cyberneticzoo.com

"Rodney", the Self-Programming Robot is based on the book How to Build Your Own Self-Programming Robot by David L. Heiserman [TAB, 1979].

ByRamiro Molinaon September 18, 2013
 This book is geared towards those that have good knowledge of electronics and are willing to jump into a project that involves CPU based control. It outlines how to build a wheeled robot controlled by an Intel 8085 CPU, programmed by hand in binary using an array of switches that bumbles around a room on its own.

ByBenjamin Graylandon November 26, 2000
If you have an interest in robotics, and a decent knowledge of electronics, then this book is certainly worth reading. Despite its age, the information it provides is applicable today.
Heiserman tells of his own robots, specifically Rodney, who can program himself. One example given was of Heiserman handicapping Rodney, by scratching his processors and removing one of his wheels – Rodney learned to move about efficiently in a short period of time, with no assistance. Similar anecdotes are spread throughout the book.
But most importantly, the book tells the reader how they can construct a robot similar to (or exactly the same as) Rodney. Schematics, wiring diagrams and so forth fill a large portion of the book – providing a clear method for construction.
Overall, this is certainly an interesting book. Even if you don't plan to build yourself a robot, the anecdotes are both entertaining and amazing enough alone.

Classes Of Robotic Self-Learning. Source: here.

It is useful to define intelligence as in robotics according to David L. Heiserman 1979 in regards to the self-learning autonomous robot, for convenience here called "Rodney".
    (1) ALPHA CLASS

    While Alpha Rodney does exhibit some interesting behavioral characteristics, one really has to stretch the definition of intelligence to make it fit an Alpha-Class machine. The Intelligence is there, of course, but it operates on such a primitive level that little of significance comes from it. ….the essence of an Alpha-Class machine is its purely reflexive and, for the most part, random behavior. Alpha Rodney will behave much as a little one-cell creature that struggles to survive in its drop-of-water world. The machine will blunder around the room, working its way out of menacing tight spots, and hoping to stumble, quite accidentally, into the battery charger.

    In summary, an Alpha-Class machine is highly adaptive to changes in its environment. It displays a rather flat and low learning curve, but there is virtually no change in the curve when the environment is altered.

    (2) BETA CLASS

    A Beta-Class machine uses the Alpha-Class mechanisms, but extends them to include some memory – memory of responses that worked successfully in the past.

    The main-memory system is something quite different from the program memory you have been using. The program memory is the storage place for Rodney’s basic operating programs-programs that are somewhat analogous to intuition or the subconscious in higher-level animals. The main memory is the seat of Rodney’s knowledge and, in the case of Beta-Class machines, this means knowledge that is grained only by direct experience with the environment. A Beta-Class machine still relies on Alpha-like random responses in the early going but after experiencing some life and problem solving, knowledge in the main memory becomes dominant over the more primitive Alpha-Class reflex actions.

    A Beta-Class machine demonstrates a rising learning curve that eventually passes the scoring level of the best Alpha-Class machine. If the environment is static, the score eventually rises toward perfection. Change the environment, however, and a Beta-Class machine suffers for a while, the learning curve drops down to the chance level. However, the learning curve gradually rises toward perfection as the Beta-Class machine establishes a new pattern of behavior. Its adaptive process requires some time and experience to show itself, but the end result is a more efficient machine.

    (3) GAMMA CLASS

    A Gamma-Class robot includes the reflex and memory features of the two lower-order machines, but it also has the ability to generalize whatever it learns through direct experience. Once a Gamma-Class robot meets and solves a particular problem, it not only remembers the solution, but generalizes that solution into a variety of similar situations not yet encountered. Such a robot need not encounter every possible situation before discovering what it is suppose to do; rather, it generalizes its first-hand responses, thereby making it possible to deal with the unexpected elements of its life more effectively.

    A Gamma-Class machine is less upset by changes and recovers faster than the Beta-Class mechanism. This is due to its ability to anticipate changes.

Robotics: Robot Intelligence: An Interview With A Pioneer
Posted here on 2008-06-06 @ 19:28:20 by r00t.

A short and informal email interview with a pioneer in the field of hobbyist robotics, David L. Heiserman.

Mr. Heiserman is the author of six volumes on the subject, published by TAB Books over a span of 11 years, from 1976 to 1987. These books describe, in detail, several robotics and simulation projects he developed during those years. Each was written and designed in such a manner as to allow the reader the ability to follow along and construct each project themselves.

