A plea for roboticists

[Donut]

Originally Posted by Alan Anderson

How many points was hanging on the bar worth in 2004? At least one robot could do it autonomously.

Okay... that was worth 50 (I remember 190 could do it). Still, there weren't multiple alliances that did it, like at the Championships last year.

[LauraN]

I'm doing some "create your own mechanical engineering major in robotics" at MIT, and I've gotta say, I think it's pretty lame that getting a mechanical engineering degree at MIT barely requires that you know how to turn on a computer. Of all the undergrads here who must be seriously interested in robotics, I know of exactly one other who is getting a degree in MechE- they are all strictly EECS people.

I didn't know the first thing about computer science until I took an introductory class in it this semester, and I'm STILL intimidated by it. I remember wiring the control panel for my team's robot in high school and passing the project off to our programmer guy. I was utterly amazed when the thing actually worked. The programming was quite literally a black box for me- I had NO IDEA how it worked.

Part of that is because computer science is so abstract- I had no idea how to do it, I couldn't literally see it, and from what I heard it was supposed to be really hard- like only really smart, hardcore hackers could do it or something. Anyone who's been programming for years is probably laughing at me right now, but it's the honest truth. I think a lot of people feel that way- so teams with limited programming skills are unlikely to obtain more and even those that have skilled programmers are unlikely to give them a whole lot of support if the other members don't understand what kinds of pressure they're under.

I completely agree with you- programming is much more important in robotics than is reflected in FIRST, and that should change.

But quite honestly, I think there's a problem with computing in general- there's a pretty big speed bump between the utterly clueless and the novice programmer, and that's more than half the battle. This is a big concern of mine, so excuse me if this gets a bit preachy, but I think it's our responsibility as mentors, engineers, and technological innovators to make computing, programming, and computer science more accessible to the average person before we start expecting rookie teams of high school students to build entire robots from scratch and program them with autonomous state estimation and mapping algorithms.

[DonRotolo]

sigh.

For two years out 'autonomous mode' has consisted of 'tuen on this motor for X seconds, then that motor for X seconds, then do this.' In other words, pure dead reckoning.

We're going to try a new design approach this year, and in that approach there are 4 weeks for programming. Not sure if we can pull it off, but will try.

Don

[Gui Cavalcanti]

I've switched majors to a Robotics Concentration under the General Engineering degree at Olin. It's working fairly well, and I just wanted to let other interested college students know what the coursework looks like:

Core Engineering Requirements:
Solids and Structures
Dynamics
Mechanical Design
Software Design
Controls

Core Math Requirements:
Nonlinear Dynamics and Chaos
Partial Differential Equations

Electives:
Robotics
Artificial Intelligence

Related General-Requirement Olin Courses:
Materials Science
Vector Calculus
Linear Algebra
General Mechanics & Calculus

I've also done a few independent studies in controlling brushless servomotors, building snake robots, and building a compliant biped. Feel free to drop me a line if you have any questions about designing your own robotics major - my approach is fairly mechanically heavy, and don't think that's a requirement by any means.

[Gdeaver]

My son is taking ME and I just can't see how one could take all the required courses and also take courses that would apply to robotics. At least in a 4 year program. In 4 years he will only have 2 elective slots that could be applied and that is with 18 credits per semester. Now if you extended the undergraduate program to 5 years then there would be time for robotics subjects or like others said get your ME in 4 and a graduate degree in automation for six years of study. Don't skimp on the core base in your base curriculum.

[Erin Rapacki]

Originally Posted by Adam Y.

Part of the reason also has to do with the fact that they dont hire people directly out of college. I'll have about 1.5 years of work experience but I don't think that would be enough to get the job right out of college.

Its unfortunate you think that way.

