Let's Make Robots! | RobotShop

Full scale human torso on a hexapod, how to start planning?

Hey everyone!

I'm quite new to he scene. I am basically somewhat qualified in AI concepts and Programming, yet I totally lack any engineering skills. Would like to learn them en passant though.

So, what I am trying to build is a "true scale" human torso (well, basically a Spine, a Rib Cage [for Style reasons], a Head (cams and speaker), and sooner or later two arms (for budget reasons I'd like to add the arms later on, I'll just need to prepare the other parts so they can be added)).

The spine should blend into some kind of "walking" mechanism. I have been thinking about snake-like movement patterns, but that loos quite complicated and I gues it does not really work with "high" weights, given I dont want to add a combustion engine. Therefore I think a hexapod base would be quite reasonable, since I dont really feel qualified for dealing with balance issues which would come with a two legged robot.

As a counterweight to the torso and as actual place for an at least mediocre embedded machine, I'd add a "Spider Abdoment" on the "tail" side of the hexapod.

So, thats the plan. My problems:

(1) How do I plan something like this? In a technical way I mean? I dont really understand, how metal parts are joined with each other,or how servos are "implemented". I thought about modeling the whole thing in a 3d program first (I'm experienced in Blender, but guess I should use something that can give me the volume of my objects), in order to find out, what the total weight would be, depending on (2).

(2) Which materials to use? For various reasons (Power supply, cash, servos, ...) I would like it to be at least somewhat "lightweight". The spine probably needs to be quite solid. Aluminuim (dont like that unhealthy stuff though!)? Even steel? But steel is not only expensive but also quite complicated to process, right? the hexapod base and legs are probably the most crucial part when it comes to weight. they gotta support torso and the machine, and themselves :D. So, ... Aluminum? Wood? Gotta endure the weight though.

(3) and this is probably the combination of scale and material choices: The Servos. I am not really confident about my estimation, but I guess about 100-150 kg are about realistic? For the whole construction I mean, so a single leg would be about...  4-6 kg maybe? Can I even do this with regular servos within reasonable budget limits?

So yea, as you see, I dont have too much of a clue, especially I try to find out if 1. all this is realistic, 2. how to start planning. Your help would be much appreciated!

(feel free to tell me, that I am delusional, I will probably listen to your opinions, if you ave any advice about shrinking all this down in one way or the other, or splitting it into moduar steps. Basicallyone of the only requirements is the size(!) I simply like big things.)


Comment viewing options

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

Sounds like a great combination torso - hexapod. The choice of materials is going to be critical as every gram is going to be supported by the motor shaft/coupling. Do you have the ability to 3D model? How is you math on torque requirements? 3D print a scale model?

My hexapod is sitting idle (3 years now) due to killing the power regulators. 

I like this idea.  I think it will be really cool.  I'm visioning the creature from "Alien."

First I will give some general notes on how to engineer something like this.  Then I will give some specific notes to your numbered questions.

Use the lightest weight materials you can for every part.  That may mean using different materials in different places, which is fine. Lightweight materials can often be surprisingly strong.  A long time ago I saw a TV commercial where they built a bridge out of carboard and drove a car across it.  Consider that until the 1940s most airplanes were built mostly from lightweight wood.  Look up the "Spruce Goose."  Styrofoam weighs almost nothing (it's mostly air) but can be very strong.  Especially if covered with a thin, rigit material as a "skin."  Check out "foam core boards" like Sintra http://www.sintrapvc.com/ at your local hobby or craft store or online.

"Backward planning":  Work from the end result back.  In other words, start at the furthest part.  In your case, that would be the head.  Design it as light and small as you can.  Then, design the lightest, smallest spine you can that will hold up the head in every orientation you desire.  From there, go to the torso, or whatever has to hold up the spine.  Again, make it as small and light as you can.  As you can see, each piece working from top to bottom will need to be bigger and stronger to hold up the parts it has to carry.  Make each part only as strong (and heavy!) as it has to be: no more.  The feet and legs of course will have to be the strongest.

Try to put the heaviest parts (batteries, motors, etc.) as low as possible.  Avoid a "snowball" effect where placing a part higher means everything below it has to be stronger and heavier, which makes the parts below that heavier, and so on.

Now, your specific questions in order:

1. I don't recommend using any type of cad program to start.  I recommend you get some materials and play with them, experimenting to see what works.  Get some hands on experience with the materials and what they can do.  Once you have a really good idea how you want to proceed, a cad program to create the design is probably a good idea.  I wouldn't bother with calculating volumes to find weights, etc.  That's probably way more trouble than it's worth.  You aren't building a starship.  And finding/calculating the densities and volumes of your chosen materials isn't going to be easy unless you buy them from people selling to starship builders at starship prices.

2. As I mentioned, the lightest weight materials you can.  I strongly suspect that either a 3 foot length of 2" diameter PVC pipe or a 3 inch square column of styrofoam with a thin plastic skin would be plenty strong for your spine.  Or perhaps some thin "craft plywood" glued into a column about 3 inches square.  Will the spine be rigid or will it be in parts that have to move with joints between them?   That will make a BIG difference.  You can often find really nice building materials pre-made for you at the hardware store.  Take a look at PVC pipe and electrical conduit, both plastic and steel.  Steel conduit can be fairly light and VERY strong.  It's also not too hard to work with, especially since it's already in workable pieces.

3.  If you build as light as you can, I think you can come in well under 100kg.  But not light enough to use standard servos.  Servos big enough to operate the legs on this thing are available, but you aren't going to want to pay for them.  They are in the neighborhood of $300 each.  Decent DC gearmotors and driver boards should be easily available for around $50 each, which is still quite a bit but a lot better than $300.  They will need more software, but that's your strong point :-)


"I am not really confident about my estimation, but I guess about 100-150 kg are about realistic?" That's actually an incredibly heavy hexapod robot. You need to understand the concept of torque for each joint and measure twice and cut once to avoid spending considerable amounts of parts which either don't work at all or burn / break because they are used beyond their limits.

You should try something smaller scale first. A kit like the one below would be around $1,700 USD and it does not use "smart motors", nor have a full computer onboard.



Lynxmotion SES Kit