Hello,
I stumbled across the use of an inverted pendulum to model walking gaits. I thought I might play with this. Unfortunately, the links I have found either lacked the math or had mathematics that was over my head.
Can someone recommend a good "for beginners" page? Better still, can someone describe how I might use this to design a new walker (preferably quadrupedal) or to measure the performance of existing walkers?
Thanks in advance,
J Wolfgang Goerlich
Hi Wolfgang. Use of the inverted pendulum model is common in modeling walking vertebrates, while the spring-mass model is commonly used for modeling running vertebrates - esp bipeds, but also generally applicable to quadrupeds, etc. Robert Full has done a huge amount of work in this area, and his papers are not too math-laden. I have a bunch of links to his relevant papers on this page ....
- dan michaels ==================
Hi Wolfgang. Use of the inverted pendulum model is common in modeling walking vertebrates, while the spring-mass model is commonly used for modeling running vertebrates - esp bipeds, but also generally applicable to quadrupeds, etc. Robert Full has done a huge amount of work in this area, and his papers are not too math-laden. I have a bunch of links to his relevant papers on this page ....
- dan michaels ==================
Hello Dan,
Quite the web site you have there! Good information. For the moment, I cannot get to Full's site either.
I am wondering how to use IP or SLIP to design walkers. Thus far, I am building BEAM-style four-legged devices. The ol' "bend the legs until it walks straight" method works but makes the engineer in me uneasy.
Do you have any tips on how to apply IP or SLIP? How could one use it to determine the leg angle/length/design?
Thanks again,
J Wolfgang Goerlich
Hi Wolfgang. Full's site works sometimes and not others. Keep trying. His papers are all online. The one's referenced on my page are the ones I found most useful. Hope you find the info on my legs-pages of some interest. They seem to get a lot of hits.
The IP and S-M models are useful representations of real walkers, but of limited use for robotic walkers. This is because robots generally have nowheres near the #DOF's as real animals. Eg, the IP model actually decribes a freely-moving mass under the influence of gravity, whereas in servo/motor-driven robots, the legs are generally not freely-moving, but under motor power 100% of the time. So the movements are completely specified by the motor trajectories. Inertia and gravity don't play much of a role in this case.
Real animals get upwards to 70% energy savings by being able to store energy in the tendons, joints, and muscles during foot-fall, and recovering it during pushoff into the next step. Few robots - at least hobby-variety - ever do this. Also, the S-M model is probably even less applicable to robots than IP, since few actually ever run or have spring-loaded feet. However, there is always the next generation of robots which will incorporate these features. For now, I don't think you have to worry about either models in building a walker.
- dan michaels
Ok. Then, perhaps the following is a better question. Can you recommend a good model or design tool for designing quadrupeds?
J Wolfgang Goerlich
By far the majority of walking robots I've seen are hexapods, but bipeds have also gotten to be very popular lately. However, every design seems to be different, except for the few that are sold commercially. I know of no actual models or design tools specifically for this. Some people use general mechanical CAD tools for their layouts. The best I can say is to scan over the various designs in order to get ideas of how to do your own.
I have collected a lot of links to walkers on my pages, and there are links there to so-called "catalogs" of walking robots, etc. Check especially
If you have some money, the single best thing you can buy is a milling machine, since you can then make your own cool parts. However, I've noticed a lot of people producing what I consider to be "over-designed" walkers, which are very beautifully made, but with very heavy aluminum pieces.
The single biggest problem and concern with a walking robot is power-to-weight ratio. The thing has to be able to lift its own weight. The heavier the parts, the larger [and more costly] the motors/servos required, and the larger [and heavier] the batteries need to be. This is a much more difficult problem than for wheeled robots, since those don't have to "lift" the batteries. This is a similar problem to that present in electric model airplanes. They're cool and they fly great, but the batteries run down after 5 minutes or less. In addition, a lot of people building biped robots are using big powerful digital servos, and spending $500-700 just on servos alone.
- dan michaels ==================
PolyTech Forum website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.