DESCRIPTION OF SUBSUMPTIVE CONSTELLATIONS...2
BENEFITS OF SUBSUMPTIVE CONSTELLATIONS...7
Third Wave Compatibility....7
Ageless Qualities of War....9
Low Cost Options...12
Space Tug Service...15
The Most Important Mission: Doctrine...16
As the size of satellite designs shrinks, so does the cost involved in making them (14: –;15:–). But requisite technology and cost are not the primary emphasis of this paper. The emphasize is on utility of a new concept rather than the size and shape and materials of an individual satellite design. The recurring motif in this proposal is a recommendation to use "lots of little" versus "few of big." The paper is divided into three areas: Description of a Subsumptive Constellation, Benefits of Subsumptive Constellations, and Proposed Military Missions.
Secondly, in a SPACE design, accomplishments of the system are larger than the scope of individual satellite’s programmed tasks. Most satellites have been flown as individual remote control sensors that download data to earth stations or as communication echoes, which repackage what they collect from the Earth below. They live and operate and are controlled as an individual. SPACE is more analogous to a distributed, remote, automatic robot. Because of the inherent ‘uselessness’ of robotic sub-systems away from the entire robot, attention is drawn to the accomplishments of the whole. For example, a wheel does little by itself. When it becomes part of a propulsion system, a larger goal can be accomplished. But when it is integrated into an entire robot, moving the robot from here to there, it has been subsumed into the overall design. Attention drifts to what the robot can do, not what the wheel can do. An unlucky picnicker, inundated with ants, never speaks of having trouble with an ant. The problem is always with ants (plural). The entire colony has identity. Likewise, killing an ant does little; the entire colony must be dealt with. In this second case of subsumption, the identity of an individual is subsumed into the colony--not destroyed, but subsumed.
The last critical component is emergent behavior. Artificial Intelligence researchers propose truly subsumption robot colonies that “work together assembling simple structures–even though no single robot has any idea of what the group is trying to accomplish” (17:50) . Consider again the ant colony. Given the description of a six-legged, two millimeter long, dual-articulated body moving pieces of dirt around, there is no way to predict the construction of little caved-in volcano shaped objects around the entry to a colony (ant hills). The construction of a circularly symmetrical earth mound emerges from lower-level, stimulus oriented, act/react activity. More germane to satellite operations, imagine the utility of robotic satellites that could “chase obstacles, but stay back from them a little bit.” (8:430) In a classical system, the burden of programming a concept of “chase,” determining what constitutes an “obstacle,” determining how far a “little way” is, becomes an involved programming task. In a subsumptive robot architecture, this behavior of the entire robot would emerge from subsumptive competition between a multitude of sensors. The behavior approximates artificial intelligence better than many highly complex computer systems that take a long time to “think.” (3: –;4: –) A robot demonstrating this emergent behavior has been actually built (8:480). Tasks were hand compiled into a measly 256-gate logic array. Not even a tiny microcomputer was used. The same holds true for Squirt, a 1 in3 robot that demonstrates covert movement toward loud activity (8:482). This, then is the “magic” of subsumption: the realization of a desirable behavior without a priori design of the behavior, and without the associated cost–in dollars or design effort.
All three characteristics claimed above (tight binding, system vs. individual accomplishments, and emergent behavior) are exhibited. There is no serialized computer program interpreting the world and commanding final movement of an actuator. What does exist is parallel processes ranging from LIFT LEG HIGHER, through MOVE FORWARD, to TRACK PREY, with minimum or no communication between tasks. The synergetic nature of the legs’ operation should be evident. One leg does not walk. Many legs walk. And lastly, the emergent behavior of walking does not exist at the leg level. Each leg has no knowledge of what walking “means,” let alone a method to accomplish it. Walking is an emergent rather than a programmed behavior.
Whereas the first subsumption architectures had glory in locomotion, the second is a system of relatively immobile robots for planetary exploration. They accomplish their goal only as a collective constellation. “As many jobs for planetary explorers consist primarily of collecting data, robots that carry around extra bulk and weight offer no benefits from their size The whole escalating problem of large motors requiring large batteries, in turn necessitating larger motors and so on, could be eliminated if we [reverse our thinking].” (8:484) After developing a case for both sensors on a chip and for the concept of locomoting an individual silicon wafer, the author says, “One application for such gnats is to spread them out over a large area of the planetary body, and let them signal their position if and only if they find some condition is met locally. The orbiter then gets a density map of the likelihood of that condition by watching for signaling gnats.” (8:485) To support the warfighter from space, think of a swarm of sensors for soil water permeability communicating back to the mother ship. If the enemy found and removed a sensor module, they would have only a small wiggling mirror. They gain minimum intelligence value, and removing any one sensor would barely put a dent in the overall mission. We do not need to defend the physical item of our technical accomplishment, and we need not worry about suddenly loosing the comprehensive system of information collection. This is the quintessential essence of SPACE war in an Information Age.
