Excited about getting into it.

[Xasher]

I am an ametuer roboticist. I mostly design robots from the ground up but saw a great oppurtunity in Pleo. I have developed walking robots for the Trinity firefighting contest in Hartford CT and seen a few Aibos converted for the contest. I imagine my Pleo will be entered into the contest next year. The contest requires that a robot autonomously finds its way thru a mockup of a house, finds a candle and puts it out. I have a few questions though.
I see they use a flashable IC chip in pleo for his brain. Is this chip snapped into a slot or is it soldered onto a circuit board. I have actually programmed similar IC chips and designed a few circuit boards for IC chips similar to Pleos. I have a Hexadecimal compiler and was wondering if it was possible to completely reprogram the chip in Pleo ? Also are there any free I/O ports on the IC chip that more sensors could be added to pleo and programmed ? I realize Pleo has a ton of sensors so I don't suspect there are.
Also I know that Pleo has IR sensors in his nose I was wondering if they could be reprogrammed to measure distances so that Pleo could use them for object detection. I have used similar IR sensors in some of my robots and use them to measure distance. Does Pleos IR sensor have a send and recieve so that you can use them in this fashion ?
One more thing are Pleos motors similar to servos in the sense that you can digitally issue position commands to them to determine how far the servo will move ? I'm guessing this is the case since somewhere in the forums I saw something Pleo having a motor controller.
As you can tell I am very excited about diving into Pleos program. Thanks in advanced for anyone who responds to this post.:D

[PleoPet]

Selective comments and advice (based on my own research, please keep asking UGOBE for more technical info):

> I see they use a flashable IC chip in pleo for his brain.
Everything is soldered into place. The days of removing chips from a socket to reprogram them are over !->
There are two ARM chips: http://www.aibohack.com/pleo/tech1.htm
They can be 'reflashed' using software (running from the SD card, no need to remove the skin or do any hardware mods)
The main ARM chip that interprets the Pawn program has 256KB of Flash ROM. The head ARM chip has 32KB of Flash ROM for the camera, microphone, IR and other things.
The most common programming (supported by the PDK) will be Pawn programming which doesn't change either of these chips (the Pawn program runs from the SD card, or optionally the internal 4MB data store).

> ...wondering if it was possible to completely reprogram the chip in Pleo ?
Technically yes
It's up to UGOBE if they want to release that much technical info. BTW: That's not in the list of things you can do with the pre-announced PDK. http://www.pleoworld.com/developer

> Also I know that Pleo has IR sensors in his nose I was wondering if they could be reprogrammed to measure distances...
IIRC: there is no intensity on the IR reflection signal. Even if there were, it makes a poor distance sensor (only a rough indication of size or distance of reflected object).
> ...I have used similar IR sensors in some of my robots and use them to measure distance.
Accurate IR distance sensors are built specially, like the Sharp IR line. You could add them externally to the robot. http://www.acroname.com/robotics/info/articles/sharp/sharp.html
BTW: The serial port on the bottom can be used to add external logic (without taking off the skin)

[UgobeTyler]

To add to PleoPets comments:

> I see they use a flashable IC chip in pleo for his brain. Is this chip snapped into a slot or is it soldered onto a circuit board.
> I have actually programmed similar IC chips and designed a few circuit boards for IC chips similar to Pleos. I have a
> Hexadecimal compiler and was wondering if it was possible to completely reprogram the chip in Pleo ?

The ARM7 processors, as PleoPet mentioned, as soldered in place, along with the 4MB DataFlash where we store the default Pleo personality. They are both flashable, and the code to flash new images is built into the existing firmware. The initial PDK release will likely not contain enough information to create a whole new firmware for either ARM, but that doesn't mean somebody could not do it. The initial PDK will focus on using Pawn scripts to control Pleo.

> Also are there any free I/O ports on the IC chip that more sensors could be added to pleo and programmed ? I realize
> Pleo has a ton of sensors so I don't suspect there are.

We have pretty much used up all the I/O available on the two ARM7 processors. As you say, there are already a ton of sensors in Pleo. If you have any specific ones you are interested in, we would love to know about them.

