>>27069 Have to post in this thread, because I discovered a video of a Japanese bipedal robot that has toes :D At the tail-end of the vid, of course: https://www.youtube.com/watch?v=QHjd-O8S0nI

Potential method for developing legs https://www.youtube.com/watch?v=i1xM61WBOBQ

>>33868 Thanks, Mechnomancer! Hope you're having a relaxing break rn. >toes I think many robowaifu designers may overlook how important the the ankle/heel/ball/toe flexion-complex is to good ambulation for bipeds. Essential. Cheers, Anon. :^) >>33876 Antie FTW! I loved that part of the story too. :^) >line-loaded, camshaft-driven That's actually a really interesting concept, Anon. Thanks for pointing it out. I especially like the fact the guy realizes how important it is to keep the 'outboard' weight down, and focus the mass inside the bot's torso. Cheers, Anon! :^)

>>33898 >relaxing break rn Not really, burned out a power converter in the non-waifu/mek project and faffing around with chonky atv tanktreads. > many robowaifu designers may overlook how important the the ankle/heel/ball/toe flexion-complex is to good ambulation for bipeds The flexing of the toes helps form the equivalent of the femisampiens leg parallelogram than allows the robot to push forward without being flatfooted.

>>33902 >Not really, burned out a power converter in the non-waifu/mek project Oh no! :/ >and faffing around with chonky atv tanktreads. Sounds pretty cool tbh. >The flexing of the toes helps form the equivalent of the femisampiens leg parallelogram than allows the robot to push forward without being flatfooted. That's a cool way to think about it, Anon. I'd love to eventually offer a Steampunk Sadie robowaifu model/kit, that looked full-on authentic OG 1800's retro. :D

>>33915 >Sounds pretty cool tbh. Yes and no. Yes in that I now have what looks like a miniature nasa shuttle mover. No in that I need to slap a buncha 3d printed parts on 'em to make the treads stay on their guide wheels (the yellow and orange bits). Without my mods probably would've been crappy to have on an ATV. Also need proper wiring techniques otherwise the splices heat up after a minute or so of driving lol. But figuring out how to digitally control a 2000lb motor is nice (never mind controlling 4 of them), because I can probably look into reciprocating walking mechanisms for lighter robots, like an owo version of zaphod daederik's steam powered man, or a spider-like platform for SPUD to ride around on.

>>33962 Dang, that looks too cool, Mechnomancer. Be careful on that thing, lol. >muh le 2000lb motor Lolwut? >or a spider-like platform for SPUD to ride around on. I'm going with the pony motif for dear robowaifu Sumomo-chan rn. It'll be her 'hoers', and she'll be it's rider. But even the pone will be able to climb up with her onto a wheeled zipper platform for scootin' around fast (for example, going with Master to the park for picnics).

>>33969 >muh le 2000lb motor >Lolwut? A 2000lb winch motor, each motor pulls 15A for a total of 180 watts.

>>33972 Oh, OK got it. Heh, I thought you were saying the motor weighed a ton. :^)

Bipedal character design. https://www.youtube.com/watch?v=7_LW7u-nk6Q

>>33988 related: https://www.youtube.com/watch?v=_0F3l-skLos

Thanks for all the YouTube links, but please add some some description when posting. Ideally more than one line. These videos have a text description on YouTube which can be copied. With ">" you can quote it, though make sure to put this in front of each paragraph when doing so. Cheers.

>>44139 (Moved to more on-topic thread) Consider this: freeform roller skating is hard for humans that can walk naturally without effort, and walking naturally without effort is hard for robots. Motorized feet seem much easier https://youtu.be/tEuMF_tb1nI Motorized shoes https://youtu.be/MqoGB6TTcGk

>>44140 I believe that "roller skate" wheels are a great idea Take Galatea for example. Gen 1 robowaifu, no active balancing. Balancing is achieved through weight distribution and the width of the base. Gen 2 is the roller skate feet, a midpoint between Galatea and full-on walking. She'll wheel around like Galatea, but with active balancing, and she can go up and down stairs. If she moves like a grandma, well it's only when stepping up and down. Imagine a hoverboard, but each foot is a hoverboard, and the robowaifu doesn't lean to go forward or backwards, she just does. Plus, the motors can add resistance to the wheels, so it's like trying to balance on roller skates, but on carpet instead of smooth floors.

>>44137 >Holowaifus are rare enough as it is compared to robowaifus without having them skate too. Not sure I understand this perfectly, Anon. Can you restate this please? >I also previously asked Mechnomancer about tendon-actuated spider walkers. Or maybe I just thought I did, I don't remember. The topic of spooders has been a longstanding one here on /robowaifu/ . >But I wondered if that would be easier than embedding motors in the leg joints. My most-basic opus here is the singular imperative: REDUCE MASS Did I also mention: REDUCE THROWN WEIGHT :) --->Why is keeping mass (weight) low so important? ( >>4313 ) --- The basic corollary agenda all of us should also be pursuing IMHO is just this: move as much mass as possible into the lower thorax volume (ie, the pelvis). This will assist every.single. other aspect of robowaifu kinematics, essentially. >tl;dr So yeah, locating actuators inside the torso and transferring actuation forces out to the limbs by some means (likely via cables or rotating rods) is both a sound idea, and likely also an essential design approach for success during these early robowaifu years.

>>44138 >If you motorize the skates, it can definitely be a good compromise for walking, while still being able to go up stairs and having active balancing. Yes, the little cars are motorized, and I definitely plan to capitalize on that fact.

>>44139 >Is it better to motorize the skates or to have their wheels rotate freely and be powered by the rest of the bot's movements, like a human wearing skates? Ideally, we'd have some kind of clutch mechanism to allow us to put the drive wheels in 'neutral', much like a manual transmission in an automobile I'm still shocked some men don't know how to drive one lol. However, these simple kits have direct-drive straight from the axle of a geared, brushed DC motor. Cheap and simple, but no clutch. This will be redesigned once we move on to pro-skater gear: then we can free wheel as well.

>>44140 >Consider this: freeform roller skating is hard for humans that can walk naturally without effort, and walking naturally without effort is hard for robots. Honestly, having inlines on and gottagofast is much easier physically than running is. Very smooth motions involved relatively. Sure you have to balance yourself, but thats true of standing and walking as well. Getting used to the extra dynamics of easy motions (ie, rolling along) is the hard part. Once you're over that little hurdle, then the rest is a snap. >Motorized feet seem much easier Yes, I think so too. Even using Heelys is easier than walking if you're on a very smooth surface.

>>44142 POTD I like where you're going with this, Anon! >If she moves like a grandma, well it's only when stepping up and down. Haha :D >Imagine a hoverboard, but each foot is a hoverboard Yes that's it you've hit on it fren GreerTech. <---> Cheers, Anons. :)

>>44147 >Not sure I understand this perfectly, Anon. Can you restate this please? Just that there are many fewer examples of holowaifus in sci-fi than robowaifus. That means there are fewer rules established for how they have to behave.

>>44151 Maybe, but a baby can walk, and most able-bodied humans can walk with no effort or difficulty, but they still have to learn and practice how to balance on skates. Walking has the benefit of friction. Idk, I'm not completely certain Dammit Chobitsu, I'm a robowaifuist, not a physical therapist! >>44155 Arguably Cortana is more popular in the West than a (positive) robowaifu character.

