Close

March 2023 Update

finallyfunctionalfinallyfunctional wrote 03/11/2023 at 01:31 • 11 min read • Like

In this update I'll be going over several devices I've been experimenting with over the last several months. Then I'll talk about next steps.

Leg Exoskeleton

In my last update I said I was trying to focus on making a new VR locomotion setup using a leg exoskeleton design. Well, that isn't going as planned. I made a few designs to try out. These differ from the first exoskeleton design I tried out a while ago. I wanted a design that would run along the user's leg so that supporting sitting would be easier and the user could cross their feet without the exoskeleton on one leg colliding with the other leg.

Here are a few of the designs I made.

With this design I wanted to make something that was easy to adjust to curve around anyone's butt. Each of the individual links can be rotated to make different curves.

This design was less customizable but had a simpler design. The lengths could still be adjusted by using bolts of different lengths.

The 3D printed parts in both designs are made out of TPU.

With both designs the idea was there would be some flexibility to move back and forth as the user's hips swing, but still be rigid enough to keep the user in place.

Both designs felt awkward to strap my thigh to. It didn't really provide much resistance to me bending or moving my leg around like I'd hoped. Sitting would cause the strap around my thigh to tighten very uncomfortably.

In the last post I listed a series of benefits a leg exoskeleton could provide. Now I'm not so sure about several of them, because of the testing I did with these designs as well as some other testing with other devices I will explain later in this post.

I think I'm going to pause creating a leg exoskeleton, and if I start again it's going to be something like my very first design, which doesn't run tightly along the user's waist, butt, and legs.

A Simple Slidemill

I wanted to see if I could make a slippery surface like a slidemill has, but avoid some of the issues slidemills run into. Slidemills suffer from feeling like you're running on ice and lack stability as you run. I wondered if I could use a slippery surface because it could simplify the design of a device, but also have the device have a way to make running feel more natural and stable. Oh, and hopefully make something not as loud as slidemills are.

My thought was that if the device could always be trying to keep the user's foot centered better than the bowled surface of a slidemill, it would provide stability. As the user brings their foot forward, there will be some resistance they push through, but once they put their foot back down on the platform it won't slip because the device will help the user pull their foot back. The devices I was thinking of were the leg exoskeleton and a device I'll talk about in the next section.

Here is the slippery surface I made.

The materials

To make a slippery surface, I tried out a variety of surface materials and lubricants. The surface materials were acrylic, a few types of wood, HDPE, UHMW, and a metal sheet. The lubricants were silicon spray, heavy grease, a few types of furniture polish, sex lube, vegetable oil (liquid and spray), and dish soap with and without water.

The acrylic sheet seemed to be the best surface type, followed by the HDPE. With acrylic I was not only able to make a slipperier surface than the rest, but it was also quiet. 

The best lubricants were the lube, dish soap with no water added, and a furniture polish spray. The lube was the best, but dish soap was almost just as good. Since dish soap is the most widely available and cheap, I did further testing with it.

I have done a few test sessions where I walk on an acrylic sheet covered in dish soap for up to 15 minutes. The soap didn't dry out in that time. In further testing I'm going to see if while playing it needs to be sprayed with a little bit of water every now and again. Between play sessions, I'm wondering if I can just leave the soap on the platform, let it dry out, then the next time I play just take a few seconds to spray it with a little water. 

I was able to get in and out of my harness easily without slipping around. And it wasn't very messy. It just took a few seconds to wipe the bottom of my shoes or feet with a towel. 

I talked about the flooring material and lubricant I used, but haven't mentioned what shoe materials worked the best. 

I tried out a bunch of my pairs of shoes. I found the pair that worked the best was an older pair where the bottoms were smoothed out in several places from years of wearing them. These are the shoes I used in the video.

I tried walking barefoot, and this worked just as well as my old pair of shoes.

I tried out a TPU slider I printed. This slips well, not as well as my old shoes or barefoot, was still pretty slippery.

After learning walking barefoot works well, I searched online for materials that are similar to skin (hope I wasn't put on a list by searching that) and found that tattoo artists actually practice on silicone pads because they're like skin. I ordered some silicone pads and they work pretty well too.

So in summary, what worked the best was acrylic as the floor material, lube or dish soap as the lubricant, and worn shoes, barefoot, silicone, or TPU as the shoe material.

My experience using the slidemill

It does not feel nearly as good as using my regular VR shoes. The slidemill felt much more unstable than my VR shoes. The slidemill was much more tiring to use and my ankles felt like they got a workout after using it for not even 15 minutes. It was still fun to play around with. It's neat being able to walk with this simple setup. It was fun trying to do some strafing even though it was awkward. And this slidemill was much quieter than I remember the Virtuix Omni being and I think seems quieter than slidemills I've heard in videos.

A custom shoe slider may make it better

I'm thinking I'm going to make a VR shoe/slider hybrid. It will basically be a slider that will have wheels on the bottom. My thought is that the wheels will guide the direction the foot goes in better and provide some stability, but the slider will still be able to slide sideways.

