Slightly revised version of talk given on Nov 30 by Jay Hauben



Download 45.97 Kb.
Date25.04.2016
Size45.97 Kb.
#19335
(Slightly revised version of talk given on Nov 30 by Jay Hauben)
Anti-Aircraft Fire Control: Wiener and the Germ of Cybernetic Insight
In the last quarter of the twentieth century a new communications medium, the Internet, emerged with a growing effect on most aspects of human society. The Internet functions as a globally distributed interconnection of people, computing machines and communications media. The TCP/IP protocol upon which it is based makes the Internet essentially a self-organizing feedback system. This huge interconnection of humans and machines is often referred to as cyberspace. This is not coincidence. The Internet was inspired, nurtured and developed based on cybernetic insights.
In the actual events that launched and nourished the Internet, a prominent role was played by the American psychologist and engineer J.C.R. Licklider (1915-1990). Licklider envisioned a great leap for human society based on a tight coupling and networking of people and computers. He did much to infect others with his early enthusiasm. He also set in motion in the US a public sponsorship and funding mechanism that brought the communications network he envisioned into reality. In the 1960s, Licklider published seminal articles such as "Man Computer Symbiosis"(1) in 1960 and "The Computer as a Communications Device"(2) written with Robert Taylor in 1968. Looking for the intellectual roots of these papers, of Licklider's vision, and of the Internet itself, serious researchers have been drawn to cybernetics and the work of Norbert Wiener.(3) For the connection between Licklider and Wiener, see the article in this volume by Ronda Hauben. Here we will step back one step further to look for some sources of Wiener’s cybernetic insights.
Wiener began his graduate education hoping for a doctorate in biology. His eventual doctorate was however in the philosophy of mathematics in 1913. He then traveled to Europe where, then as now, the intellectual climate was more stimulating. However the First World War came and his father called him home.
When Princeton University professor Oswald Veblen gathered mathematicians in 1918 to help in the American war effort, he offered Wiener the position of computer. In the US at least, the people who in WWI and WWII calculated artillery range tables were called computers. Accurate tables required the point-by-point solution of differential equations and even required original application of new mathematical theories. Wiener enjoyed the company of other mathematicians and found the practical use of mathematics a confirmation of its importance.
Wiener began his teaching and research career at the Massachusetts Institute of Technology (MIT) in 1919. He distinguished himself with original contributions in mathematics and in the connection of mathematics with physical systems as in his study of Brownian motion. The 1920s and 1930s during which Wiener matured witnessed in the US and in much of Europe two significant network developments. One was for electrical power distribution. Such networks had to be carefully planned to avoid instabilities and power surges. Electrical engineers could model such networks mathematically but it often took two or more months to solve by hand the resulting differential equations. Wiener worked with the MIT electrical engineer Vannevar Bush who was building analog devices to solve the differential equations, which modeled these networks. These machines were able to solve the equations in less than one tenth the time. Wiener would say later that he witnessed in Bush’s lab the usurping by machines of human mental functions. Bush never patented any of his inventions, instead making descriptions of his work widely available. Bush also had a student assistant Claude Shannon whose masters thesis demonstrated a one-to-one correspondence between Boolean algebra and the components of the other important network development.
That other was the spread of dial auto-connect telephone networks. The dial system replaced human operators with electromagnetic relay circuits. The on/off switches processed the information entered into the system by the dialer to set up a path for voice exchanges. These robust, fast information processing systems caught the imagination of many thinkers and inventors like Konrad Zuse in Germany and George Stibitz and Howard Aiken in the US who sought to concentrate such switches to build relay calculators. Relay based telephone systems were reflected in the 1940s in the work of Warren McCulloch and Walter Pitts who were seeking to model the human nervous system.
During the 1920s and 1930s in Europe and the US, the introduction of new machinery and manufacturing processes, the spread of electric power and telephony, and military preparations especially naval gun control lead to the use of more monitoring and control devices. Particular use of such devices was in auto-pilots for ships and planes. Observers who saw for the first time saw auto-pilots automatically steering ships often remarked how uncanny it seemed that a ship was being successfully steered without human intervention. Companies like the Sperry Gyroscope in the US, the Siemens, the Askania and the Anschütz companies in Germany and the Royal Aircraft Establishment in the UK produced auto-pilot and auto-stabilizers of increasing sophistication that made possible the expansion of commercial aviation. Auto-stabilizers in particular relieved the pilots of having to make constant adjustments. They used gyroscopes to indicate level flight and reacted to deviations by adjusting surfaces to return to level flight.
