Posts Tagged ‘Grey Walter’

M. speculatrix – Scanning: It makes all the difference

scanning – a form of behavior (sniffing, looking, listening, palpating) by which a sensory stimulus is sought or expected, and which is guided by an expectancy of input instead of a future internal state (goal). The search must be broad in the sense of looking everywhere, but narrow in the sense of being specific as to the characteristics of the anticipated stimulus, which implies that the search is guided by memory of past experience, as distinct from a predicted goal or automatic guidance by a set point. – from Biographical Sketch: W Grey Walter by Walter J Freeman, Encyclopedia of Cognitive Science (2003)
 

Discussions on Child Development (In one Volume 1971) Ed. by J. M. Tanner &  Barbel Inhelder, p31

When asked "How does it keep after the light ?"
GREY WALTER responded:
"If you have a scanning device, a rotating photo-cell, then once it gets on to a line it will tend to pick up that line again within a reasonable time. The same unfortunately is true of self-guided missiles. If a missile is aimed at London and interrupted by some counter-force, it is easy to make it take up the line again, after the perturbing force has been circumvented. This is characteristic of quite a simple system without storage."

and again p34-35

"Fig. 5 [not shown here] has a lesson for the philosophers among us, if there are any. This is the solution by a two-element model of what has been called the dilemma of Buridan's Ass. The dilemma, as conventionally stated, is that of a creature which does not possess free will, when it is faced with two exactly equal and equidistant stimuli; it is then not able to choose between the two alternatives and will die of hunger before it decides which of these two alternatives to accept as a stimulus.
This particular model has a scanning device in order to have freedom of movement. It also has a very interesting property in that it lives in Bergsonian and not Newtonian time. Newtonian time is reversible; the Newtonian solar system could be run backwards and it would be exactly the same as it is now. But Bergsonian time cannot be run backwards; all organisms live in Bergsonian time; time for us, as we all know, has an arrow. That arrow points to the grave;
that may be deplorable, but it has one important advantage, that with it we can solve Buridan's dilemma. We have the creature again starting at the top of the picture. There are two candles which are about equidistant ; they may be precisely so. The creature is released at the top and it happened to see the light on the left first. The spatial symmetry is not also symmetrical in time, so the effect of a scanning machine is to solve all dilemmas which involve symmetries in space. One of the signals must be seen first, and the one which is seen first is the one visited first. The creature explores the possibilities; when it gets close to the first light its moderation mode is invoked and it wanders across and does the same thing with the other light; it forms a figure-of-eight diagram. So our two-element model, which is essentially a reflexive mechanism richly inter-connected with a scanning device, can solve the dilemma of Buridan's ass, and would not die between sources of nourishment."


Wartime goal-seeking missiles (Craik, Wiener) vs Pavlov's Reflex approach (Grey Walter).

The Living Brain, W. Grey Walter (1953)

“ The first notion of constructing a free goal-seeking mechanism goes back to a wartime talk with the psychologist, Kenneth Craik… When he was engaged on a war job for the government, he came to get the help of our automatic analyser with some very complicated curves he had obtained, curves relating to the aiming errors of air gunners. Goal-seeking missiles were literally much in the air in those days; so, in our minds, were scanning mechanisms. Long before (my) home study was turned into a workshop, the two ideas, goal-seeking and scanning, had combined as the essential mechanical conception of a working model that would behave like a very simple animal.”

Given that inspiration, Walter was later to write

“ the model may be made into a better ‘self-directing missile’ by using two photo-cells in the usual way, but it will be a worse ‘animal’, for though it will keep more closely to its beam it will have to be aimed roughly in the right direction and will not ‘speculate’ – that is, spy out the land – nor will it resolve Buridan’s dilemma.’

Survey of Cybernetics ed. J. Rose (1969)

For many years, Wiener's thinking and my experimentation kept converging and overlapping, almost without any direct collusion or collaboration. His notion of scansion by the alpha rhythms matured at almost exactly the same time as we had observed some sort of space time transformation in analyses of real brain rhythms. Oddly enough, the artificial animals that I built at this time incorporated an elementary scanning receptor, white his model did not. My `tortoises' are not confused by the dilemma of Buridan's donkey, while his `moth' would trundle confidently midway between two targets, missing both. Wiener refers to some of the differences between the two models in both The Human Use of Human Beings' and in the second volume of his autobiography,2 but without apparently appreciating the significance of space time transforms as a possible resolution of the `determinism free will' dilemma. This oversight seems all the more remarkable because of his intense interest in the reduction of philosophic verbalisms to operational hypotheses.

From The Human Use of Human Beings Norbert Wiener (1950 ed.)

I have recently received a letter from Dr. Grey Walter of the Burden Neurological Institute at Bristol, England, in which he expresses interest in ’the moth’ or ‘bedbug,’ and in which he tells me of a similar mechanism of his own, which differs from mine in having a determined but variable purpose. In his own language, ‘We have included features other than inverse feedback which gives to it an exploratory and ethical attitude to the universe as well as a purely tropistic one.’  The possibility of such a change in behavior pattern is discussed in the chapter of this book concerning learning, and this discussion is directly relevant to the Walter machine, although at present I do not know just what means he uses to secure such a type of behaviour. The moth and Dr Walter’s further development of a tropism machine seen to be at first sight exercises in virtuosity, or at most, mechanical commentaries to a philosophical text.’

