Shop OBEX P1 Docs P2 Docs Learn Events
Fill the Big Brain - Page 17 — Parallax Forums

Fill the Big Brain

1141517192059

Comments

  • HumanoidoHumanoido Posts: 5,770
    edited 2011-04-04 21:05
    Mike G:

    >Where is the Brain Channeling in the OBEX?

    More on Brain Channeling

    Brain Channeling is not an Object in the OBEX. BC is a technique. The resources to create BC are found in the OBEX, such as the ones you mentioned.
    • PING distance reading
    • moving a servo
    • sound input
    • saving to EEPROM
    • sensor touch input
    BC uses one Propeller chip for each channel and makes no attempt to mix Cogs with other channels at this time. This is open for development. BC is another successful experiment as defined in previous posts - its concept is to report data to a chip that is overseeing and managing with code to send packets over the net. The code slot for this system is available for further development. Check the index link in post #1. You may find other references and new information about Brain Channeling, its techniques, and some ideas for developing it.
  • HumanoidoHumanoido Posts: 5,770
    edited 2011-04-04 21:19
    Humanoido wrote: There is example code posted to send and receive data using the Brain Base and the Brain Spans for Propeller to Propeller communications. The BUS is also released as a schematic.

    Mike G wrote: I beleive you're are referencing the 1/2 duplex and full duplex code? This architecture will work fine in Prop to Prop or Prop to STAMP communication but not so well in a large parallel network.

    Why not? Half duplex works in every BSS project and some have over 20 parallel processors - it also works with 320 Propeller Cogs.

    For contribution, I wrote a 1Mpbs 1/2 duplex code that frames a command packet and shoots it off to a mutli-drop serial bus. It's in the OBEX in the AX-12 projects. Still, that would not work well in a parallel system. It's a master slave setup.


    Why not have a master with n number of slaves in Parallel?
  • HumanoidoHumanoido Posts: 5,770
    edited 2011-04-04 21:36
    Mike G wrote: What is a neural matter injector? Do you have a code example? How do you "recycle pins for communications"? Does this mean multi-purpose IO?

    More about the Neural Matter Injector

    The sources below include all the known posted information about the Neural Matter Injector. The first developed unit was known as the MNI. The state of the current unit is known as NMI or simply BMI.

    There are some interesting stories about the BMI and one doctor of medicine has taken an interest and asked me numerous questions about potential machine to human brain apps! Perhaps this will lead to something applicable to humans in the future.

    p20
    396 Brain Breakthrough - a Multi-Propeller Neural Injector (MNI) - Case of the evolving hybrid

    397 Brain Matter Injector BMI - Machine injects & distributes neural matter

    p23
    453 NMI's as Discrete Brains - Neural Matter Injectors are multi-purposed


    It is known that on Propeller chip startup (read the bootup procedure in the main Propeller manual), the high numbered pins are employed. After boot is completed, these pins become available for use. There are 32 I/Os for use. You can also tri-state these pins for more applications (see other threads for more information).

    The term Multi-purposed I/O can have several meanings. It can also refer to recycling an I/O pin.
  • HumanoidoHumanoido Posts: 5,770
    edited 2011-04-04 21:54
    Mike G wrote: »
    All my Prop projects have a similar architecture where data is passed through the HUB. Nothing new just using the Propeller as intended. I have no idea how to write code for the Brain as the interface is not defined. To contribute to the project, I would expect clear specifications that outline how code must behave in the brain. Things like basic object communication within a Prop and with external Props. Without specifications I'd be wiring the Brain from scratch.
    Mike G, this is a darn good idea! While I have posted schematic drawings of several interfaces, the information is somewhat scattered and truly disembodied. Now that the drawing program is up to speed on the learning curve, I would look forward to making it a detailed project, giving new schematics rather than hand sketches or links and making options available for contributing various portions of programming. Stay tuned! It's a really big project, but things will come together in small completed sections over time.
  • HumanoidoHumanoido Posts: 5,770
    edited 2011-04-05 04:27
    Brain Demo Day
    Scientists in China get a glimpse

