The best Hobbiest robot, What is it?

-------------- No, you just need to learn what they know about this stuff. And actually, that really ain't much of the totality that they are taught.

Any sincere hobbyist can learn this kind of stuff a piece at a time off the net, easy.

------------ Not really, and there shouldn't be either.

Why? Because there is no sense to building a kit if it teaches nothing. You seem to want to avoid learning anything. Have you asked yourself why? Are you just trying to "wind up" with "cool robot" and save money? Just buy a finished one if that's the case, and a good service plan.

-Steve

Reply to
R. Steve Walz
Loading thread data ...

I agree with Earl, Dave. I know you said you are neither a software nor hardware engineer. But you don't really have to be either to jump in and build a unique robot to your specifications and capabilities. Enough of the process has been "commodity'ized" that you don't have to build everything from scratch. But as you have already discovered, while there are numerous complete robots that are available for purchase, they can carry somewhat of a sticker shock. But not necessarily unjustifiably so.

Case in point - I had my rover at a Hamfest not too long ago and a gentleman approached me wanting to buy it:

formatting link
I mumbled something about it not being for sale but the fellow was quite adament that he was leaving with that robot and how much did I want for it. I hadn't really considered selling it so I didn't really even have a price worked out but I did a quick calculation in my head of just the parts alone that I could remember, added about 50% to cover any that I forgot, and tossed that number out. The fellow was not quite so adament any more :-) I think he was looking to spend around $200, but that would not even cover 3 of the 4 motors, let alone the motor drivers, electronics, radios, joystick controller, wheels, custom machined hubs, dome, extruded aluminum, PVC, batteries, wire, connectors, switches, etc, etc. Add it all up and this little guy is over $1000, not even counting construction time.

Were this robot for sale commercially, it would be a lot more than that. Liability insurance alone would probably be through the roof.

Even at the very low end, it's really hard to do much for under $100, once you purchase controller, motors, motor drivers, sensors, and chassis. You should probably expect to have at least $300 in a basic robot, assuming buying parts new. E-bay and surplus houses can help to lower that, but as always, buyer beware - sometimes it's marked down for a reason.

As you saw, complete commercial platforms start at around $1000 and go up from there. By complete, I mean flip on power and it's ready to go, everything included, not just chassis + motors.

There are a number of semi-kits out there, though. These are typically a chassis + motors and sometimes motor drivers. Examples include the kits from Budget Robotics:

formatting link
Also take a look at Rogue Robotics:

formatting link
Their "ATR" base looks pretty intersting. While I don't have any direct experience with the Rogue Robotics kits, I have purchased a few of Gordon's kits at BudgetRobotics and they are superb. Typically they come with R/C servos for motors + motor-drivers and you add the controller and any sensors and anything else you want.

Building entirely from scratch can be challenging, especially if you've never done it before. I'd recommend starting with at least a semi-kit (base + motors) as that will solve many of the mechanical problems right out of the gate. It sounds like you are wanting voice recognition which I'm not too familiar with with, but that sounds like it might imply a PC for running the recognition software, so your main challenge is selecting a semi-kit base that you can purchase that is large enough to carry a PC motherboard like a mini-itx, enough battery power to run it for an hour or so in addition to having enough juice to move around and act like a robot. Like Earl suggested, PC's make horrible motor controllers, so plan for a "microcontroller" for the low-level controls such as controlling the motors, reading sensors like bump sensors, ultrasonic and infrared obstacle sensors, etc.

Of course, if you can find a Hero 2000 that is in reasonable condition at a yard sale or something where the current owner doesn't know its real value, that'd be pretty sweet :-)

Something else you might take a look at is the White Box Robotics "PC on wheels" platform:

formatting link
These have supposed to have been available now for like a year or something but I still don't see a price or a place to order. They look nice, though. They kind've remind me of Huey, Dewey, and Louie from Silent Running :-)

-Brian

Reply to
Brian Dean

That robot that you made was real nice. I can see that alot of time went into it.

I would spend probably up to $2000 for a robot if it did everything that I wanted. For me Voice recognition is imperitive. There is just something awesome about controlling a machines actions with the spoken word. It doesn't have to understand a conversation, though, I don't expect that. I am thinking along the lines of simple one or two word commands. With this you can break tasks down with a branching structure and do complex things with simple words.

Example:

me- "robot"

Robot - "yes"

me- "guard"

Robot - "did you say guard?"

"yes"

Robot - "guarding"

I can't program at all so me writing the software would be totally impossible, I would need help there. Mechanically I am fine as I said I am a machine designer and design mechanisms all day on 3D CAD.

