Perhaps this is somewhat of a weird task for your machine, but it would be cool to develop a weather forecasting module. I think a machine of your size could handle a watered down version... You know something like a machine that could take in a bunch of environmental variables from multiple sources(Air pressure, temperature, wind speed/ direction, humidity, dew point, sunlight..etc) It would then remember all of these readings(from all over a small area, perhaps the size of a small state) and slowly it would "learn" the weather and would forecast it based upon the previous inputs and by what happened when certain inputs were present. The machine couldn't obviously forecast to the extent that NOAA is capable of, but it would be cool. Speaking of NOAA, you could use them as a source of weather data. They freely supply a bunch of different weather data for your area on their website. "DIY weather forecast device". I can see it now in the next Popular science issue:)..I am also somewhat of a weather buff to. I used to want to be a tornado chaser when i was a kid, that's what i get for being born in Texas:)
Hi Humanoido,
Trying to answer your first question about the "Big Brain" I'm closer with the ressponse #5 from "Whit"
So here is my perfect robotic "Big Brain"
Brain1: Survivability e.g. by returning to a recharging station, or keeping itself in light where it could recieve recharging through solar panels or tracking the solar panels (at this point I can contribute)
Brain2: Ability to move autonomus and avoid obstacles, and also to choose the best way (Maze problem)
Brain4: Ability to take a variety of measurements (use any kind of sensors) (I aggre with "Ravenkallen" )
Brain3: remote comunication with it master (ability to send data and receive commands)
Brain5: Ability to perform task using a robot arm
Brain6: Ability to transform Itself (change shape or leave some mechanical parts in order to achieve something).
Actually I dream this brain for my rover.
I suppose that your goals are more sophisticated and slightly different but you can take into account something of the above.
E.g. Have you ever thought your robotic "big Brain" to have a mechanism that can add or remove the aproppriate card in a specific socket?
Perhaps this is somewhat of a weird task for your machine, but it would be cool to develop a weather forecasting module. I think a machine of your size could handle a watered down version... You know something like a machine that could take in a bunch of environmental variables from multiple sources(Air pressure, temperature, wind speed/ direction, humidity, dew point, sunlight..etc) It would then remember all of these readings(from all over a small area, perhaps the size of a small state) and slowly it would "learn" the weather and would forecast it based upon the previous inputs and by what happened when certain inputs were present. The machine couldn't obviously forecast to the extent that NOAA is capable of, but it would be cool. Speaking of NOAA, you could use them as a source of weather data. They freely supply a bunch of different weather data for your area on their website. "DIY weather forecast device". I can see it now in the next Popular science issue:)..I am also somewhat of a weather buff to. I used to want to be a tornado chaser when i was a kid, that's what i get for being born in Texas:)
Ravenkallen, this is definitely food for thought, and a project that's possible. Reading net data is now a reality for the Propeller chip so it's doable in some fashion. I see that Parallax now offers many sensors that are atmospheric related. I'm using some on a very amateur space program exploring the lower atmosphere. The brain could certainly handle ground based exploration as well and perhaps process telemetry from the probe or satellite. Thanks for the NOAA weather tips and we should see how this data can be read and entered into the Propeller. Albeit risky, storm chasing is a very exciting hobby! Though, I'd rather watch it on TV on the Discovery channel.
I think you're underestimating the complexity. I see code executing in a COG on every Prop that deals with bus arbitration and some kind of packet protocol. You'll need a way to process and organize the disparate EEPROM data. What kind of bus do you envision? Full-Duplex could be a little tricky. Half-Duplex?
Mike G, good point, I do tend to oversimplify. If I can keep it simple, then I can probably understand it and make it doable. Processing and organizing the disparate EEPROMs will be assigned to the master or a sub master. I haven't decided yet on how the self enumeration will take place which in turn will lead to the slave addressing but that's the first level to get to the EEPROMS. In the SEED supercomputer a pin was dedicated to an RC circuit for such purpose. I'd like to try something different in the brain. Sure, the brain has 2 wires which can do half or full duplex. Additional pins are available for 8-bit parallel. The singe wire and the two wire bus is now working though I'm studying several communications packages and already wrote/tested a very simple communications software - simple enough to fit into one cog and have left over space for main code for the master. As you said, maybe it's too simple, it may need more development, but for now the goal is to get things working.
