasons this could prove uneconomic.

A possible compromise solution would be to use long, spider-like legs to span between the tops of posts, a foot or two above the soil surface, or even just low mounds of gravel. Providing this much infrastructure would not only prevent tracking and compression of the soil over most of the area, but it would help the machine locate itself in the field, since the posts or mounds would have known, static locations.

While such machines might move more slowly than if they were equipped with wheels running on rail, the logistics of having several working the same field would be simpler, since they could just walk around each other.

cascading distributed network

Another such idea (taken through initial development as a thought experiment), in this case one that you'd have to be a chip hacker or microcode programmer to actually implement, first saw the light of day years ago, on The WELL, and then more recently in a topic in the Robots Podcast Forum (since closed).

This one is about very efficient addressing and message passing through a processor network having arbitrary topology, using only the minimum necessary number of bits for each step in a path, and automatically generating a return address, which can also serve to identify the source of the message.

It's recently occurred to me that this idea might be particularly applicable to robotics, where machines might have a separate processor to control every major joint and sub-system, and need to pass messages directly between them without going through a central switch, to keep latency manageable.

Such a network could also accommodate situations where hardware needed to be hot-pluggable, added and removed as the situation required, since newly attached hardware would automatically acquire predictable addresses and, in the case of removal, remaining hardware would always have return addresses for use in sending "cannot deliver, that path is closed" messages.

examples (and the limits) of design through imagination

At the beginning of March, 2009, two such ideas (designs or simulations running inside my head) had been taking up cerebral resources for some time, weeks or months, so, since they weren't going to be getting any better in the absence of something more tangible, either a CAD model or a mockup, neither of which I had time for, I decided to offload them to one of my blogs, in the hope that someone else might benefit.

The first is essentially the miniature equivalent of inserting an air hose through the tread of a tire at a very shallow angle, nearly tangent, to create a dust barrier via the resulting airflow, with the idea of using it to keep dust off of camera lenses and the like.

The second had its origin in the knowledge that the closer you get to the pivot point of a lever the more force is available. Applied to a robotic manipulator, this means that the outer tips of the 'fingers' should be more sensitive and delicate than segments closer to the 'wrist' (the point of attachment to the supporting arm). Conversely, it also means that those inner segments might be used where more force is needed, as in clipping through the stem of a woody shrub. Inconveniently, stems in need of clipping come at odd angles, so if a shear only operates in a single plane that plane may need to be rotated as much as 90 degrees in moving from one clipping to the next, which might require repositioning the entire machine, which could slow down the operation considerably. Giving the manipulator a set or semi-rotatable digits, that can pair in two different X-shaped configurations, 90 degrees opposed from each other, could provide as many as six shear planes without any rotation of the manipulator unit as a whole. This would allow a pruning robot to move from one clipping to the next with a simple repositioning of its digits.

Further Introduction

Not mentioned in my intro is that I received a Bachelor's degree in biology in 1980. I'd hoped to return to school for a second degree in engineering, but that never happened, and I spent several very hard years essentially trying to punch my way out of a cognitive bag composed of academic categories, and the emotional baggage I attached to each.

The resolution I found came through the discovery of General Systems Theory, itself an academic category, but one that points to the general applicability of a collection of fundamental concepts. Thus armed, I approached learning with renewed confidence.

It wasn't long after this that I began to become obsessive about computer processors and software, always with an eye to how they might apply to robotics, since I was already interested in mechanizing and scaling up horticulture. Being possessed of a vivid imagination at least with regard to machinery, I built many machines and set them running in my mind, frequently sharing descriptions of these designs with whomever would listen.

For me that was the missing ingredient, collaboration. With no one to share my enthusiasm, it was wet blankets wherever I turned. It's only recently that I've begun to feel like I might have found my tribe.

But I'm not a tinkerer; I'm out to change the world, by replacing big, dumb machines with smaller, smarter (wiser!) ones, beginning with agriculture.