The design and study of control systems and in some cases the communications needed to make the control remote, led to much practical and some theoretical experience. In many cases the practical work was based more on trial and error than on the budding theory. Hermann Schmidt (1894-1968) found in his work in the German pattern office that one third of the classes into which pattern applications got put contained inventions based on control technology. However, there was a confusing variety of terms used to describe the various uses of such technology. Wiener had some knowledge of control technology from his contact with the Servomechanisms Laboratory at MIT.(4)
The 1930s also saw apprehension in Europe of another war. Polish intelligence realized that the German naval code had become undecipherable. They assumed correctly that the messages were being encoded by machine. They set out to break the code but could not until the mathematician Marian Rejewski arrived at an equation that fully represented the Enigma machine that the Germans were using. Based on Rejewski’s equation, the Poles built machines to do the necessary work to decode the messages. In Czechoslovakia, Antonin Svoboda thought to combine analog circuits with optical input to better aim anti-aircraft fire. Svoboda was a mathematician and engineer trying to model the electrical distribution networks. He saw the possibility of applying his circuits to the defense of cities from air bombardment.
The vulnerability of the UK to aerial bombardment was answered by an appeal to science for a solution. Money and resources were poured into radar research.
Sadly, the war people feared came in 1939.
The problem of anti-aircraft fire control was a problem of immediate importance to the British because of the superiority of German airpower and the vulnerability of cities to aerial bombardment. Sensing that the war was threatening the accomplishments of European civilization, by late 1940 Wiener immersed himself in the study of the anti-aircraft fire control problem. Julian Bigelow, an engineer was assigned to Wiener as his assistant and partner. Wiener and Bigelow first took up to analyze the problem and determine if it could be represented and solved mathematically in a way that could be realized as Wiener would say “in the metal”.
Because of the great speed of an airplane, an anti-aircraft gun must be aimed ahead of where the target is at the time of firing. The amount and direction ahead must be estimated quickly and accurately, a task outside the natural human capacity. Where to aim is based on knowledge of how the plane has been traveling and where it is likely to travel in the time the shell takes to reach it even if the pilot takes evasive action. Wiener was able to contribute to the solution of this prediction problem partly because he had previously developed the equations to be solved when knowledge in one region is used to predict behavior in another (Hopf-Wiener). Wiener had also had close touch with the work at MIT with analog computers and had done work with one of his students Y. W. Lee building electric networks that performed desired operations specified mathematically.
Wiener and Bigelow analyzed the problem as a single system with human and machine components: the aircraft, the pilot, the tracking mechanism, a future-predicting mechanism and an aiming and firing system. They had in mind the early systems that had human not radar trackers.(5) Wiener argued that no system would make perfect prediction since the particular evasive action that the pilot might take could not be known in advance. But the pilot was constrained. Psychologically he wanted to get out of the danger zone quickly and he would not want to make maneuvers that might destabilize his plane. To Wiener this meant he could treat the pilot’s evasive actions by averaging over an ensemble of actual trajectories and finding the most probable. This was reminiscent of his Brownian motion analysis. The prediction problem then took the form of having as input trajectory data from the plane’s observed motion, taking into account the most probable evasive maneuvers, and producing as output where the plane was likely to be in a time in the future when the fired shell could burst in its vicinity. However there would be the inevitable errors introduced by the humans doing the tracking. Wiener and Bigelow recognized that if those errors could be treated as noise they could be filtered out.
Applying themselves diligently to the problem and calling into play much of Wiener’s deep understanding of mathematics gave them confidence they could solve the problem of more accurate anti-aircraft fire control. They expressed their theoretical solution to the real problem in the form of a mathematical equation similar to:

