The robot below built with Lego Mindstorms and using a Nokia N95 Cellphone with an ARM processor (same type of chip as in the iPhone and Google Phones) is able to solve a 4×4x4 rubiks cube puzzle.

While the application is quite trivial (as in the robot can only do one thing and it is a simple algorithm no real AI here) it is a cool example of what some basic robotics can do.

I hope to post some photos and/or video of some of the robots where I am studying in the next few days at the latest later in the month when we have a robot soccer match between ourselves and The University of Newcastle.

Hat Tip: WhaleOil

From the wonders of xkcd

One of my papers this semester is focused on Google Android Mobile Operating System.

The brief for my final assignment is:

Your task is to write any application you like. The are no restrictions on what your application can do but it should show of the capabilities of the platform and be well written.

Marks will be awarded for interesting applications that make good use of the Android platform.

Make sure your application works on the emulator but I will also test it on a real device.

For this assignment I decided to create a Blog Reader that reads the RSS XML feeds off blogs to display them in a Android Application.
I have spent around three days coding this assignment. I will not release the code until after the assignment has been marked, however here are some screen shots of the work

My Application Sitting in the Android Application Menu on my phone

The applications home screen

Adding a new blog feed into the application

Viewing a list of blogs

Removing a blog feed

Viewing a list of blog posts

Viewing list of blog posts

Viewing single post

Another week, time for another video.

This one shows changes that have been made to the intersection controllers within the Traffic Simulator.

The give-way controller makes all the cars give way to all the other roads connected into the intersection with a minimum of 2.5 seconds between vehicles.

The traffic lights controller makes all the cars on the green road go at 2 second intervals and makes the others queue with a minimum wait time of 16 seconds.

The round-a-bout controller makes each car give way to the right and then go with a minimum of 3.9 seconds between each car on a road, but multiple cars from multiple roads can cross the intersection at the same time… hence the chaos on the video.

This one turned out much better than the first one, enjoy.

Silly computer has taken almost two hours to process this single video into a usable format but at least it can show some of my work.

Okay a few posts ago I mentioned that I had been successful in setting up my laptop as a virtual supercomputer.

Here are hopefully some relatively clear instructions on how simple it was to do.

Setting up a MPI development system on Ubuntu Linux

1. Download Ubuntu – check out the latest version at www.ubuntu.com

2. Intall the required mpi files.

   1. Open a new terminal window.

   2. Enter: sudo apt-get install libopenmpi1

   3. Enter: sudo apt-get install libopenmpi-dbg

   4. Enter: sudo apt-get install libopenmpi-dev

   5. Enter: sudo apt-get install openmpi-bin

   6. Enter: sudo apt-get install openmpi-common

   7. Enter: sudo apt-get install openmpi-dev

3. Test the mpi install by compiling and running a simple program

   1. mpicc testfile.c -o testfile

   2. mpirun -np 2 ./testfile

      1. Where np is the number of cores * number of processes on your system.

      2. e.g. on a two core laptop np should be 2.

4. If you get an error regarding ssh when you enter the mpirun command install ssh

   1. sudo apt-get install ssh

5. If you get requested for your password everytime you run mpirun set up a stored RSA key control

   1. cd $HOME

   2. ssh-keygen -t dsa

   3. cp .ssh/id_dsa.pub .ssh/authorized_keys2

Further reference:

http://www.csc.cam.ac.uk/academic/practicals.html

http://www.math.umbc.edu/~gobbert/mpi.html

Okay, now there is no excuse for any geek or anyone with a keen interest in computers to not try out linux.

For the past year and a bit I have been playing around with Virtualbox and virtualization, the major problem here however is the speed of booting, and drivers etc as it is in an emulated/simulated environment.

Enter WUBI. Wubi is an installer that allows you to install Ubuntu and Linux Mint on an NTFS partiation and use the windows boot loader so it does absolute nothing to windows at all. And to make matters better you can access your windows files straight out of the box.

I have now had a WUBI Ubunutu set up for the past week and it is brilliant.

I am now trying a new distro called Linux Mint which is aimed at end users.

If you want some help getting it installed or just want to try it out just contact me, it is brilliant just really awesome.

Okay, the traffic simulator project has been slow going over the past few weeks, just not been very focussed and playing with DLA and other assignments too much.

This morning/arvo I changed some of the code and layout of how cars were being generated. Now instead of a car know what road it is on it has been reversed, so each road has its own collection of cars to manage. With some changes to how the graphics are produced this now means that cars are now rendered at the same time as each road. A that allows perfect placement of cars onto roads rather than having to do a mathematically approximation. Clear as mud?

Yes I am still making DLAs just trying to make the maths that little bit nicer, and get that all impressive performance boost.

This is a 6,000 point balanced model, it took just under 1.5 hours to generate, compared to 4+ hours for a 5,000 point unbalanced DLA on Sunday

The maths to create this footage kept me up to the wee hours of the morning. But I am pleased with the outcome.

Lions and tigers and bears. Roads and lanes and cars. Oh my.

Still a long long way to go, but here is the first screenshot of the new 3D traffic simulator for my honours project.