Tuesday, July 21, 2015

Asymptotic Robotics

This post is a discussion of how robotics might be used in an alien civilization, and how this might impact our possible interaction with them or detection of their existence. Recall that the presupposition is that the alien civilization has lasted a long time, long enough to get past the early days which might be chaotic, and long enough to have learned about all there is about science and engineering. Robotics is one of the fields that we divide science and engineering into, but they may not. It is conceivable that there are such fundamental differences between mobile robots and controller robots, which do some task while remaining in one place, that the field split apart. There may be huge sub-specialties which are focused on the materials used, the energy sources, the motion actuators, the sensors, the computer or computers used to manage it all, and intangible parts as well. There could well be a branch of science and engineering that relates to software controls, which ensure a robot stays within some bounds of behavior, especially mobile ones.

Also recall that what we call soft sciences now, such as economics and political science, have on the alien planet long since been turned into hard sciences, meaning all the fuzziness has been dealt with, and the sciences are quantitative and measurable. The limits on them are the limits on omniscience, as discussed elsewhere. With this knowledge comes social calm, and alien civilizations have this as their theme for the later millennia. One alien civilization may achieve it much faster than another, for various reasons buried in their histories, but if science and engineering are allowed to accumulate, they will reach the pinnacle of asymptotic technology, which includes the knowledge of star-faring. Of course there are obstacles to that, and a society may fall prey to one that eliminates its ability to retain what it has learned, and descend into a dark age. Perhaps it will be temporary, or permanent, as there are traps into which an alien society might fall which don’t allow recovery, such as the establishment of anti-science memes, scarcity problems, genetic drift away from intelligence, and certainly others. These are not the alien civilizations we might meet or detect long before we move to the stars, if we ever do. These halted or collapsed civilizations are certainly interesting, and should be discussed, as if they are present in the galaxy, alien civilizations which do travel the stars may meet them, and interact on some basis. This may influence what we might detect on exo-planets, once our astronomical instruments achieve the next generation or two in capability.

In a different post we discussed asymptotic genetics, and introduced a word for convenience, intello, meaning an intelligent organism grown for some purpose. One basic difference between intellos and mobile robots is that intellos are grown as cellular objects, and robots are assembled from separately constructed parts. If you consider one of each, the intello is much less expensive to produce, provided the DNA or whatever coding has already been done. A robot also has design work that precedes its construction, which might be comparable to the coding for an intello. Both would have been automated to a large degree, and perhaps to simply a level of asking for specifications.

An intello is grown somewhere, as a farm animal might be, except that it requires training more extensively. A robot needs manufacturing facilities, one for each part. Manufacturing might be done by something descended from our latest manufacturing innovation, 3-d printing, in that each part is built up by deposition. But each part is still built individually, and then assembled. For a lot size of one, it would certainly seem that an intello is less expensive to produce.

For a larger lot size, the facilities would be amortized over the whole production quantity. This would reduce the per unit cost of the robots, but not so much the intellos, as their production takes a dedicated facility that lasts first over the duration of gestation, and then over the duration of growth and training. So, on a cost basis, the alien civilization might use intellos for some tasks, where individual specialization was needed, and robots for tasks where there was a large number of identical processes to be done. In an alien civilization, production and recycling tasks have a repetitive nature and are done to meet the needs of the whole population, therefore in mass, so robots would appear here. Facility care, if there were large numbers of unique facilities, might be done by intellos. Anything that required specialization, such as household service, might better have intellos.

The cost of an object is also affected by the cost of disposal, especially where there are no dumps, only recycling. A robot can be repaired, or if too many parts have reached the end of service life, can be disassembled and the parts easily separated for recycling and remanufacturing. However, here we have a large number of specialized facilities for the different classes of parts. An intello just dies, and the body is recycled. This would push the cost factor toward intellos, provided there is a balance in one area: longevity. The cost of something which provides a service for a long time is better expressed as a cost per year of service, rather than overall costs. Other costs are converted to per year costs. If a robot can be built to last a hundred years without going overboard on construction costs and materials, but a intello can live for three hundred, the cost advantage tips its way. Some Earth animals live for two hundred years or more, so the possibility of three hundred or even more is conceivable. Recall that alien society has become stable, meaning that the task breakdown and composition remains constant over any lifetime.

Maintenance and repair costs also might favor intellos. Intellos can be designed like the natural life on the planet, to be self-maintaining and self-repairing. Specialized repair facilities are necessary for robots, although if a large stock of replacement parts are available for the more common robots, maintenance and repair would be simple. All in all, the cost basis might favor intellos more than we might initially guess. Intellos also have self-diagnostic nerves, whereas to build a robot to accurately sense its own status down to the level that a normal nervous system would requires a on-board communication network within the robotic body far beyond what we imagine today.

A cost basis is not the only reason for selecting one or the other; there may be task requirements that eliminate one or the other, or establish a preference. The members of the civilization may prefer to interact with intellos. Thus, even with a small cost disadvantage, they would be used there. Robots have a size advantage, as they can be made larger than biological limits. Large construction jobs might need robotic work, although the construction jobs in a world where recycling is done to a very high degree, are re-construction jobs. Nonetheless, size may matter in this area. Since cells have a finite size, there may be some roles on the other end of the size spectrum as well, as robots can be designed to be quite small.

There are environmental effects as well. There are cells that can withstand radiation much better than others, having a rapid protein and DNA repair capability. Robots can be designed out of radiation-resistant materials. We might expect that isotopic separation has become routine in the alien civilization, and materials built out of isotopes most immune to radiation effects could last far longer than ones we could build today and longer than any biological creature. Thus, work on nuclear energy facilities would likely be done by robotics, not intellos. These facilities might occupy a significant fraction of an alien city’s total space, and provide many tasks.

Robots are also much less sensitive to temperature, and can be designed to be largely unaffected by vacuum or high pressure. Tasks where these are involved will be wholly robotic.

Are there some tasks which have such specialized requirements that one of the two options are excluded? The neural network that grows in advanced organic creatures has a degree of complexity that will press robotic designers to match. If so, tasks which are not specialized but require a high degree of motion control and local and global sensing might be intello-only. Perhaps music playing on traditional instruments is an example. Could a robot be built with enough fingertip sensitivity, along with motion control, to match what an organic being could do?

Space traveling is a unique requirement, and star-faring is even more extreme. It is very obvious that building habitats for organic creatures is quite costly, compared to building and flying robots. Our current baby steps into space have shown this. There seems to be no reason why this would be different for an advanced alien civilization for interplanetary travel. But star travel has one unique difference. It lasts longer. Can robots be built to be as self-repairing as an organic being over the long period that a starship would travel? It is an interesting thought experiment to compare a possible starship, a probe not a colonization ship, controlled by an organic entity, with a longevity greater than that of the travel time, by regenerating cells continuously, with a starship controlled by a robotic entity that has to be able to repair parts of its controller itself periodically. Cellular growth is built into the organic entity, but is an add-on for the robotic one, meaning that there is more checking on status, more reliability questions, and more likelihood of mission failure. As far as we can tell now, it seems certainly possible that any pre-colonization ships would be controlled not programmatically like a robot, but by neural processing, or perhaps by a clever combination of them. This has implications for the size and the thermal signature of any interstellar probe that we might search for inside our solar system.

No comments:

Post a Comment