Stateside
Posts: 66
Joined: Sun May 13, 2012 12:26 pm

Re: Using RPi as a Standard Robotics Interface

Thu May 31, 2012 3:46 pm

Hacking the Fluke . http://calicoproject.org/Hacking_the_Fluke
The Fluke is a small electronic board that contains wireless Bluetooth, camera, IR (infrared) sensors, LEDs (light emitting diodes) and an ARM microprocessor.The Fluke sits between software running the Myro API and a robot. In fact, you can talk to a robot (such as the Scribbler) directly.Some of the following byte codes can be used without the Fluke, as they are Scribbler commands.
The Fluke handles camera commands, obstacle commands, and brightness commands.The Scribbler handles everything else (sounds, movement, IR, battery, etc).
The Fluke (and underlying robots) are controlled through sending byte codes, or messages. The interface is used in the following files:
Python, host-side interface
C#, host-side interface
BASIC, Scribbler-side interface
The byte-codes are: ............

User avatar
speculatrix
Posts: 51
Joined: Thu Apr 05, 2012 8:29 am
Location: France
Contact: Website

Re: Using RPi as a Standard Robotics Interface

Thu May 31, 2012 6:49 pm

Just to add my two-pennorth...

I'm interested in using the Pi for robotics in the same way I'm currently using a BeagleBone - for high-level logic.

Not everything in robotics is about hooking up to sensors or split-second timing. I'm currently using a BB with Python code for the robot's more intellectual processing. It communicates via serial with two Arduinos.

One Arduino runs the motors and handles obstacle avoidance (IR proximity sensors & rangerfinders plus contact switches). It has a limited level of autonomy - eg, moving away from obstacles.

The other Arduino handles various sensors, plus some servos for a moving 'head' that carries rangefinders etc.

The Arduinos and BeagleBone pass messages where necessary - eg, the BB might tell the motor controller 'turn left 90 degrees' but leaves it up to the Arduino to figure out how to do that. In return, the motor unit might send a message to the BB that it has just turned 30 degrees left in order to avoid something (a fact the BB software can build into its mapping).

Up to the current robot, I'd only used Arduinos. Adding the BB allowed me to start writing far more extensive software, without having to worry too much about saving a byte here and there. (That sort of optimisation is all very well and you can argue it's good practice, but it's a chore and gets in the way of the important and interesting stuff). And I can write in Python, my favourite language.

In addition, I've finally managed (just about) to get the Kinect working with the BB (something I'd definitely like to do with the Pi). That's something I couldn't have done with the Arduino.

I like the idea of a module robot design using distributed processing. So the idea of something like the Pi or BB being the main 'brain' of a robot commanding microprocessor-controlled, black box-style modules (for motors, sensors, etc) is very appealing to me.

Stateside
Posts: 66
Joined: Sun May 13, 2012 12:26 pm

Re: Using RPi as a Standard Robotics Interface

Thu May 31, 2012 7:50 pm

I have reached out to the Myro Users group.The rely is as follows:

Interesting question:

Is it possible to replace the IPRE Fluke with the Raspberry pi?
http://www.raspberrypi.org/faqs
If so it would help reduce the hardware costs to schools and perhaps more
schools could implement computer training using robots?

It should be possible (as you can use the Scribbler directly connected to
a serial cable, although that use might not be completely tested).

But don't forget about Fluke's camera, one of the best parts.

Although, it may be cheaper to bundle a Raspberry Pi with a webcam, and
perhaps even some kind of wireless connection. Interesting idea.

If someone makes a do-it-yourself Fluke-like thing, please let us know!

Stateside
Posts: 66
Joined: Sun May 13, 2012 12:26 pm

Re: Using RPi as a Standard Robotics Interface

Fri Jun 01, 2012 4:32 am

speculatrix wrote:I'm interested in using the Pi for robotics in the same way I'm currently using a BeagleBone - for high-level logic.

I'm currently using a BB with Python code for the robot's more intellectual processing. It communicates via serial with two Arduinos.

One Arduino runs the motors and handles obstacle avoidance (IR proximity sensors & rangerfinders plus contact switches). It has a limited level of autonomy - eg, moving away from obstacles.

The other Arduino handles various sensors, plus some servos for a moving 'head' that carries rangefinders etc.

The Arduinos and BeagleBone pass messages where necessary - eg, the BB might tell the motor controller 'turn left 90 degrees' but leaves it up to the Arduino to figure out how to do that. In return, the motor unit might send a message to the BB that it has just turned 30 degrees left in order to avoid something (a fact the BB software can build into its mapping).

