Need Guidance on What Kind of Robot to Build For Project

Hello,

I want to build a robot so that I can experiment with robotic navigation and sight. I want a fairly stable robot. So far I have narrowed the design down to five types of robots: humanoids, hexapods, quadrupeds, two-wheeled robots, and four-wheeled robots. For my purpose, which of these types of robots would be the best (and the most interesting). I am on a budget, so please list the options from best to worst so that I can figure out which robot type to look for (as in a kit).

Thank you, George

What’s your budget?

$150 dollars or so. Please note that I already have an Arduino, so you need not count that cost in.

So you need some motors, perhaps servos (or modified continuous rotation servos) or stepper motors You need some controllers/drivers for the motors You need a power supply, or a big battery pack You need some kind of chassis to secure all the parts on to. You need sensors. You need connecting wires for everything.

I'd start by finding motors & drivers - I think those will be the most expensive parts. These would all be interesting: humanoids, hexapods, quadrupeds but need the most parts and thus higher cost.

Four wheeled robot would likely be the most stable platform to build up from. Add a multijoint arm after you get t he base moving around and avoiding things. Or 2-wheeled self-balancing robot, add remote control so you can drive it around. That requires interesting things like gyros and accelerometers too.

I found examples of the parts I need for all of these robots (although the humanoid robot is very simplistic). They are all within my budget. Just to let you know, in addition to the Arduino, I also have one servo, a photo resistor, and a PIR sensor.

I suggest you start out simple and inexpensive to start your project. Below is a simple bot ptoject that might be of interest. The most difficult part seems to be making the wheels, which pie or pizza pans from the dollar store might make an easy substitute.

http://inventorartist.com/rolly-bot/

Thank you. But for me assembling the wheels for that robot would actually be harder than creating a wheeled robot with a sophisticated turning mechanism.

You can buy an RC vehicle for next to nothing at a thrift shop. After removing the radio, you will still have usable motors and a steering mechanism.

That requires a lot of hacking, though. I am not exactly tool-rich. Implementing sensors and the Arduino board would also be rather difficult. For better context, though, I will show you all the options I am looking at for building the robot.

  1. Dagu Mini Robots Kit - $59.95 (includes parts to build a basic 4 DOF humanoid or hexapod)
  2. Simple Hexapod Chassis (no electronics) - $99.95 (but includes servos, 12 DOF hexapod)
  3. Playful Puppy 2 Quadruped Robot KIt - $146.95 (includes everything I need, 11 DOF quadruped)
  4. Dagu 2WD Beginner Robot Chassis V2 (blue) - $14.95 (2-wheeled robot)
  5. Dagu 4WD Chassis - $25.95 (4-wheeled robot)

All of these kits are from RobotShop.

Go on Amazon and look up "Redcat Racing" - ah, what the heck - here's a link:

http://www.amazon.com/Redcat-Racing-Blackout-Waterproof-Electronics/dp/B00O9MTTDO

Check out the videos, check out the comments - it ain't a Traxxas E-Maxx - but you won't be paying that kind of cash, either. It seems like a good deal.

Since it's a standard hobby-grade RC vehicle, it uses a regular ESC (for a brushed motor), and a regular servo for steering - and a regular 2.4 GHz radio system. It comes will all the parts - plus batteries, radio gear, charger, etc - but you can put more money in if you want or need (I would recommend getting another battery or 2, plus a smart charger - and maybe a banana plug (redcat) to tamiya battery adapter - which is more standard - do some research here, so you know what you need to get).

You can then hook the ESC and steering servo up to the Arduino easily, and control them using the Servo library.

There's your basic robot. Add on another servo, plus a bracket, plus an ultrasonic sensor - for a panning sonar system - all of that can still fit on the Arduino. You'll have to figure out a way to mount it to the front of the vehicle - ideally fairly solid.

The car's a beefy beast - fairly fast, fairly tough. Plenty of spare parts and hop-up parts available too.

I know you said that you didn't want to hack on something, or use tools, etc - but you're going to find that if you want to do robotics, you're going to need to get into knowing how to do things like that. It just comes with the territory.

Also - check out Servo City in the meantime, too - if you have the money - they sell a ton of compatible parts, gearing, brackets, and aluminum extrusion and such to all you to easily build robots (again, if you have the cash - the stuff isn't cheap, but it is high quality):

http://beatty-robotics.com/mars-rover/

That's a "mars rover" built (partially) with what they sell - Beatty posts here every now and then; he's really dedicated to teaching and instilling in his daughters the importance and knowledge of a STEM education - a great guy and dad, by all accounts - and his kids seem to have taken to it, too!

Lastly - RobotShop is a good place to go to review things - but at times, their stuff seems like it is really marked up or something - I would use them as a reference only, and only buy from them if you can't find it elsewhere cheaper. Many times, you can - if you look.