I was an intern at iRobot last year in Home Robots; while there, I made some friends on the G&I side and my senior project is in G&I... and get this, I'm not a Mechanical/Electrical/Software engineering student either! I'm Industrial Engineering (who wants a Mechanical Masters, but I'm not qualified to do hard-core design analysis work yet). My senior project is all mechanical, I learned how to machine/prototype/hack-design in FIRST... so that's enough for this round of project prototyping. My mechanical project partner is doing the hard-core analysis, but I'm coming up with the effective material-handling methodologies because, well, my frame of reference from FIRST and working at HSES, HSSI, DEKA, PW, Gillette, and iRobot has giving me a well-rounded view (of course, well rounded people are pointless). Actually, the reason why I'm a roboticist for a respectable company at such a young age is because I'm interested in the field, I have a hunch on where it will go in the next few years, and I'm crazy passionate about doing anything I can to push that field further into new products and development.

iRobot has an intern-to-full hire program to it where if you're right out of college and they don't know you, they'll hire you on as temporary for a few months (intern/contractor) then put you on full time if they like you.

It's worth a try; if you want the recruiter's information: PM me. He's always looking for talented/passionate people no matter what your academic/experience level is. The way they hire is: if you're smart and they like you're personality; they'll enjoy working with you and teach you what you need to know.

Act like a passionate roboticist, and you'll get paid to be one .

[Mikell Taylor]

Adam Y --

Another vote for being hired right out of college. Look for the right companies. Be in the right location, too -- California, Pittsburgh, and Boston are the big areas in the US for roboticists. If you search them out, and you're good/passionate, they're wicked excited to have you.

Also (for everyone!), don't underestimate the power of FIRST. Two of my three job offers out of college, all absolute dream offers from wonderful robotics-based organizations, were due to my FIRST involvement and the networking that happened here on the forum and in the organizational stuff I do. Leverage everyone in this community!

[Adam Y.]

Originally Posted by Erin Rapacki

Its unfortunate you think that way.

I was an intern at iRobot last year in Home Robots; while there, I made some friends on the G&I side and my senior project is in G&I... and get this, I'm not a Mechanical/Electrical/Software engineering student either! I'm Industrial Engineering (who wants a Mechanical Masters, but I'm not qualified to do hard-core design analysis work yet). My senior project is all mechanical, I learned how to machine/prototype/hack-design in FIRST... so that's enough for this round of project prototyping. My mechanical project partner is doing the hard-core analysis, but I'm coming up with the effective material-handling methodologies because, well, my frame of reference from FIRST and working at HSES, HSSI, DEKA, PW, Gillette, and iRobot has giving me a well-rounded view (of course, well rounded people are pointless). Actually, the reason why I'm a roboticist for a respectable company at such a young age is because I'm interested in the field, I have a hunch on where it will go in the next few years, and I'm crazy passionate about doing anything I can to push that field further into new products and development. I have no idea what other odd fields I can come up with but there is definately more.

iRobot has an intern-to-full hire program to it where if you're right out of college and they don't know you, they'll hire you on as temporary for a few months (intern/contractor) then put you on full time if they like you.

It's worth a try; if you want the recruiter's information: PM me. He's always looking for talented/passionate people no matter what your academic/experience level is. The way they hire is: if you're smart and they like you're personality; they'll enjoy working with you and teach you what you need to know.

Act like a passionate roboticist, and you'll get paid to be one .

Thanks. I did not know that. Every single EE job description I looked at for IRobot said at least two years of experience except for the internship on their website. In all actuality I will only be missing less than 4 months by the time I graduate.

(of course, well rounded people are pointless).

I would actually have to say robotics is a field where being well rounded is a must. I've bought a psychology book because of the fact that three robotics other books made reference to that very book. Another one of my books has a chapter entirely devoted to theories in animal behaviorism.

[Erin Rapacki]

Originally Posted by Adam Y.

I would actually have to say robotics is a field where being well rounded is a must. I've bought a psychology book because of the fact that three robotics other books made reference to that very book. Another one of my books has a chapter entirely devoted to theories in animal behaviorism.

Hehe, I only threw that in there because I heard it on Prarie Home Companion and couldn't stop laughing!

