Most likely that we (and I think nature too) model leg motion as two legs, and then all others are multiples of these two base legs, i.e., a quadruped moves (although not always) two legs at the same time, doing the same motion. a hexapod 3, octapod 4, you get the analogy. So what would you do with the odd leg out?
Surely some form of ripple or wave gait, with (in the case of a pentapod) two legs raised while three are down. Not that i’m saying natures got it wrong I just think that there is untapped potential here, particularly on a pricing front, as someone who can’t accord all of the servos for a Hexapod (18), but can accord more than a Quad or Scout (12), could create something special. Plus a pentapod would offer increased stability over a quad due to the lesser need to shift the COG.
Doing a quick search i’ve found this half finished robot.
http://thingiverse_beta.s3.amazonaws.com/renders/78/56/ff/8f/a8/3049264954_252c75605d_o_display_large.jpg
If something like this was made into a kit, it could always be marketed as a star fish.
Just my thoughts. Does anybody else have views on this? Perhaps Robot Dude?
I’ve been thinking back to what kurte mentioned (about bush robots), and have come up with a rough idea of how one could be made and be a viable robotic platform.
If you’ve seen the 1998 film “Lost In Space” then use the image of the spiders that attack the ship as a starting point. (I would show you a picture but I can’t find one on google ). The robot would be a tripod of sorts, with large feet to better support its weight whilst moving. The feet themselves wouldn’t just be a flat piece of metal, they would be a set of three 3DOF toes that would assist when moving over rough terrain. Relating it back to bush robots, this would be a bush with branching factor 3, taken to degree 2.
Having this setup offers a few advantages over a traditional robot, as it can lift one of its legs and use its toes as fingers to grasp and move objects, with the leg itself acting as an arm. The robot wouldn’t have to be immobile either, as its other two feet can be converted to a set of 6 legs allowing it to walk on them with a wave gait (provided that 5 toes can support the robots weight + payload).
In total such a robot would require 36 servos minimum, with an extra 3 being possible to add a roll to the wrist joint. All of this complexity is probably unneccessary as you could quite easily add a robot arm onto one of Lynxmotions hexapod platforms and get the job done, but the whole idea of a robot that is a little bit different appeals to me and this certainly fits the bill. I don’t think it is a good idea for a first build (although I would like to build it), but I definately think it is possible to simulate it. If I did go for this I would look at building it in stages, starting with a CH3-R (its legs becoming the toes on two feet).
What do you people think, particularly seeing as you’ve built robots and have an idea of the capabilities and limitations of Servos and SES. Or maybe I’m being too inventive as the cost of the servo’s alone makes this project problematic.
I chose to adopt a different modeller to that I’ve used previously, and have spent most of the time trying to faithfully recreate the SES parts. They are all to scale relative to each other, but I doubt they are the same as the real pieces.
The idea is that it should be able to use that foot to act as a hand, although the custom piece is a bit too large to allow smaller items to be picked up.
I’d appreciate any comments on the design or its feasibility, particularly if there are things that cannot be done or that could be done in a more efficient manner.
-1
P.S for now imagine three of these connected to a CH3-R chassis, until I get around to modelling it.
I have been having thoughs about the Branched Tripod I mentioned earlier, and I’ve come to the conclusion that it is just far too complex and expensive to build, and would only really work in simulation. I do like certain features of it, such as its legs made from AL5D’s and having a foot/hand on the end. If anybody has any ideas of how the design could be altered to make a large legged robot (doesn’t have to be a tripod, actually I’m leaning more towards a quadruped now) then I’d like to hear them.
What if a quadrupod could be made with unusually long legs with an AL5 series arm on it, but rather being mounted on the top of the robot it were mounted underneath the robots chasis. Perhaps even the arm could function as a leg. Imagine Innerbreeds latest quadrupod with an arm mounted underneath it.
that sounds like a very interesting idea. with longer legs and a wide chassis you could have plenty of play on the arm. i also think that adding the arm/LEG in the center of the body would be quite easy to add to the IK equations. im guessing that when/if used as an extra leg its main reason would compensate for the center of gravity shift between the other legs.?
the other thing you need to think about is the purpose of this robot. just simply having a hex with an arm would give you stability all round. but it would be expensive from my point of view.
a quad would be ok, the arm would act in ways mentioned above. Cog shift!
a tripod wouldn’t be very efficient as its stability while in motion is scary alone.
i would think about designing a quadruped that once it arrives at its destination it can then rotate its legs into a tripod configuration and then use the other leg to manipulate a given object. maybe that leg could fold further out by having an extra DOF with maybe a series of mini servos acting as a small gripper.
It could be used to do that. It might even be able to pick up the robot and turn it.(If the gripper were to be big enough and the arm was sturdy enough)
That would be a pretty good idea. It is true the underchasis would be kind of be impractical compared to an arm thats right side up, but I suppose thats why this post is called “unusual configurations”.
Related to what you two have been saying, how about the reverse? Rather than having a walking robot with an arm over or under it, how about a walker with manipulators on its legs but with an extra leg when needed, a bit like a tail. So in the case of a quadruped this tail could be in the air moving around (perhaps to aid COG shifting), then if the robot wanted to do what innerbreed suggested and shift to a tripod stance, one arm would be free and the tail could replace it as a leg, still allowing the robot to walk. Alternatively if two objects need to be connected to each other the tail and back two legs could become a tripod, leaving the front two free as arms.
I’d certainly say that this is an unusual configuration of robot, but if I was to be asked what its purpose is or what it can do over simpler (and cheaper) robots I couldn’t answer that. I just think its cool to combine all these abilities into one robotic platform.
there is only one floor with my suggestion of having a quad that can tripod out and then use the remaining leg/arm. it cannot carry anything in that manipulator. no biggy but then it always come back to what you want the robot to do.
to see a quad maneuver into a tripod and use its other leg would be very cool. you could always incorporate a inventory bay on its back for carrying objects.
I guess you could but it would require segway like balancing to achieve a stable tripod walk. An alternative approach (and perhaps a cheat really) is to design some complex hybrid leg that can not only act like an arm, but also have a deployable wheel to allow the robot to drive whilst carrying something. I’ve been trying to think about how this could be achieved but I fear it may be too complex, heavy and expensive to be practical. Essentially what I was thinking was a wheel system like a mars rover, whereby it can turn each wheel on the spot.
So as you say, the design of the robot is entirely dependant on what you want it to do.
I got the picture from wikipedia on an article for the Bioloid Robotis kit. Every time I try to attach it a message comes up that says “the message you tried to attach is invalid”.