Category Archives: Apec

The 2012 APEC event will take place at the Disney Coronado Springs Resort, Orlando, Florida on February 5th-9th 2012. On the evening of Monday 6th February there will the be 2012 APEC micromouse contest. Registrations are open now. You can … Continue reading →

Continue reading →

Continue reading →

This year I had the opportunity to enter the US micromouse contest held as part of the APEC conference. The 2010 event was held at Palm Springs in California. I had two mice, a high level of confidence and I had never been to the United States before. Why wouldn’t I go?

The best option seemed to be a direct flight from Heathrow to Los Angeles. Although this was 11 hours on the plane, the flying part was over and done in one hop. Virgin Atlantic were perfectly acceptable as a carrier and I once again enjoyed the relative luxury of a seat with nobody next to me on both journeys. In LA I hired a car. While I hadn’t been sure about the wisdom of that before going – I had never driven on the right before – it turned out to be an essential extra. Driving conditions were much better than I had expected and I was completely unprepared for the scale of the country. the hotel, which had seemed so close on the map was actually two miles from the convention centre and more like three miles from downtown Palm Springs. The drive to Palm Springs from LAX was about three hours with fairly heavy traffic for the first half of the 120 mile trip. On arrival, I was soon checked in and squared away. It was still pretty early, I needed food and the coffee supplied in the room was truly dreadful so I set out to the store across the road. This road was a relatively quiet street but, given the space available, someone had planned ahead for high volumes of traffic so it had four lanes plus the central lane. Even without jaywalking laws, it was sensible to cross only at the intersection.

Predictably, sleep patterns meant nothing and I woke at about 05:30 the next morning still pretty tired. As soon as it was light, I set out in the car to get some driving practice while it was still quiet. It didn’t take long to locate the Convention Centre, where the contest would take place, and the airport, where I was to pick up Harjit later that day. Palm Springs is a rather attractive place and sits at one end of the Coachella Valley with mountains dominating the skyline to the West. My hotel room was about as good as they got for the morning view. The header picture is what I could see from the balcony at 06:37 on the first morning. Palm Springs is comprised of quite large blocks which are either occupied and thus often very green with lush, irrigated grass around the edges – or which are as yet unoccupied in which case they are just bare desert with little but sandy grit and sparse, scrubby native plants. There seems to be little logic to this. Perhaps development leap-frogged the parcels of land that had been bought up by speculators in favour of cheaper lots further out. No doubt it will all get filled in soon enough. One thing the place is not short of is golf courses. Several huge, gated communities exits built in and around full size golf courses. The constant presence of the mountains means there is nearly always an interesting view and I guess it is their rain shadow that keeps the town mostly dry.

By the time I got to the Convention Centre, the maze had been set up and several Singaporean competitors were already hard at work adapting their mice to the vagaries of the APEC maze. the venue itself was suitably impressive and the maze generally very good. During the contest, a tent-like frame carries a video camera to that the audience can get a clear view of the mice – many of which are now just too small to see over the walls. During practice, this had a high-speed camera mounted by one of the Singaporean teams so that they could get a good look at their mouse behaviour and fine-tune the turns.

After a basic check of the mouse, I went out to collect Harjit from the airport. While he spent the afternoon working on his mouse, I chose to bunk off and have a look at a couple of local attractions – the Palm Springs Air Museum and Ruddy’s General Store.

The Air Museum boasts the largest collection of flying WWII aircraft and it was a really great opportunity to have a close look at some aircraft I felt somewhat familiar with as a result of a childhood making plastic models. Most spectacular was the opportunity to climb inside a B17 Flying Fortress.

Ruddy’d General Store is a museum in downtown Palm Springs which is a replica of a typical 1930′s general store. The contents are al genuine products from the period, many of which had been retained by a liquidator and stored in his basement for over 40 years before being made available for public viewing. The store is a truly fascinating glimpse into a period of recent history. There are a surprising number of familiar everyday brands as well as a good number of product that, quite rightly, it is probably illegal to sell now – like the bottle of 1000 Digitalis pills for the heart.

On the Monday we concentrated entirely on the contest with the entire day spent tuning and adjusting the mice in the Convention Centre. Decimus was behaving badly and could not cope with an entirely acceptable step in the maze floor in a critical square. After some scratching around for a solution, I added weight to the front of the mouse to bring the balance forward enough to overcome this problem. The Singaporean teams beavered away shaving small amounts from their already impressive run times. the high-speed camera worked really well for them allowing the handler to see where a particular turn was executed a little early or late or perhaps had a slightly inadequate turn angle.

in the wake of Kato-san’s success in Japan with a four-wheel mouse, it was interesting to see that a number of similar entries had already been constructed in Singapore. Assuming they did not really embark on this until around the time of the Japan contest in November, it is quite a significant feat to design, build and run a competitive mouse in such a short space of time.

Somehow, I never seem to know what happens during a contest. I am so focussed on my own mouse that I lose track of what the others are doing until the end of the contest and suddenly I don’t remember seeing them all run. It was not clear to me that the four-wheel designs had a clear edge over the more conventional designs. However, when I asked, there was a real feeling that we would be seeing a lot more four-wheel designs on the coming year.

My own mouse, Decimus, failed to find the center of the maze which was more than a little embarrassing. It kept adding in walls that were not there and crashing into walls that were there. This is something to do with the forward-looking sensors but I do not know what. It seems unlikely that the addition of weight to the front of the mouse would make a huge difference but the initial sensor alignment is done with the mouse tipped back on its rear skid. Really though, this would only put the emitter spot about 4mm lower on the wall and should not be a problem unless it meant that the reflectivity of the walls caused large differences in reading. I should test this to be sure but, whatever the reason, Decimus failed to perform on the day.

I do remember that the Singaporean mice were impressive. Almost faultless and very fast. The maze was a typical APEC affair with a long path consisting of a couple of long straights and a very twisty section. This kind of maze gives mice the opportunity to really display their strengths (and weaknesses). Looking at the timings, I would have to say that their mouse speeds are comparable with the results they achieved in Japan but may well still fall short of the performance demonstrated by Kato-san’s Tetra.

On the tuesday, with the contest done, there was an opportunity to play tourist so, in the morning, I took a ride up the Palm Springs Aerial Tramway with Harjit Singh. This is something would recommend to anyone visiting Palm Springs. The 15 minute ride takes you a mile up toward Mt San Jacinto to the North West of Palm Springs. That is a mile in altitude. The angle of ascent makes it feel like going up a flight of stairs. When you get to the top you find you have gone from warm sunny desert to cold, snowy mountain. If you don’t want to go skiing or hiking, there isn’t a fat lot to do at the top except enjoy the view but that alone is probably worth the cost.

Harjit had to be back on a plane home at lunchtime and after that I went for a walk around the shops of downtown Palm Springs to get a couple of gifts to take home and scout out a place for dinner. there are plenty of typically tourist places and not much really out of the ordinary – except perhaps the fact that I met several English people. Maybe they just stood out because of the accent. I imagine they were there because of the APEC conference so it should not be too surprising.

Wednesday gave me a whole day so I took myself off to the Joshua Tree National Park. This is a huge (by English standards) desert area to the north East of Palm Springs. Here the Mojave Desert meets the Colorado Desert. there is not much there now except scrubby plants, wind-carved rocks and the famous Joshua trees. That said, it took four hours to drive through – partly because I had to keep stopping the car for more than a few wow-lookit-that moments. By far the most impressive of these was the view from Keys Point.

Looking out across the Coachella Valley you can, haze permitting, from the Salton Sea to the head of the valley North of Palm Springs. In the distance is Mount San Jacinto another half mile higher while the area around the Salton Sea is below sea level. Stretching out in a ribbon along the valley floor is the ridge that marks the line of the infamous San Andreas Fault. Locals tell me that it will be a rare week when they feel no earthquake activity and the survey maps of the area are covered in epicentre marks. I did not detect any activity during my stay but there were a couple of sizeable ‘quakes in the mountains around the area during the previous month.

The desert was fascinating but it marked my last full day so it was back to the hotel for dinner and to pack my bags for the morning. The plane was not due to leave until about 6pm so, rather than take the direct route back to Los Angeles, I decided to head directly for the Pacific coast then drive up the Pacific Coast Highway into the city.

The first part of the journey went to plan and the long, snaking road out of the valley along Route 74 gave another series of tremendous views before I found myself touring through the mountainous region to the coast. Not far out of the way was the factory of Sherline Machine Tools. As I have an interest in model engineering, it seemed too good an opportunity to miss so I called in for a look. The folk there were incredibly friendly and very helpful. They have a lot of very impressive exhibits and there is an opportunity to see and try their full range of machine tools. By then, I was concerned about the possibility of being late so I set out again with just a brief stop at Dana Point. I stopped here because the hotel receptionist suggested that, if I were to stop in just one place, this should be it.

I was suitably impressed but lacked time to do the place justice. A temporary road sign indicated that the PCH was closed further up so, rather than take a chance, I just hit the freeway and motored on back to LA to drop off the car and catch my plane home. Shame to miss the coast road, Laguna Beach and all that but perhaps another opportunity will present itself one day.

All in all I had a great week and I have to thank Dave Otten and Harjit Singh for providing excellent company and local knowledge during my first visit to the USA. I very much look forward to a return visit one day.

For anyone interested, there are pictures of the area and the micromouse entrants available on my PicasaWeb galleries here:

Joshua Tree National Park California February 2010
Palm Springs 2010 Album
Palm Springs Air Museum
Ruddy’ s General Store – Palm Springs 2010
APEC 2010 Micromouse Contest

Continue reading →

The twenty fourth annual APEC MicroMouse Contest was held at the Palm Springs Convention Center in Palm Springs, California on February 22, 2010. A record total of 24 mice were registered for the contest, including 22 foreign entries. BR10SW from Nanyang Polytechnic in Singapore came in first with the best score. Excel-7 from the Institute of Technical Education in Singapore came in second. MITEE 11 from the Massachusetts Institute of Technology came in third. Excel-7 from the Institute of Technical Education in Singapore had the Fastest Run. Ghost Gasper from the Nanyang Polytechnic in Singapore received the Best Student award. All the contestants are listed in the table below together with their best score.

Cash prizes were awarded again this year. BR10SW received US$500 for first place. Excel-7 received US$250 for second place. Sakura Nezumi San received US$125 for third place since MITEE 11, the third place winner, was entered by the contest organizer. Ghost Gasper received US$500 for the best student entry. Excel-7 received US$150 for the fastest run.
The contest was held on Monday night after the exposition so that everyone at the conference could attend. To handle the audience of approximately 200 people, an aerial view of the maze was projected on a large screen behind the judge’s table. The contest was run on a maze imported from Korea.

New this year was an electronic scoring system that timed the mice with infrared sensors at the edge of the start and finish squares. The scoring system superimposed the timing information on the overhead view of the maze, so that everyone could see it in real time. Unfortunately a programming error displayed the wrong score for each run. The run time and maze time were correct however, so it was possible to calculate the correct score for each run from the archived data on the computer.
Gerardo Molina prepared the maze design once again. APEC has developed a reputation for very difficult maze designs. This year’s design had two paths to the center, 108 squares long shown in blue, and 110 squares long shown in red. Most mice used the shorter blue path.