There is a robot club in my area but unfortunately seems to focus on bug, soccer, and line tracer bots mostly for kids, this really is not my interest.

I have seen the White box robots. That actually looks to be the closest thing to what I am looking for. My only problem with those is since they are running Windows or Linux or something like that on a PC motherboard, they will be very in-efficient on power. I think the better rout would have been to base them on a 400 MHz X-scale Pocket PC PDA instead. This way they could always be on standby like a PDA is normally and not have harddrive motors running and sucking power. All software would be stored on a 2-Gig compact flash card, real lean and mean. You could even go nuts and have two CF ports. One for firmware the other for storage.

What would be reasonable for a HERO 2000 with an arm in working order? Assuming the guy knows what he has.

Thanks again Dave

Reply to
DaveC

Well yes, I kind of do want to wind up with a "cool robot" or one that does what I want anyway. This way I could get into actually experimenting with the robot itself and spend my time that way. I honestly don't want to mess around with the "grunt work" for a year and still have a poor result. I would have no problem with the mechanical end, just the electronics design and code. I would like to buy a finished one but I haven't found one that does what I want.

In my case I would like something where all of the low level grunt work is done and I could just add to it or modify it to my needs. With a kit I could end up spending lots of time, money and still end up with something that falls short of what I want. I could find that I lack the skills to program it (which I do, I am not an experienced programmer) and end up with a half finished bot that is useless. Or I could find that with the kit I bought it simply lacks the ability to go as far as I want as it is too limited. I don't expect the impossible either. I would be happy with a robot that is basically an updated version of a HERO 2000.

Dave

Reply to
DaveC

formatting link

Reply to
blueeyedpop

If you had looked a bit 'deeper' (actually, the next line down...) you would've found this tell tale line:

"The Scorpion is sold only in conjunction with an ERSP purchase."

The ERSP is an SDK for application development.

Reply to
Joe Squid

Dave,

The problem is that if you don't know the skills needed to do exactly what you want and you are unwilling to learn the skills needed, you have to make compromises. The trick is to play to your strengths. Put some effort into the areas you are good at and look for partners on the net or locally why have strengths in other areas.

Take another look at the ER-1. The cool thing about the ER-1 is not the robot, it is the software. The controlling system is written as an open architecture, it is designed to allow modules to be easily added in nearly a plug and play fashion. There are modules already written to do what you want. Some come with the kit, some are written by other people. If you do a little looking, you will find that there are modules that either do voice recognition and speech or interface with other PC programs that do these things. The hard part has been done.

Now, take a look at the parts in the kit. You know that X bar is not the lightest or stiffest way to make a frame. You have the skills to make a frame and a shell that are light enough for the motors being used in the kit, but looks like an updated version of the HERO that you want, BUILD IT!

At that point, you will have a bot that looks like you want it to and can do most, but maybe not all, of the things that you would like to. Next yo show it off. Find web communities that talk about the ER-1 like:

formatting link
. Show them what you have done. A lot of the guys working on the ER-1 are programming types. The fact that you are able to do so much structural work will impress them. Get to know these guys. Offer to trade with them; you will make frames that they have been dreaming about, they will make program modules that you have been dreaming about. Interface, share, trade, learn.

Eventually, you will find someone who has been porting the ER-1 software to a smaller, more power efficient controller such as a CE device or a SBC. Add them to the group of people with whom you are networking. You or other people you now know, probably can help them with problem that they are having. Interface, share, trade, learn. In the end, you will have the bot you wanted and so much more.

Paul

DaveC wrote:

Reply to
Catman

Thanks.

Depends on the buyer :-) Seriously, the HERO 2000 is somewhat of a collectors item and carries value simply because of that. From rumors I've heard there were only about 3000 or so sold so they are somewhat rare.

As far as trying to put a value on it - consider the electronics in it are over 20 years old, trying to get one working and stay working might be a chore. While I don't have personal experience with a HERO

2000, from the photos I've seen they appear to be very well built mechanically - the arm looks very nice, far better than the toy arms you can find based on hobby R/C servos.

I'd put a higher value on the base than on any of the electronics. I think electronics could be replaced with superior modern components and boards for not too much money. But good base designs are harder to come by.