    The brain demo went well today but there was an interesting twist. Various code was loaded to demonstrate numerous functions. The portability was obvious. The add-on brains were also shown along with real time demos of the injectors and processors. Parallel processors were demonstrated. Some translated comments were about the complexity of the Brain and the technology. This is good for Parallax and a boon for the Propeller chip. Other meetings may be scheduled at the University of Science & Technology. Oddly enough, an interest has arisen in one part of the Brain - the Neural Injector. It seems medical doctors have more interest in this device on human brains as compared to Robot enthusiasts. I had to explain it works with a machine brain and not a human brain. Then I started thinking from their perspective. What if... it could actually work on a human brain... and after thinking about, the BMI is far from that app. Whether or not there's a purpose for injecting a few machine neurons into the human brain remains to be seen.
  • Mike GMike G Posts: 2,702
    edited 2011-04-05 06:57
    Sorry about that. I should write more information. However, it's very clear to me. What part seems confusing?
    The title "Fill the Big Brain" is misleading when compared to the content. I'd expect that the US40 Boe-Bot project was successful because the Big Brain was "Filled" and controlling the Boe-Bot. That's not the case.
    Brain Channeling is not an Object in the OBEX. BC is a technique. The resources to create BC are found in the OBEX, such as the ones you mentioned.
    So, there's no Brain channeling project just a concept.
    Why not? Half duplex works in every BSS project and some have over 20 parallel processors - it also works with 320 Propeller Cogs.
    Tthe slaves only respond to commands master commands? How do slave devices interrupt the master if something important comes down the pipe like the motors are stalled. Is the master continually polling all 20 salves serially?
    Now that the drawing program is up to speed on the learning curve, I would look forward to making it a detailed project, giving new schematics rather than hand sketches or links and making options available for contributing various portions of programming.
    Sure a drawing is good. However, you need to actually produce code that demos how objects interact in the Big Brain. If you produce base functionality, more folks will be interested in contributing.
  • TtailspinTtailspin Posts: 1,326
    edited 2011-04-05 09:06
    I believe Humanoido was asking for input...
    Original Post: If you had a big brain for your robot, what would you put in it?
    I think some similarity to a human brain is good, for example, put in
    vision processing, memory, and a robot autonomic nervous system.
    Ok, your turn, fill the brain..

    This is the important part of the first post...
    Ok, your turn, fill the brain..
  • HumanoidoHumanoido Posts: 5,770
    edited 2011-04-05 18:06
    Mike G wrote: »
    I'd expect that the US40 Boe-Bot project was successful because the Big Brain was "Filled" and controlling the Boe-Bot. That's not the case.

    Actually that would be a lowly task for the US40 since Boe-Bot already could do pre-programmed motions with one BASIC Stamp.

    The Boe-Bot Brain came along first using the US40. The current Big Brain is a spin off of that project. Refer to the previous post on this topic. A lot of my projects are conceptual. This means they are for my use and hobby pleasure. They may not fit your goals or needs in your projects. They may not fit your idea of the project. It's not supposed to. It's offered as-is and free. You can prospect and find gold in these projects or move on. I set my goals and meet them. It may or may not match your goals. The goals of that project were stated in two recent posts. I simply cannot help you with your expectations on my projects.

    Except this project is different and open development. Your input, ideas, programs, drawings, plans, details, and various contributions will be considered and integrated. Prepare to be assimilated..
  • HumanoidoHumanoido Posts: 5,770
    edited 2011-04-05 18:07
    Mike G wrote: »
    So, there's no Brain channeling project just a concept.

    Brain Channeling is operational. Read related posts. Refer to the Brain Index, link found on page 1. BC is at a simple testing level now and still under development.
  • HumanoidoHumanoido Posts: 5,770
    edited 2011-04-05 18:09
    Mike G wrote: »
    Tthe slaves only respond to commands master commands? How do slave devices interrupt the master if something important comes down the pipe like the motors are stalled. Is the master continually polling all 20 salves serially?

    The Brain has this ability to reconfigure itself. If the Brain wants to use a Master - Slave arrangement, then yes, it can do as you ask. Motor stalling or some motion control anomaly and error checking is simply handled by the Brain Stem.