>>44156 Bones GreerTech, what have I done!? :DD

OK, so I decided to do a sketch of what I think would be a cheap and doable humanoid robot leg that can both walk and drive; it may be applicable to other robot designs too. Just pretend this is drawn much better than it actually is. The waifu is supposed to have a curvy, sleek look; think KOS-MOS from Xenosaga, but pretend her thigh-highs/lower legs have been replaced with this leg design. In rolly mode the robot gets around on large wheels that are built into the foot structure, instead of having separate wheels like roller skates; the robot's upper body also shifts slightly forward, while the angle of the legs to the ground becomes shallower, going from near 90 degrees to more like 75-80 degrees. Big off-road wheels such as Segway wheels are better for crushing through terrain than small roller skate wheels that extend off of a shoe, and when you're a robot, you don't have to have a completely human-like foot structure, so you can have a giant wheel built into your foot. It also has 2-3 smaller, unpowered "training wheels" as toes in front of the large main wheel to help balance; these can be roller skate wheels. The leg uses reverse leg joints for greater stability, similar to some Paralympic sprinters, but when walking, the mass shifter and pulley system are used. I also included this random Metal Gear fan art I found so I could use it to illustrate a very important feature, the U-shaped bars on the back of the feet. This is how you enable it to both walk and drive. Most robot leg designs that use wheels on the feet either have no alternate mode or they have the wheels retract like airplane landing gear when the robot needs to walk, but I think the U-shaped bar might be a better solution. The bar goes out past the back of the big wheel, enclosing it; the robot puts the U-shaped bar down to contact the ground to walk and raises it up to let the wheels touch the ground in order to drive. You could also put springs somewhere on the leg, maybe the thigh, to serve as an energy return system and save more battery power when walking.

>>44163 POTD I like it, HoloAnon. You've thought through issues and suggested solutions. The sketch is a nice touch as well. Do you think you will focus some time on fleshing out this plan in more detail? Good luck.

>>44164 Well, I thought this would be helpful for somebody else like you who's actually planning to build a robot. I don't think I have the facilities to do it myself, at least not yet, and I'm still focused on trying to do a holowaifu; this is just something I randomly came up with. This is somewhat similar in principle to some ideas that have been floated by others regarding robowaifus using wheelchairs or walking sticks to get around because they have trouble walking normally; it's using a robot leg design based on a type of prosthetic used by people who've lost their legs. Since you wanted to do a skating robot, I decided to produce a design concept based on that. It's also partly motivated by my interest in mecha, because many notable mech designs like the AT-ST use reverse joint legs.

>>44165 Thanks! I appreciate the gesture, HoloAnon. >reverse legs I don't consider this a viable choice for robowaifus -- too inhuman. Ofc, there are in fact good technical reasons for such a design approach, so it would be a reasonable choice for ancillary companions such as /monster/ might crave. >bigger wheels; walking rails Both good choices. I admit my primary mobility goal for these early designs rn is simply getting around Anon's flat; comfortably for all involved. But for picnics up on mountain tops, a variety of interchangeable robowaifu feet would be in order I think -- including Segway-like ones. Good ideas, Anon. Cheers. :)

>>44168 I have to ask what you were planning to do for the walking function. Were you going to use retractable landing gear wheels, or just put them in park and have her walk on the wheels?

>>44171 These are cheap & straightforward 1:48 gearing, TT brushed motors with smol rubber wheels directly attached to the output shaft. Examples include: https://www.amazon.com/Motor-Leads-Gearbox-Shaft-200RPM/dp/B0D8H89XDY https://www.amazon.com/AEDIKO-Motor-Gearbox-Shaft-200RPM/dp/B099Z85573 These are only slightly backdriveable, and generally want to stay in place when unpowered. Since I'm aiming for very low mass for these first prototypes, I'm hopeful that no additional braking will be required: just deactivate them, and the 4 tires of each foot should serve well enough as a stable platform (rember rn I'm just going for basic traversal over level surfaces within Anon's flat) for walking. We'll see as the design advances how long this simple arrangement remains effective! :)

>>44172 I guess that'll be a good place to start. Designing the bot to have swappable feet is a good idea too, so you can just swap out for the design I suggested whenever you find a situation where the default feet don't work well.