E.g. Have you ever thought your robotic "big Brain" to have a mechanism that can add or remove the aproppriate card in a specific socket?
NikosG, Wow - you have lots of great ideas! That would be cool if the brain could have a lot of spindly arms that could act as a card changer and also serve as legs to move around and fix things. The toes on the legs could be small hands to reach and grasp things, control things, and perform some tasks.
Software Overview
It's talking (the brain) but now another co-software is being developed to up the complexity so the brain can load its own software (self loading) and do self modifying programming. Self loading makes it easier to stick code into 20 or more boards and self modifying program is good for recycling code space as RAM is somewhat of a premium at this point. It's a balancing act with the Propeller cogs. More on this after testing.
Hardware Development
Hand sketched a circuit to lock in a method for a hybrid design today. This can one wire single duplex communicate at high speed or accomplish two wire full duplex at high speed. There's communication in a single cog with programming space left over for both methods and it can hardware adjust for speed as well. I look forward to trying out the designs and testing and reporting back here.
Brain2: Ability to move autonomus and avoid obstacles, and also to choose the best way (Maze problem)
NikosG, another good one - there's staid algorithms for maze solving and a brain this large could store and utilize a rather sophisticated one. I think this could do new things for maze solving robots in terms of achievements. Maybe it could solve a maze faster or more efficiently. It could certainly interface to more sensors. A sensor at each side, and one in front would be easy to add. How about a very special sensor - a high up telescoping mirror could look at the maze from top down and ascertain a solution in a different manner. In contests however, make sure it's not against the rules.
Brain6: Ability to transform Itself (change shape or leave some mechanical parts in order to achieve something).
Actually I dream this brain for my rover.
NikosG, I think maybe you read this link 4D Morphing Computer (with CoProcessor).
I believe there are many untapped venues for a shape shifting brain. The concept undoubtedly works with software as well as hardware. If the brain carries its own tools and parts, it could shape shift into a new configuration, by perhaps removing one assembly and attaching another. Various assemblies can store in a spare brain cavity.
Humanoido, would you be so kind as to post your 1 and 2 wire communication demo code? I'm working on a Spinneret project and one aspect is communication with several other propeller controlled sensor devices. Don't want to reinventing the wheel.
I imagine you're using some kind of a bootstapper to "Self Load" code into the multiple Props? Maybe similar to the ZigBee bootloader blog/post? Is there anyway that we could get a look at that code as well? I was thinking it would be pretty cool to update firmware by HTTP using the Spinneret.
Mike G, I'm as anxious to post code as you are to see it but it's not quite ready for human consumption but definitely stay tuned and as soon as it's a done deal you'll see it and can download it from the post and use it for your projects!
Another thing I should mention is that software currently uses BASIC Stamp commands beginning at the first level of the Brain Stem which filters upwards through the Brain Base and on through the Brain Span - I wouldn't want you to wait for the code only to find this type of programming and be disappointed. I expect the programming will be three tier eventually so more development will happen.
The self loader, as mentioned below, is still under development and requires board design changes to work. I hand sketched a design and it's under construction. It uses the hybrid design mentioned earlier. No software yet for this. If all goes well, in my mind, there will be bus single wire, two wire, parallel interfaces, along with self loading, eeprom and crystal propagation. The first model brain however doesn't need all that. We can leave some room for future development. The first model brain can be like the first model T car.
Humanoido wrote: Software Overview: It's talking (the brain) but now another co-software is being developed to up the complexity so the brain can load its own software (self loading) and do self modifying programming. Self loading makes it easier to stick code into 20 or more boards and self modifying program is good for recycling code space as RAM is somewhat of a premium at this point. It's a balancing act with the Propeller cogs. More on this after testing.
