Techfest 2016 – UK Micromouse Contest

By | June 27, 2016

The UK micromouse contest took place on June 25th in Birmingham. The venue was slightly different with the event being located in Birmingham City University’s new Curzon building. Read on for the results of the contest.With only nine entries this year and none from China or Singapore, this was a slightly reduced field compared to previous contests. However, there was still a distinct International feel with an entry from the Netherlands and a visitor from Japan who, unfortunately, was not able to bring his mouse.

For my part, I ran Decimus 4E largely unaltered since the APEC contest earlier this year. While there remains plenty to do with Decimus 4E, I feel it is time to move on and build a new, fan-assisted mouse. I hope to have it ready fro Taiwan in September. Failing that, it has to be ready for Japan in November. It is increasingly clear that there will be no way into the top five places without some kind of vacuum fan.

Because there are no real developments on my mouse, there is not much of a story to tell this year. I simply ran what I had and, aside from a couple of crashes because I was overconfident about sensor calibration, Decimus 4E ran pretty well.

Heats

The heats maze was reasonably short with a path length of about 750mm over a choice of two main routes:

UK2016-maze-heat

The qualifying scores were as shown below

Decimus 4E Peter Harrison 12.03
Mouse X2 Derek Hall 21.97
Mouse X Derek Hall 22.18
Isambard II Martin Barratt 24.94
PicOne Turbo Jim Chidley 26.31
PicOne Jim Chidley 35.14
Fab 1 Derek Hall 38.38
FREDA Ian Butterworth 52.37

We also record the fastest run times:

Decimus 4E Peter Harrison 4.58
Mouse X Derek Hall 9.45
Mouse X2 Derek Hall 10.25
PicOne Turbo Jim Chidley 15.41
Isambard II Martin Barratt 19.39
PicOne Jim Chidley 24.81
Fab 1 Derek Hall 31.49
FREDA Ian Butterworth 39.54

Often, in the UK contest, the score-based results do not show the fastest mice but favour the mouse that can get the total time of the first and second run as low as possible. Fast searching is rewarded.

Finals

The finals were held in the afternoon and, since no other maze had been designed, we chose to slightly modify the All Japan 2012 qualifier maze. It was the easy way out but would also allow comparison of run times with the results of the Japan contest to see how the UK performance stacked up.

The maze was a bit longer with three basic paths of lengths of around 11000 mm. Choice of path will be very dependent upon the algorithm used for this maze. I know that Decimus 4E used the cyan path below but I cannot recall where other mice went.

UK2016-maze-final

The final results were:

Decimus 4E Peter Harrison 16.83
Mouse X Derek Hall 24.10
Fab 1 Derek Hall 26.92
PicOne Turbo Jim Chidley 31.78
Isambard II Martin Barratt 39.24
“Fat Cat 3C” Tim Foden 41.24
Mouse X2 Derek Hall 50.84
PicOne Jim Chidley 58.18
Kaasjager Coen Roos 184.64

and the fastest run times were:

Decimus 4E Peter Harrison 6.99
Fab 1 Derek Hall 12.00
Mouse X Derek Hall 13.55
PicOne Turbo Jim Chidley 22.39
“Fat Cat 3C” Tim Foden 23.04
Isambard II Martin Barratt 34.57
Mouse X2 Derek Hall 42.48
PicOne Jim Chidley 58.18
Kaasjager Coen Roos 158.39

ALL Japan 2012

For comparison, the All Japan 2012 qualifier maze differed only in the last couple of cells around the goal and produced results of (Names are translated by Google – apologies to any Japanese readers if they are badly wrong):

Robot name Name group name Time
Excel-8b Khiew Tzong Yong Institute of Technical Education 6.342
Wa-Tiu Shih-Wei Chao Lunghwa University of Science and Technology 6.386
Snow wind 5 Nakajima FumiTakashi 6.499
Tushi2 CHANG, CHIN-CHIA Southern Taiwan University of Technology 6.533
Hulk-2 Hung Chung Yuan Institute of Technical Education, Singapore 6.544
Yukikaze 4.5 Nakajima FumiTakashi 6.570
Excel-8a Khiew Tzong Yong Institute of Technical Education 6.624
Hulk-1 Hung Chung Yuan Institute of Technical Education, Singapore 6.646
TPK Chuan-Sian Fu / Jheng-Yan Syu Lunghwa University of Science and Technology 6.733
Tsubasa Sheng-Hung, Lin / Sin-Mao, Fu Lunghwa University of Science and Technology 6.753
Prying mouse 7CL Koichi Kojima 7.529
[YU-05] BRAVE Yutaro Uchida UEC Robomeka workshop OB golden generation 9.606
Turtle 32F YU, TSUNG-YU / WU, YU-CHENG Lunghwa University of Science and Technology 9.965
Faded glory midnight sun Baba Yusuke Mechatronics Research Club 10.727
MITEE 13 David Otten Massachusetts Institute of Technology 10.777

Note how the top 11 mice were all within half a second of each other. The route in Japan was about 200mm longer than that in the UK contest. At an average speed of 1.7m/s that is a difference of about 0.12 seconds. I crashed repeatedly and only managed three search runs.

Wall Followers

There is a senior wall follower event using almost the same maze but with modifications to ensure they can find the middle. Their final times were:

E-rat-ic David Hannaford 39.12
Lefty Ken Hewitt 42.01
FREDA Ian Butterworth 52.89
Kitty Martin Barratt 69.64

I am afraid I do not know how long the course was. Worryingly, some of the wall followers seem to be able to manage better average speeds than my mouse can when searching. It is not clear how except that they do not waste much time turning around in dead ends. I need to fix that.

And finally …

Many thanks to Chris Evans and Tony Wilcox and their helpers at Birmingham City University and to the University for making all this possible.




Leave a Reply