Wednesday, April 25, 2007

Team Zoolander, The Saga Comes to an End

First let me tell you about the team name. The robot is programmed to follow a line and during the testing phases (actually, the day before the competition) the robot was successfully making the right turns, but for some reason when it got to a left turn it would go straight or actually go right. Hence, since the great male model Zoolander struggled with the left turn, at the last moment we unofficially dubbed ourselves Team Zoolander.

Here we are. From left to right, I'm first, my name is Tom Kruger, then Jack Stites and James Murray, and of course Jack is holding Zoolander. The event was held in the Union Ballroom at the Olpin Union Center nestled right smack in the middle of the University of Utah campus. The competition was part of Mechanical Engineering Design Day which is were the department gets to trot out all the cool projects they have been paying for. All the senior design projects are situated in a gallery adjecent to the ballroom (to the right in the photo above). This year there was everything from a mountain bike for people who have suffered spinal injuries to a powered long board to on and off-road race cars. No kidding. Anyway, enough about them, this epic is about the rise of Zoolander.

If you would direct your attention to the next photo you will see the track itself. Since the design of the robot is based on what the track was expected to be, I will explain the track first. It is designed such that four robots compete at the same time. Each robots must navigate the course on their own by either following the black line that can be seen the bottom left or by following the walls. The object is to get from the starting block to the two step pyramid in the middle, climb the pyramid, and stay on top before two minutes have elapsed. The trick is that in each robot must overcome four obstacles: turf, gravel, sand and water. Also, if the first to get to the top of the steps gets knocked off, then king of the hill rules apply and whoever is on top at the end of the round wins.

In order to tackle the step we devised a multi-degree-of-freedom system that we believed would make the competition tremble in their very shorts. The fancy jargon "multi-degree-of-freedom" just means that it can do lots of stuff and look really cool doing it. Unfortunately it also means "very complicated". The credit goes to Jack for being able to put the robot together. Between his ability to build and my ability to create strong descriptive solid models (computer models) we were able to develop the idea. You should read my previous posts for more information on the design process and for an example of a solid model. Its in that post where I talk about Autonomous Solutions' Chaos which is the poster child for this type of system.

Once we got the bot built we found that it was not strong enough to stand the way we thought we wanted it to. Also, it wasn't high enough off the ground to even tackle the smallest obstacles. To compensate we placed a large idler wheel right behind the sprockets. This raised the bot enough that we thought it would be able to get things done. For the most part we were right, but we still had struggles. Also visible in this picture is the battery. We were given an RC car battery for the project but we needed something lighter. Enter the battery from the RC18T. This little guy gave us better performance than the big one we had been issued. It was nearly one third the weight and it was a nickel-metal-hydride which is widely accepted as superior to nickel-cadmium which is what we were given. Why it's better is a whole blog by itself and I would rather treat that in a blog about RC racing rather than robot racing. Maybe I'll get to that one. On the other hand, maybe not, its a snooze-fest.

In the last paragraph I kind of flew past the point that the robot was too heavy to move the way we wanted it to. We didn't find this out until after six months of design and build. I don't think I've ever felt such intense frustration. I was ready to bag the design and throw together another chassis that we could at least use for line following. Jack, on the other hand, wouldn't hear of abandoning the design. No matter what problem we hit he always thought we could get around it. And we could to. If we had no lives and vast coffers we could make it work. Just look at Chaos. I said it in a previous post, but Chaos is the effort of bankrolled professionals with graduate degrees and years of experience. Zoolander is the best effort of three ME undergrads who were willing to make sacrifices to make a design work on a sparse budget. By sacrifices I mean grades. I can think of at least three assignments that didn't get turned in because we were working on the robot.

Before I talk about the competition outcome I'd like to take this opportunity to poke fun at one of the TA's. I apologize for the quality of this picture, but it was kind of taken on the sly. You see, this guy has had these headphones surgically attached to his head. I've never seen this guy without them and he's usually got at least one ear in them. As you can see, this was in the middle of the competition and he was a judge, but that wasn't enough to get him to leave them in his backpack.