    The Master and Slave operate according rules. The rules of the Master Slave concept were developed with the Basic Stamp Supercomputer BSS. Both M/S can listen or talk. They do not sit in a state of constant polling, rather they do tasks and decide when to talk and when to listen. Sometimes they play tag to accomplish a plate of work. There is a demo of this on youtube with the BSS video. They must all follow the rules of the net.
  • HumanoidoHumanoido Posts: 5,770
    edited 2011-04-05 18:20
    Mike G wrote: »
    Sure a drawing is good. However, you need to actually produce code that demos how objects interact in the Big Brain. If you produce base functionality, more folks will be interested in contributing.

    Good idea. Who is going to produce the demo code?
  • Mike GMike G Posts: 2,702
    edited 2011-04-05 18:52
    Actually that would be a lowly task for the US40 since Boe-Bot already could do pre-programmed motions with one BASIC Stamp.
    Then what is the purpose of putting the US40 on the Boe-Bot in the first place? I'm not getting the connection.
    Brain Channeling is operational. Read related posts. Refer to the Brain Index, link found on page 1. BC is at a simple testing level now and still under development.
    I went to page to page 1 and found the index link but was unable to find anything about Brain Channeling.
    The Brain has this ability to reconfigure itself.

    How does this work? Is this a hardware reconfiguration or software?
    Both M/S can listen or talk. They do not sit in a state of constant polling, rather they do tasks and decide when to talk and when to listen.
    If you're not using polling, how does bus arbitration work in the Brain system? What if two devices try to talk at the same time?
    Sometimes they play tag to accomplish a plate of work. There is a demo of this on youtube with the BSS video. They must all follow the rules of the net.
    I looked at the uTube video and the code. A master STAMP places a message on the a multi-drop bus addressed to one of the waiting STAMPs. All STAMPs are in a wait state until one STAMP recognizes its ID. The addressed STAMP then preforms a functions and returns a response to the master. That's a polling scheme, right?
    Good idea. Who is going to produce the demo code?
    I'm confused... I thought that you would be providing the framework since this is your Brain Child.
  • HumanoidoHumanoido Posts: 5,770
    edited 2011-04-05 21:51
    Mike G wrote: »
    Then what is the purpose of putting the US40 on the Boe-Bot in the first place? I'm not getting the connection.
    Mike G, now you're just not paying attention. I answered this question in post 481, 478 and finally 480.
  • HumanoidoHumanoido Posts: 5,770
    edited 2011-04-05 22:48
    Mike G wrote: I went to page to page 1 and found the index link but was unable to find anything about Brain Channeling.

    Brain Channeling Sources

    Go to page one. Follow the thread link to page 16.
    http://forums.parallax.com/showthread.php?124495-Fill-the-Big-Brain/page16

    Then use your browser's search function, found in the edit menu, using the word "channel" ...
    Note the post numbers.

    332 Opening Up Brain Channels - Brain is capable of multi channels
    482 More on Brain Channeling
    490 Humanoido re: Brain Channeling

    The index is now fully updated as of this posting.
  • HumanoidoHumanoido Posts: 5,770
    edited 2011-04-05 23:18
    Intro to Interfacing Reconfiguration

    The Brain has this ability to reconfigure itself. How does this work? Is this a hardware reconfiguration or software? Use the search words on index for interface..hybrid..bus etc. for more information. (Find the index from the first post or on page 16 in the thread)


    The brain is being wired and tested with several different interfaces. So far, work is proceeding on several spin-off technologies.

    The Brain Interfacing Spin-off
    1. AM Algorithm Machine,
    2. 4D Morphing Computer,
    3. Basic Stamp Supercomputer
    4. Brain Matter Injector
    5. SEED Machine
    The interfacing includes
    • BUS
    • Serial
    • Parallel
    • Injection
    • Combo
    • Hybrid
    • Reconfig
    The combination of this is a new Hybrid that goes beyond the AM. (has more functions and elements and the potential for greater applications)

    The new feature being developed is reconfiguration RECONFIG. At first, it was designed for a human to reconfigure the interfacing with wiring. A few months later and the idea to let the machine do the config by a Propeller chip was born, thus ushering in a self configuring brain.

    Exactly how will this be employed? It's really task orientated upon the object to solve, or within the section of the Brain that is executing thinking. It is entirely possible, according to designs in the making, to run several of these hybrid interfaces at the same time. What will this achieve? It could be remarkable! This will be examined in greater detail in the future.