Oh - one more thing - if you ever do want to "hack" on a junk toy RC car like jremington proposed - check out this thread:

http://forum.arduino.cc/index.php/topic,86883.0.html

Good luck!

You misunderstood me somewhat. I meant that I would not take a part an RC vehicle if I had to smash into the frame. I prefer an RC vehicle to have screws so that I can unscrew it and put in components. The other difficulty would be rewiring the motors. I would have to unsolder the motors from the control board (a tool I don't have) and figure out where everything should connect to. The car you mentioned seems a little too pricey for what it is, and is too fast for what I want it to do (navigation: i.e., exploring an area and being able to create a map of that area in memory to use later). If I decided to use the car you mentioned, I would need to gear down the speed and do a significant amount of rewiring. It seems that legged robot is best for what I want it to do. Thank you for your comment on buying robot parts, by the way. I will do some more research to see if I can find better prices.

I want to build a robot so that I can experiment with robotic navigation and sight.

It seems to me you need to decide the sort of environment the robot the device will operate in, and the sorts of things it will encounter first.

For instance, rough terrain may require a more complex stabilised camera platform than, for instance, an ice-rink. In either case, the traction requirements will be somewhat different.

Right now the robot will be operating inside the house, on carpet and tile, but it would be nice if the robot could travel on rough terrain as well.

I found an old RadioShack “RoughRider” for $4 at Goodwill. I then found out where the motor inputs were and now I can control it with a 5V signal from a Arduino. Really, just start with a RC car.

goodinventor: You misunderstood me somewhat. I meant that I would not take a part an RC vehicle if I had to smash into the frame. I prefer an RC vehicle to have screws so that I can unscrew it and put in components.

Well - hobby RC cars are designed to be taken apart and rebuilt - they are actually perfect for your needs (if a wheeled vehicle will work for you).

goodinventor: The other difficulty would be rewiring the motors. I would have to unsolder the motors from the control board (a tool I don't have) and figure out where everything should connect to.

Only with a toy RC car - but this is something you should strive to learn! But, for a hobby RC car, everything is basically a plug together design. No soldering is involved. You will, however, need to purchase a small set of tools sized for the parts on the hobby RC car (no way around the tools, I'm afraid - if you want to build robots, get used to using tools - and get used to getting hurt, because it will happen - heck, just the other day I nearly took off a finger using a right-angle grinder with a cutting wheel).

goodinventor: The car you mentioned seems a little too pricey for what it is, and is too fast for what I want it to do (navigation: i.e., exploring an area and being able to create a map of that area in memory to use later).

Actually that car is dead cheap compared to most hobby RC 4WD truggys - have you ever priced a 1/10 scale Traxxas E-Maxx or Revo? Prepare for sticker shock!

goodinventor: If I decided to use the car you mentioned, I would need to gear down the speed and do a significant amount of rewiring.

No - you shouldn't. That's what the ESC on the car is for - to allow you to go as slow or fast as you want. These cars can run really slow - or go quite fast.

goodinventor: It seems that legged robot is best for what I want it to do. Thank you for your comment on buying robot parts, by the way. I will do some more research to see if I can find better prices.

Legged robots bring in a whole 'nother host of issues, especially if you are dealing with this as your first robot. You're already up to your ears with the mechanics of a simple platform - then you want to add localization and mapping to the mix (not an easy task by half!) - and you want to add gait pattern and dynamic leg/foot placement (not to mention forward and inverse kinematics!) to the mix?

It sounds like you are tackling a project much larger than your experience level - then again, you haven't told us what that is (outside of not being able to wield tools) - so maybe I am wrong...?

:D

No, I have experience with using certain tools and have them (like wire strippers and screwdrivers), but I do not have experience with soldering (nor do I possess a soldering iron). The reason why I thought a legged robot would be better is because legged robots can travel through obstacles (especially big ones) like stairs much better than wheeled robots can.

The question is, how will a robot, even a legged one, climb stairs? It will have to be big enough and also probably have the dimensions of the stairs hardcoded into it.

Thank you. But for me assembling the wheels for that robot would actually be harder than creating a wheeled robot with a sophisticated turning mechanism.

What I suggested is essentially hot gluing a servo horn to the center of a pie pan. If that is difficult for you, your project may not go anywhere.

Actually, zoomkat, that's doable. What is hard is assembling the complicated wooden wheels shown in the picture. As for making a robot climb stairs, I only said that it is possible to make a legged robot climb stairs. I never said it was easy; in fact, I am not working on that problem just yet. First I am going to try to just get the robot to walk.

So, are you saying that making a walking robot is not a good idea? The problem with wheeled robots is that they cannot traverse rugged terrain very well, especially stairs.