The robot picured in the last picture was the one that took second place. It was the only one to successfully climb the stair, but took second due to some operator prodding. The winning team didn't climb but got all the way to the step and made a solid attempt at a climb without operator intervention.

All in all, no robots completed the entire course autonomously. It's strange actually, because I saw a team complete the course a week before. The problem is the excess of infrared light. You see, fluorescent lights do not give IR, but sunlight reeks of it. The ambient light levels in the Union Ballroom were off the charts compared to the lab and there just wasn't time to compensate due to the disorganization of the project and competition as a whole.

Disappointment about the competition was expressed by many at the event. While my disappointment in the project as a whole is undeniable, I am pleased with our robot. Now looking back it is obvious that our design was doomed from the beginning. It's the sort of situation where you don't realize what you don't know until you try to do it. We poured everything we had into Zoolander. We gave it all we had and it still didn't work. If it had worked at 80% effort then that's all we would have given it. Instead we tackled design obstacles we never even dreamed about at the project conception. The best way to learn about hot is to get burned, and the best way to appreciate the value of a simple design is to kill yourself making a difficult one. The lessons I have learned about hard designs will serve me more than winning the competition.

Monday, April 16, 2007

The Home Stretch

Well, we have almost made it to the end of the semester. For better or worse it will all be over in less than 2 1/2 weeks. The robot competition, however is sooner than that. Next wednesday, April 25, is the Mechanical Engineering design day, which means thats when the robot has to be done. Let's see, today is the 16th, so that leaves only 9 days. It's time to work a miracle. The design has come across several stages since I last posted, so I'd like to showcase the current model. First, we replaced the track arm manipulation motors with servos which are easily seen in this photo:
At first the servo was directly attached to the largest gear that you see there, but it proved inadequate. This was a surprise to us since the servo is so strong that you can't really force it by hand. Anyway, we didn't want to lose the cost of the servos so we decided to hack them. Servos use a little device to tell where the arm is so that it can be controlled by the radio transmitter (which we are not using either). What we did is take out that little guy and replace it with the same type of component, just hopped up on steroids. It's the thing just behind the big gear with the wires coming out of the top. This thing tells us where the track arms are just like it would have worked in a standard radio system. The trick here is that the robot brain is what really needs to know where the servo arm is. That way we can give it a routine to perform beforehand and it can do it by itself.

Another change we made was the addition of another roller on the top of the arms. The idea behind this is reduction of the force required to lift the bot. It decreases the amount of leverage that is necessary. Here's a good picture of them:
Between these and the servos, we should be able to make the arm concept work.

Another challenge we are tackling is that of teaching the robot to navigate. Really all it has to do is follow a line or follow walls. The ones that follow walls do so by having little feelers on the sides. This method is much easier but having the track arms makes it impossible for us to use this method. Our line following array is built, but it's kind of hard to interpret the signals it is providing to the computer. Basically, the array consists of five little eyes that can tell the difference between black and white. They will be following a black line on a white background and the robot must be able to steer according to what eye is seeing the line. We should be able to at least have something by next Wednesday.

This has been a good project. If I had it to do over again I would certainly not make the project so elaborate. It has gotten several oohs and aahs for its craftsmanship, but that won't matter if it doesn't do what its supposed to. Wish us luck.

Monday, April 09, 2007

New Babies Galore!

Well, the last few weeks have seen two new babies in born our family. My brother's wife had a little girl two weeks ago, and my wife's sister had one towards the end of last week. It's fun to see new babies, but my wife gets a bigger kick out of it than I do. I like children, but when they just sleep, eat, and poop they aren't that interesting. When I have my own it will be different.

In the last few weeks I have discovered the best part about newborns (at least I think it the best part because it gets so much discussion) are the spitups and various other bodily expulsions. The detailed descriptions of the color, consistency, and frequency over the dinner table really lighten the mood and provide top notch entertainment. I don't think I would have enjoyed my blade roast at all if I hadn't heard about the spitup that bore a strong resemblance to thin oatmeal. Pass the potatoes.