A person more familiar with tiny satellites and space operations carried the proposal even further to gnat robots with no locomotion capability. This time, the goal was control of forest fires here on Earth (16:783). Swarms of temperature sensors could be deployed by airplane to catch in the higher branches of a forest’s trees and signal noteworthy temperatures to an overhead platform. Any readings out of the ordinary could trigger deployment of more precise, densely distributed sensors, allowing the fire control command post to remotely track the extent and intensity of a fire.
But the arrangement is still clunky because each satellite is an individual. For instance, when uploading adjustments to the Selective Availability feature, ground controllers must wait for each satellite to pass overhead. Simple geometric limitations create a 6 hour delay, and pragmatic issues of our nation’s constellation command and control (C2) capability can increase this several times over. The burden of day-to-day control operations for all 24 satellites in the constellation through normal AFSCN channels precipitated dedicated GPS antennas and support electronics at two remote transmitter sites. Cross-link capability is expected circa 1996; but, conceptually, that will only make tri-function satellites (nav signals, NUDET, and comm relay). The real break to a subsumptive constellation would occur if each satellite were no longer important in and of itself, and did not need dedicated communication efforts.
A second example is the present day ALERT mission of the 11th Space Warning Squadron that has subsumed the Cold War DSP constellation. The original mission is not destroyed, but it is taken up inside of what has quickly become the showcase accomplishment of theater ballistic missile warning. ALERT is a good example of synergetic data, but DSP’s heritage in the Cold War and 1960’s satellite operation concepts shackles individual satellites with myopic functionality and near zero automation or intelligence. Still, the example is illustrative because, although the fundamental character of ALERT requires individual behavior of each satellite, the collective character of capability exceeds that of any individual satellite. The synthesis of all the satellites (or at least the data) is significant enough to give it an entire new name, a new squadron, and a new mission. It is not the same DSP. New behavior has emerged.
From the Roaring Twenties to the Freedom of the Sixties, quality stood opposed to quantity because units of measure were physical items. In today’s world, the majesty of physical assets are not the measure of a nation’s influence nor a military’s power. In this Third Wave of society, thousands of individual facts become the torrent of information that powers a nation. The Tofflers document growth in quantity of information, software, and computers while simultaneously speaking of demassification. The center of power is not in the creation of a “battle-star” satellite, but in the duplication of many satellites that form the constellation. One constellation replaces one thousand satellites. The apparent retrograde to less capable space vehicles is really a huge step forward with the technical quality of the constellation because of the environment of today’s wars.
Holley’s second point is concerned with doctrine. Our nation has no doctrine for using diversified constellations of the Information Age. Production quantities of tiny satellites will provide the physical assets to develop that doctrine. As Holley points out, facts, tactics, technical detail, and trial runs will only occur as we have the weapons (satellite vehicles) to exercise (19). A large burden for contemporary space operations is that each vehicle, even within a family, is an individual. The concept of a Technical Order Manual, so pervasive in Air Force aviation operations, is unheard of for a series of satellites because each satellite has individual operating qualities. This very character has shackled research and design engineers to space operating consoles.
A second way SPACE fits into a 21st Century concept of space power is in its value of persistent presence versus defense of physical assets. The Air Force Strategy Division at the Pentagon works to expand “traditional concepts of military presence to include not only the physical merits of forces, but also the virtual advantage...” This “virtual battlespace...is not defined in terms of traditional, centralized, geopolitical boundaries, but in terms of a decentralized, global web of networks.” (18:–) Just like Dr. Brooks proposal for communal investigation of Mars (8:485), NASA recognizes the power of presence with its New Millennium missions, the ultimate goal of which “is an armada of tiny spacecraft to give humans a virtual presence in deep space” (20:6). The benefit of SPACE is the collective presence of the swarm; the loss of any individual space vehicle is of secondary importance.