Later, from Weiner’s expanded 1952 edition

“here let me mention some earlier machines of Dr. Walter, somewhat similar to my ‘moth’ or ‘bug,’ but which were built for a different purpose. For these phototropic machines, each element carries a light so that it can stimulate the others. Thus a number of them put into operation at the same time show certain groupings and mutual reactions which would be interpreted by most animal psychologists as social behaviour if they were found encased in flesh and blood instead of brass and steel. It is the beginning of a new science of mechanical behaviour even though almost all of it lies in the future.’

from The Physics and Chemistry of Life, A Scientific American book, (1956), p254

….. In a TV set, scanning of the field goes on continually whatever the picture may be; in certain radar sets designed to control artillery and in many target-seeking projectiles, a scanner  is set to search for targets, but once an echo has been received the scanner stops and swings the gun or missile into a position of best attack. This simple system was incorporated into the toy robot, Machina speculatrix, which we made years ago to see how "scanning" would affect behaviour. In systems such as these, the more active and excited the system is, the less regular and rythmic the scanning cycle becomes. So perhaps within our heads we carry a bundle of target-seeking tissue-in origin primeval, but in function as penetrating and as precise as any imagined, even in the realms of science fiction. Here we can discern at work the organ of selection and imagination, first stages on the road to learning, understanding and foreseeing the shifting patterns of the outside world-and all contained in a cupful of tepid, pinkish-gray, electric jelly.
…..

from The Tortoise Against Modernity, Andy Pickering (2001)

The tortoise was thus a homing device, capable of pursuing its target, a light source, in a cluttered and difficult terrain. Here we can recognise a version of the basic cybernetic connection between purpose and feedback from the environment, and one is immediately reminded of the primordial referent of Norbert Wiener’s cybernetics—the autonomous antiaircraft weapons system Wiener sought to construct during World War II (Galison 1994).8 Two differences between Walter’s robots and Wiener’s are worth noting. First, the tortoise was more active and lively than Wiener’s device; the tortoise would prowl around, scanning the horizon for lights, while Wiener’s autonomous weapons merely waited for a target to present itself. And secondly, while Wiener’s ‘guns’ simply lined themselves up on a target, the tortoises displayed unusual and unexpected forms of complex behaviour.

Sep 27 1956?. [WGW response to letter.]

“….You are quite correct in your interpretation of the mechanical lay-out. I am afraid it is not as effective to drive from the rear wheels since this does not provide scanning through 360° together with movement towards the source of light when the scanning stops.  You will realise what I mean if you consider the situation when the creature is moving forward and the only available light is directly astern. If the drive is on the rear wheels only then as  the scanning proceeds, the front wheel will get into position when it is at right angles to  the direction of travel so that no movement is possible and will then rotate further until the model is moving say hard aport with the photocell looking backwards. If it then catches sight of the light, it will continue to move away from the source in a circle and will not tend to pull itself in on the beam. With the arrangement as described however, signals will be picked up from any bearing and the model will eventually always pull in and approach the source of moderating light.

1956 [WGW response to letter from Australia.]

…”Thank you so much for sending me your letter [in archive] and the sketch of your little automaton. As far as I can gather, your ideas are quite independent of mine in detail but seem essentially convergent.
Our models also scan 360° and a light signal turns off the scanning motor just as in yours; the photocell and front wheel being on the same spindle. I do not however, turn the other wheels so that the model has an interesting cycloidal gait.
There is a fairly full description of the general scheme in my book “The Living Brian” published by Duckworths in 1953, Appendix 2, but not a detailed photograph.
I think our drive system is simpler than yours since the motor turns with the whole steering gear and photo-cell but as I say, the principle seems the same and it would be rather interesting to discover whether this may be in fact the simplest method of making a freely moving self-governing target-seeking device.
Our toys are a little bit more complicated in another sense; the amplifier has two stages and the effect of bright light which is sufficient to operate a relay in the first stage and shows that the model will diverge from a really brilliant source. The avoidance of obstacles is also achieved by connecting the output pack to the input of the amplifier so that it multivibrates slowly.”


Scanning

From  Biographical Sketch: W Grey Walter,  Walter J Freeman, Encyclopedia of Cognitive Science (2003)
 

An autonomous robot embodies the principles of goal-seeking and scanning that characterize animal behaviour. Grey Walter used his wartime exposure to radio detection and ranging (RADAR) to build a simple 'brain' that endowed his artificial 'turtle' with complex adaptive behaviors.
Biomedical Engineering
In the decade before the war Grey Walter had already done important work in a field we now call biomedical engineering by his discoveries in electroencephalography (EEG), a medical procedure in which the oscillating fields of electric potential on the scalp and in the brain are measured and interpreted. His first achievement was to identify correctly the source of the alpha rhythm (8-12Hz).
During World War II Grey Walter was involved in radio detection and ranging (RADAR). This experience with electronics  enabled him to build equipment for EEG research at the Burden Neurological Institute laboratory in Bristol.. One of their achievements was automated spectral analysis of EEG traces. Walter used his skills in analog electronics to conceive a device built by engineers that displayed the frequency content in an EEG trace, Walter extended his temporal spectral analysis of time series to spatial analysis by conceiving a bank of amplifiers connected to an array of 22 oscilloscopes. This advance enabled him to show not only the amplitude but the phase difference of each trace of the alpha with respect to the others, by using cinemas of the oscilloscopes. With his 'toposcope' he visualized the spread of alpha waves across the surface of the brain. Walter proposed that the alpha represented 'scanning' by the brain in search of local centres of activity when none was present, and that it stopped when a 'target' was found in the cortex.

Autonomous, adaptive robots
Walter's greatest achievement stemmed from his wartime experience with electronics (Figure 1). Guided missiles with proximity fuzes then were one of two very active foci of interest, the other being devices for scanning the horizon for targets to be identified and intercepted. The scanning mechanism he helped develop was known as the 'plan position indicator', consisting of the point of light created by an electron beam that moved from the center to the edge of the oscilloscope screen and created a bar like the spoke of a wheel. The spoke rotated counterclockwise at the refresh rate of the screen. A likely radar target appeared as a bright spot, giving its direction and distance. This device is in widespread use today, for example, in ships, submarines, and air traffic control centers. It was the basis for Walter's toposcope when applied to alpha waves.