-Brian

Reply to
Brian Dean

formatting link
> The ER1 in the basic kit model selles for around $300, but you still need

Reply to
mehaase

Well thanks and there is some good advice here. I looked at the ER-1 again but it just isn't there yet. For one thing it only allows for the sensors that come with it. You can't add anything like sonar, infra red, bump sensors etc to it. It is basically an as-is system. Also it requires alot of computing power to run it. This means it will suck power too fast. I want a robot that can be on standby and run at least the brain, for hours maybe even days, not minutes. If the software was made so that it could run on 400 Mhz PDAs like a PPC it could run for a very long time as it would use CF and RAM instead of hard drives. Also PDAs can be in sleep mode for weeks between a recharge and still be running apps such as calandars etc. PCs require too much power just to run the OS even before a single robot command is run. It just isn't practical. Just look at the AIBO, it has voice rec and vision and they don't need a big laptop to do it. It is because the hardware and OS is dedicated to robot functions, not running all kinds of useless Windows functions that have no benifit to the robot like ER-1 has. If AIBO had a big battery and cut down on all of those leg servos it would last a real long time.

So I search on. I am hoping sooner or later someone will come up with a good and efficient system. Dave

Reply to
DaveC

What about one of the gumstix modules or system for a controller?

formatting link
Alex

Reply to
Alex Gibson

The ER1 has 15 Analog inputs, 8 Digital inputs, and 8 Digital outputs to interface sensors, switches, motors, etc., via the Hobby Ports. You do have to do some tinkering, however, to hook up something like the Sharp infrared sensors to the Hobby Ports. There is also a Command Line API which allows you to control the robot, including the Hobby Ports, from code you write.

If you wanted to pursue it, I'm sure the folks at Evolution.com would be more than happy to provide some free technical support showing how to hook up such items to the robot.

Just wanted to add this in case you weren't aware of it.

Reply to
jc-Atl

I was reading on their forums that many people were mad because ER didn't, and wasn't planning to, allow access to the hobby ports. Maybe that has changed since though.

Reply to
DaveC

You can't access the hobby ports through their GUI, but you can by using the socket based API protocol. I know because I wrote some software to do just that, which is published on the ER1 forum.

I "think" you can access it through the ERSP; but I don't don't if you can still get a copy of that when you buy the 'bot from them. Others that got the ER1 early were able to get the ERSP, though.

I'm going to check out the forums to see what you mentioned about others being mad about access to the ports. I think they should provide access to the hobby ports, too. Maybe it has changed since I got the latest download of the GUI !!! ( wishful thinking, probably ... )

Reply to
jc-Atl

It appears this is correct:

formatting link
There is not yet official access to the hobby ports through the GUI. This is what I would need because I am not a programmer. I suppose if enough coders hack around enough their may be some stuff useable to us non-coders.

I don't have this robot yet, I am looking at different things still. I am thinking of this or the Whitebox robots when/if they become available. I like the idea of the Whitebox ones looking more HERO like and the fact that you don't need a pricey high-end laptop for it.

This is one area where I think they kind of "blew it" a little with the ER-1. The fact that you need such a high resolution screen just to run the software seems kind of limiting. They should have a version without all of the "fluff" with the pretty GUI windows and have a version that would run on lower spec laptops with 800 x 600 screens. If this could be done than smaller, lighter, and more power efficient laptops or sub-notebooks could be used. If they got real efficient with their code you could even run it on PDAs like the x-scale 400-500 Mhz PPCs or Palms. These PDA versions would be a bit more trimmed down in the vision dept. or at least would respond a bit slower. Don't forget the AIBO has vision and voice rec. and it only runs at around 500 Mhz.

My ideal bot would be one that could be on standby and only need a charge once a week or so when the motors aren't running and in sleep mode. This way it would always be ready and you could have it "awaken" at certain times and/or to a sound or voice command. Having to "boot-up" your robot before you use it seems kind of limiting. Maybe I could connect a PDA to the computer switch and have the PDA turn on the laptop on the bot at scheduled times? and /or have a sound activated switch connected to a latching relay which would turn on the bot when a sound is heard? For example I could yell "wake up"or "robot" and the switch would turn on the laptop and it would start working?

Also it should be autonomous enough to go back and charge itself when the battery is low. A $200 Roomba even does that.

For robots I think the "lean and mean" approach is good, it just makes it more practical.

Anyway , I haven't decided yet as I don't have all of the facts. I suppost if those White box ones ever come out I will have something to compare. The ER-1 could be good for me too, I am not sure yet. If you have any info that would help feel free to let me know.

Dave

Reply to
DaveC

Yes the ER1 has its limitations when it comes to advanced robotics development, but Sony AIBO licenses the Vision part of the software (part of ERSP and the ER1 control software). Check this out:

formatting link

- PJ

Reply to
PJ

PolyTech Forum website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.