Its been another busy week at work and at home, but finally the “slow boat from china” has delivered a new robot platform for me to work with.

2 wheel drive robot kit assembled with battery box attached

I’m hoping to use this robot platform as the basis for a “build a bot” event at one of the forthcoming Paignton Pi Jam’s, the idea is to create a “robotic solution” that is reasonably priced an has functionality (such as line following / maze running / object avoidance). The micro controller of choice will be an Atmega chip commonly found in Arduino Open Hardware reference boards, although the final layout and component configuration has yet to be worked out, and we may even add a pi into the mix at a later stage (just like marco 2).

The platform itself is relatively cheap coming in at £7.21 directly from China, which is the only catch as you have to wait approximately 7 working days (possibly longer) for it to arrive, and you are at mercy of the prevailing humour of the Tax and Customs officer. The current regulations seem to allow up to £36 for items marked “gifts” (which this doesn’t fall under) or £15 for other items (which I guess this does) before paying import taxes.

So the kit arrives as a “bag of bits” that need to be assembled:

“Newbot” kit comes as a bag of bits…

The instructions are in Chinise (naturally) but luckily the diagrams are mostly servicable and the Youtube video on the website was even more informative in terms of assembly strategy.

I had mine pretty much assembled in 20 – 30 minutes, although I did have to reassemble one of the metal blocks attached to the motor because I had assembled it upside down.

I was a little disappointed that there was a hairline crack in the acrylic base plate. Whether this happened during manufacture or during transit we will never know, and the only other annoyance was that while there were holes in the base plate strategicly placed for adding line following / IR object detectors and encoder wheel sensors, none of the hole configurations lined up to a standard ardunino system board. Finally although wires are provided for the motors, you will need to attach these yourself, which realistically means a bit of soldering. As the “build a bot” session is intended to be a “plug and play” event for younger robotics engineers to enjoy I will need to factor in some time to pre-solder the wires to the motors.

“Newbot” base with caster wheel attached and showing motors with wires attached (soldered).

Overall I am happy with the purchase as this provides a very reasonable platform at £7.21 for experimenting with robotics and computing. Hopefully when I have had some time to work out a final layout we will be able to put on a “build a bot” session at Paignton Library.

Finally some more pictures:

“Newbot” assembled with white nylon standoffs attached.

“Newbot” with fomex plate placed on top of standoffs.

Please note that these pictures may not be representative of the final design and configuration.

Looks like the Raspberry Pi foundation has hit the 5 Million mile stone.

http://www.raspberrypi.org/five-million-sold/

If you in Torbay or South Devon (or even beyond) the next Torbay Pi jam is 14th March

http://www.dcglug.org.uk/torbay-pi-jam/

Lets see what people are doing with their Pi’s (or similar devices)

At the Raspberry Pi Jam on Saturday the 14th I was working on creating a cluster supercomputer with a piece of software, GNU Parallel. It was successful, and I did a test with the prime number generator Paul made. It seems to require me to copy all files involved to all computers involved. However, it appears around 40% faster when using two computers instead of just one.

It appears that GNU Parallel has trouble finding the amount of cores on each computer, which is likely the reason for Paul’s inital results, which showed not using parallel to be faster. For all computers it says ‘Could not find the amount of cpu’s on 192.168.1.x, using 1’. It is possible to specify to parallel how many cores computers have, so I will research this.

At Minecraft club on Wednesday, we attempted to link Parkfield’s netbooks together with their switch. We discovered something extremely odd; the Linux netbooks there, along with one of the laptops from Raspberry Pi jams, do not have OpenSSH Server (sshd) installed, even though it is meant to be installed by default on Linux. One of the netbooks wouldn’t let us install it either, but that didn’t really matter – it just meant we had to use parallel on that machine instead of another. We didn’t get time to run a test, but we should have time next Wednesday.

Although so far we have just used GNU Parallel with a python program, it can be extremely useful. For example if you have a password-protected archive with important files, but you forgot the password, you can decrypt the folder much quicker with more processing power.

The Torbay Pi jams have had a mention on the Torbay CDT website

http://www.torbaycdt.org.uk/index.php/events

So a big thank you to them for this.   Please re-tweet so we can spread the word on the PI jam and the work of the Torbay CDT.

Devon and Cornwall GNU / LInux user group North Devon Meet up

Saturday 21st Feb 2015 @ 14;00

Place,  The White Hart,, Holsworthy, Devon

Monthly meet up for members of the Devon and Cornwall GNU / Linux user group or anyone in the area who is interested in a social meetup.  Grab a drink and have a chat about free software, Linux, tech or what ever comes up.

The unfortunate grove socket incident with the motor shield / gove breakout means I have to wait for new cables / connectors to arrive before I can look at adding sensors. First priority will be line following an “old classic” for robots and frankly pi + arduino + motor shield is probably over kill for this, but then again I am trying to squeeze as much functionality into a single robot as “robotically” possible.

While I “wait” … I have swapped out the Pi B for my only Pi B+ board so that I now have 4 USB ports (without needing to clutter the already busy platform with more hardware like a USB hub). Now I can have WiFi dongle, coms link from Pi to Arduino and Webcam input, and still have one USB port free for what ever evil genuis plan takes root in my brain.

So now I can use the upgraded python scripted http server / robot controller (glorified serial coms down USB port) to control robot from any web browser (that supports javascript) on any device that can connect via the wifi hotspot (hosted on the robot’s Pi B+ board). Although the webcam is mounted upside down, I have managed to flip the picture using pygame.transform library (yes that’s right I’m using pygame libraries in a robot!), it only displays snapshots (usually after processing a command eg forward, reverse, left, right, stop). Next tasklet is to get this streaming a constant flow of updated images, I dare not call it video streaming as there seems to be a bottleneck in the processing somewhere, so http page refreshes are very slow (something else to work on).

I have also ordered some similar motors / wheels for a new build project which will be the upsizing of the previous frapper bot project, so for those that have seen it before a marco sized frapper bot type design using spare bits of foamex board for the platform. Unlike Marco’s motors the new “kit” comes with 8 pole magnet “hubs” and “hall effect” sensors which will allow me to control and regulate the speed of the individual motors. At the moment Marco turns to the left when going forward because the right motor is going faster than the left even when setting them at the same speed. Even if I manually trim this, as the batteries drain so the trimming comes out of alignment again, so some form of control algorithm is required. The Hall sensors will detect 8 changes in the polarity of the hub as the shaft it is attached to spins around 1 revolution, thus being able to determin both speed and distance of the individual motors. The trick is to use that information to regulate speed, and recognise a stall situation.

I would like to say that I had to work really hard to get the webcam version of the web based control system working, but I had actually already tested this on my PC prior to replacing the B with the B+. I only had to make a minor tweak to the PC version (the image flipping mentioned above) to accomodate the fact that I had to mount the camera upside down. This is really one of those times where you make up for hardware issues with “smart” software.

Enough talking !!! Now for some Marco 2 pictures:

Marco 2 next to Tesco Hudl Tablet used for remote control testing.

And another picture showing Marco 2 next to Tesco Hudl Tablet.

Tablet shot showing controls (left) status information (middle) and webcam image (right).

Top down view of Marco 2 showing motor shield with Arduino Uno board sitting underneath.

Rather blury view (I am not a photographer) of front mounted webcam as stated mounted upside down.

Top Down view of Marco 2’s Pi B+ board.

Wider angle top down view of Marco 2.

Side view just showing battery pack (used to power the Pi B +) tucked between the controller platform base and the motor platform base.

Another side view showing USB connectors (on the far side). The battery pack on top is used by the motor shield to provide power to the motors.