The concept of a machine that would define a goal and seek it by scanning resonated with his interest in brains as biological systems that evolved through learning from the consequences of their own goal-oriented actions. He undertook to incorporate these two cognitive operations, goal-seeking and scanning, into an electronic 'toy' that would simulate these most basic characteristics of animal (and human) behaviour. The outcome was fully spectacular, though at the time its significance was not recognized, and his device has been all but forgotten. It was a roving machine so life-like, as he described it in his book, "The Living Brain", that an old lady who felt pursued by it ran upstairs and locked her door. He named his device Machina speculatrix in order to distinguish it from passive devices, such as his earlier conception of M. sopora that incorporated Norbert Wiener's principle of stabilization of machine performance by negative feedback ("Cybernetics"), and from W. Ross Ashby's 'Homeostat' that extended the principle of the stability of biological organisms by introducing adaptation through learning. These stable models used what the Harvard physiologist Walter Cannon called "homeostasis", but unlike plants and sessile automata, M. speculatrix was continually on the prowl in search of its designer-endowed goal: moderate illumination. His threewheeled vehicle, which came to be known as "Grey Walter's turtle", had two motors, one for progression by the front wheel dragging  trailing the hind wheels like a child's tricycle and one for turning the front wheel. Its drive system, batteries and 'brain' were mounted on a chassis. Above it he hung a carapace (shell) from a centre pole, so that it could swing inwardly from any direction and contact the chassis. This contact operated a switch, so that if the tortoise hit an obstacle or encountered an incline, it would stop, back up, turn, and eventually move around it or avoid it altogether. This 'receptor' gave the device the sense of touch, information about the direction of gravity, and the means to explore objects in its environment by touch. In a single charge cycle it could explore nearly 100 square meters, dealing with obstacles by pushing them aside or going around them, though sometimes straying too far and being found "starved to death" behind a couch. When it regained its hutch it turned itself off and took nourishment from electrical contacts on the floor. As a means for detecting the internal state, Walter fixed a marker light on the carapace that stayed on when the turning motor was on but went out when turning stopped. When the tortoise encountered its own light in a mirror, it stopped and oriented to its own light, but stopping turned out its light . Then it resumed circling, saw its light again, and stopped. This behaviour continued until it had passed the mirror. If it encountered another of its own kind, attracted by the other light, a stately dance ensued of bumping and backing. Walter thought that these behaviours expressed self recognition and recognition of conspecifics. These complex and not fully predictable behaviours of exploration, negative and positive tropism, discrimination, adaptation to changing internal and external environments, optimization, and stabilisation of the internal medium were done with a very simple brain: two miniature valves serving as 'neurons', two mechanical relays, two capacitors, two receptors, and two motors.

Evaluation and Summary
They were the first free-ranging, autonomous robots capable of exploring their limited worlds. The essence of an intelligent machine is that it has within its brain a capacity to conceive desired future states, and it has the degrees of freedom needed to create and adapt its actions in pursuit of those goals in the unpredictable circumstances of the immediate and remote environments. These flexible brain functions that enable simple systems to function in infinitely complex environments are not achieved by rule-driven symbol manipulation, which is at the heart of cognitive science and conventional artificial intelligence. Moreover, Walter emphasized analogue electronics to simulate neurodynamics at a time when most of his colleagues such as John von Neumann were developing digital computers to implement symbolic logic and deep arithmetic algorithms. His devices were the forerunners of currently emerging machines that are governed by nonlinear dynamics, and that rely on controlled instability, noise, and chaos to achieve continually updated adaptation to everchanging and unpredictable worlds. He can well be said to have been the Godfather of truly intelligent machines.


". . . the driving wheel is fixed at whatever angle it was when the light was seen, and the scanning of the horizon by the eye also stops of course. At the same time. . . the driving motor is turned up to full speed. The model stops looking slowly round and hurries toward the light. However, unless the light was seen when the eye happened to be facing straight ahead, the angle at which the steering came to rest at the moment of sighting will deflect the model gradually away from the light. When the deflection is so great that the activation level of the photo-cell falls below threshold, the Relay 2 opens again, the scanner starts up, the drive is reduced to half speed and the model is re-positioned, this time so that the light is more directly ahead. This process of progressive orientation is an important part of the behaviour mechanism. It is cumulative-every time the model steers itself slightly off-beam the momentary operation of the steering-scanning mechanism brings it back more nearly on course and it ends up with a heading on-beam. The process often looks clumsy, because the eye seems to veer away from the light and then the scanner has to make nearly a whole rotation to bring it back, but inevitably with each such operation the model gets itself into a better position to bear down directly on its goal. The aiming-error is steadily reduced as the goal is approached. " Walter (1960)


Better than Scanning – Random

On several occasions WGW mentions an improvement of the rotary scanning approach.  In a response to Ivan Sutherland’s letter dated 7th Jan 1960, WGW says 

“ As an extension of this principle [rotary scanning] it is interesting to consider what would be the consequence of producing a purely random rather than rotary scanning so that the angle and therefore the heading of the creature  were a purely random function all the time. Observation  of certain animals suggest  that their movements might be considered as a random vectorial  addition and this suggestion might be more realistic in comparison with real creatures than in the rather monotonous rotary scanning of M. speculatrix.”

As it turned out, scanning as method for both drive and steering proved difficult to construct by others that were inspired by Grey Walter's tortoises.  As I've added new blogs to the rest of the cybernetic zoo , this feature has become apparent. There is an article written by Ivan Sutherland solely about steering ["Stability in Steering Control", Electrical Engineering April 1960]. I have reproduced this article here.


Scanning Analysed

Before visiting the Bristol Robot Labs in 2009, I was  in dialogue with Ian Horsfield who provided some footage of one of their replica tortoises in motion.

I picked some points to plot, the centre of the photo electric cell, and a point just forward of the rear wheels in an attempt to reproduce the original trace photos of Elmer and Elsie. The original Elsie scanned in a counter-clockwise direction, and Elmer in a clockwise direction.  The geometry of the trailing wheels is such that when scanning in the counter-clockwise direction, the tortoise creeps to the right. It cannot scan and remain stationary.  The two replicas both scan in the counter-clockwise direction.

A similar time-lapse pattern as seen in the original ELMER/ELSIE photos.

A manual plot of some points during a scanning cycle showing the rear wheels according to the direction of the scanning (rotating) photo-electric cell.


Kenneth Craik had a formative influence in modern experimental psychology and was one of the first to explore principles common to brains and machines.

Craik was only thirty-one (in 1945) when he was tragically knocked off his bicycle, by someone carelessly opening a car door, and run over.


