Page 2 of 3 FirstFirst 123 LastLast
Results 11 to 20 of 28

Thread: My Advanced Realistic Humanoid Robot Project

  1. #11

    Re: My Advanced Realistic Humanoid Robot Project

    Here's a little update on the project. Finally got the hand bones joined and ready to rig!



    Also, here's a link to the hand video where I demonstrate its range of motion:
    https://www.youtube.com/watch?v=mRTtyT2cWaI

  2. #12

    Re: My Advanced Realistic Humanoid Robot Project

    Careful! You could scare someone with that! :-)

  3. #13

    Re: My Advanced Realistic Humanoid Robot Project

    Robot rib cage clay sculpt done.


    <code>
    Attached Thumbnails Attached Thumbnails Click image for larger version. 

Name:	rib-cage-clay-sculpt.jpg 
Views:	301 
Size:	83.3 KB 
ID:	6406  

  4. #14
    Join Date
    Dec 2012
    Location
    Los Angeles, CA
    Posts
    860
    Images
    25
    Rep Power
    95

    Re: My Advanced Realistic Humanoid Robot Project

    With all this extra aesthetics how do you propose actuation will work? That has to add quite a bit of extra weight. I checked out the hand video as well, are they intended to have the ability to grasp something? You mention mg996r servos in your project page pulling on strings. Those are rather low grade hobby servos that won't have anywhere near what you would need I feel.

    You also mentioned in your project page using fifty 7.4V 30C 5200mAh LiPo batteries, that alone would be roughly 30 pounds. That is a lot of extra weight there too.

  5. #15

    Re: My Advanced Realistic Humanoid Robot Project

    @KevinO - I am operating under the assumption that the design of the human body's musculo-skeletal system is efficient and flawless and that I will be able to install enough servos to mirror the muscular system of the human body. As far as aesthetics, I am assuming the ribs aren't just for aesthetics. I guess I could remove every other rib and get similar results but lose quite a bit of strength... I doubt removing 6 ribs would matter much in the end as it would be like 1lb or less difference in weight (bones are hollow fiberglass when completed). Yes, the hands are intended to be able to grasp as they will be connected up to servos that pull string that move the different bones the same as tendons would. The mg996r servos can be multiplied until enough of them combine to form the forces I need. Would you say I can have less weight and use less space but have equal strength with a more expensive servo? The huge amount of LiPo batteries is needed to power the huge amount of servos. I don't see any way around that. I guess the whole challenge is how to fit enough batteries and servos to get enough power to operate the humanoid. The problem is the more servos and batteries you add the more you need to add to lift the added weight of the servos and batteries you just added and on and on it goes but there has to be some ideal balance you have to find. The math behind it all is beyond what I'm willing to attempt so I am planning to just "wing it" and then improve on what I come up with and solve problems as I go...

    Also, I am considering using more powerful and expensive servos in place of the cheap hobby ones if and when I find the hobby ones not being sufficient (like for the legs). Plus, I am planning to upgrade to more expensive and stronger servos over time as funding permits while using the cheaper ones to see how far they can take me.

  6. #16

    Re: My Advanced Realistic Humanoid Robot Project

    I am operating under the assumption that the design of the human body's musculo-skeletal system is efficient and flawless
    Flawless? Not quite. But I agree -- it's very efficient! If you could mirror that, you'd be in good shape.

    and that I will be able to install enough servos to mirror the muscular system of the human body.
    The efficiency and energy-to-weight ratios of the best experimental robot actuators/batteries compared to biological muscles/metabolism is 5:1 at best, and usually much worse (probably 20:1 on those MG996R servos.)
    Note: Electrical motors can get excellent efficiency and power-to-weight, especially at medium sizes (70 pounds,) but the rest of the system bogs you down (gearing, batteries, cooling, ...)

    I *highly* recommend you actually do the weight/torque/power kinematics math of your project so that you know exactly what the challenge is while you're trying to tackle it! You may for example be able to trade speed for torque (with gearing) and trade runtime for weight (with smaller batteries, or a tether/wire.)

  7. #17

    Re: My Advanced Realistic Humanoid Robot Project

    I can see how our body's muscles have much more strength compared to their volume and weight than any servo motor. This means I have to use way more mass and weight to achieve similar strength. So my robot will look like a big strong body builder but have the strength of an 8 year old girl. Although it encourages me to see robots like Asimo seeming to pull off strength to weight ratios quite well considering it isn't all that slow (down gearing) and isn't all that weak (can run) and yet it is all fitting in a compact package. So I know it can be done. So you are saying that some servo motors have a better energy to weight ratio than others. How so? Why? Does that have a direct correlation with their price as well? I imagine servos using higher voltage are more efficient no? These sorts of things I need help on as I have no practical experience. You say some batteries/servos can be 5:1 with human muscles but my batteries and servos are 20:1? That means there are batteries and servos out there FOUR TIMES better than the ones I chose. I would like to know how that is possible and which ones I should be buying to get better results. Thanks for your advice so far!