    Software switching allows The Brain or the programmer to use a choice of interfacing. Hardware switching allows for a semi-permanent machine. In some instances it is more convenient to reconfigure with switching manually.

    Not investigated are the myriad of programming techniques when engaging multiple interfaces in parallel functioning (at the same time) or in various combination.

    Work is progressing on the HYBRID. Some wires can double up and some wires can be more efficiently removed from the whole. This will be investigated in more detail after the individual interfaces are more tested and have more code and programming examples.
  • HumanoidoHumanoido Posts: 5,770
    edited 2011-04-05 23:24
    Mike G wrote: »
    If you're not using polling, how does bus arbitration work in the Brain system? What if two devices try to talk at the same time?
    Two devices on BUS cannot talk at the same time because they must follow the BUS rules of the Neural Net. If a Slave cannot follow the rules, he or she will be immediately removed from the system and punished with a lot of schooling and tests.
  • HumanoidoHumanoido Posts: 5,770
    edited 2011-04-05 23:43
    Mike G wrote: »
    I looked at the uTube video and the code. A master STAMP places a message on the a multi-drop bus addressed to one of the waiting STAMPs. All STAMPs are in a wait state until one STAMP recognizes its ID. The addressed STAMP then preforms a functions and returns a response to the master. That's a polling scheme, right?
    There are many names for this interesting setup.
    • Party Line
    • Bus
    • Talk/Listen
    • Poll
    • Master/Slave
    This is well covered in the Basic Stamp Manual. There's tons of information in the Handbook of BASIC Stamp Supercomputers.

    You may want to look at SEED. It can listen in to all who speak! It consists of ten processors which evolve into life forms. They have addresses, talk and listen to neighbors, nap, sleep, dream, memorize, recall, all using TINY AI. The interesting thing is some of these develop personalities which was unexpected. The life forms living at the base of the Tower are very industrious. Those at the top seem to end up quite lazy as I recall..
  • HumanoidoHumanoido Posts: 5,770
    edited 2011-04-06 00:03
    Mike G wrote: »
    I'm confused... I thought that you would be providing the framework since this is your Brain Child.
    LOL! Ok it's definitely my brain child. I have to teach it everything from door one. So the open source development has become a kind of follow along the Brain development, which is ok, since the OSD was a test to see how well it worked. We've had a lot of really great input at the beginning which filled the initial purpose, which included what to put inside the Brain, so a lot is based on this information. However as it goes now, you are correct, I will provide a lot of glue to make it all come together. It will be exciting to post all the demo code when it's ready. But now, I'm diluted rather thin and working on one Brain aspect at a time. In the future, I envision a Brain, with postings and enough information so that guest programmers can have some time on it. A lot needs to happen before that point is reached. For now, I'm working to get the Brain up into a very high processor level and enjoying this stretch of wiring. But this may change tomorrow..
  • prof_brainoprof_braino Posts: 4,313
    edited 2011-04-06 16:51
    And that, sir, is known as plagiarism.

    LOL! I see a slap delivered with a velvet glove.

    @Phil & erco - I did not get the impression that this forum post was to be used as the basis of a doctoral thesis or that Humanoido was representing any of the referenced work as his own.

    Please keep up the good work Humanoido. Thanks for sharing what you find, I am interested and enoy seeing it. And when you get the chance, put in good word for me with our new AI overlords.
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2011-04-06 19:00
    I did not get the impression that this forum post was to be used as the basis of a doctoral thesis or that Humanoido was representing any of the referenced work as his own.
    It doesn't matter. Not giving credit to someone else's work when quoted or paraphrased -- regardless of this venue's informality -- is not to be tolerated. But I have not persisted in my screed, since Humanoido's behavior seems to be on the mend -- despite his innate inability to acknowledge his own culpability, a character trait I've come to expect from him.

    -Phil
  • HumanoidoHumanoido Posts: 5,770
    edited 2011-04-06 20:18
    Concatenating Brains
    Unexpected Brain surplus creates a mammoth Brain!

