The walker challenge at Robogames is less than a month away. I have seen many cool robots on this forum that could beat my sorry quadruped. Come race!
This is interesting. Where is this happening?
I think most of here are more incline towards multi-leg locomotion, precision and elegance of the walk, instead of speed. 
Perhaps the open category is more suitable. I have seen some very fast hexapod video in youtube. There is another similar event, but smaller, in Singapore. They don’t have as many event and competition as this though.
By the way, quad and hex are competiting in the same category? Apparently hex will overpower quad 
I think the Robogames where in San Francisco, CA.
Tnay,
If you have a graceful or interesting robot there may be a competition that suits you better.
My robot is not graceful. But I think it can go across my lawn- not sure yet. My goal is a walker that can go up and down curbs, but I am not there yet. Lynxmotion has been a terrific resources for mechanical parts.
Right now I am the only person registered for the walking race and my robot is not fast at all. I would rather lose a race then win by default. I think there are a lot of people on this forum-- including some in California-- who could whip up a better walker in a few weeks. That was my hope.
This would make a fast walker with HS-645 servos. 8)
That is a beautiful robot! Well designed and with solid parts. Probably clears 5 cm with ease. Someone should race one like that!
For this walker contest, are there walls to track when walking down the arena, or … how do bots navigate? I note that it says autonomous. I’m wondering what sort of sensory would be required. Most of the bots appear to have sonar arrays.
-Fergs
My understanding of the rules is that you just need to walk in a straight line for two meters across uncertain terrain, including 5 cm obstacles. There is also a known set of 4 cm steps at the end. There are no walls.
In my experience it is harder to keep a walker going in a straight line on uneven terrain then it is to avoid walls. One of my projects has been to get a quad to walk around my house. You can see some video at trailofdestruction.com. It can be done.
My plan is to use a digital compass on the walker to vary the gait and try to keep it on a straight line. I have been able to do this with my old walker, but I don’t know if it will work with the new one because it has fewer joints and hence less to work with to make small changes in its’ gait
Two interesting developments: I am trying to add a tail, because it is proving very hard to get ground clearance with the back legs. I have never tried a tail before and I think it is not something you see a lot in hobby robots. Second, I did try my quad across the lawn today. It worked fine. If nothing else it is cool to have a walker that can go around in the outdoors.
Jim, have you experimented with any other manufacturer’s servos?
I wonder if a faster servo would truly speed things up, or if the bottleneck of the lag is actually in the IK calculation and servo PWM outputting…
The servo I wanted to experiment with are the same servos that I use in my RC application:
ACE Hobby or Team Associated (same company) DS1015.
It has 200+ oz/in of torque, and has a transit time of about 0.1 / 60 degrees.
This thing is faaaaaaaaaast and strong, which is why I use them in my 1/10th 4wd buggy…
Would love to see a hexy with these servos scurry across the floor…
Unfortunately, they are roughly $90/each, 18 of them would be almost $1800…
I’ve experimented with futaba before but not lately. Hitec has fast servos too. Xan has shown how fast the Phoenix can be. It do not believe the limit is with the IK, but with the servos themselves. Note, the hexes described in this thread are 2DOF and have no IK involved with the gait. In fact the SSC-32 has a 12 servo gait engine built right into it. Hope this helps
I am no expert on these matters but I do not believe speed of the servos is the limiting factor. The lifting servo for all these robots are geared down for power, either with a mechanical lever or some sort of extra gears. That slows the leg down relative to the speed of the servo.
A stronger servo needs less of a mechanical lever to lift, and therefore lifts faster. Also, the more strength you have relative to weight the more push off you have for your servo to operate closer to its no load speed.
Their can also be a bottleneck in the sense that different gaits work in different ranges of speed and you can’t just speed them up and assume they will work the same only faster. This is even more true for quads then hexes, which have an aspect of instability.
Also, as I understand it inverse kinematics is the process of knowing where you want a foot to be and adjusting the joints in the gait to adjust for that target. People use IK in video games some times to keep the feet of animated characters on the ground as they walk on uneven terrain.
A dynamic gait is one that can change speed, or properties such as leg lift or stride length smoothly on the fly. I think the dynamic gait is more important for keeping the robot in a straight line and getting it to go faster in this limited case.
I only use Hitec servos and the Lynxmotion parts are the ones that work the best.
Fast and strong servo is not the primary key to successful walking robot. There are some really fast hexapod in youtube, but not that efficient. There are noticeable slip on the legs when they walk.
Even with static gait, when it walks too fast (fast leg motion), the reciprocating motion of the leg will result in substantial amount of moment and vibration, cause slip on leg, disturb the body forward motion, and steer the robot away. One way to counter that is make sure there is sufficient friction force on the leg and able to resist the moment reaction.
To incease friction, there are only 2 ways, use a material with higher friction coeefficient (sand paper, anyone ), or add weight to the robot…
You are absolutely right about the slip. I think it would be really cool to design better feet to get better traction.
Hi George, My reply was to Tom’s post where he was talking about 3DOF legs not 2 DOF legs. Your comments are true, they just don’t pertain to his question and my reply.
Static gate is one that can be stopped at any time and not cause a fall.
Dynamic gate is one that if stopped at the wrong point will cause a nasty fall.
Hexapods are usually static as there are always at least 3 legs on the ground. Stable eh?
Bipeds gaits can be both depending on speed, but are usually always Dynamic. Stopping the gait in mid stream prevents the “up” leg from catching the bot as the momentum of the bot is pushing forward.
Quads can be both as well.
Much appreciated! 8)
My brat is static. With a low CG and large feet, it does very well. 
I get you-- I misunderstood.
Thank you for the explanation. I’m glad I prefaced my remarks by admitting I don’t know much… : ) It’s great to be able to talk to people who can explain these things.
Tonight for the first time-- with 9 days to go-- I finally cleared 2" with my back legs. I always have a hard time getting ground clearance with my back legs. I think I can go higher 4", maybe 6"? But 5 cm (2") is the key for the contest.
An odd thing about the robot is that it is 2 1/2 DOF. That may not be the right term, but there are 10 motors for four legs. Originally I had six motors for four legs. This worked well except that I had trouble getting good ground clearance with the back legs. I have seen so many better robot designs on this board that it is kind of embarrassing. But it is also kind of fun to stumble forward without a plan to follow.
wtb pics
After a lot of experimenting with feet, I find the feet that work best seem to have an approximate mean time between failure of 90 seconds. I am using little boards where before I was using broom handles.
It seems like the lynxmotion biped feet will work well and be more sturdy. I’m a little worried about traction because metal can be smooth.
Clearing 2" obstacles:
youtube.com/watch?v=da1tcMhW8IU
This is the path where I would get stuck a lot with six motors:
youtube.com/watch?v=AqeoA0U9rqc
I don’t know if this proves I am crazy or not crazy.
That is maybe the most unique quad I have ever seen. I never would have even thought to try it that way. 
It works surprisingly well!