Wow, I can’t believe this is the first post to the forum!

Well here is the question.

What are the advantages of building your own mobile robot platform from scratch and do they outweigh the advantages of buying a complete package of a mobile robot platform.

Things to consider:
1.) Money
2.) Time
3.) Skills

Recently, I bought the Blackfin robot from Surveyor Corp. and am so excited to receive it in the mail. I can focus on the software implementation and not have to spend so much time on low-level hardware issues. Plus later I can upgrade my Blackfin to have stereo vision, wireless image transmission, and GPS.

The problem is that a friend of mine suggested that there is more satisfaction gained from building your own from scratch. His reasoning is that you can upgrade it later and you don’t have to keep spending time on low-level hardware issues. I don’t disagree with him. There is a lot of satisfaction from building your own.

I personally think it is best to have a vendor made robot platform as well as build your own. You get the best of both worlds then. So this is what I’ve decided to do.

I’m going to keep my Blackfin robot from Surveyor Corp. and then start building another robot from scratch that will fit my ideal requirements for the features that I want.

Same thing goes for a humanoid robot. Buy one from the Internet and also build one yourself. Then you don’t have to spend forever on low-level stuff. You can already get started figuring out the weakenesses of pre-made systems and then make your own platform even better.

Just my 2 cents.

Anyway have anything else to add.

Thank you,

Melanie

On one extreme, if you get something like the lego NXT robot and use their graphical programming language, are you really learning robotics? Or one of the many servo-based kits that let you make humanoid or car-like toys. If your goal is to make a machine that can accomplish a task (make a walking robot), then I’d go ahead and by a COTS (commercial off the shelf) product. If your goal is to learn something, then maybe choose a portion of the project to build from scratch.

I say this having gone down the hard path way too many times… For instance, trying to build a simple position controlled slide from “scratch” turned into a collasol project. I did everything from machining brackets to hold bearings, to writing low level control code (PID), designing a custom pcb, fighting noise, the whole 9 yards. I learned all sorts of stuff doing this:

-about making flexible couplings where the motor attaches
-about suppressing noise from motors (or at least trying)
-about resonant frequencies
-how non-linearities throw a wrench in linear control design
-about home-made pcbs oxidizing and failing

All of this stuff is neat to know, but it’s not like getting past those hurdles made the end project more impressive. I basically reinvented the wheel over and over, and my wheels turned out pretty lopsided. I figure I spent about 10x the time and money trying to make things from scratch instead of just buying sub-systems that worked out of the box, even if they seemed outlandishly expensive to start with. Had I bought COTS stuff in the first place, I would have had more time to work on the “big picture” ideas.

I think it comes down to your goals and what you want to get out of the project. If your end goal is to make the damn thing work, starting from scratch is probably a bad idea—especially if you’re getting paid. If you’re trying to learn about robotics, then starting from scratch can be great. But even in that case, how deep do you want to go? You could bit-bang your own serial communication, make logic gates from transistors, grind your own motor shafts, anodize your own aluminum in coolers of acid in your garage , write edge detection or blob detection code for vision, etc etc. There’s a lot of pride to be gained by doing all these things, but at some point you have to pick a topic to learn, and buy the rest of the parts from the store. And in the end, are you going to get more out of re-writing a PID control loop for the zillionth time, or maybe spending that time trying to do something creative and maybe even novel?

Your post makes many valid points.

I recently learned that if you make a robot from scratch people are more likely to contact you for consulting or other services for their custom projects.

Once I finally understood this I can understand why people want to make things from scratch especially robots.

Buying robots pre-built is lame. Check out some of the robots my team and I have built from scratch: http://hardware.gt-ieee.org/past-robots

(I’m going to update that site with some new pics soon.)

Just my (biased) 2 cents.