NikosG, another good one - there's staid algorithms for maze solving and a brain this large could store and utilize a rather sophisticated one. I think this could do new things for maze solving robots in terms of achievements. Maybe it could solve a maze faster or more efficiently. It could certainly interface to more sensors. A sensor at each side, and one in front would be easy to add. How about a very special sensor - a high up telescoping mirror could look at the maze from top down and ascertain a solution in a different manner. In contests however, make sure it's not against the rules.
This would be very cool!!! but I haven't done any experiment with artificial vision yet. I suppose that I must convert the simple image from the camera to useful information. Maybe a CMUcam can do this.....?. Is one of my future plans.
Anyway I have done something about mazes using my BoeBot, IR sensors and algorithm but I'm not very contented. I'll publish a project soon.
This would be very cool!!! but I haven't done any experiment with artificial vision yet. I suppose that I must convert the simple image from the camera to useful information. Maybe a CMUcam can do this.....?. Is one of my future plans.
Anyway I have done something about mazes using my BoeBot, IR sensors and algorithm but I'm not very contented. I'll publish a project soon.
NikosG, the extent of the CMUcam experiments that I did was on Boe-Bot and a Penguin robot where it could follow a colored ball. My favorite maze solving routine was the most simple using the right hand rule. Top down would be handled in a much more complex way by outlining the maze using some contrast edge detection and then solve the maze before walking it. I look forward to seeing your project!
Work begin on the LED mod. The photos show a yellow wire lead to pin 24 for tests. This re-enables use of the Parallax Propeller Proto Board's power LED and re-purposes it as a data light.
Power is fed to the VSS and Vdd junctions from the pin connector.
The procedure requires scratching off the board resin before soldering from the back side. Wire wrap was not necessary.
Test Software
Here's some code used to test the LED modification on the Parallax Propeller Proto Board. The modification is described above in several posts.
It also runs on the Parallax Demo Board by changing the pin from 24 to 23. This is the reason for the LED ghost correction.
{{
The Brain Blob
LED Mod Test on the Brain Blob Project
Blinks LED on pin 24 using the Parallax Propeller Proto Board
Use pin 23 for demo board
by Humanoido
Friday December 31, 2010
}}
CON
_clkmode = xtal1 + pll16x ' 80MHz clock
_xinfreq = 5_000_000 ' 5MHz crystal
LED = 24 ' LED mod is on pin 24 Proto Board, pin 23 demo board
delay = 40_000_000 ' Approx 1 second on and off
PUB Main
dira[LED] := 1 ' Set LED pin as output
dira[22] := 1 ' correct ghost led on demo board
outa[22] := 0 ' correct ghost led
outa[LED] := 1 ' Initialize pin on
repeat ' Loop below
blink ' Do Private Method to blink LED
PRI Blink ' Private Method to blink LED
!outa[LED] ' Toggle state LED
waitCnt(delay+cnt) ' Delay
Power Regulator Modification
This modification to the LM1086 CS-3.3 is now completed on all 20 Parallax Propeller Proto Boards.
The mod involves desoldering the voltage regulator's middle leg to disable all circuits behind the device. Power is fed to all boards from an external power source. This reduces power consumption by disabling the power LED and two regulators.
This disables the power LED, however, the small modification described in the previous posts enables it to be recycled and function as a data light.
Over the holiday the data light will be installed in as many boards as possible. Later, testing will be a separate project. Is there any simple way to determine the wire gauge that fit into this tiny hole by the power slide switch (near the LED) in the Parallax Propeller Proto Board? (see photo in previous posts, with the yellow insulated wire)
I'm lining up the next project that will follow the LED mod installs. The decoupling capacitor will be installed into all boards and then tested. The question is, what is the best code to test the propeller functionality with and without a decoupling capacitor??? My guess is the acid test is hooking up a scope to see the signal conformity. Since I travel, a large scope is not possible. It looks like the small Prop Scope could fill this requirement. Anyone have experience with this and recommendations?