>>44173 That all makes sense. Thanks for the good advice HoloAnon, cheers! :)

Unitree roller skates It seems like we predicted the future https://youtu.be/srPz8TRpZ_8

>>44371 Very cool. Obviously this is 3D CGI, but the basic premise is sound. I predict not only our own robowaifus will have these as optional 'feet', but so will most Chinese robits as well. Thanks, GreerTech! :)

>>44372 Upon second glance, I think you're right

>>44373 You mean about it being CGI? Yeah it clearly is (they always screw the lighting up). >tl;dr This is a so-called 'Comp' edit (short for composite). Typically some sort of stand in moves around in an IRL environment, then motion-tracking is applied to the footage plates to derive the original practical camera motion. That tracking data is then applied against the 3D CGI assets (the virtual robit, in this case) to obtain a so-called 'Match Move'. In high-end work, this matchmove footage is further refined with: motion-blurs; (possibly) rotoscope work; lighting tweaks; color grading; and general mastering to produce final. It's a long & expensive hand done process by skilled 3D film technicians. Cheap work like this probably skips all those steps and it shows. <---> However the kinematics sim here is close to spot-on, so you can expect the real Unitree skatebots should perform basically in this manner once they're being manufactured. Its simply a matter of time.

>>44374 I was wondering how they got it so agile compared to the WhistlinDiesel video we saw (and other clumsy Unitree videos). I thought it was just really good programming and optimization.

>>44373 >>44375 >I thought it was just really good programming and optimization. Believe me, it will be. * But they still have a ways to go yet. Looking back though, it will seem like all these superhuman robits everywhere suddenly appeared 'over night'. :) --- * For example, confer this real footage of another company's (clearly focused on military application contracts) running robit. Its very real, and it's performance is already quite remarkable in 2026: https://trashchan.xyz/robowaifu/thread/26.html#1464 My guess is that the company will be asking ~US$1M for these things once they are mass-produced.

Good discussion of leg articulation/actuation models and design. https://news.asimov.inc/p/how-we-built-humanoid-legs-from-the

>>44163 I thought I'd bump this thread because I just found a much better-drawn picture of my concept here completely by chance. It doesn't have the toe-wheels on the front, but the basic idea is here.

>>44516 ...woah

>>44516 Having just two wheels is actually more efficient rolling-resistance-wise though (if more unstable). It strikes us as being more awkward if we visualize our own bodies in it's place (cf. the odd but efficient heelys), but a robowaifu need have no such feelings. Sorry but I'm spoilering your post. :D Cheers, HoloAnon. :)

>>44542 You really think she doesn't need the toe-wheels? I guess if you have swappable feet it's a non-issue because you can try it both with and without toe-wheels.

>>44544 No! I'm not saying that at all. If you're going full-tilt skate wheels, then 4spool inlines simply can't be beat. However, rn my mind is focused on maximal efficiency of batt life, so keeping the drag coefficient down to the minimal is a priority for me rn. As to your design ref ( >>44516 ), I can see pursuing this once the cart waifu project is rolling see what I did there? :) and I can use it as a mobile boom system (cf. common robit overhead gantrys atm) to begin dev'g bipedal cardboard waifus...I mean to have full 2x trucks + 4wheels ea. as the basic footgear for her.

>>44546 >cardboard Why cardboard? Shouldn't you pick something more durable?

>>44547 Well, A) I have a metric sh*tetonne of cardboard saved up now, after @Kiwi's ebin post on the topic, and B) I want to do fast IRL prototyping of leg forms that are basically throwaway essays as we find economical & sturdy approaches. I hope that makes sense, Anon. I'm sure once some pics from the process come out, then it will be more clear still. >tl;dr It's cheap & durable (as long as you keep it dry).

>>44549 I guess you could try to create some kind of composite with it to improve its durability. But if it's only for prototyping it might work, maybe even moreso for skating waifus because you're at higher risk of having them crash and damage themselves.

>>44550 Exactly. That's it you've got it, Anon. Cheers. :)