I'll make a guide

I think this slidemill setup could be cheaply made. A piece of wood to screw the acrylic to and some pipe and pipe fittings to make a baby walker style support rig. Wrap a towel or two around the waist ring. It might not require a 3D printer at all. This setup isn't what I prefer but I imagine others may like it because of how cheap it is and similar to consumer treadmills.

String/Rubber bands walker

In this section I'll be talking about a device that is very similar to the University of Tsukuba's stringwalker.

I was wondering if a device like this could be made that utilizes the slippery floor I talked about in the prior section, and that could be passive instead of motorized. Instead of strings, could elastic bands be used? I created a design that would allow me to use rubber bands, tension them, have them automatically wind up, adjust to different stride widths, and be able to turn.

After I finished the design, I realized I didn't need to build it to test out the basic concept. I realized i could use the ring left over from my free standing support rig to do a quick, hacky test.

Here is that test. I'm not strapped into my ceiling mounted harness.

I first show walking backwards. At 0:47 I try walking forwards but it doesn't work very well. I wondered why walking backwards worked way better than forwards, so I turned around at 1:23 and walked forward. This time it worked better.

One thing you may be thinking is that the bands don't allow you to lift your feet very high. To resolve that I would create a shoe slider that uses the mechanism I show here where the slider would stay on the ground but the user can still lift their foot.

I've been talking about devices that have a mechanism the user pushes against to move their feet away from the center, then the device helps them bring them back. After doing this test, I have doubts that idea is as good as I thought it would be. The bands here didn't provide as much stability as I was hoping. I still had to more focus and use energy to make sure I don't fall over than I would've liked. If I added more bands so that they could pull my feet to the center with more force, to provide more stability, it would feel more and more like walking through thick sand or mud. I was hoping there could be a nice balance between stability and the force the device exerts, but that doesn't really seem to be the case. So I doubt this "bandwalker" device, and the leg exoskeleton with springy joints, would provide a locomotion experience I would be happy with.

All that being said, this "bandwalker" device might not be terrible as an alternative to a slidemill if I could work out the kinks. I could keep the force the bands exert on the user low at the expense of stability, so that you have about as much stability as a slidemill provides. With this device a rigid safety structure might not be needed. Instead you could use a loose safety structure, like a vest with a rope going to a swivel bolted to the ceiling or a pipe overhead. Additional bands that aren't attached to sliders that are positioned on the outside of each foot may make strafing easier than on a slidemill.

I haven't decided yet if I'm going to continue with this "bandwalker" idea.

I tried making another omniwheel

There are many design considerations to take into account when trying to use omniwheels in a VR shoe. They need to be small, handle a high load capacity, and be quiet or be able to be used on a noise dampening surface. I attempted to make another omniwheel that I hoped could meet these requirements.

These omniwheels are 1.5in in diameter. I had each of the 3 rows connected via a geartrain with the idea being that the rollers in contact with the ground at any given time would be evenly distributed over the area the wheels take up. Each roller uses 3mm roller bearings so they should be able to handle a person's weight. 

These wheels ended up being very bumpy and could not handle an uneven weight distribution.

The Omnisphere

I described the omnisphere in this post, which is just another omnidirectional wheel but in a spherical shape. I made a quick VR shoe design to try this wheel out and was disappointed with the results.

Two of the spheres broke during my testing. Under my weight, the wheels were very bumpy when I tried to move my foot side to side. Forward and backward movement had a lot of friction.

Calf-mounted motorized caster

Previously in this post I described a concept for a motorized caster mounted to a VR shoe that runs up along the backside of the user's calf.

I made a quick design, using parts I already had. This design is large and if I continue with it I will make it smaller.

The linear bearings did not slide up and down along the rods very nicely when lifting my foot. The caster itself did not swivel very well when I changed the direction my foot was going. I'm probably not going to continue with this design, but in the future I may mount electrical components to a frame running along the user's calf so that they do not need to take up space in the shoe.

Next Steps

I'm about ready to stop trying to make omniwheels and just stick with regular wheels, but I have one more attempt in me. I'm going to see if I can design a small mecanum wheel. 

I'm going to try to make a cheap slidemill for the reason's I listed above, but I probably won't use it myself very much as I prefer my VR shoes. I'll also make the hybrid VR shoe/slider. I may end up using the slidemill more if the hybrid VR shoe/slider turns out to work better than I expect it to.

I still plan on making a new freestanding VR support rig that I think will work better than the last one.

I have some minor improvements to make to my VR shoes, like curving a portion of the bottom under the heel like I curve a portion under the toes. I'll start looking into adding one or two small motors to the design, or instead adding tracking. I'm not sure yet.

Further out, if I have a motorized VR shoe that isn't omnidirectional but can turn or spin, I may start looking into using a technique like redirected walking to gradually correct for any drift from the origin as the user walks, and maybe even possibly support some amount of strafing. We recently discussed some ideas on how to do this on my discord.

In summary, in the near term, I'll be making a slidemill, sliders for that slidemill, a new freestanding rig, a mecanum wheel, and minor improvements to my VR shoes. Further out is tracking, motorizing my VR shoe, and methods, similar to redirected walking, to correct for drift and support strafing. I'm not planning on trying to mess around with entirely new devices anymore.

Like

Discussions