Brain Blob (Dangerous Growth Spurt?) Checkout these latest pics of the Brain Blob
I knew it would happen. As the New Year's Holiday began, and with steady work on the project, I noticed something different about the Brain Blob. Once again, it had grown larger, now consuming the greatest portion of the Parallax laboratory project table. I was surprised, as you can see in the photo, the larger span of this entity seems out of control as the boards are coming together and multiplying in some odd linear X and blobular Y directional way.
The latest Brain Blob growth is shocking.
When will it stop? The Blob on the radio
program - WLS Chicago, continued to grow
until it consumed the Earth and everyone
on it.
Stay tuned to this channel thread and see what
happens with the Brain Blob. No responsibility is
claimed for this project and we disavow any
knowledge of its existence at this time.
Decoupling Capacitor The board assembly was completed a short time ago by installing all of the decoupling capacitors. These are 10nF inserted into pin sockets across from pin 24 socket at Vss and Vdd.
Testing Phase
Completed the testing phase of all boards. Have 17 boards that work perfect. One board has an LED that won't light and two boards have the LED lit all the time but the code still blinks it brighter and dimmer. The next work phase is to troubleshoot three boards.
Troubleshooting Phase This will be a checklist for troubleshooting the three boards.
Class I - LED is Dead
( ) Check the pin socket
( ) Check solder joint on back side
Class II - LED is Shorted On but Still Blinks with Code
( ) Check for short on front side printed circuit board ( ) Board number
( ) Check for short on back side trace ( ) Board number
( ) Check for short on front side printed circuit board ( ) Board number
( ) Check for short on back side trace ( ) Board number
When I first came across your posts, I though wow these guys really have it together. Connecting many STAMPs or Propeller together to make a supercomputer, wow, that is a serious endeavor.
I spent several hours reading posts and following links from the 40 board Propeller skyscraper to the STAMP SEED supercomputer. I have to conclude that you have created nothing more than a bunch of distinct and separate boards stacked one on top of the other. I looked at the STAMP super computer. It's just a bunch of STAMP boards connected to pin 0.
Sorry for the rant. I'm just mad at myself for spending time on this. I wish you the best.
Sorry for the rant... I'm just mad at myself... I wish you the best.
Remember, these hobby machines are designed to satisfy my purposes. You could also say the Propeller chip is just a bunch of connected cogs. You would be right. Parallax put it out there and by itself it does nothing - but in the hands of a person that can wrap their mind around it, new worlds are opened.
The BSS follows a standard of Master-Slaves concept explained in the Parallax BASIC Stamp Book, as does the SEED and the US40 machines. Pin zero one wire serial half duplex interface bus - absolutely yes. A bunch of Propellers, and BASIC Stamps, connected together - absolutely yes. Distinct and separate boards only stacked - absolutely not. (Even if it was, that would be ok - refer to the discussion on art and visual representation.)
The boards are all connected, software driven at the same time, and talking to each other. Each machine has a specific purpose. Reviewing, for example, SEED software will show multiple Stamps listening at the same time and responding (talking) and processing data simultaneously in true parallel. Programs run at the same time as machines demonstrate various concepts of multi-processing, self enumerating, true parallelism, self evolving, memorizing, AI, determinism, aggregate memory, and so on.
Troubleshooting Results All boards are Operational
Board one had an LED that would not light - it was open from the front side board plated through hole to the back side, resulting in an open circuit and no LED response. The fix was scratching and removing the dielectric resin on the front side and re-soldering the through-hole wire on both sides.
Board two had the LED constantly on, but driven to increased brightness. Troubleshooting led to the trace under the second voltage regulator which had a tiny remnant trace fragment shorting out to the adjacent trace. The solution was to remove the trace.
Board three had a tiny short circuit surrounding the solder island at the back side. This was resoldered, solder wicked, and finally a knife was used to sever the ill connect point. Two out of the three boards required a microscope to detect the troubleshooting points of interest.