I like the idea of a module robot design using distributed processing. So the idea of something like the Pi or BB being the main 'brain' of a robot commanding microprocessor-controlled, black box-style modules (for motors, sensors, etc) is very appealing to me.
That is exactly how the IPRE Fluke board controls the Scribbler robot. In stead of 2 Arduino boards it uses a "Propeller " board that contains Microcontroller- P8X32A-Q44 multi-core processor with eight processing cogs. (cores). which can operate simultaneously at your will. Using the Propeller is like employing a team of up to eight workers for a project; the team members can work in parallel on given tasks and coordinate as needed to achieve a common goal. Being truly flexible and efficient, they can share, shuffle, and dedicate to duties, quietly wait for events, start and stop, and direct each other as needed.

Cogs have exclusive access to their own internal memory and unimpeded access to the System Clock and all 32 I/O pins. Each cog tracks I/O pin states with its own input register and influences pin outputs using its own output and direction registers; the collective of cogs determines an I/O pin's ultimate direction and state.

Cogs share access to main memory in a round-robin fashion through a central hub. Among other things, main memory is used in this fashion for coordination between cogs.

In addition, each cog contains hardware to assist with certain high-speed, repetitive tasks such as signal detection and video generation.

http://www.parallax.com/portals/0/help/P8X32A/QnaWeb/

http://www.parallax.com/Store/Microcont ... fault.aspx

I am a novice at Electronics so sorry I could not summarize the above in my own words like you did.

User avatar
mkopack
Posts: 242
Joined: Mon Nov 07, 2011 8:46 pm

Re: Using RPi as a Standard Robotics Interface

Fri Jun 01, 2012 11:55 am

That's true for the scribbler 2 (red one). The original (blue) scribbler didn't have that sort of processor on it...

sv507
Posts: 1
Joined: Sun Jun 17, 2012 7:33 pm

Re: Using RPi as a Standard Robotics Interface

Sun Jun 17, 2012 8:58 pm

sorry to hijack this topic but ... but how does raspberry pi compare to scribbler2 +fluke? for robot implementations?


my own interests are in computer vision. So what i was interested in was how to do image processing on a robot CHEAPLY!... so the idea of using fluke to send images to process on main computer sounded like quite a good approach- unfortunately I have yet to find any projects using fluke that do this.
I was interested in how people thought this would compare to using raspberry pi directly on board... I saw fanjita's link...anyone can point me to other raspberry pi robot implementations? or any alternative approaches - under $200.

mande01
Posts: 6
Joined: Sun Jul 22, 2012 12:53 am

Re: Using RPi as a Standard Robotics Interface

Sun Jul 22, 2012 1:01 am

@ speculatrix do you have more information on your project? As in all of it!
I tried your web site but it was down, I'm only starting at the robot stuff, but I would love to try and recreate what you are doing! At least the rpi and Arduino motor control,
I would like to have the rpi online controlling the robot.
Thanks,
Derry
Derry.Manley at good old gmail.com

User avatar
speculatrix
Posts: 51
Joined: Thu Apr 05, 2012 8:29 am
Location: France
Contact: Website

Re: Using RPi as a Standard Robotics Interface

Sun Jul 22, 2012 7:20 am

mande01 wrote:@ speculatrix do you have more information on your project? As in all of it!
Unfortunately, the robotics project has had to take a back seat while I concentrate on other things - such as earning a living. -=sigh=- What I have been doing has mostly involved tinkering with the control systems.

Although I've been playing around with the RPi as the main Command & Control (C&C) device, I'm still more inclined to use a BeagleBone for that purpose. The latter is less fussy about the power supply, and has four UARTs for serial comms with other controllers (ie, Arduinos).

I've developed a messaging protocol for passing commands, alerts and data between devices. The way things stand at the moment, I foresee having three devices - BeagleBone as C&C equipped with wifi & maybe Bluetooth and running the main Python code, one Arduino Mega for motor control and environment sensing (collision & proximity detection, compass, maybe GPS, temp, ambient light level) and an Arduino Uno for controlling another sensor package, probably on a rotating head (sonar, IR rangefinder, thermal mapping).

I've also been playing with the main code (in Python) but no progress that can really be shared yet.

So it's all going rather slowly...

cswanson
Posts: 7
Joined: Wed Sep 25, 2013 5:40 pm

Re: Using RPi as a Standard Robotics Interface

Tue Oct 15, 2013 10:58 pm

How can the MyroAPI be modified to interface to the Pi instead of the IPRE Fluke?

Stateside wrote:I have reached out to the Myro Users group.The rely is as follows:

Interesting question:

Is it possible to replace the IPRE Fluke with the Raspberry pi?
http://www.raspberrypi.org/faqs
If so it would help reduce the hardware costs to schools and perhaps more
schools could implement computer training using robots?

It should be possible (as you can use the Scribbler directly connected to
a serial cable, although that use might not be completely tested).

But don't forget about Fluke's camera, one of the best parts.

Although, it may be cheaper to bundle a Raspberry Pi with a webcam, and
perhaps even some kind of wireless connection. Interesting idea.

If someone makes a do-it-yourself Fluke-like thing, please let us know!

Return to “Automation, sensing and robotics”