Yes, they have job descriptions online, but they also have job descriptions they'll create if they have the right person sitting in front of them saying they want to work there. Send an email with a resume, you might get lucky. Its worth a phone call with a recruiter - so start selling yourself!

[Bongle]

I know this thread is a solid year old, but since it has been bumped and people are reading it...

That being said, I've really wondered what it is like to be one of the programmers, I'm just glad I'm not one of them, as our team is very evil to them and gives them about 3- days to program the thing.

This post isn't directed at you specifically, it is hopefully directed at team leaders who have the power to allocate people's time: PLEASE don't do this.

At the very least, try to plan so that the programming team has access to the RC, OI, sensors, batteries, and other electronic kit right up until they are installed on the robot. At the very least that enables the programmers to make sure that their outputs make SOME sense given certain inputs, but there is really no substitute for on-robot testing. Most of the components the programmers need to test with are mutually exclusive from the mechanical components of the robot, so they can have their own special box.

If the programming team is already using the robot testing something and you have something to do, ask them what they're doing. If they are testing or implementing something vital (arm control, autonomous modes, drivetrain) and you're there to tighten bolts or something, ask yourself "do I want to sacrifice an autonomous mode for $TASK?". If the answer is no, then let them do their thing. For some reason or other, programmers always seem to at the bottom of the robot-access priority list. I've been delayed from autonomous coding and camera testing because bumpers needed their bolts tightened.

Without fail, every year since stack attack (2003) I've been asked/told both as a student or a mentor "Why wasn't that tested?" "Why don't WE have a killer autonomous mode" "What was that bug?" "It was the programmer's fault" "Why was the old version still on the robot?". Programmers cannot program correct programs without testing. Testing doesn't mean turning the robot on 5 minutes before a match (at that point, you're already doomed, especially with the slowness of IFI Loader). Whether at a competition or during the build period, programmers need time with the robot to make sure everything works.

There aren't many ways the robot software can fail that DON'T doom you in a match. Your robot might be uncontrollable, it might not move, it might ram into a wall real hard. Motors might turn in ways that software safeties were supposed to stop them from turn. Autonomous mode might not work. Your scoring thingy (arm/shooter/capper) might not work or break itself.

So, reader, if you're in a situation where you have to decide whether to build for one more day or give the programmers the robot for a day, please give that day to the programmers.

If that is impossible: At the very least, make a set schedule and say that "builders build from 8:00 - 11:00, then programmers get it all lunch with no-one bugging them". You won't believe how much more productive everyone is when there aren't builders coming in every 5 minutes tightening, measuring, mounting, and moving things. Every time someone comes in, we've potentially got to shut the robot down, move computers, unplug cables, and then put everything back. That is wasted time for the builder, and wasted time for the programmers. If people are coming in all the time, then that scarce minute/hour/day you've given the programmers to get things done is being wasted.

[indieFan]

A couple of comments based on the previous posts:

1. I was told not to worry about how many years experience a company says they want for a job applicant. They will fill it with anyone who is "qualified" to suit their needs. If you look through classifieds, it will often say a BS in EE or X years equivalent experience.

2. Use your FIRST connections. Chances are they can get you information about recruitment opportunities at companies you're interested in, if not try to get you hired themselves.

3. Don't think that your first job will be your last job. I was explicity told by a professor that company loyalty is unheard of these days. Most people will move on after 5 years or so.

4. The broader your background, the better off you will be when dealing with robots, especially industrial robots which I'm beginning to believe are all but completely ignored around these forums. I work with several types on a daily basis. One I have to be the mechanical/electrical eyes and ears to relay information about crashes to the programmer. This robot is a "one-off" machine, so there is no precedent for any of the issues that come up. The other robot I have to not only be the mechanical/electrical eyes and ears, but I also have to be the programmer since they are 15-20 years old and nobody else is around for coding anymore.