  8. #18
    Join Date
    Dec 2012
    Location
    Los Angeles, CA
    Posts
    860
    Images
    25
    Rep Power
    95

    Re: My Advanced Realistic Humanoid Robot Project

    Have you looked into actual robotics servos? Those servos are simple PWM servos with low resolution and at a stall torque 10 kg·cm. That is super, super low for something even a tenth of the size of what you are planning. I have an OS-5 humanoid with Dynamixel MX-106T's in the load bearing joints and that isn't major overkill for his size. Check them out, they have a 102 kg·cm stall torque which is 10 times stronger than the ones you have chosen.

    http://www.trossenrobotics.com/p/mx-...bot-servo.aspx

  9. #19

    Re: My Advanced Realistic Humanoid Robot Project

    Good link Kevin I'll keep servos like that as my secret weapon I can upgrade to later. For now, too expensive for me...

  10. #20

    Re: My Advanced Realistic Humanoid Robot Project

    it encourages me to see robots like Asimo seeming to pull off strength to weight ratios quite well
    First, if you look at Asimo, it does not have the full range of motion of a human. Second, did you see the budget for that project? Do you have a similar budget?

    Does that have a direct correlation with their price as well?
    Oh, God, yes! For three reasons:

    1) Weight scales by cube of volume. Thus, something that's "twice as big" will have eight times as much weight. (Related to "why can't an elephant jump?")

    2) Robotics equipment is manufactured with different volume and different methods than hobby gear, which means the economies of scale you get for cheap plastic Chinese stuff don't apply to humanoid robotics. More people want to bash some sand with an RC car than want to pay for a humanoid robot, much less a full-scale humanoid. This adds a significant multiple to the cost comparison.

    3) As you scale up beyond "toy" in size, durability and safety becomes really important. This adds to the requirements for engineering, which adds another multiple on top of the cost. Additionally, specifications actually mean something, as opposed to some rather inflated claims made in the toy world.

    If you go back to high school physics, and look up the formulas for torque, it's calculated as distance times force, where the force comes from gravity acting on weight.
    If an upper arm is 40 cm long, and the weight of upper arm plus lower arm is 8 kg, then the necessary torque to just hold the arm against gravity (in the shoulder) is 40*8 = 320 kg.cm. How much are those hobby servos rated for? Do they actually achieve that rating continuously, or is that maximum peak on a good day, and overheating after two seconds?
    If you now decide that the elbow, wrist, and finger actuators can simply be 5x heavier than for a human, then you're suddenly looking at 1600 kg.cm for the shoulder actuators.

    Now, if you have mechanical manufacturing capability, you can choose to trade off speed for torque, with gear boxes (assuming you get something better than the RC hobby servos to drive them.) Hobby servos already have a gear box that typically make the output turn at about 60 RPM, which means you move 60 degrees in 1/6 of a second. If you're OK with moving that distance in two seconds, you can get a factor 12x in torque. Of course, the force on the gearbox and output spline will still be massive, so you'll need significantly beefier gears than what you find if you unscrew a hobby servo.

    Anyway: All of this is wide open for learning if you do some Googling! I highly recommend looking into other humanoid projects (Asimo was one; the DARPA Challenge was recently another; the Willow Garage PR2 has a lot of documentation available, etc.) Then look at their budges, what the available engineering documentation is, and what challenges they really solved, for comparison.
    And do the math on your force and torque requirements!

Thread Information

Users Browsing this Thread

There are currently 1 users browsing this thread. (0 members and 1 guests)

Similar Threads

  1. Giger-OP - An Open Source Research Humanoid Project
    By DresnerRobotics in forum Humanoids, Walkers & Crawlers
    Replies: 9
    Last Post: 06-05-2014, 09:52 PM
  2. Humanoid Project
    By abhishek001 in forum Humanoids, Walkers & Crawlers
    Replies: 2
    Last Post: 03-02-2014, 10:57 PM
  3. Question(s) University humanoid robot project, need controller suggestions
    By Jayws in forum Arbotix, Microcontrollers, Arduino
    Replies: 10
    Last Post: 05-05-2012, 11:48 PM
  4. Question(s) choosing platform for project (biped, torque control servo, advanced kinematic)
    By sergeyko in forum Robotics General Discussion
    Replies: 7
    Last Post: 05-16-2011, 01:02 AM
  5. Incredibly realistic Hexabot Movement
    By Matt in forum Humanoids, Walkers & Crawlers
    Replies: 16
    Last Post: 08-05-2008, 07:16 AM

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •