Robotics: Part 2

It's quite weird to talk about failure, because it is human nature to avoid failure at all costs, and failures are regarded as something negative. However, failure is the key element that makes success rewarding. Parts 2 and 3 of the robotics blog entries will be about robotics projects that are breakthroughs in their own rights, in terms of innovations, ideas and concepts, but are somehow unsuccessful in arena of competition. They may have failed to earn awards, but they deserve a full mention, because such projects paved the way for the experience and knowledge we have today, and the successes that followed after.

National Robotics Competition 2007 Open Category (National Level)
Theme: Science Fiction Story
Rules: 50% of the project MUST be Lego, Lego controllers only.

This is the follow-through project from the state level project featured in the previous blog entry. It is listed here not because it is lousily built, or badly designed, but because it simply did not win, due to various circumstances experienced before and during the competition.

Anyway, this is the centerpiece of the project, a new and improved transformer:


After the state level competition, we decided to stick to the 'transformers' theme, but because the stakes are higher, we decided on a complete rebuild, rather than a simple improvement on our existing project. Furthermore, since the regular category teams were no longer participating (none of them made it to the national levels), we had nearly all NXT and RCX controllers, along with about 20-30 motors and about 40 sensors at our disposal. Our hands were basically itching to try out something bigger and better, without the flaws of its predecessor.

One of the flaws of the original transformer is the ability to transform. From the previous article, you can see that the folding mechanism is rather impractical, as it is quite difficult for the motors on the base to push the heavy robot body up. We did not want that, and threw out the idea of extension mechanisms, opting for a fixed body height and fixed base. That simplified matters. Our strong robot base this time is very strong and reliable, with 8 caterpillar tracks driven by 4 separate motors through a set of reduction gears.

The body and arms, however, were built by a lot of trial and error. This was the original configuration of the body:


As you can see, 3 seperate NXT controllers are mounted on top of the body, and the arms are basically built to fold out, with an NXT on it. There is a vertical rack (bar with gear teeth) on the center of the body, as we initally wanted a small head module to climb up and mount on top of the robot body. This design had many flaws. First off, 3 NXT modules and 8 motors meant that the top of the body is VERY heavy, and the centre of gravity of the robot is raised significantly, causing instability. Secondly, the middle section is much too fragile to withstand the weight imposed by the top body. It was difficult to reinforce the middle body as any brackets would directly interfere with the climbing module.

Arms:


Each arm had 3 motors and an NXT controller. The motor (marked 'A') is designed to fold the arms out, for the building (dormant robot) to transform into a city protector (activated robot). Folding out wasn't too much of a problem for motor 'A', but once the arms moved, all hell broke loose. The plastic shaft connecting the motor 'A' to the NXT and arms simply broke when the arm moved too fast, or changed direction too quickly.

2 weeks before the competition, Kenneth, the robotics trainer in charge of Sabah paid us a visit, and told us something that we didn't know beforehand. First off, open category is not just about the robot, how the mechanisms are or whether there are any innovations in design etc. It's also about aesthetics. It is a whole project, with a set up, a scene, a poster, basically something that attracts attention. This was something we didn't really realise. Another thing was that we were far behind schedule. Other schools, with support from teachers, members of the public and paid engineers were already completing their projects, nearly ready for the competition. We were quite far from that stage. Upon hearing this, we went back to the drawing board. The complicated body, with climbing module, heavy body and arms were dismantled. In its place was a more refined, but less radical design. No climbing module, just arms that fold out.


The new arms were 2 motors each, and each motor's is connected to the moving part via reduction gear. This produced a slower but more steady movement which, according to Kenneth looks much better compared with an acute, quick movement. It also puts less strain on the attachment between the arms and the body.

With the climbing module idea abandoned, the structure of the body was beefed up heavily. After building for so many months, the 6 of us had lots of experience on how to strengthen structures, and we made the body rigid and strong. We used a lot of “L” bars and so-called “lambs” (these are the names we used for certain commonly used parts), and we nearly finished the whole stockpile of these parts. In the next few projects we were involved in, the usage of such parts to form strong structures became almost a recognizable feature in all our robots. Anyway, this is what the whole body and base looks like:

Port View:

Starboard View:

Kenneth said something about aesthetics. That meant scenery. For our robot that is rigid and doesn’t extend up, that meant the robot was a building, and the scenery was urban scenery. We had only a week left, and because the school never embarked on such a large school project, they did not give us any money to buy materials, and we didn’t know how much they can reimburse us. Thus, we had to improvise. For some reason we had a rather large piece of thin plywood at my backyard, so that was sprayed using a spray can (self bought), and used as a base. The buildings were plain grey cardboard, obtained from my father’s friend who runs a printing press. At this point we were a little worried about the ‘50% lego parts’ rule, so we made building frames out of Lego, and covered the frames with cardboard. Poster paints provided street markings and plasticine provided roundabouts.

Today this looks like rubbish compared to the latest open category works of the school. But remember, all this was done with minimal cost. The latest robotics projects easily cost RM100+ in terms of materials. Our materials cost less than RM10.

Now comes the sad part. Why we lost. After 3 years, I can more or less explain the whole mess objectively, but the blame does not lie in a single person, or any group of people. In fact, there should not be any blame, as all of this provided a valuable lesson or organizing and planning in the future. However, this story obeys Murphy’s Law, possibly to the last word, making it an interesting story.

The usual procedure for competition teams representing Sabah was that the air tickets and itinerary would be booked and organized by the Sabah State Ministry of Education. Possibly due to an overly lax Government officer or miscommunication, our tickets weren’t booked, and we didn’t know that till the last possible moment. Fortunately they did tell us 45 minutes before AirAsia shut its booking window, and we were able to book 2 separate connecting flights to the competition venue, Kuala Terengganu, via a stopover in Kuala Lumpur, albeit arriving just in time for the competition, while everyone else arrived a day beforehand to relax and prepare. This meant 2 things. We won’t have as much time to prepare our booth compared to the other teams, and the night before the competition will be spent resting on the cold marble floor of Kuala Lumpur International Airport Low Cost Carrier Terminal. You might call it ‘resting’ but the end result was that none of us got as much as a wink of sleep. We arrived at the competition venue like a bunch of swollen-eyed zombies, and by that time, almost everyone else was set and ready.

Our project utilized Bluetooth connections between NXT controllers so that the whole demonstration could be synchronized. The connections kept failing for some reason (possibly due to dozens of other Bluetooth devices trying to connect into our system). Time was running out, and basically our tired minds failed to come out with a solution. This was the scene as we were frantically making our project presentable:

The presentation broke apart, as we had to control the robot manually by switching them on and off at intervals. That failed to impress the judges, and so it was hello, failure.

“Everything that can go wrong will go wrong.” – Edward A. Murphy, Jr.

Looking back, there were many things which we could have done to avert disaster: we could have assumed that the state ministry is awfully inefficient, and pestered them to book our tickets days ahead; we could have had many solutions to solve the Bluetooth connection problem; we could have lobbied for more funds and support from the school for the aesthetics of the project. But bygones are bygones.

For the next few projects, we arranged for project timelines to have more time allowances for contingencies. We arranged for artistic volunteer groups to help out in scenery making. We made sure all our travel arrangements are fixed and ready, and developed techniques that allow NXT-to-NXT communications without Bluetooth, lest the connection fails. We also made sure that funds were sufficient in future projects.

That's all for now.