February 5, 2011

On Automated Defenses, as Considered by Alexander Leandro Dela Fuente

When I was growing up, my parents had a rough list of required cultural material. Noteworthy among these, at least for this article, was the Alien trilogy. I was naught but a wee child at the time, but I trusted my parents and it really seems like they thought there was something in these movies that was worth the potential trauma.

To this very day, I still don't know what that something is. I did enjoy the films, but whatever deeper lessons necessitated ignoring a clear R-rating (Was it R-rated? It should have been.) remain out of my grasp.

I remember this scene, however:

It was an action scene, sure, but in light of our lessons in Psych 135 it also serves as an eye-opener about the detection and discrimination capabilities of these weapons. Aliens had the Colonial Marines relying on a motion-detection system for their sentry; once activated, the sentry guns were programmed to target and fire upon any moving objects in their field of vision. It's been a while since I last saw the movie (my siblings and I enjoyed the series enough for multiple viewings), but I think there was some mention of only firing at organic matter. This may be an implanted memory, of course, but I have no means of checking at this moment. In any case, the inclusion of an instrument that can discriminate between organic and inorganic matter seems to be pure science-fiction; in the unlikely scenario that it did make an appearance in the movie (not just in my imagination), such a system would beg the question of possibility. Can an automated non-contact scanner differentiate between organic and inorganic matter? Would it matter? I do not know, but it certainly sounds infeasible.

Wikipedia says modern sentry guns do not have the detection/discrimination capability of those depicted in fiction; however, a bit of searching in Google Scholar reveals that several groups are hard at work on disproving that apparently poorly-referenced claim. One such group, the Section L01, Turret Trackers Design Team under Dr. David Keezer, proposed a low-cost sentry gun, the Fully Autonomous Sentry Turret System. Three months later, they came out with a report on their working prototype. The prototype covered a greater field of view and acquired targets faster than initially proposed, at the cost of a lower effective range. To minimize the incidence of so-called friendly fire, a two-step discrimination system was used. The first mechanism relied on color recognition, the second on the constant transmission of an IR signal from individual allied personnel. Predominance of a specified safe color, on a scale of 0:255, identified a target as an allied individual; as I understand it, this hews closer to the trichromatic theory than it does to opponent processing theory. As for the IR signal, I don't think any specific neurological function can serve as a directly comparable system; at the very least, however, this might be the equivalent of an uttered safe word, much like those used in World War II. (American soldier says "Flash," confused German doesn't respond with "Thunder," gunfight ensues.)

Most of the article detailed how the proposed Fully Autonomous Sentry Turret system could be assembled and used with easy-to-find materials. While I was convinced by the effectiveness of their system, I sure do hope any adoption for military use will utilize more advanced discrimination systems. I know actual autonomous object discrimination is still far beyond the powers of present-day technology, but surely there are better ways to prevent friendly fire than the fulfillment of a monochromatic requirement and the use of an easily-replicated infrared signal. Human lives, after all, are at stake.

No comments:

Post a Comment