Let's Make Robots!

Big company robots making it look easy?

G'day LMR (Im Australian what did you expect? :))

I am new to this site and Im finding my time here enjoyable and uplifting. I have been intersted in electronics for as long as I can remember and four months ago began making things with an Arduino Uno, be it small gadjets or robots.

One thing I have noticed about robotics is how complicated it is (which I quiet enjoy), with all the designing/building/programming it can take anywhere from months to maybe even a years. I am posting all the things I make on YouTube for a sort of video diary, so I can see where I started and how far I have progressed, as personal encouragement.

Lately my friends have been going onto my YouTube channel and coming to me asking questions about the things I made, like how long it took to make etc. On one occasion a friend of mine asked if it was possible for me to make a walking humanoid robot like Asimo, I answered with a no, telling him how I was new at this and wouldn't even begin to understand where to start. My friend replied with "Come on, it cant be that hard, look at all those robots on TV and stuff, asimo wouldnt be that hard to make, its like four years old". Immediatly I felt something inside of me die, I couldnt even imagine how many years of designing/planning/building/revising/programming/debugging it took Honda to make a humanoid robot that could even walk.

The main thing I'm getting at here is how contant exposure of top end robots on the media makes the public less excited about the little things and lose appreciation for the true intelligence possesed by the people who make the marvels.

That was my first forum post and it felt good :)

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

Only the sky is the limit chickenparmi :-)

Well, people like your friend does not have the knowledge to judge what's doable with ressources you have at home or you can buy as a "normal" guy. There are limits we (LMR and other hobbyist robot builders) which we can not achieve, at least not within the budget we have available.

I mean, robots like the ones on LMR are made of materials and parts we can buy in common shops or we can manufacture by ourself. With this materials we build sometmes the most incredible robots but on the other side outsiders does not see the technology behind. 

Peopel keep asking me where on ALF the microcontroller is. When I explain them that there is not such thing but only an analog circuit they mostly say "Ok, that's cool" telling me that they not understand how this is possible :-)

Seeing ASIMO and seeing some bipeds here on LMR or other websites will lead to the conclusion that's all the same backend. He (ASIMO) walks, grabs things etc. such as the "homemade" bipeds do. On the outside it's partially true since they don't know better and this is good for us because in their eyes we are smart :-)))

Telling them it's just a matter of money will make them believe that there is a gap between ASIMO and your homemade robot, even if the behavior is quite the same. I do not know the price of, lets say one servo, ASIMO is using but i guess it's a bit more than a standard Futuba S3003. There are digital servos on the market for 600 $US per piece (and I guess that's not the end). Just ask your friends who can afford the 600$US x 18 to build a humanoid robot. Besides this the hardware is not complete yet. Servos like this need a controller, power and a programming to work proberly. 

I don't say that ASIMO could be build in the basement of an Australian house but if the budget is there why not? See the Space One project. NASA does need much more money to shoot a rocket into space than these private funded companies. So why not believe that a robot like ASIMO could be manufactured by hobbyists. It's all just a matter of time and money. Technology get's cheaper everyday (see the price for processing power - 1GHz CPU 1999 and now) and with this it's more feasable to build more intelligent systems in your basement. 

Wow, there are 1000 more reasons but I will stop here for now :-) 

I don't say that ASIMO could be build in the basement of an Australian house but if the budget is there why not?

I think there's a challenge to do almost exactly that running right now-both on LMR and through DARPA. I know MyBlackImpala60 is working on one.

On the other side of the argument I think that some good things can come out of it. Perhaps it gives us something to aim for? Everything for us robot builders is a challenge (wise words as always from Oddbot), but perhaps the challenge for us is to 'fake it' (kingart2) and see if we can build something to what the untrained eye see's as 'state of the art'. One thing that Protowrxs mentioned is the 1%, I think we as a community learn much more from each other than just how to build robots or use tools, building robots changes our aspect on the world. That previous statement might sound a bit dramatic but understanding how to program a robot to walk gives us more knowlege about how we walk for example, we start appreciating the little things that the human body is capable of is what I'm trying to say. And maybe one/two of my friends actually understand the work that is needed to make a robot, the rest, just think I'm a magician faking it :). 

 

Be proud to be part of the 1% who think the game "naughts and crosses" (tic - tac - toe) is only mentally challenging if we can make a machine play it as well or better than a human.

 

Your post has inspired a new challenge: http://letsmakerobots.com/node/35296

 

If your friends think it's so easy then get them to enter this challenge.