However, the books aren't plans so much as they are guides. They form a complete encyclopedia for a compelling subject of study, which Mr. Heiserman has termed "Robot Intelligence" and/or "Machine Intelligence":

Build Your Own Working Robot – #841 (ISBN 0-8306-6841-1), HB, © 1976
How to Build Your Own Self-Programming Robot – #1241, (ISBN 0-8306-9760-8), HB, © 1979
Robot Intelligence…with experiments – #1191, (ISBN 0-8306-9685-7), HB, © 1981
How to Design & Build Your Own Custom Robot – #1341, (ISBN 0-8306-9629-6), HB, © 1981
Projects in Machine Intelligence For Your Home Computer – #1391, (ISBN 0-8306-0057-4), HB, © 1982
Build Your Own Working Robot – The Second Generation – #2781, (ISBN 0-8306-1181-9), HB, © 1987

I first read these books as a boy in grade school, and continued to study them periodically through high school. As an adult (now almost 35 years old – where did the time go?), I collected the set for my library. Along the way, I wondered what Mr. Heiserman did with his robots, and whether he planned on publishing anything more about them or his experiments. This interview and other email conversations with him have helped to answer these  questions.

PG: What, and/or who, inspired you to pursue the research of machine intelligence?

DH: I saw the robots in sci-fi films of the 50s and 60s, and I wondered how it would be possible to build one.

PG: Was Buster the initial platform for your research, or were there prior (but unpublished) platforms and/or systems you used prior to Buster?

DH: There was a prior version in 1963. I can't remember the name, but it was strictly radio controlled — vacuum tubes, no less.

PG: During the period your books on robotics and machine intelligence were published, TAB Books seemed to provide a haven for similar authors. Did they provide or do anything special to encourage this?

DH: No.

PG: Were you ever in contact with any of the other robotics experimenters (published by TAB or otherwise) during the period your books were published?

DH: No.

PG: Rodney seemed to anticipate the experiments carried out in "Robot Intelligence" and "Machine Intelligence". Were these projects inter-related?

DH: The books are pretty much a technology-based sequence. I had no idea about doing machine intelligence when I did the book on Buster.

PG: Did you ever bring together the software concepts developed in "Robot Intelligence" and "Machine Intelligence" with an actual hardware platform, or did you view the software environments you created as a better avenue for development of your ideas on machine intelligence?

DH: "Projects" was an attempt at hardware implementation, but I was more interested in computer simulations by this time. I never published my work for several weak reasons; one of which was that I was beginning to catch so much nasty flack from the amateur and quasi-professional AI community. I won't go into all of that, but let's just say I am enjoying some quiet satisfaction today.

PG: Why was the decision made to create the second generation Buster as a "hard-coded" robot, rather than continue with programmable machines as represented by the earlier Rodney?

DH: Well, I think it was because I was losing a segment of people who were not sophisticated enough to do any programming.

PG: What are the major differences between Buster as described in the original "Build Your Own Working Robot", and the Buster described in "Build Your Own Working Robot – The Second Generation"?

DH: Second Generation had better hardware designs.

PG: Whatever happened to Buster (I-III)?

DH: Buster I is somewhere down in the crawlspace of my house. The others were scrapped or given away a long time ago.

PG: What about Rodney?

DH: I gave him to a high school science class. I imagine it is gone.

PG: Do you have any current photos of Buster and/or Rodney (assuming they still exist)?

DH: No.

PG: Were any other later hardware platforms built (but left unpublished)?

DH: Rodney had a short-lived expression as a commercial product sometime in the early-to-mid 80's. It was the RB5-X, manufactured by RB Robot Corp in Golden, Colorado. I was rather well compensated for the work, but the company and my compensation soon evaporated.

PG: Are you still involved in robotics and/or machine intelligence as a hobby or otherwise?

DH: No. But I like to tinker with my own version of artificial neural networks.

PG: Do you intend on writing any further books on robotics in the future?

DH: Not as a hobby machine. Over the years, I've used my models of machine intelligence to play with ideas about extraterrestrial intelligence.

PG: Are there any thoughts or advice you would give to today's robotics and/or machine intelligence enthusiasts?

DH: Let a machine think for itself. Let a community of machines think for themselves and share their knowledge and skills.

But keep your hand on the plug.

I feel that Mr. Heiserman's work is still relevant for today's robotics hobbyist, especially for those interested in machine learning. His techniques and programming methodologies can be easily applied to modern microcontroller and PC-based systems. There are many avenues available to explore in this research, and Mr. Heiserman has forged a path ahead of us to follow. If you are interested in robotics, you owe it to yourself to pick up a volume or two of his books, and explore.

Andrew L. Ayers, March 2008

The RB5X Connection:

Heiserman also wrote some software for the personal robot RB5X.  From an interview …

RN: Did you ever consider taking any of your robot designs commercial as kits or assembled robots?