    In developing several test models of the Brain for running experiments, a surplus of brains has built up. In terms of completed Brains including breadboard modules, ie Brains, the lab now owns a total of five. One is relatively comparatively small (only 40 Cogs) but the others are quite massive (one is 216 processors). It was unexpected that an additional four Brains used for testing would become available after testing. What to do with the extra brains? First let’s take a look at what we currently have operable.

    Primary Brain
    23 boards (21 Propeller, 2 BS)
    168 Propeller Cogs
    21 Propeller chips
    2 BS Processors
    170 processors
    Secondary Brain
    1 breadboard
    5 Propeller chips
    40 Cogs
    40 processors
    Tertiary Brain
    1 breadboard
    14 Propeller chips
    112 Cogs
    112 processors
    Quad Brain
    1 breadboard
    12 Propeller chips
    96 Cogs
    96 processors
    Pent Brain
    High density breadboard
    27 Propeller chips
    216 Cogs
    216 processors
    Total processors 634
    (170, 40, 112, 96, 216)

    These boards are currently setting end to end and coupled with color coded interface extension wires. For Breadboard retention, a rack is available to accomodate all breadboards. The 634 Brain is currently the largest working Propeller machine of its type. It includes 79 interfaced Propellers and additional Stamps in a Multiplicity Modular Arrangement.

    The MMA includes the Brain Stem, Brain Base, four Brain SPANS, and four Brain extensions. Additional various testing will likely continue with the subsectional Brain portions. It is not yet decided how these will integrate in the original system in terms of size and physical placement.

    Do we have an plans for another test Brain? Yes! The project is doing well, in the testing phase, in terms of both hardware and software, we look forward to seeing the introduction of additional massive Brains that will eventually be absorbed into the system and make one really mammoth Brain.
  • HumanoidoHumanoido Posts: 5,770
    edited 2011-04-06 21:56
    Brain Machine Biology
    A symbiotic union of technologies

    In biology, as in a biological brain, there are biological matter, cells, and matter based on genetics. In a machine brain, the materials are completely different, yet the electric brain can still offer workings of genetics, templates of cells, and biological simulations.

    The Brain worked on here is definitely a machine, consisting of various hardware and electronic components. It contains elements of copper, tin, derivatives of plastic and polymer, phenolic, photonics, light, glass, aluminum, brass, rubber, silicon, junction materials, nickel, paper/wood, ink, and resin.

    Why the comparison? When you think of a robot, do you always think of metal only? No. When you think of a machine, do you always think of gears? No. As the Earthly years of evolution progress upward, so does scientific technology.

    Robot technology is no longer limited to hardware only. We are beginning to see biological matter working its way into machines. A robot may have hardware components and yet connect to biological neurons from a rat’s brain. A monkey may move objects using a monkey brain and a machine robot arm. Nano machines are becoming so small, in some experiments they go into the human body to correct medical disorder. Humans can use brain thought and a computer to move virtual objects or control robots. When will this human brain control interface with electric brains?

    Therefore, the next step is the symbiotic union of biology and machine. This includes the brain. Now we are working to simulate neural matter and we can inject it into the Brain and distribute it. It is not too far fetched to induce some actual neural matter or components of some biological nature.

    Is biology the only technology that can integrate into the Brain? No. We have photonics - light, color, heat - thermodynamics, the physics of space-time, and other technology. Remember muscle wire? Apply a current through the wire and it contracts just as a biological muscle contracts. Now new gel materials are available to accomplish the same.

    The evolution of this technology will continue. It is proposed that a future brain be comprised with many technologies in a symbiotic union, a hybrid, a machine no longer based on electrics and computing technology only, but rather a futuristic device that transcends from one evolutionary technology base step to another.
  • HumanoidoHumanoido Posts: 5,770
    edited 2011-04-08 07:10
    Brain Breaks 100 Processor Chip Barrier!
    Setting a new record for the largest Propeller chip machine

    It's been an epic journey to elevate the number of Propeller chips in the Brain so to better use the method of creating and disseminating more machine neural matter. After concatenating a series of smaller Brains and larger Brains successfully, a record was achieved today. The previous Brain assemblage held 79 Propeller chips. Today that number of processors has exceeded by a quantity of over 100.

