And here's the walking machine we named "Sticky". Each of the three platforms contained a pair of articulated legs. Each leg slid up and down at the end of a rod I called the hip. The hips swivelled back and forth and each platform, as you can see, moved with respect to the other platforms.
This diagram shows how I used a cable to turn the walking machine. As one end wound up, the other end payed out, so the body could curl in either direction. For this, I used twine, but added springs to take up a bit of the slack. It worked pretty good, but the thing was laterally weak at the joints of the body segments. I eventually figured to mount the segments on pancake spinner joints, but that's a project for another time.
Here's how the feet worked, sliding up and down the legs. The legs were shafts fixed to the ends of the swivel hips. I used twine again to move the legs up and down in the shaft. I was a bit tricky setting this up without the cord interfering with the system, but eventually, I got it to work quite well.
The pair of legs on each hip were twined together in such a way that the same motor would send one to the top of its run while the other leg dropped to the bottom of its run. Stringing this twine around was really fun, especially discovering that I had to keep it from interfereing with twine moving to the other leg, but still secure it in such a way that it didn't slip.
One of the techniques I used to make a special widgit. In this case, it was a bolt that had a smooth, round shaft in the middle (the twine passed over this part of the shaft and I was afraid the threads would groove the twine to one side or another or, worse, wear the twine out). So I chucked the bolt and spun it while applying a file to a specific section of the threads. In a matter of minutes (okay, I went through a few bolts before I got the hang of it), I had my custom machined bolts. Way cheaper than buying a lathe!
The finished walking machine worked better than my previous project. I was able to cause the hips to swivel back and forth under computer control (did I mention getting the interface to work?!), but the motors controlling the feet simply weren't strong enough, so the feet didn't go up or down.
A next gen of this would probably use gearboxed DC motors and I would place limit switches at the extremities of the legs and hips. The gearboxes would provide the kind of force I'd need to lift the body clear of the ground.
Oh, an interesting footnote to this: As designed, the walker would work upside down, too. Yay!