DLH: I never did it on my own initiative, but Rodney appeared on the market as RB5-X. It was advertised as educational tool, and we had a couple of RB5s running around in the science center here in Columbus. The company was RB Robot, Inc., in Golden CO. When RB when bankrupt, someone else bought the rights and inventory. I don't think the machine is around anywhere these days. I was just a token consultant for the company, anyway.

David Hieserman had already built "Buster" the robot, but was developing "Rodney" the "Self-Programming" robot at the time. RB5X software utilized "Rodney" technology.

The RB5X robot comes with what the company calls Alpha and Beta level self-learning software. This "Artificial Intelligence" software, developed by David Heiserman allows your RB5X robot to learn from it's experiences.

Self-Learning Software / Artifical Intelligence
The RB5X comes complete with "Alpha" and "Beta" levels of self-learning software, which which empowered the robot to absorb and employ information from its surroundings. Developed by leading robotics author David Heiserman, this software allows RB5X to progress from simple random responses to an ability to generalize about the features of its environment, storing this data in its on-board memory.
Self-Learning: This small, first step toward true "intelligence" enables the robot to learn from its own mistakes. For example, you could set the RB5X down in a room and let it roam about randomly. It will probably run into walls several times, perhaps a desk, and maybe even a person. As it rolls around the room, it will "learn" in its own computer-like fashion where the obstacles are in a room, thus avoiding them in the future. The self-learning software are on "Alpha" and "Beta" levels, which were developed by the robotics author David Heiserman for the purpose of giving robots a simple way to "learn" from their experiences, somewhat like humans do.

See other early Mobile Robots here.

An early example of multiple robotic creatures operating together. Other than light and touch sensors, there's no other apparent interaction with them. Possibly an early but simple example of swarm robotics and collaborative robots.

English translation of article text:

To emphasize wont in machine control, a British firm [Mullard] of electronic devices has created these mechanical mice playing hockey on the ice. Each mouse is equipped with a photoelectric cell. Circuits and polarized magnetic lines of force, located under the floor, move the mice to the hatch into which they let the ball.

Source: La Tecnica Illustrata, March 1959.

Per dare risalto ai suol controlli per macchine, una ditta britannica di apparecchi elettronici ha realizzato questi topolini meccanici che giocano a hockey sul ghiaccio. Ciascun topolino e munito di una cellula fotoelettrica. Circuiti polarizzati e linee magnetiche di forza, situate sotto il piano, fanno muovere i topolini verso la porta nella quale devono far entrare la palla.

The Mullard logo.

See all the Cybernetic Animals and Creatures here.

ROBOTICS: Featuring An Automated Pavlovian Dog!
 
Developed many years ago, in the "Pre-IC Age" these Robot Rovers could simulate such Classical Pavlovian Responses as: CONDITIONING, EXTINCTION, SPONTANEOUS RECOVERY, LEARNING CURVES and HIGHER-ORDER CONDITIONING.

Three-deck stepping-relays comprised the main elements of the dog's memory. A few transistors were used for "eye" and "ear" sensors, plus a "tail-wagging power amplifier."

Frederick W. Chesson

I knew of the April, 1961 "troubles" at SJ, but it was only when I was working in the Middletown area c 1961-69 that I regularly commuted through Berlin and got regular glimpses of the place and heard about it from fellow workers that I had any inclination to wonder what went on there. In that general period, I had developed the "Automated Pavlovian Dog" teaching-machine (also on my web site) that led to a connection with the Psychology Dept. at Wesleyan University. The "dog" was shown there and to a number of schools, hoping to build up my psych-lab construction business. I also attempted to interest Mr. Francis, knowing of his background, but by then in the late 1960s he had become excessively suspicious of my innocent motives, that resolved to "keep tabs" on what further incidents went on at SJ…which were all-too forthcoming over the next few years until the place finally closed for good.
Fred W. Chesson. E-mail 15 April 2006