Comments
Trying to answer your first question about the "Big Brain" I'm closer with the ressponse #5 from "Whit"
So here is my perfect robotic "Big Brain"
Brain1: Survivability e.g. by returning to a recharging station, or keeping itself in light where it could recieve recharging through solar panels or tracking the solar panels (at this point I can contribute)
Brain2: Ability to move autonomus and avoid obstacles, and also to choose the best way (Maze problem)
Brain4: Ability to take a variety of measurements (use any kind of sensors) (I aggre with "Ravenkallen" )
Brain3: remote comunication with it master (ability to send data and receive commands)
Brain5: Ability to perform task using a robot arm
Brain6: Ability to transform Itself (change shape or leave some mechanical parts in order to achieve something).
Actually I dream this brain for my rover.
I suppose that your goals are more sophisticated and slightly different but you can take into account something of the above.
E.g. Have you ever thought your robotic "big Brain" to have a mechanism that can add or remove the aproppriate card in a specific socket?
It's talking (the brain) but now another co-software is being developed to up the complexity so the brain can load its own software (self loading) and do self modifying programming. Self loading makes it easier to stick code into 20 or more boards and self modifying program is good for recycling code space as RAM is somewhat of a premium at this point. It's a balancing act with the Propeller cogs. More on this after testing.
Hand sketched a circuit to lock in a method for a hybrid design today. This can one wire single duplex communicate at high speed or accomplish two wire full duplex at high speed. There's communication in a single cog with programming space left over for both methods and it can hardware adjust for speed as well. I look forward to trying out the designs and testing and reporting back here.
I believe there are many untapped venues for a shape shifting brain. The concept undoubtedly works with software as well as hardware. If the brain carries its own tools and parts, it could shape shift into a new configuration, by perhaps removing one assembly and attaching another. Various assemblies can store in a spare brain cavity.
I imagine you're using some kind of a bootstapper to "Self Load" code into the multiple Props? Maybe similar to the ZigBee bootloader blog/post? Is there anyway that we could get a look at that code as well? I was thinking it would be pretty cool to update firmware by HTTP using the Spinneret.
Another thing I should mention is that software currently uses BASIC Stamp commands beginning at the first level of the Brain Stem which filters upwards through the Brain Base and on through the Brain Span - I wouldn't want you to wait for the code only to find this type of programming and be disappointed. I expect the programming will be three tier eventually so more development will happen.
Humanoido wrote: Software Overview: It's talking (the brain) but now another co-software is being developed to up the complexity so the brain can load its own software (self loading) and do self modifying programming. Self loading makes it easier to stick code into 20 or more boards and self modifying program is good for recycling code space as RAM is somewhat of a premium at this point. It's a balancing act with the Propeller cogs. More on this after testing.
This would be very cool!!! but I haven't done any experiment with artificial vision yet. I suppose that I must convert the simple image from the camera to useful information. Maybe a CMUcam can do this.....?. Is one of my future plans.
Anyway I have done something about mazes using my BoeBot, IR sensors and algorithm but I'm not very contented. I'll publish a project soon.
Work begin on the LED mod. The photos show a yellow wire lead to pin 24 for tests. This re-enables use of the Parallax Propeller Proto Board's power LED and re-purposes it as a data light.
Power is fed to the VSS and Vdd junctions from the pin connector.
The procedure requires scratching off the board resin before soldering from the back side. Wire wrap was not necessary.
Here's some code used to test the LED modification on the Parallax Propeller Proto Board. The modification is described above in several posts.
It also runs on the Parallax Demo Board by changing the pin from 24 to 23. This is the reason for the LED ghost correction.
{{ The Brain Blob LED Mod Test on the Brain Blob Project Blinks LED on pin 24 using the Parallax Propeller Proto Board Use pin 23 for demo board by Humanoido Friday December 31, 2010 }} CON _clkmode = xtal1 + pll16x ' 80MHz clock _xinfreq = 5_000_000 ' 5MHz crystal LED = 24 ' LED mod is on pin 24 Proto Board, pin 23 demo board delay = 40_000_000 ' Approx 1 second on and off PUB Main dira[LED] := 1 ' Set LED pin as output dira[22] := 1 ' correct ghost led on demo board outa[22] := 0 ' correct ghost led outa[LED] := 1 ' Initialize pin on repeat ' Loop below blink ' Do Private Method to blink LED PRI Blink ' Private Method to blink LED !outa[LED] ' Toggle state LED waitCnt(delay+cnt) ' DelayThis modification to the LM1086 CS-3.3 is now completed on all 20 Parallax Propeller Proto Boards.