5. Re: Programming the Bots
If you're not a rookie team, then consider using an older robot or removing required components and using those to work on your code. Bongle alluded to people having to come and tighten screws when the programmers were trying to get stuff done. My question in response to this is: Has it occurred to you that this might be an issue of safety? What happens if the bolts come out due to vibration? What seems insignificant to programmers may actually be very significant to the mechanicals and vice-versa.

indieFan

[Cuog]

I'm glad this thread has been brought back up as many of the points Gui has made a still very valid. In addition to many of the things he has said, I definitely would encourage everyone in FIRST to walk out of your comfort zone and learn another aspect of the robot, just to know it. I have done this and been very happy about it. When I joined I started as mechanical because thats the way my team usually works, and I had a large amount of experience doing the mechanical side of things. Then last year i decided to make a switch over the electronics and programming to see how that part of the robot is built. Once you know a good amount about how everything works, you can design soo much better.

[artdutra04]

Sorry to hijack this thread...

Originally Posted by Bongle

I know this thread is a solid year old, but since it has been bumped and people are reading it...

This post isn't directed at you specifically, it is hopefully directed at team leaders who have the power to allocate people's time: PLEASE don't do this.

At the very least, try to plan so that the programming team has access to the RC, OI, sensors, batteries, and other electronic kit right up until they are installed on the robot. At the very least that enables the programmers to make sure that their outputs make SOME sense given certain inputs, but there is really no substitute for on-robot testing. Most of the components the programmers need to test with are mutually exclusive from the mechanical components of the robot, so they can have their own special box.

This can be solved extremely easily: mount the FRC controller on a Vex Robot chassis and start programming code from day one. Sure, you may not get your exact coefficients needed for certain functions and the like worked out, but you can definitely get the algorithms perfected for doing such things as a CMUcam searching for the light and driving towards it, or using gyro sensors to make sure your robot can automatically correct itself if it's knocked off course.


Getting back onto the topic at hand, over the past year I've become much more involved in the three areas (mechanical, electrical, programming) of robotics than in past years. And I can say that I love being able to do all three fields. Whether I'm programming an absolute-positioning, gyro-based holonomic code for my Vex robot in EasyC or innovating such things as variable-diameter wheels (both of which are ongoing projects of mine), I really love robotics.

Earlier today I received my acceptance letter from WPI; I'm glad that they now offer a degree in robotics, as well as the fact that their campus location is less than an hour away from the Boston hot-spot of robotics/high-tech companies. While a degree in robotics may seem too specialized at the moment, I foresee a very bright future for robotics in the 21st century, and I would love to be a part of it.

[Donut]

Originally Posted by Bongle

At the very least, try to plan so that the programming team has access to the RC, OI, sensors, batteries, and other electronic kit right up until they are installed on the robot. At the very least that enables the programmers to make sure that their outputs make SOME sense given certain inputs, but there is really no substitute for on-robot testing. Most of the components the programmers need to test with are mutually exclusive from the mechanical components of the robot, so they can have their own special box.

Originally Posted by artdutra04

Sorry to hijack this thread... This can be solved extremely easily: mount the FRC controller on a Vex Robot chassis and start programming code from day one. Sure, you may not get your exact coefficients needed for certain functions and the like worked out, but you can definitely get the algorithms perfected for doing such things as a CMUcam searching for the light and driving towards it, or using gyro sensors to make sure your robot can automatically correct itself if it's knocked off course.

Building off of these, just give your programmers SOMETHING to test with. It can be a Vex robot, it can be a previous year's FRC robot, it can be the current Robot Controller attached to a block of wood, just give them something to run some code through. We worked out 80% of our camera code for our 2006 robot on our 2005 one so 2006 could be worked on while we tested.

If worst comes to worst, they can test the drivetrain with it propped up on blocks while you screw on a few bolts to the arm. Everything needs to get finished, so compromises need to be made on both sides. And yes, all parts are important and very interesting; I love thinking of mechanical designs though I mainly program for the robot. Try investigating something new for a new idea on how everything really works.