The experiments with dogs relating to Classical Conditioning by Dr. Pavlov, earning the Nobel Prize for Medicine and Physiology in 1904, have been simulated over the years, culminating with today's extensive computer programs.
    The robot dogs shown in the photograph were developed by the author in the early Sixties, when the teaching-machine "fad" was approaching its heady zenith. At the time of the design, relay logic still had a cost advantage over the contemporary RTL gates, but some transistors were employed for the "eyes" and "ears" of the automated canines.
     Pavlov's experiments into Classical Conditioning underly much of modern learning theory; hence, if a robot, android, or humanoid is to learn, it is desirable to know what conditioning is all about. On a basic level, Pavlov rang a bell, then fed the dog, measuring the animal's response by the amount of saliva generated. After a while, the bell alone could evoke a salivatory reaction. On a human level, do our mouths not water at the mere aroma of a tasty pie? Or even at the verbal cue: "Dinner's ready!"…? But should the announcement prove false or premature, our anticipatory responses will diminish markedly. They can, however, be readily restored, along with our faith in human nature.
    Thus, the electro-mechanical dog was designed to perform the following simulations, which will be examined: conditioning (learning), extinction (forgetting), spontaneous recovery, higher order conditioning, learning curves, memory of stimuli occurrences, and stimuli hierarchy.
    In operation of the simulator, pressing the RESET switch puts the robot dog at an untrained level (electronic brainwash!). Salivation being somewhat difficult to imitate, the response to feeding was represented by having the dog wag its tail, a readily observable act of canine satisfaction. To hold the interest of younger students, the feeding stimulus was simulated via a plastic bone having a concealed magnet. When the magnet end of the bone was in proximity to the dog's "nose," a reed switch was closed, activating a tail-wagging power transistor and solenoid.
    Via a microphone and photocell, the dog could "hear" and "see." Normally, the audio stimulus was dominant, activating a Schmitt-trigger delay for a preset time interval. If the food stimulus was presented during this period, an AND gate caused this coincidence to be recorded by the Conditioning Event Counter, a ten-point stepping relay. (Today's equivalent probably being a CMOS type 4017 decimal-decoded counter chip.) Thus, when a preset number of coincidences had been registered, a relay flip-flop circuit caused the dog to now wag its tail to the sound stimulus as well as to food.
    So long as occasional sound-food coincidences, (reinforcement), occurred, the conditioned state would be maintained. But after another preset number of sound-stimuli without food following, (anticoincidence), say five, the flip-flop resets the dog to an unconditioned state, and it must be retrained.
    Sometimes, the experimenters found their animals would recover their condition, (spontaneous recovery), without any apparent external action.
This is similar to being given a telephone number in the afternoon, then forgetting it by night, only to have it suddenly come to mind the next morning, apparently released from some buffer-storage in the subconscious.
In the simulator, the spontaneous recovery function could be cut in and its "latent period" set by a potentiometer. Should normal conditioning then be re-established before it can act, it is reset for future use. Once it has acted, however, it is of a one-shot nature; following a second extinction, true conditioning must follow for the SR circuit to be reset.
    After conditioning and extinction, Pavlov found that his dogs not only relearned faster, but that their conditioned response was more resistant to extinction. This learning curve holds true in human education, as anyone who has learned a mathematical equation or foreign language will agree. Learning something the second time around nearly always is quicker and seems to stick longer as well.
    The learning curve simulation required multi-level stepping-relays in the original model, whose pick-off points were determined in connection with the original settings for conditioning and extinction counts. Thus, the original number of four coincidences would be reduced to three and then only one, while the anti-coincidences for extinction might be increased from five to six or
seven, and then to eight or ten.
    When the living dog has been very well trained to salivate to the sound of the bell, it was found that the bell as well as food could be employed to condition him to a new stimulus, such as light. This is called Higher-Order Conditioning, and represented the simulator's highest accomplishment, being activated by the learning curve counter.
    While the above model and its concepts are quite elementary, they still furnish a base upon which increasingly diverse and subtle forms of learning behavior may be simulated and explored. It has been found, for example, that conditioning is more resistant to extinction when every trial stimulus is not always rewarded. Such variable reinforcement scheduling, could lend itself readily to microprogramming applications.

"Way back in the dim '60s, in the midst of the Teaching Machine Era, I came up with a Pavlovian Dog demonstrator. (Two were actually built)  Responding to food (magnetic bone) sound and light stimuli, the concepts of Conditioning, Extinction, Learning Curves, Spontaneous Recovery and Higher-Order Conditioning were presented. (All done with relay logic and memory back then!)"
Fred W. Chesson

The Interface Age1978 article does not include the images above.

Extract from 1980's article:

The articles present experiences in interfacing and programming a SUPERKIM single board computer for the control of a Lour Control ET-2 robot shell. The ET-2 (Experimental Transmobile with 2 drive motors) consist of a three level frame powered by two separately driven wheels and balanced by a free caster.

Part 2 adds the sensors to give it true 'feedback'.

The SUPERKIM controlled ET-2 robot is an excellent, moderately priced system to which the robotics experimenter can easily add more sensors and other equipment.

The contact sensors … can be used to demonstrate obstacle avoidance behaviour in a suitably prepared environment.

see Robotics Age pdf's here  and here .