gh(t)= f(t- )dhW() t 0 [Masini, 1990: 185]



0

f(t-) is the input data from the tracking mechanism. Wh() is the weighting function that took into account the averaging over the possible paths and included the filtering of the tracking errors. And gh(t) is the predicted future position of the plane an interval h later. Solution of this equation for the unknown Wh() would not be easy. Wiener and Bigelow were confident they could solve it and build an apparatus that would perform the weighting Wh to consist of a cascade of Wiener-Lee electric circuits. But first they needed data on the actual paths bomber pilots took and on the actual behavior of human trackers.


Wiener, Bigelow and a technician Paul Mooney designed and built a simulator of the human activities required for piloting a plane and for tracking a plane. The simulator consisted of a projector shining a spot onto a wall or ceiling. The projector was given a motion that carried the spot across the ceiling and back with a rapid, smooth non-uniform motion. Human operators were brought in and given the goal to follow the spot using a lever coupled to a second projector. Bigelow had been a pilot. He took responsibility for making a mechanical apparatus that would give the lever the feel of an actual airplane controller. Wiener and Bigelow felt the apparatus with its lag simulated the stress pilots or trackers would experience. Both projectors were coupled to a trace recorder that recorded the motions given to the lever parallel with the motions given to the first spot. In this way Wiener and Bigelow produced curves that they could analyze. They compiled data that allowed the computation of auto and cross correlation functions. Each operator had his or her own characteristic behavior but all in common worked to achieve the goal by what appeared to be constant trial and error. Each would move the lever a bit one way, see the result, and then move it a bit differently to make the result more successful. Wiener and Bigelow had seen a very similar dance performed by the anti-aircraft guns that were controlled by the servomechanisms used to aim them. Servomechanisms are devices that monitor and stay true to a goal by sensing deviations from the goal and adjusting the input signal to diminish the deviation. Both the human and the mechanical situations were describable in the same way. In engineering terms both were acting as negative feedback systems.
The correction motion was at a much higher frequency than the average motion that the operators were trying to achieve. Such extra motion could be filtered out of the input f(t-) by a properly constructed Wh(). But Bigelow and Wiener still needed path data of actual bomber flights to compare with their experimental data and to use as input to test the anti-aircraft predictor they were going to construct. The best they could get was data at US military centers. Wiener and Bigelow visited these centers and studied the data. Using the data, they were able to arrive at their full solution to the anti-aircraft fire control problem.
Wiener and Bigelow prepared a demonstration of their anti-aircraft predictor. Observers of the demonstration remarked it was uncanny how the predictor could indicate one second before it happened where a curve would lead to. But what was needed was a successful prediction 20 seconds later since it took a shell about that long to reach a high flying plane. Comparing their predictions for 20 second leads to the results of simpler methods showed they had not achieved the great improvement they had hoped for. Wiener candidly suggested that his project no longer continue because the improvements he knew would be possible would require more time and resources than should be diverted from other war work.
But Wiener and Bigelow were excited anyway. They had seen in their work a very deep similarity between the behavior of the human operators in their simulation and the behavior of servomechanisms used for example in gun control. They also saw a deep similarity in the automatic radar tracked systems that were being developed at MIT and Bell Labs and the human nervous system (6). The automatic systems used tracking radar also based on feedback to generate the input that was processed by an analogue computer whose output was feed directly to the servomechanisms that controlled the gun aiming. The tracking radar and the human trackers played the same role, almost as if each were a servomechanism. Both introduced similar high frequency correction activity that could be filtered out. The totally automatic system with its sensory input, processing and aiming and firing output seemed almost life like.
As a mathematician, Wiener was struck by how similar were the auto correlations he calculated for the behavior of the operators of his simulator and for servomechanisms. Wiener and Bigelow began to suspect that negative feedback was involved in purposeful human action. To test this suspicion they wondered if the “hunting” motion that occurs when there is faulty or excessive feedback in a servomechanism had a counterpart in humans. They raised this question with Wiener’s collaborator and friend the physiologist Arturo Rosenblueth. He immediately confirmed that there was such a pathology called purpose tremor. A person with this disease might show no abnormality when at rest, but when attempting to lift a glass to drink from, it would swing wider and wider until the water had spilt. Rosenblueth also indicated that it is known that often the disorder can be traced to a malfunction of the person’s cerebellum, which controls voluntary muscular activity. Wiener reports that Rosenblueth’s answer confirmed for him that feedback played a large role in voluntary human activity.
This and other similarities led Wiener, Bigelow and Rosenblueth to consider whether goal seeking and purpose had a more general role to play in nature than they had been allowed and whether there was a generalization possible that had humans and servomechanisms as sub classes. Darwin had observed the variations within and between species and concluded that the boundary between humans and other animals was not sharp. Wiener’s work on the anti-aircraft predictor showed him human and machine behaviors that were similar. He wondered if these similar behaviors might be seen as manifestations of a general theory, a theory of control and communication in the animal and the machine.
Also, up until this work, the servomechanisms for the control of gun turrets were assumed to belong to power technology rather than communications technology. What dawned on Wiener and Bigelow was that the action of the motors could be conceived valuably as communicating the aiming parameters to the turret and hence that the motors and the computers controlling them could be treated as communications devices. Wiener wrote that this point of view made him "regard the computer as another form of communications apparatus, concerned more with messages than with power."[Wiener, 1956: 264] Viewing the striking analogy between the workings of an automatic anti-aircraft system and that of a living organism helped Wiener began to regard the brain and the nervous system and automatic machines in much the same light. Out of such considerations a new synthesis emerged which Wiener eventually termed cybernetics (from the Greek word for "steersman"). As the communications and engineering consequences of Wiener's new ideas were worked out, he began to predict that the series of analogies between the human nervous system and the computer and control systems would lead to the possibility of a very high level of automation.(8)
In 1944 at Princeton University, Wiener gathered a group of neuro-physiologists, communications engineers, and computing machine people for an informal session to layout some of this thinking. There was willingness on the part of the members of different disciplines to learn what others were doing and to see the striking similarities. Encouraged by this gathering, there was support for the launch of two series of similar interdisciplinary sessions, one in New York City and the other in Cambridge, MA. Wiener worked out his new synthesis in, Cybernetics or Control and Communication in the Animal and the Machine (The Technology Press, 1948; MIT Press, 1961) and later popularized it in The Human Use of Human Beings (Houghton Mifflin, 1950).
Wiener's work raised an important question. What should be the relations between humans and machines in the age of automation? He called for an "independent study of systems involving human and mechanical elements to decide which functions should properly be assigned to the two agencies, human and machine."[Wiener, 1964: 71] Wiener also worried that automation would lead society to unbearable unemployment unless it was carefully implemented with full concern for the working people. Communication was the unifying thread in Wiener's synthesis. He concluded that "communication is the cement of society. Society does not consist merely in a multiplicity of individuals meeting only in personal strife and for the sake of procreation, but in an intimate interplay of these individuals in a larger organism."[Wiener, 1965: 326] It was in the strengthening of this larger organism via the improvements in communications that his hope that the problems also generated could be solved lie. He therefore sought to "bring to the attention of all the possibilities and the dangers of the new developments."[Wiener,1956: 308]
After WWII, cybernetic ideas from many sources in the US, the UK, Germany and elsewhere, began to be known and discussed in scientific circles. Licklider attended the gatherings in Cambridge, Massachusetts. He brought to them his relevant experience gained from research in psycho-acoustics. His papers, mentioned above, carried on the work, giving his answers to the question of the relation between humans and computers and the importance of communication. Eventually, the insights that came partially from the work of Wiener and Bigelow on the anti-aircraft fire control problem fueled new thinking in most areas of human activity and thought. Licklider’s vision of an “Intergalactic Netrwork”, an interconnection of all the computing communities and people in the universe, can be seen in today’s Internet.
It is the basis of cybernetic insights in the real world system of humans and machines and the application of those insights to create the Internet among other human-machine systems that demonstrates the legitimacy and strength of cybernetics.
Bibliography:

Masini, P. R.: Norbert Wiener 1894-1964, Birkhäuser Verlag, Basel, 1990


Wiener, N.: I Am A Mathematician: The Later Life of a Prodigy, The

MIT Press, Cambridge, Massachusetts, 1956


Wiener, N: God & Golem, Inc.: A Comment on Certain Points Where

Cybernetics Impinges on Religion, The MIT Press, Cambridge, Massachusetts, 1964
Notes:
1) Licklider, J.C.R.:"Man-Computer Symbiosis", In: IRE Transactions on

Human Factors in Electronics, Vol HFE-1, March, 1960, pp. 4-11.

Also reprinted in: In Memoriam: J.C.R. Licklider: 1915-1990,

Report 61, Systems Research Center, Digital Equipment Corporation,

Palo Alto, California, August 7, 1990, pp. 1-19. (http://memex.org/licklider.pdf)


2) Licklider J.C.R., and Robert Taylor: "The Computer as a Communication

Device," In: Science and Technology: For the Technical Men in



Management, No 76, April, 1968, pp. 21-31. Also reprinted in: In

Memoriam: J.C.R. Licklider: 1915-1990, Report 61, Systems Research

Center, Digital Equipment Corporation, Palo Alto, California, August



7, 1990, pp. 21-41. (http://memex.org/licklider.pdf)
3) See for example, Chapter 6 in Hauben, M. and R. Hauben, Netizens: On the History and Impact of Usenet and the Internet, IEEE Computer Society Press, Los Alamitos, 1997, (http://www.columbia.edu/~hauben/netbook); and Chapter 8 in SEGAL, Jérôme: Théorie de l’information: sciences, techniques et société de la seconde guerre mondiale à l’aube du XXIe siècle Faculté d’Histoire de l’Université Lyon, Lyon, 1998, (http://www.mpiwg-berlin.mpg.de/staff/segal/thesis/).
4) For this period see e.g., Bennet, S.: A History of Control Engineering 1930-1955, Peter Peregrinus, London, 1993; Dittmann, F.: “Aspects of the Early History of Cybernetics in Germany”, in: Trans. Newcomen Soc., 71,1999-2000, pp. 143-154: and Mindel, D. A.: Between Human and Machine, Johns Hopkins Press, Baltimore, 2002.
5) See illustration between pages 214 and 215 in Baxter, J.: Scientists Against Time, Little Brown and Co., Boston, 1946.
6) See illustration on page 133 in Buderi, R.: The Invention that Changed the World, Simon & Schuster, New York, 1996.






Download 45.97 Kb.

Share with your friends:




The database is protected by copyright ©essaydocs.org 2022
send message

    Main page