TRIVIA:

In the reference discussions on child dev , when asked about the name of Grey Walter's personal tortoise, he responded:

"This one is called Olga, the beautiful spy. This spying machine was made for this type of demonstration with a transparent and rather glamorous coat, and a well-proportioned body!" 

    On several occasions Grey uses the term 'spying' instead of 'scanning'.

    The only reference I can find  to a spy called Olga c1940's is Olga Chekhova: alleged Hitler's Favorite Actress and a Russian Spy.


    Tags: , ,

    1956 – Mechanical Animal – William Robert “Bert” Sutherland / Ivan E. Sutherland – (American)

     This copy of a letter from 1957 describes the first "Mechanical Animal" built by the Sutherland brothers, Bert and Ivan.

     

     

    Here’s a transcript of the letter sent from Ivan E. Sutherland to Grey Walter in 1957:

    Nov 10. [IES  to WGW]
    “Dear Sir:
    Early last month I had sent to you two copies of a paper entitled “An Electro-mechanical Model of Simple Animals” which was submitted by my brother, Bert (William R. Sutherland), and his close friend, Mac (Malcolm G. Mugglin), to their department of Electrical Engineering. Perhaps a little of the history of that paper would be of interest to you.
    I am now a Junior (3rd year) at Carnegie Tech, also studying electrical engineering – in this and many other things I have followed the lead of my brother. Bert is two years older than I, recently became married and is now on active duty as an officer of the U.S. Navy. Our interest in mechanical and electrical things probably comes from our father, a Civil Engineer from New Zealand: Ph. D. from London, but our first good luck and stimulation came when we met Edmund C. Berkeley in 1952.
    Mr. Berkeley took an interest in the work that we had already done, namely a simple adding machine, and encouraged us to continue, both by suggesting problems and by providing funds for their solution. During the period October, 1952 to June, 1955 we worked under the guidance of Mr. Berkeley. We did a major protion [sic] of the work on a mechanical maze solving mouse similar to one constructed by Claude Shannon of Bell Labs. During the latter part of this same period, Bert left home for college, and I continued our work alone.
    During this contact with Berkeley’s organization we often saw “squee”, his mechanical squirrel; this was our first contact with the species of mechanical animals. Our next contact came when we read your The Living Brain. We were both interested in all the things you have done, but most familiar with the mechanical and electrical aspects, and most interested in your Machina speculatrix. Can you imagine the joy of two young people reading about important work accomplished far away in a field they were just becoming part of?
    It was no surprise to me when Bert suggested, about Christmas of 1955, that we build a mechanical animal also. On page 45 of Bert’s thesis is a picture of the first crude result. When this first model was finished, about May, 1956, Bert for some reason lost interest in the project for a time. During this period, May to December, 1956, I continued work on the second model, the one which finally became the subject of the paper sent [to] you.
    About Christmas 1956, Bert decided to write his thesis. By the end of January I had finished making the frames, motor mounts etc for the models shown in the various pictures; these Bert took over, assembled and used as a basis for his work. Mechanically these machines were good; electrically they were incomplete, as the thesis shows. They had two big drawbacks however: the wet battery needed constant care, and by the way cost us many pairs of pants through acid holes; the machines were cumbersome and heavy.
    At the moment, Bert is busy with his new wife and the Navy, so I am in charge of our project. To get around the two drawbacks mentioned I have constructed a third type of beast. This new model, commonly called “beastie” because of its smaller size, uses dry cells for power, is entirely operated by transistors and proves to be the best we have yet accomplished. However, although I have the mere construction problems fairly well met, I have not yet obtained any results from this latest model The problems which were not yet solved in when Bert’s paper was written are still not solved.
    Perhaps by now you are wondering just why I should write this letter. It is a sort of news report, an information carrier rather than a questionnaire. I examine what we have done: we have a rather nice looking machine which will respond to light and avoid obstacles in a rather crude sort of way. We have a great many possibilities for future work. I examine what I think we should do next: proceed with communication and learning as interesting behaviour. Perhaps making the machines (I’d like to build more of the “beastie” type) play tag might be a good start. We need a better obstacle strategy.
    Building these machines has been, to say the least, an education in itself. I have found time and time again that to us the problems of actual design and construction were fairly straightforward; the decisions such as I face now of what to do next are more difficult. Perhaps you have some ideas. I am, of course, curious to know what you think.”

    CHALLENGE

    I've tried to track down Bert Sutherland's thesis to obtain a picture and further details of this "beast", but without success. Maybe an American out there could find this information and I would happily publish it here. From the above article, the thesis was completed early 1957.

    Grey Walter’s Tortoises – the video clips

    In my research for all things Grey Walter and his tortoises, I have uncovered five (5) video clips available on the internet.

    Of the five, I have downloaded four of them, the 5th has been allusive for some time, having not been able to re-locate it again after spending many hours trying. It is not so bad, as this video clip is a newsreel clip of the IBM exhibition "A Computer Perspective" and is a walk-by of the static model. I took a still image of it at the time, though.

    The other four I will try to re-locate on the web as I did not keep the link when I downloaded them sometime ago.

    When I describe tortoise behaviours in a future post, I will refer to these video clips.

    (not re-located on ina.fr yet!)

    A "Computer Perspective" film was produced (1971). Its been loaded into Youtube by the Eames Office. Thanks to Domenico in notifying me of this. The  tortoise is 7:22sec into the  clip.

    TRANSCRIPT:

    Bristol's Robot Tortoises Have Minds Of Their Own

    In a simple villa on the outskirts of Bristol lives Dr. Grey Walter, a neurologist, who makes robots as a hobby. They are small and he doesn't dress them up to look like men – he calls them tortoises. And so cunningly have their insides been designed that they respond to the stimuli of light and touch in a completely life-like manner. This model is named Elsie and she "sees" out of a photo-electric cell which rotates about her body. When light strikes the cell driving and steering mechanisms send her hurrying towards it. If she brushes against any objects in her path, contacts are operated which turn the steering away, and so, automatically, she takes avoiding action. Mrs. Walter's pet is Elmer. Elsie's brother, in the darker vest. He works in exactly the same way. Dr. Walter says that his electronic toys work exactly as though they have a simple two-cell nervous system, and that with more cells, they would be able to do many more tricks. Already Elsie has one up on Elmer. When her batteries begin to fail, she automatically runs home to her kennel for charging up, and in consequence can lead a much gayer life.