    Expanding material was placed on a series of solderless breadboards and semi-permanently held in a tower formation. After designing the High Density Breadboard, connecting a maximum number of Propellers became a piece of cake. HDBs are definitely a great design fully capable of running the Propeller at reliable clocking. Neural Distro is in parallel and simultaneous, and extremely fast. More will follow regarding this new powerful brain.
  • HumanoidoHumanoido Posts: 5,770
    edited 2011-04-08 07:22
    Brain Propeller Chip Reliability
    Parallax Achieves 100% Reliability

    Out of all Parallax Propeller chips tested, and used to construct more than five brains of assorted array sizes and configs, inclusive of extended burn in periods over hundreds of hours, not one Propeller has failed.

    This is a great credit to Parallax Inc. proving the reliability of the Propeller chip. This type of reliability is unheard of in the world's industry in which components can easily top ten percent fallout. The Propeller chip has consistently achieved 100% reliability.

    In various experiments, the Propeller was under clocked, over clocked, run with and without crystals, and run with all 32 ports per chip kept active and loaded, with varying software. The Propeller remains 100% reliable. We definitely want to thank the Parallax company for this great achievement.
  • HumanoidoHumanoido Posts: 5,770
    edited 2011-04-08 07:46
    The Next Generation Brain NGB
    A quantum leap of processors


    Work has accelerated on the incredibly massive NGB Next Generation Brain. With unprecedented stunning achievement of processor density, the Brain will increase in size by a factor of 1,000 times. Why? Higher numbers of machine neural processors more readily establish greater quanta of neural materials in machine arrays. This will lead to the potential modeling of small animal brains and/or the construction of a great machine intelligence.

    Once the technique was established for "quantizing" the Brain, it led to larger and larger Brain arrays.
  • HumanoidoHumanoido Posts: 5,770
    edited 2011-04-08 07:52
    A More Manageable Method of Brain Connection

    Rather than use various connectors, the simplicity of connecting multiple Brains is achieved by single pre-made color coded wire cables. This allows Brain processor arrays to be stacked on top of another, mounted end to end, or folded. The boards, if mounted, will fold, thus establishing higher densities. Cable mounts can take the form of single ends for dis-assembly and reconfigs. This is an easy way to leave the cables open ended and then reconstruct the modules at a moments notice.
  • HumanoidoHumanoido Posts: 5,770
    edited 2011-04-08 08:27
    Brain Containment
    Recycling Appliance Machines

    Depending on how large the Brain machine grows, a small discovery is made regarding potential Brain containment. There is essentially an unlimited supply of 220-volt vertical standing massive air conditioners available for a song and a dance. Delivery included. Thousands of these are scattered about at several locations in the city, in various styles and conditions of workability . These are perfect for the largest Brains that stand approximately 5 to 7-feet high and run 2-feet wide and deep and can send their brain-waves over wireless. The cooler apparatus already comes in tower format, has cooling compressors and coils, side mounts and spare room inside various compartments for mounting Brain boards.
  • HumanoidoHumanoido Posts: 5,770
    edited 2011-04-08 09:06
    Brain LED Array Redesign
    New circuit helps offload Propellers


    The LED output circuit is redesigned. In an effort to more offload the Propeller chip array, a 10K ohm resistor is now in use for each high brightness sub-sized LED. Experiments began with 2.2K and continued with standard increments up to the current 10K. For LEDs, the 2.2K ohm resistor is now replaced with 10K.
  • HumanoidoHumanoido Posts: 5,770
    edited 2011-04-08 09:18
    Experimenting with the Brain Modular Array

    On this coming Saturday, experiments will begin with the modular arrays. The various interconnections are under test. The results of this will lead to a more effective, useful, and reliable interface design. The next series of tests will use manual and automatic switching of HYBRID Interfacing. The Modular Array is an important aspect of the Brain because it allows the connection and interoperability of additional Brains.
  • HumanoidoHumanoido Posts: 5,770
    edited 2011-04-08 09:28
    Brain Hosting at Max

    The Brain is now fully developed for Brain Hosting. The EXO can accommodate approximately fifteen additional exterior boards and around 16 additional interior boards. This represents at least another 31 boards. Hosting is now possible using "Proximity" and cable connections. Proximity is almost unlimited in the numbers of processors it can append.
This discussion has been closed.