First, consider the effects of intangible Clauswitzian elements of luck, friction, the unknown, and danger to space operations. One author from the Center for Aerospace Doctrine, Research, and Education shows that “complex operations in space, regardless of how scientific, have been punctuated by surprises that underscore our inability to anticipate all contingencies, even when the only foes are human error and the inanimate space environment.” (9:11). He gives examples including the Voyager 2 study of outer planets, Magellan mapping of Venus, the decay of the Solar Maximus satellite, and the Mars Viking landers. He finds that “when we examine space operations, we find that they exhibit uncertainties that are not totally unlike those ascribed by Clauswitz to war.” (9:10) This “fog of space” may be mitigated by one of the big advantages of SPACE. Response to dynamic aspects of the environment is the precise quality documented as the forte of subsumptive architectures: “generating robust behavior in the face of uncertain sensors, an unpredictable environment, and a changing world.” (5:1228)
Not only can SPACE peer through the “fog of war”, it can do it better and faster than previous architectures. Military strategists speak of making decisions inside the time scale of the enemy’s decision loop. Making decisions on a fast time scale is the second strongest forte of subsumptive architectures (5:1228). One of the original authors in the area of subsumption architecture was initially intrigued that insects, with only primitive neural bundles, exhibited complex flight operation in real time (17:48). They’re tuned for fast response to critical situations. Think of a cockroach. They’re fast. They respond to what’s important. This automated, self-reliant, synergetic survival instinct is a desired quality in manmade subsumptive architectures, yet it is overlooked by the space community. One author writes that, “[ASAT] survivability techniques accomplish one of three purposes: they make the satellite or system hard to find, hard to hit, or hard to kill.” (28:28) But contrast this with the analogy of ants discussed earlier. They are not hard to find, they are not hard to hit, and they are not hard to kill. Yet their persistent antagonizing is hard to get rid of. That image of persistent presence is what SPACE could provide.
A second freedom would be the capabilities for graceful system enhancement or degradation. Binary failures of heavily invested space vehicles give planners fits. The long suspense time of getting on a launch schedule, the ambiguous failures of a vehicle experiencing long-term effects of space weather degradation, arguments of launch on need vs launch on schedule, all complicate the issues of simply replacing a broken satellite. With SPACE, architectural utility issues are simple and less stressful. There will never be a need to “Launch NOW!” because the loss of a single vehicle will only be a small fraction of system capability, and because of flexible lanch vehicles and orbits, no strong competition for facilities occurs when replenishment is desired.
Commerical ventures depending on large constellations recognize the importance of mission and hardware evolution. Motorola is on the verge of launching Iridium, a revolutionary constellation of communication satellites (25:–). Iridium is built with new processes and new management, not new technology. They are planning hardware obsolescence and mission changes on a two year time scale. "You do not achieve the lowest possible satellite size and mission cost by a priori assignment of requirements to the satellite design." (30:792) Let the satellite evolve to fit the goal. Motorola know that only the broad strategic goal will remain the same: to provide global communication.
SPACE exploits all three of the “critical technologies” envisioned and advocated by a panel of experts in Spacecast 2020, upon which SPACE and other future systems will depend (27:201). First, it exhibits high performance computing–not in the form of some massive, central processing unit, but rather in the form of distributed, parallel processors that are compatible with Third Wave technology. Secondly, it will use micro-mechanical devices introducing the distributed, subsumptive concept (29:120). Lastly, SPACE uses advances in material technologies to accomplish small size, integrated packaging, and low weight.
SPACE uses frontier technology recognized by others, but possibilities in the area of C2 set it distinctly apart. Imaginative use of SPACE depends on command of a constellation, rather than commanding a bunch of satellites. The synergy of a plentiful satellite architecture allows designers to investigate non-traditional C2 options that do not depend on the serialized 1-on-1 satellite upload/download umbilical cord binding current satellites. With an on-orbit constellation, transmissions from the ground don’t have to be to any vehicle in particular. Transmissions could be “in the blind” to the entire constellation. The space vehicles become comm-silent listeners in space, and thus the enemy is deprived of any SIGINT opportunities. Cross-links may exist between satellites so each other’s actions become a sensory input, but this may not be necessary. In some scenarios, the goal is accomplished by competitive behavior of many individual satellites. Enzymatic Satellite Pairing (ESP), introduced on page 16, is but one example.
In addition to eliminating traditional control, eliminating traditional tracking of individual satellites may be possible. We do not really care where the satellites are, as long as they collectively accomplish what they are suppose to do. They needn’t threaten later launches as “debris.” They could be directed to give clear path routes to other launches as easily as they could be directed to rendezvous and assist payloads. If they are used to actively de-orbit existing debris, the tradeoff for putting many in orbit becomes beneficial in light of what they remove from orbit.