    TRANSCRIPT

    Now meet Dr Grey Walter of Britain . Why the torch? Well, here's the reason – its Toby, a mechanical tortoise with an electronic brain which functions like the human mind. Toby's  head, or rather 'magic-eye' is a photo-electric cell constantly revolving until it picks up the strongest source of light, to which it is then attracted. In this case an ordinary electric torch guides the mechanical tortoise in any direction its inventor chooses. It can also negotiate obstacles.  When it hits an object, the pressure on the shell causes a short circuit of the photo-electric cell mechanism, and the tortoise moves at random until it is free of the obstacle. With a stronger source of light placed in position, Toby is attracted to the lamp  in the same way as a moth is attracted by light. Now, the front wheel of the tricycle undercarriage which is coupled to the photo-electric cell motor is turning on its axle while its two back wheels remain static and the tortoise attempts to get still closer to the light. A small syringe is being used to inflate the tyres, and with its shell removed, the inner workings of the complicated mechanism of Dr Walter's brain-child and the immediate affect of light on the magic-eye can be seen. Toby's probably getting tired and hungry by now for light to Toby is like food to any ordinary animal. And that light in his hutch never fails to bring him home,  without a torch, too.

    Note: It is interesting that the tortoise is called "Toby" by the narrator. I don't know if this is journalistic license or whether, in fact, after the "CORA" circuit was cut out, the tortoise was actually called "Toby" thereafter.

    This segment is from the movie "Future Shock". It is 17:27 minutes in. The unit next to the tortoise is not CORA, but another unidentified model, used as a prop for obstacle avoidance.

    TRANSCRIPT: [Orson Wells – Narrator]

    Step by step, the body parts grow disposable-like products we use and discard. This quaint English village, a remnant of permanence in a    future-shocked world  is the home of neurophysiologist  Grey Walter . He's one of the many scientists leading us towards the ultimate replacement –  Artificial Intelligence. Twenty-five years ago he pioneered the development of behaviour machines.

    [WGW] This looks rather as though it was a childs'  toy, and I suppose it might be, but in fact it's a rather serious model of my ideas of behaviour. And it behaves in a complex way with all kinds of behaviour modes only having two elements compared with our ten billion in our brains, but its behaviour is finely?  complex. Now you see it hesitating a moment,  and then the body out of ?  and on its way slowly and by a rather devious path, right into its hutch down here. And so, rather like us, it has a sense of ?? which although its such a very simple  toy, but not really just a toy, a model of behaviour . END

    (in French)

    Site is quirky. http://www.gaumontpathearchives.com/index.php?urlaction=docListe  You need to register to view the clip. Last time I tried registration took around 24 hours. If you still can't see it in the English version, use the French version. Search criteria is CONGRÈS DE LA CYBERNÉTIQUE TORTUE . Date of clip = 18/01/1951: running time =
    57 secs.  Clip also shows the chess automaton of Torres y Quevedo.

    CHALLENGE: If a French speaking person would like to offer a translated transcript, I will publish it here.


    See also my other posts on Grey Walter and his Tortoises here.

     

    Tags: , , , , , ,

    1951 – La Tortue Cybernetique (Cybernetic Tortoise) – Paul-Alain Amouriq (French)

    In late 1951, Paul-Alain Amouriq, a Frenchman then aged 17, built a cybernetic tortoise inspired by Grey Walter's as published in a French science magazine Science et Vie (February 1951). Several years later Science et Vie became aware of Amouriq's tortue, and Pierre de Latil visited him and the subsequent article was published in the March 1954 issue.

    It uses all-relay logic, no vacuum-tubes are used at all.

    Photo courtesy A.-P. Amouriq 2009.

    STOP PRESS – 9th Oct 09 –Wonderful news – Paul A. AMOURIQ saw this post about his "tortue" and posted the first ever comment to my new blog. Thanks Paul.

    He has sent through some further information, plus a picture of what 'tortue' looks like now after some technology modifications. See new info at bottom of this post.

    Note: The picture of the Tortue below is facing to the right. The single steering wheel is at the rear of the tortoise.

    Photo courtesy A.-P. Amouriq 2009.

    Attirée par la lumière et sachant contourner l'obstacle, la tortue de P.-A. Amouriq est le mieux agencé des robots cybernétiques.

    P.-A. AMOURIQ REPLACE LES ROUES AVANT DE LA TORTUE, QUI SONT SIMPLEMENT PORTEUSES

    UN LYCÉEN A CONSTRUIT UN ANIMAL ARTIFICIEL


    Un animal artificiel de plus, et fabriqué par un jeune homme de dix-sept ans ?…
    Simple « bricolage » d'amateur imité des précédents, sera-t-on tenté de penser.
    Mais d'abord quand Paul-Alain Amouriq, alors élève de « mathelem » à Louis-le-Grand (il prépare aujourd'hui sa licence ès sciences) le construisit, c'était il y a deux ans, et nul n'avait encore donné de descendance aux fameuses « tortues » de Grey Walter.
    Ensuite cet engin autonome et automatique n'a rien d'une improvisation. Bien au contraire, il est calculé ; c'est le mieux construit, le mieux agencé des diverses « tortues », le seul qui ne soit pas réalisé avec de simples pièces de « meccano ». Il faut préciser que le lycéen bénéficia du concours d'une grande firme d'appareils de mesures électriques dont son père est directeur.
    Tout entière usinée en duralumin, sa machine n'est pas une simple copie améliorée des « espèces » antérieures d'animaux artificiels. Elle présente plusieurs dispositifs originaux.