Style X, Ghostly, and Ghost Gasper were developed by students from Nanyang Polytechnic in Singapore. The robots use an Analog Devices gyro sensor for measuring of turning angles. Each robot has an on-board digital signal processor to solve the maze and control the motion. All these robots participated in the Singapore Robotic Games 2010 this past January with Ghost Gasper emerging as the First Runner-up.

X580, BR831X, BR10, BR10SW, Black Mouse, Final and Rush are designed and built by staff from Nanyang Polytechnic. Black Mouse has 8 infrared sensors and a gyro sensor for navigation and it emerged as the Champion in the Singapore Robotic Games 2009. BR10 and BR10SW are new separate designs with 4 wheels. Final is a light weight mouse, weighing approximately 90g. Rush runs on two Faulhaber DC motors and it came in first in the All-Japan Micromouse Contest 2009 – Expert Class category last November and first in the Singapore Robotic Games 2010 this past January.

Knight and STM are designed by students at the Institute of Technical Education in Singapore. Excel-7 is designed by staff at the same school. The name STM actually stands for Short Term Memory. Sometimes the micromouse will reset by itself when the students are practicing in their own maze, so they named it short term memory (forget everything). It is just a coincidence that the name is the same as the ST microelectronics MCU. Excel-7 does actually use the STM32F103 MCU.

Decimus and Decimus 2 are designed by Peter Harrison of the UK. Decimus is a conventional wheelchair mouse using Faulhaber 2224 motors controlled by a dsPIC30F6015 processor running at 32MHz. The processor was chosen because it is cheap, provides high performance, has a good set of peripherals, and uses readily available, free tools. This is his first mouse driven by DC motors. It can manage accelerations of up to 4m/s2 and corners at variable speeds depending on the turn radius. By his estimate, it is half as quick as the leading S.E. Asian mice and could rank inside the top 15 if he could just stop it from crashing. Decimus 2 is an evolution of Decimus using a slightly different sensor configuration and Faulhaber 1717 motors, making it significantly lighter. Both mice make use of an on-board LCD screen for interaction with the user.

MMIT is built by a team of students from Vellore Institute of Technology, a University in South India.

MITEE Mouse 11 is from the team of David Otten and Tony Caloggero, staff members at the Massachusetts Institute of Technology. It came in 7th at the All-Japan Micromouse Contest last year.

DEXTER, SNIPER, and XETA 4 are built by a team of students from Vellore Institute of Technology. They use an ATmega microcontroller and are programmed with an advanced flood fill algorithm. They use DC servo motors.

Zeetah V is designed and built by Harjit Singh with some help from Pierre Hollis. It uses a STM32F103 CPU with 512KB flash and 64KB of RAM. It has six triangulation based sensors, the same technology as the MITEEx based mice as well as a ADXLRS610 gyro and KXSD9 tri-axis accelerometer. Also included are 4MB of on-board serial flash, a beeper, and a four character display for debugging and data logging. Power comes from two LiPo 140 mAh cells. The motors are MicoroMo 1524T006SR with IE-512 encoders. The mouse weighs 100 g and measures 119 mm x 75 mm.

FED is built by a team of students from Mahatma Gandhi Institute of Technology, an affiliate of Jawaharlal Nehru Technical University, in Hyderabad, India.

Indian Robochargers is built by a team of students from Vellore Institute of Technology, a University in South India.

Sakura Nezumi San is built by Shunsuke Sakura, a student at the Tokyo University of Science. The name includes his name “Sakura”, “Nezumi” which means mouse in Japanese, and “San” which means third. It uses DC motors with encoders and measures 75 mm x 100 mm x 25 mm. It is powered by a 2 cell Li-Po battery rated at 240 mAh. There are no filters in the 4 optical sensor circuits; all the filtering is done in software. In addition to the optical sensors, the mouse also monitors one gyro and the battery voltage. This is his first DC mouse, made partially with information collected from the web.

This report is abridged from the original written by the contest organiser, David Otten . The complete, original report with the individual run times is attached below…

Continue reading →