For survivability and efficiency, self-initiated orbital maneuvers may occur. A low-thrust, high-Isp propulsion system could provide long-term propulsion, and a more traditional motor could be used on missions and in situations only as required. If we had to know the location of a highly maneuverable satellite or group of satellites, an echo ping from the ground could be initiated in a manner similar to an aircraft DME transceiver. For the slow or non-movers, tracking information could be obtained through the normal Space Surveillance Network sites.
The best illustration of remote sensing opportunities comes from a comparison to the current repertoire of weather satellites (GOES, DMSP, TIROS,). These huge, expensive satellites offer super high resolution and calibrated absolute sensors, but a lot less is required by most weather services. Almost all of the Air Force aviation activities are serviced with analog WEFAX-like transmissions. The broad coverage with course resolution is more than enough, and often desired for reasons of interpretability, refresh rate, and transmission time. For accessibility reasons, simple omni-directional, low bandwidth images are chosen over the high-data-rate signal, obtainable only with a narrow-beam trackable antenna. With modern technology, WEFAX-like images can be acquired and sent back to Earth with payloads weighing in at ounces instead of 100’s of pounds.
Large constellations of small satellites show tremendous promise for communication systems. The military’s recent moves in the area of communication satellites tend toward severe complexity (e.g. MILSTAR), even while the on-orbit workhorses are still mostly simple transponders. The operational systems, as opposed to research and test systems, appear to the user as unadorned as reflective spheres and dipoles (12:–). They give a stronger signal, but not more capability. Soldiers often do not need much more than repeaters above them. There is a tremendous market for horizontally integrated “low end” communication services--basically omnipresent repeaters-- above the battlefield. Omnipresence of simple communication capability is the guiding star of commercial systems on the horizon such as Teledesic and Iridium (25:–).
The advantages of futuristic tiny satellites capable of the same jobs as behemoth satellites are classical: no overflight conflicts, duration aloft, and survivability via remoteness. And the disadvantages of satellites such as revisit delays and launch cost are largely negated by tactical launchers of swarmed constellations. The characteristics of any one satellite are not new and different, but rather “the architecture of such a system is the novelty.” (27:195) The military might of such an architecture comes in the replaceability and pervasive presence of a system tactically launched or stored on orbit, ready to go. One satellite is cheap; many together are power.
The difficult political facet of these low cost missions is that they are low cost. Congressional budget handlers love them, but the companies depending on them for survival are concerned. Long program commitment times desired by the industry, are not issued by the government (24:–). “It’s not like making soft drinks where you can predict people are going to buy X number of bottles each year” (26:10). And the threat is that the military will not buy a steady stream of inexpensive satellites. With such short lead budgets and short coat-tails, these projects can be as inexpensively terminated as begun. A comprehensive military program embracing the new paradigm, gaining momentum, would give manufacturers motivation to innovate solutions.
Degrade or Destroy operations against a satellite often conjures up images of explosives or nimble, defensive dodging of an attacking satellite. But an alternative “...concept involves the development of a series of satellites designed primarily to protect high-value orbital systems. These satellites could escort the high-value orbital systems and provide protection to inhibit an adversary’s ability to detect, identify, track, or destroy our space assets.” (27:200). A similar idea to defend against natural threats such as meteors or orbital debris has been around for at least 5 years (23:5). SPACE would excel in this area.
Specifically, consider a “wolf-pack” technique: attack a large prey by overwhelming in number rather than size. Brilliant Pebbles, born of active SDI research in the 1980’s is a close analogy, but no comparable system is in place today. Even those “plans barely scratched the surface of possibilities.” (29:115) Brilliant Pebbles required many interceptors for many potential re-entry vehicles. It was a one-on-one relationship. The advantages in technology of today allow maneuverable, actively defensive satellites. The TAOS satellite launched by Phillips lab was a research prototype testing some of these capabilities (10:–). With intelligent defenses, a one-on- one attack may not be good enough; there are too many degrees of freedom that allow escape. With multiple attackers there may be no escape path. A destructive collision, shattering orbital debris everywhere, is not necessary. For instance, a target satellite could be actively towed or pushed into a decaying orbit. Less dramatic actions could include simply blocking view of the Earth or attaching an ECM pod to the target satellite.
ASAT often leaves a bad taste in the mouth of the American public because of pristine, virgin images of space. Although some avoid ASAT designs as ignoble or barbarian, the door is not closed on developing this capability. Many people consider Earth protection from meteors as a similar problem. and protecting the Earth from attack of natural debris or solar system visitors has gained respect from industrial, academic, and military think tanks (27:20).