    Une inspiration puisée dans « Science et Vie »

    C'est en lisant l'article consacré par Science et Vie aux tortues électroniques de Grey Walter que P.-A. Amouriq, comprenant tout l'intérêt de ces premières applications de la cybernétique balbutiante, voulut réaliser une nouvelle «espèce ».
    Son engin aurait trois roues dont une à la fois directrice et motrice. A la différence d'Elmer et d'Elsie, premières nées de l'espèce Machina speculatrix, la roue directrice ne serait pas à l'avant, mais à l'arrière. Ainsi les virages seraient mieux « pris », l'engin risquant moins d'accrocher, par exemple, le chambranle d'une porte au-delà de laquelle il vient d'apercevoir une attirante lumière.
    Deux sensibilités : un sens nuancé, la vue, et une sensibilité plus fruste, celle des chocs.
    Les organes de la vue sont deux cellules photoélectriques qui balaient l'horizon.

    Qu'il en existe deux et non une seule, peut donner des réactions beaucoup plus subtiles aux perceptions visuelles.
    La somme des courants perçus par les deux cellules représente une appréciation de  'intensité lumineuse ; à cette fin, les cellules sont montées en « parallèle », mais leur montage « en opposition » confère aussi au robot l'appréciation de la direction de la source lumineuse.
    L'influence de la cellule de droite l'emporte si celle-ci perçoit plus de lumière que la gauche ; ou inversement. Si la lumière se trouve juste entre les deux « yeux »,  les deux influences se balancent. Quant à la sensibilité aux chocs, elle se traduit par une augmentation de l'intensité dans le moteur de locomotion brusquement bloqué, puisque les roues ne peuvent plus avancer. Voilà donc la bestiole conçue, sensible à l'intensité et à la direction de la lumière ainsi qu'aux chocs contre un obstacle. Restait à décider de quelle façon elle réagirait — car une machine ne pourra jamais « connaître » que par les sens et agir que par les organes dont nous l'aurons dotée.
    Paul-A. Amouriq gratifia son engin d'actes simples : avancer à plein régime ou à demi-régime, tourner, ou reculer (un seul régime).

    Un comportement très avancé 

    Voyons comment ces actes sont reliés aux sensations que provoquent, par leur présence et leur absence, la lumière et les obstacles. Admettons qu'il n'y ait pas de lumière. C'est un cas particulier d'une perception équilibrée des deux cellules. Le moteur de direction, lequel est naturellement sensible au déséquilibre des perceptions lumineuses, n'entre pas en jeu et la marche est rectiligne. Mais l'absence de courant dans le circuit affecté par l'intensité lumineuse détermine l'introduction d'une résistance dans le circuit des batteries alimentant le moteur de direction ; ce courant d'alimentation faiblit donc, et la marche est ralentie. En même temps, une lampe s'allume à l'avant du robot. Elle symbolise un état dé prudente investigation. D'ailleurs, si la lampe est assez puissante, elle peut par son reflet avertir de la présence d'un obstacle.

    Recul devant l'éblouissement, mais sur choc, demi-tour

    Admettons maintenant qu'il y a au moins une lumière. L'animal se dirige alors vers elle. Mais, dès que l'intensité lumineuse devient trop forte, il recule jusqu'à ce que l'intensité de cette lumière (ou d'une autre) soit de nouveau attirante.
    Les cellules photoelectriques — du moins celles utilisées ici — ont un défaut : elles sont moins sensibles lorsqu'elles perçoivent de la lumière depuis un moment. Mais ce défaut s'est trouvé salutaire : il confère à l'engin une nuance de comportement hors programme : l'animal domine en partie sa réaction de recul et s'approche de plus en plus du danger.

    En cas de choc

    Que se produit-il dans le cas d'un heurt ?
    L'excès de courant dans le moteur de locomotion commande immédiatement la marche arrière, ainsi que l'inversion du courant des cellules. Cela pendant cinq à six secondes — le temps de parcourir 30 à 50 cm. Puis c'est de nouveau la marche avant, mais le parcours n'est plus le même.
    Si l'obstacle est heurté pendant un recul, alors intervient une désensibilisation à la lumière.
    Pendant quelques instants l'engin marche en avant à la recherche de l'obscurité, et ce n'est qu'après un repos d'une dizaine de secondes, dans un état d'équilibre des deux cellules, que la marche vers la lumière reprend.

    PLAN RÉEL DE LA TORTUE DE P.-A, AMOURIQ

    Simplicité et sensibilité

    Regardons maintenant l'ensemble du mécanisme : il ne comporte aucune lampe de radio, aucun amplificateur, mais seulement des relais d'une extrême sensibilité, mis au point par le père du jeune cybernéticien. Ainsi l'engin est-il beaucoup plus solide et possède-t-il" un comportement bien plus stable que ne l'était celui de ses devanciers.
    L'ensemble, fixé sur une plaque de bakélite, est facile à démonter du châssis. Nous ne ferons pas un parallèle entre ce robot et les tortues de Grey Walter. Du moins, grâce à la possession de deux « yeux » se comportet-il plus comme un animal que ne font les tortues de Grey Walter dotées d'un seul oeil tournant dans un seul sens, ce qui fait qu'elles ne peuvent tourner que dans un seul sens, contourner un obstacle que d'un seul côté. Avec deux yeux dotés d'un va-et-vient symétrique, la marche vers la lumière est plus décidée ; elle n'est pas pour cela absolument rectiligne car toujours, dans la pratique, l'influence d'une cellule l'emporte sur l'autre, ce qui détermine de très légers zig-zags correspondant exactement aux corrections continuelles par lesquelles procède la marche (et en fait tous les gestes) des êtres vivants.
    Quant à l'alimentation automatique, elle n'est certes pas réalisée dans ce robot.

    Mais ceux de Grey Walter ont démontré une fois pour toutes qu'un tel mécanisme était possible. Il est secondaire désormais de vouloir, au prix de complications mécaniques, le reproduire.

    Pierre de Latil

    SCHÉMA DES ORGANES DU COMPORTEMENT

    Using Google language translator:

     

    Attracted by the light and knowledge around the obstacle, the turtle P.-A. Amouriq is better organized cyber robots.