The mod involves desoldering the voltage regulator's middle leg to disable all circuits behind the device. Power is fed to all boards from an external power source. This reduces power consumption by disabling the power LED and two regulators.
This disables the power LED, however, the small modification described in the previous posts enables it to be recycled and function as a data light.
Photo relinked here
Checkout these latest pics of the Brain Blob
I knew it would happen. As the New Year's Holiday began, and with steady work on the project, I noticed something different about the Brain Blob. Once again, it had grown larger, now consuming the greatest portion of the Parallax laboratory project table. I was surprised, as you can see in the photo, the larger span of this entity seems out of control as the boards are coming together and multiplying in some odd linear X and blobular Y directional way.
The latest Brain Blob growth is shocking.
When will it stop? The Blob on the radio
program - WLS Chicago, continued to grow
until it consumed the Earth and everyone
on it.
Stay tuned to this channel thread and see what
happens with the Brain Blob. No responsibility is
claimed for this project and we disavow any
knowledge of its existence at this time.
These boards have their mods
completed.
Work now enters the testing phase
to test each of the mods on twenty
boards. This equals 40 tests and
any troubleshooting that may be
required.
The attached file can be used to test the LED data light mod.
The board assembly was completed a short time ago by installing all of the decoupling capacitors. These are 10nF inserted into pin sockets across from pin 24 socket at Vss and Vdd.
Completed the testing phase of all boards. Have 17 boards that work perfect. One board has an LED that won't light and two boards have the LED lit all the time but the code still blinks it brighter and dimmer. The next work phase is to troubleshoot three boards.
This will be a checklist for troubleshooting the three boards.
Class I - LED is Dead
( ) Check the pin socket
( ) Check solder joint on back side
Class II - LED is Shorted On but Still Blinks with Code
( ) Check for short on front side printed circuit board ( ) Board number
( ) Check for short on back side trace ( ) Board number
( ) Check for short on front side printed circuit board ( ) Board number
( ) Check for short on back side trace ( ) Board number
I spent several hours reading posts and following links from the 40 board Propeller skyscraper to the STAMP SEED supercomputer. I have to conclude that you have created nothing more than a bunch of distinct and separate boards stacked one on top of the other. I looked at the STAMP super computer. It's just a bunch of STAMP boards connected to pin 0.
Sorry for the rant. I'm just mad at myself for spending time on this. I wish you the best.
The BSS follows a standard of Master-Slaves concept explained in the Parallax BASIC Stamp Book, as does the SEED and the US40 machines. Pin zero one wire serial half duplex interface bus - absolutely yes. A bunch of Propellers, and BASIC Stamps, connected together - absolutely yes. Distinct and separate boards only stacked - absolutely not. (Even if it was, that would be ok - refer to the discussion on art and visual representation.)
The boards are all connected, software driven at the same time, and talking to each other. Each machine has a specific purpose. Reviewing, for example, SEED software will show multiple Stamps listening at the same time and responding (talking) and processing data simultaneously in true parallel. Programs run at the same time as machines demonstrate various concepts of multi-processing, self enumerating, true parallelism, self evolving, memorizing, AI, determinism, aggregate memory, and so on.
I wish you the best in your studies.
All boards are Operational
Board one had an LED that would not light - it was open from the front side board plated through hole to the back side, resulting in an open circuit and no LED response. The fix was scratching and removing the dielectric resin on the front side and re-soldering the through-hole wire on both sides.
Board two had the LED constantly on, but driven to increased brightness. Troubleshooting led to the trace under the second voltage regulator which had a tiny remnant trace fragment shorting out to the adjacent trace. The solution was to remove the trace.
Board three had a tiny short circuit surrounding the solder island at the back side. This was resoldered, solder wicked, and finally a knife was used to sever the ill connect point. Two out of the three boards required a microscope to detect the troubleshooting points of interest.