> Also I know that Pleo has IR sensors in his nose I was wondering if they could be reprogrammed to measure distances
> so that Pleo could use them for object detection. I have used similar IR sensors in some of my robots and use them to
> measure distance. Does Pleos IR sensor have a send and recieve so that you can use them in this fashion ?

We actually use the IR sensor in Pleo for "object" detection (as well as a Pleo "beacon"). As PleoPet mentioned, it cannot give very accurate distance measurement, but it can certainly the presence or absence of an object in front of Pleo.

> One more thing are Pleos motors similar to servos in the sense that you can digitally issue position commands to
> them to determine how far the servo will move ? I'm guessing this is the case since somewhere in the forums I saw
> something Pleo having a motor controller.

The Pleo motion control system is made of multiple components. Together they allow this high-level "move to position X" commands, though there are no servos in the system.

[Xasher]

Pleopet, I have a few of the sharp IR sensors. I was actually thinking of designing a brain similar to the stamp 2 and putting inside a small backpack pleo could where on his back. This would allow me to add sensors to Pleo without having to mess with him internally. The hard part would be getting any data that the backpack brain recieves to the internal IC chip so that Pleo could respond to it. I would probably write the code for the backpack so that it uses interrupt requests and sends signals through specific I/o ports, then I can hook those I/o ports to pleos brain so that he can respond to the inputs from the external sensors. Again the hardest part would be connecting to the internal IC chip in pleo in a way that his internal program could react to the external signals. That is why I asked if there were any additional I/O ports available.
The only other way I have found to utilize an IC chip if all it's I/O ports are used would be to use ports that are used to send power or information out but are not utilized for recieving signals, somethimes you can plug into these ports and rewrite the program to monitor the I/O for input as well as sending out output. It can be tricky and you have to closely monitor wether an I/O port is a digital or analogue type. The schematics for most IC chips are available to the public over the internet and that information usually is easily attained if you have the IC chip name.
Is it possible to feed sensor information to the Pleos brain thru the USB port in his belly while he is active? Plugging the backpack into the USB port would be the most logical and efficient way of doing this.

[PleoPet]

>Is it possible to feed sensor information to the Pleos brain thru the USB port in his belly while he is active?
Without taking off the skin/shell, there are two options.
The USB port, or the serial debug port (under the glued plastic cover) http://www.pleodreams.com/files/pleo-aux-port.jpg

Both can act like serial port for sending console commands and doing other things.

The USB port uses a standard connector. I suspect the USB port doesn't have power. Also the Pleo must be up and running adequately for the USB device to be detected.
The hardest problem is making a small embedded controller act like a USB 'host' (the Pleo is a USB client/slave). It is technically possible, but so far hasn't been done. You will likely need a dedicated 'host capable' USB chip (much less common than USB client chips)

Instead I recommend the physical serial debug port. The hard part is finding the connector, the rest is much easier and more reliable:
It provides 3.3Volts of power, and a direct serial port (3.3V CMOS level). It is very easy to hook up to most common embedded processors for the addon brain. If the addon brain runs off of 3.3V and can do RS-232 signals up to 115,000baud, then you don't need extra chips.
BTW: this is already working (for XBee and Bluetooth RF) http://www.aibohack.com/pleo/rf_pleo.htm

[Shagz]

I'm really excited about this too. I just purchased a Pleo, but it has not arrived. I'm so anxious that i'm already getting all the dev-tools together while i wait.

AI has always been a hobby, Previously i was attempting to do it on a PC where hardware can be anything you want. Ended up joining an AI group, Virtual Humans...

Turns out, it wasn't AI at all.. It was simply a scripted list of "if they type this, reply this", and any mention of a learning, self-sufficient, AI process ... got scared responses. Needless to say I quit the group and decided to go pursue true AI.

I think Ugobe and myself have similar goals, and the Pleo should be really quite helpfull, it all all the hardware that i don't have for "PC". (altho, i do kinda wich is had my Logitech face-recognition camera inside it's head, instead :) )