Multiple rows of intricately articulated legs are the defining characteristic of the Strandbeest, but [James Bruton] wondered if he could reduce that down to a single row using the same principles at work in a self-balancing two wheeled robot. While it’s perhaps a bit early to call his experiments a complete success, the first tentative steps taken by his (relatively) svelte Strandbeest certainly look promising.
Initially the robot only had two pairs of legs, but in testing [James] found this arrangement to be a bit unstable. By bringing the total count to four legs per side and improving the counterweight arrangement, the bot has been able to walk the length of the workshop. Unfortunately, an issue with the leg design seems to be preventing the Strandbeest from taking any backward steps.
Normally this wouldn’t be that big of a problem, but in this case it’s keeping the Strandbeest from being able to self-balance while standing still. In other words, the robot needs to keep moving forward or it will fall over. Still, [James] thinks the idea has promise and wants to continue experimenting with the bot in a larger area.
Specifically, he wants to see if the dual-motor robot can turn by varying the speed the two sets of legs are running at. If it can walk in a tight enough circle, it could keep right on marching until the power runs down. Sounds more than a little nightmarish to us, but we’d still like to see it.
Reader’s may recall [James] from this other another robotic projects, such as the phenomenal OpenDog. We don’t know where his obsession of legged robots comes from, but we certainly aren’t complaining.
The first thing Jeremy Cook thought when he saw a video of Theo Jansen’s Strandbeest walking across the beach was how incredible the machine looked. His second thought was that there was no way he’d ever be able to build something like that himself. It’s a feeling that most of us have had at one time or another, especially when starting down a path we’ve never been on before.
But those doubts didn’t keep him from researching how the Strandbeest worked, or stop him from taking the first tentative steps towards building his own version. It certainly didn’t happen overnight. It didn’t happen over a month or even a year, either.
His first builds could barely move, and when they did, it wasn’t for long. But the latest version, which he demonstrated live in front of a packed audience at the LA College of Music, trotted across the stage with an almost otherworldly smoothness. To say that he’s gotten good at building these machines would be something of an understatement.
Jeremy’s talk is primarily focused on his Strandbeest creations, but it’s also a fascinating look at how a person can gradually move from inspiration to mastery through incremental improvements. He could have stopped after the first, second, or even third failure. But instead he persisted to the point he’s an expert at something he once believed was out of his reach.
20 years ago, PCB production was expensive and required a multitude of phone calls and emails to a fab with significant minimum order restrictions. Now, it’s cheap and accessible online, which in addition to curtailing the home etching market has created significant new possibilities for home projects. Now that flexible PCBs are also readily available, it’s possible to experiment with some cool concepts – and that’s precisely what [Carl] has been doing.
The aim is to build a walking robot that uses actuators made from flexible PCBs. The flexible PCB is printed with a coil, capable of generating a small magnetic field. This then interacts with a strong permanent magnet, causing the flexible PCB to move when energised.
Initial attempts with four actuators mounted to a 3D printed frame were unsuccessful, but [Carl] has persevered. With a focus on weight saving, the MK II prototype has shown some promise, gently twitching its way across a desk in testing. Future steps will involve building an untethered version. This will replace the 3D printed chassis with a standard fibreglass PCB acting as both control board and the main chassis to minimise weight, similar to PCB quadcopter designs we’ve seen in the past.
Walkers like the Strandbeest are favorites due in part to their smooth design and fluid motion, but [Leandro] is going a slightly different way with Octo, an octopodal platform for exploring rough terrain. Octo is based on the Klann linkage which was developed in 1994 and intended to act as an alternative to wheels because of its ability to deal with rough terrain. [Leandro] made a small proof of concept out of soldered brass and liked the results. The next version will be larger, made out of aluminum and steel, and capable of carrying a payload.
The Strandbeest and Octo have a lot in common but differ in a few significant ways. Jansen’s linkage (which the Strandbeest uses) uses eight links per leg and requires relatively flat terrain. The Klann linkage used by Octo needs only six links per leg, and has the ability to deal with rougher ground.
[Leandro] didn’t just cut some parts out from a file found online; the brass proof of concept was drawn up based on an animation of a Klann linkage. For the next version, [Leandro] used a simulator to determine an optimal linkage design, aiming for one with a gait that wasn’t too flat, and maximized vertical rise of the leg to aid in clearing obstacles.
We’ve seen the Klann linkage before in a LEGO Spider-bot. We’re delighted to see [Leandro]’s Octo in the ring for the Wheels, wings, and walkers category of The Hackaday Prize.