    P.-A. Amouriq REPLACE THE FRONT OF THE TURTLE WHICH ARE SIMPLY CARRIER

    STUDENTS TO CONSTRUCT AN ANIMAL ARTIFICIAL

    An animal more artificial, manufactured by a young man of seventeen years? …
    Simple 'DIY' amateur imitation of precedents, will he be tempted to think.
    But first, when Paul-Alain Amouriq, then a student of "mathelem" Louis-le-Grand (he now prepares his BSc) built on was two years ago, and no one had yet given lineage to the famous "turtle" by Walter Gray.
    Then the autonomous vehicle is not automatic and an improvisation. Rather, it is calculated, it is better built, better organized various "turtle", the only one that is not done with simple pieces of "meccano". It should be noted that the student enjoyed the support of a large firm of electrical measuring instruments which his father is the director.
    Completely machined duralumin, his machine is not a simple copy improved "species" of previous artificial animals. It has many original features.

    Drawing inspiration from "Science and Life"

    In reading the article on Science and Life by turtles electronic Grey Walter that P.-A. Amouriq, including the interest of the first applications of cybernetics in its infancy, would produce a new "species".
    His machine would have three wheels with both a director and driving. Unlike Elmer and Elsie, first born of the species Machina speculatrix , the steering wheel would not be in front, but on the back. Thus the curves are better "caught" the craft less likely to hang, for example, the jamb of a door beyond which it has seen an attractive light.
    Two sensitivities: a nuanced sense, sight, and sensitivity crudest, the shocks.
    The organs of sight are two photocells sweeping horizon.

    That there are two rather than one, can provide much more subtle reactions to visual perceptions.
    The sum of the currents collected by the two cells represents an assessment of light intensity and to this end, the cells are connected in "parallel, but mounting" in opposition "also gives the robot the assessment of the direction of the source light.
    The influence of cell line wins if it receives more light than the left, or vice versa. If the light is just between the two "eyes", the two influences are balanced. As for sensitivity to shock, it results in an increased intensity in the locomotive engine suddenly stopped, because the wheels can not move anymore. So that the creature designed, sensitive to the intensity and direction of light as well as a barrier against shock. It remained to decide how she would react – because a machine can never "know" only through the senses and act as the bodies which we have endowed.
    Paul-A. Amouriq bestowed his gear acts of simple forward at full or half-system, turn, or back (one system).

    Conduct advanced

    Let's see how these acts are linked to sensations that provoke, by their presence and absence, light and obstacles. Let there be no light. It is a special case of a balanced perception of the two cells. The engine management, which is naturally sensitive to the imbalance of light perception, is not at stake and walk a straight line. But the absence of current in the circuit affected by light intensity determines the introduction of resistance in the circuit of the batteries supplied to the engine management; the supply current weakens, then, and walking is slow. At the same time, a lamp lit in front of the robot. It symbolizes a state of cautious investigation. Moreover, if the lamp is powerful enough, she can tell by his reflection in the presence of an obstacle.

    Decline to glare, but on shock, turn

    Suppose now that there is at least one light. The animal then heads towards it. But when the light intensity becomes too strong, it recedes until the intensity of the light (or another) is again appealing.
    Photoelectric cells – at least those used here – have a flaw: they are less sensitive when they receive light for a while. But this defect has been salutary; it gives the vehicle a shade of behaviour outside the program: the animal dominates in part of its response comes back and more danger.

    Collision

    What happens in the case of a clash?
    The excess current in the motor control of locomotion immediately reversing, and reversing the flow of cells. That for five to six seconds – time to walk 30 to 50 cm. Then again it forward, but the term is no longer the same.
    If the obstacle is hit during a fall, then comes a desensitization to light.
    For a few moments the craft forward march in search of darkness, and only after a rest of about ten seconds in a steady state of two cells, as walking towards the light resumed.

    MAP OF THE REAL TURTLE P.-A Amouriq

    Simplicity and sensitivity

    Now look at the whole system: it contains no radio tube, no amplifier, but only relays extreme sensitivity, developed by the young father of Cybernetics. Thus the gear he is much stronger and has he "behaviour much more stable than was that of his predecessors.
    The complex, set on a bakelite plate is easy to disassemble chassis. We will not make a parallel between this robot and the turtles Walter Gray. At least, thanks to the possession of two "eyes" are comparable to more like an animal than for turtles Grey Walter's with a single eye turning in one direction, thus they can only rotate in one direction, around an obstacle on one side. With both eyes with a back-and-forth symmetrical walking towards the light is decided, it is not absolutely straight because it always, in practice, the influence of a cell outweighs the other, which determines very slight zig-zag match exactly with continual adjustments by which conducts walking (and indeed all the actions) of living beings.
    As for the ADF, it is certainly not done in this robot.

    But those of Grey Walter demonstrated once and for all that such a mechanism was possible. It is now secondary to want at the price of mechanical complications, reproduce.

    Pierre de Latil

    ARRANGEMENT OF BODIES OF CONDUCT

    Photo taken at a zoo in Paris. Image courtesy A.-P. Amouriq 2009.

    STOP PRESS – 9th Oct 09

    Wonderful news – Paul A. AMOURIQ saw this post about his "la tortue" and posted the first ever comment to my new blog. Thanks Paul.

    He has sent through some further information, plus a picture of what "la tortue" looks like now, after some technology modifications.

    "Hello Reuben,
    I will try to send you a not too old video clip of the tortue. Many years ago it has been equipted with a shell, the huge condensers and the batteries have been replaced by smaller ones and the chassis could be shortened. But it spent many years in a cellar and I had to repair it, change the selenium photocells (fortunately I could find the manufacturer, then retired, who still was having a couple of cells in a drawer and who has been kind enough to give them to me)  try to make the old relays working again. The 3 sensitive (black) relays are galvanometers with one contact on the pointer and 2 adjustable platinum contacts, one on each side.
    I had seen the Elmer and Elsie of Grey Walter in 1949 in an article of the Science et Vie and that gave me the idea to try and make one. In fact I made about 4 models with Meccano pieces before making the one you know with the help of an engineer who machined the metal parts according to my drawings and instructions.
    I met Albert Ducrocq and his partner in an exibition. Their renard was very slow and I never saw it really working but just moving a few centimeters.
    With my best regards.
    Paul A. AMOURIQ"
    As IOTA is today…..
     