It's good to someone said something along this line. I think that most of us have run into this before. You spend hours building and coding something that DOES move around own its own, avoids (most) obstacles and and finds the light or dark and you get the unimpressed look from most people you know. Even getting something to balance on two wheels (but not move anywhere yet) gets the that's cool but why doesn't move anywhere feedback even from family.
Speaking of the picture enhancement, heck we've had detectives at the city where I work I come up to our IT shop with a tiny grainy scan of a drivers image wanting us to "enhance " it to poster size for a wanted poster and get angry because we said it can't be does.
It's a much more complicated world then 99% of the people out there know. It's the 1% of people like most here and the technical people of the world that know its not really "magic" like it appears but brains, hard work, and money that make the magic go round.
Of course it's only going to get MORE complicated as the onward and upward log scale of technology continues. My Dad was a technologist, he developed hard drives and helped develop storage capacities as they went from hundreds of pounds for 60 MEG on 15 inch platters of storage to the 2 inch 1+ Gig drives before passing away. I think even he would be amazed at the 4+ Tera byte drives available to the general public.
At the current rate I am wondering how long it will be before its all "magic" to me as well.
Stephen

Yeah, I knw what you mean. Plus, there's the japanese and korean humanoid fighting robots (remote controlled...) that your friends might see online and seeing those move will make one think it's not too complicated to make them autonomous. If you look at the humanoid soccer robots, you'll see more advanced ones that have more processing onboard and cameras for vision, but they are still controlled by a PC in the background. Nao and Darwin are expensive marvels created by teams of research students and teachers and I think they are pretty much as hardware evolved as they could be, the only thing that will improve over the next years is the software, that is always behind hardware developments. I think the Raspberry Pi will soon make humanoid robots more cappable, but I fear there are not enough robot builders out there with proper programming knowlege to make it happen fast. Oh well, look at the Trossen Robotics forums, they are using ROS in their robots so they will adapt the Pi fast in their robots.

I agree that the average person has no idea how challenging it can be to make a robot that can even find its way out an open doorway.  It's not until you start building a robot that you truly appreciate how "smart" people really are.  When was the last time you met someone too dumb to find their way out an open door?

Our excellent vision sensors, combined with a fast CPU and massive object identification database make all such tasks trivial.

Asimo is a marvel of engineering, while showing how far we still have to go on.  I mostly blame the programming at this point.  The sensors and the mechanics are quite good, but to build a "human class" robot is going to require a mountain of software.  I doubt I will live long enough to see it, but I have great faith that it is inevitable and will come.  The general public also has high expectations from seeing robots in science fiction, these set a high bar.

In regards to out own efforts, I delight in simply trying to make a robot that does more than the previous one.  I learned a lot from building my first robot and now that I am starting my second one I am learning more "surprises" everyday ...LOL.

My own focus is on the companion/pet type robot and in that regard I think a lot can be done to "fake it".  I am currently trying to make a robot that can find its way to different waypoints around my house.  Assume I succeed.  One day a friend comes over and I say "Bert, go to the dining room table", then I press the button on my remote for waypoint 5.  Bert wanders out the door and down the hall and into the dining room.  My friend thinks, "WOW, Bert went to the dining room table."  Of course you and I know it avoided a few walls and bumbled his way to the right spot.  Bert has no idea what a table is, but in a closed environment, much can be done.

I believe our challenge is to see how close we can "fake it" to Asimo without spending 30 years and several million dollars.

One question though chickenparmi, are you sure you are Australian?  All the Australians I know are Chinese ...LOL.

 

                        Welcome to the fun,

                                         Hal

 

 

I have been always looking into the newest achievements in the programming fronts for robotics and It is looking very promising. Computer vision and object recognisation are the two areas in robotics that I have seen come a very long way in a very short time span. The other day when I was on the YouTubes I was watching an interview with a ex physcologist who had a change of interest and it now helping to produce new AI systems for robitics, I forget the name of this person but searching through some old history might uncovner this interesting video. Its funny I/we see robots in these science fiction movies and can take reasonable guesses as to how far away this technology is, whereas the general public just raise thier expectations to what is possible for us to do. 

Your next robot sounds very interesting and I would definetly like to see it. I too am into making robotic companions, I think that is what my interests in computer vision and speech recognition came from.

And yes, Australian born, Australian bred.

It's a pleasure to met you, hopefully you will find a lot of fun here.  I have started a blog on robot #2, it can be found here.

http://letsmakerobots.com/node/35201

Speech and the ability to talk are both high on my list to.

Good luck with your projects, I think you will find the people here are fun and helpful.

Actually Asimo is a lot older than 4 years. Honda started research that ended in Asimo in the 1980's and Asimo himself was introduced in 2000. So tell your friends that Asimo is the result of about 30 years of research and many millions of dollars.

Designing robots is never easy.
If you work entirely from off the shelf parts then you need to work within the limitations of the parts.
If you design new parts then you need to put in a lot more work.

In both cases, budget, time and resources (e.g. access to a 3D printer or CNC machine) are your enemies.

If your working as part of a team then communications can also cause problems even if you all speak the same language.

 

We build robots for the challenge and to learn new things. Not because it's easy.