First, consider a marine analogy. Huge ocean-liners meant for deep water navigation do their job on the high seas exceedingly well. Once in the bays, near port, they become subservient to local tug boats, which are responsible for the finer resolution navigation. The biggest barrier to space operations is simply getting into orbit. Space tugs could be the perfect hand-off from “transatmospheric vehicles for space lift and global reach.” (26:201) Any operations beyond low orbit could be done by “small brain-big burner” tugs. During a STSchool luncheon with Brig General Vesely, 14th AF Commander, this is the exact idea he mentioned.
An armada of space tugs would provide virtual C2, redundancy, and presence in ways never before possible. Imagine launching tugs that randomly visit sensed satellites in a way Dr. Brooks robots ‘randomly’ demonstrate intelligent search. There could be three or more types of tugs. One responds to a short milliwatt pulse from a laser diode on the host, another to a medium length, and another long length pulse. If no pulse is detected, the tug proceeds to other satellites. If it senses pulses, it latches onto the host satellite and escorts it to a polar, semi-sync, or geo-sync orbit. Now carry the hypothetical further. There is no need to use laser pulses. Different patterned paint splotches on the host satellite would be sufficient. There need be no communication between the satellites at all. At this point, the control link has gone virtual! There is no real link, yet the effect is as if there were.
For lack of other precedence, I’ll call this Enzymatic Satellite Pairing (ESP). The name is meant to carry connotations of the way biological reactions work, and of enigmatic communication paths. Based on key patterns on a host molecule, enzymes or antibodies bond and proceed to accomplish their task. And so the tug would bond to the host satellite, deposit it to a desired orbit, and press on with other tasks. The actual bonding could range from mechanical or magnetic to any exotic scheme that may come along. The tugs could be launched and maintain an on-orbit presence. Like an elevator, they could wait at certain floors (orbits) until required, or they could search out required tasking. If an item is large enough to latch onto or entangle in some manner, taking it to a lower orbit would be a natural extension of the proposal. When not tasked with lift priorities, the space tugs could be used to de-orbit space junk. Preserving the pristine nature of space becomes a politically correct use of space program moneys
We must find the analogous decisive doctrinal issue for efficient flight of space vehicles in a modern, information age. Ever since the glory of the Apollo program, we have enjoyed space vehicles technically triumphant to anything else created on the Earth. But as we enter the Third Wave, the information era, we need to adjust our concepts of superior. The thesis that superior weapons favor victory “...is insufficient unless the ‘superior arms’ are accompanied by a military doctrine of strategic or tactical applications which provides for full exploitation of the innovation.” (19:19) In the Information Age, instructions on how to use the tool are more important than the tool itself. The American military’s “biggest problem” is that “it let technology drive strategy, rather than letting strategy determine technology.” (29:11)
The advantage, compared to “Battlestar” satellites, is that SPACE satellites may be as expendable as bullets are to a rifle. Put simply, they are cheap enough to lose (14:669). Small satellites, when launched in constellations of ones or twos, can not capitalize on this insignificance (24:–). It is only subsumptive, populous constellations that suffer little upon losing a contributing satellite. If some operational doctrine under test fails, knowing that failure mode is the goal of doctrinal development. If a satellite burns up in the effort or runs out of fuel in a valiant attempt to keep up with sensory input, we have accomplished the goal. During an STS site visit to the Iridium manufacturing plant at the Motorola campus in Chandler, Arizona, we asked how many Iridium launches on the Chinese boosters could fail and the process still be seen as a success. His answer: They all could fail. Motorola understands that penetration of the Chinese infrastructure is a more important constellation issue than any individual satellite.
The mission (whatever that currently may be) is not the end goal. A visionary warrior sees new capability beyond the current missions and develops new doctrine to use it. SPACE is the right tool to use. It offers a new method -- a new methodology -- a new vision-ology, if you will. When the Wright brothers offered aviation to the Army, the Army flailed for thirty years determining appropriate missions and the doctrines. As we transition to the 21st century, I don’t want my nation to squander its possibilities of leadership in near-Earth orbital systems. SPACE gives the United States an opportunity to excel.
Think of survivability issues. This will come up again later. A piezo-electric actuator could move a mirror surface to modulate a laser beam scanned from an overhead platform. Mirror modulators are already used in the adaptive optics of the Starfire optical range outside of Kirtland Air Force Base. Col Worden, during an STS Round-table discussion, noted the American public will fund space research for one of three reasons: 1) They think the science is neat (but not after Apollo), or 2) They perceive a military threat (but not after the Cold War), or 3) They’re scared. ASAT-like Earth protection schemes may be funded under #3. iv