     

    See all the Early Cybernetic Animals here.

    W. Grey Walter, Edmund C. Berkeley, Ivan E. Sutherland and the Tortoise

    Who is Ivan E. Sutherland? Ivan was born in 1938, Nebraska, USA and is a computer pioneer, inventing Sketchpad, being the first what we now call a Graphical User Interface (GUI). He also built a walking machine, but that will be the subject to another post later.

    As an under-graduate student, Ivan, with his elder brother Bert, and Bert's then close friend Malcolm Mugglin built their first "beastie".

    1957

    Here's a transcript of the letter send from IEW to Grey Walter in 1957:

    Nov 10. [IES  to WGW]
    “Dear Sir:
    Early last month I had sent to you two copies of a paper entitled “An Electro-mechanical Model of Simple Animals” which was submitted by my brother, Bert (William R. Sutherland), and his close friend, Mac (Malcolm G. Mugglin), to their department of Electrical Engineering. Perhaps a little of the history of that paper would be of interest to you.
    I am now a Junior (3rd year) at Carnegie Tech, also studying electrical engineering – in this and many other things I have followed the lead of my brother. Bert is two years older than I, recently became married and is now on active duty as an officer of the U.S. Navy. Our interest in mechanical and electrical things probably comes from our father, a Civil Engineer from New Zealand: Ph. D. from London, but our first good luck and stimulation came when we met Edmund C. Berkeley in 1952.
    Mr. Berkeley took an interest in the work that we had already done, namely a simple adding machine, and encouraged us to continue, both by suggesting problems and by providing funds for their solution. During the period October, 1952 to June, 1955 we worked under the guidance of Mr. Berkeley. We did a major protion [sic] of the work on a mechanical maze solving mouse similar to one constructed by Claude Shannon of Bell Labs. During the latter part of this same period, Bert left home for college, and I continued our work alone.
    During this contact with Berkeley’s organization we often saw “squee”, his mechanical squirrel; this was our first contact with the species of mechanical animals. Our next contact came when we read your The Living Brain. We were both interested in all the things you have done, but most familiar with the mechanical and electrical aspects, and most interested in your Machina speculatrix. Can you imagine the joy of two young people reading about important work accomplished far away in a field they were just becoming part of?
    It was no surprise to me when Bert suggested, about Christmas of 1955, that we build a mechanical animal also. On page 45 of Bert’s thesis is a picture of the first crude result. When this first model was finished, about May, 1956, Bert for some reason lost interest in the project for a time. During this period, May to December, 1956, I continued work on the second model, the one which finally became the subject of the paper sent [to] you.
    About Christmas 1956, Bert decided to write his thesis. By the end of January I had finished making the frames, motor mounts etc for the models shown in the various pictures; these Bert took over, assembled and used as a basis for his work. Mechanically these machines were good; electrically they were incomplete, as the thesis shows. They had two big drawbacks however: the wet battery needed constant care, and by the way cost us many pairs of pants through acid holes; the machines were cumbersome and heavy.
    At the moment, Bert is busy with his new wife and the Navy, so I am in charge of our project. To get around the two drawbacks mentioned I have constructed a third type of beast. This new model, commonly called “beastie” because of its smaller size, uses dry cells for power, is entirely operated by transistors and proves to be the best we have yet accomplished. However, although I have the mere construction problems fairly well met, I have not yet obtained any results from this latest model The problems which were not yet solved in when Bert’s paper was written are still not solved.
    Perhaps by now you are wondering just why I should write this letter. It is a sort of news report, an information carrier rather than a questionnaire. I examine what we have done: we have a rather nice looking machine which will respond to light and avoid obstacles in a rather crude sort of way. We have a great many possibilities for future work. I examine what I think we should do next: proceed with communication and learning as interesting behaviour. Perhaps making the machines (I’d like to build more of the “beastie” type) play tag might be a good start. We need a better obstacle strategy.
    Building these machines has been, to say the least, an education in itself. I have found time and time again that to us the problems of actual design and construction were fairly straightforward; the decisions such as I face now of what to do next are more difficult. Perhaps you have some ideas. I am, of course, curious to know what you think.”

     


    1959

    July 13 ECB to Hy Nagourney of Science Materials Center (HN) – ECB spoken to WGW on recent visit to England and discussed possible manufacture of small robots.
     
    Dec 19. IES to WGW
    Dear Dr. Walter:
    I have written a paper for the American Institute of Electrical Engineers which attempts to show how simple animals steer themselves. It considers, amongst other things, your Machina Speculatrix. This paper is soon to be published in Electrical Engineering, the monthly publication of the AIEE.
    I would like to include a picture of Machina Speculatrix with my paper. Would you be so kind as to send a picture suitable for publication?
    I am particularly interested in showing the “tricycle” [sic] type steering system which you used. A side view with the cover removed would, I think, be best. …..”


    1960

    Jan 23 [Ivan Sutherland to WGW]
    “I thank you very much for the photographs of M. Speculatrix. ….I observed the numeral 6 stamped on several parts. Was this the sixth model you have built?

    Feb 1. [WGW to IS]
    “the number ‘6’ which you saw on the chassis is, as you guessed, the serial number of the model.
    You may be interested to know that Basic Book Inc …., who run the Science Library, are proposing to manufacture these models for demonstration purposes.”

    RH-Note – its interesting to note that Berkeley was including IES in on the tortoise deal Mar 1 1961.


    1961


    Mar 1. Berkeley invites Ivan Sutherland to join Science Materials Center. Berkeley writes "We would like very much for you to be associated with us at Science Materials Center in one or more projects, including particularly to small robot project. …..we could draw on all of Grey Walter’s and all of your ideas and capacities, in order to produce small robots which would be of scientific value and instruction. I am sure that we need your help in addition to Grey Walter’s in order to make a resounding success of this project.”


    No further correspondence known of from Berkeley archive.

    I will talk further in another post about Ivan and Bert about later cybernetic models, including Franken, the maze solver.