Hi! I’ve been working on this project for over 2 years now. It’s what got me into 3D printing, modeling, and mechanical engineering. I’m working on developing a motorized “shoe” or platform, where one shoe is strapped to each foot. The motorized shoes keep you in the same spot as you walk, like a treadmill. Unlike a treadmill I want to be able to walk or strafe in 4 directions (forward, backward, left, right). The user would wear a safety harness while using the shoes.
I’ve gone through many different designs. I’m a programmer and have experience with electronics, but mechanical engineering is something I didn’t have much experience on. It’s been awesome learning so many new things. The current design I show in the below clip.
You can see that on the ends of the platform there are two sets of wheels for moving forward and backward, and in the middle there are 3 sets of wheels for moving left and right. The boxes on the top contain the motors and some gearing. I’m using powerful brushless motors you’d normally use for electric skateboards and VESCs to control them. The VESCs are controlled using an Arduino. Not shown in the clip, but I’ve programmed the Arduino to connect to my phone over wifi so I can easily do testing. In the below clip the shoe is moving me, supporting my entire body weight (about 160 pounds), and I’m controlling it using my phone in my hand.
It’s not completely smooth yet, I think because the chains aren’t properly tensioned. They vibrate a lot. I’m working out the kinks right now.
The design doesn’t use traditional wheels. It uses what I call omni spheres. I tried using omni directional wheels in a previous design, but they didn’t work out for a number of reasons. If anyone wants to know the reasons I’d be happy to go over them, but the main issue was that I need the omni directional wheels to be small to fit under the platform, and the omni wheels I designed all ended up being very bumpy as they rotated. The omni sphere design addresses the problems I ran into. Here is a closer look at an omni sphere.
It has rods going through it on the sides (picture shows only two, but I ended up using four). If you rotate the rods, the sphere rotates. So hook the rods up to a part that is turned by the motor and the sphere turns. The middle of the sphere is a roller. So when the sphere is orientated correctly the roller can turn in the direction perpendicular to the motor powered direction. So it’s omni directional in 4 directions as long as it’s oriented correctly. To know the orientation of the wheel I’m going to add a magnet that rotates with the sphere and use a linear hall effect sensor to know when the magnet passes by. So it functions basically as an encoder.
The user doesn’t actually lift the platform up whenever he takes a step forward. The shoes aren’t exactly light, and it would be extremely awkward to lift the platform with the motors being off to the side. My solution is shown in the video clip below.
So my foot is strapped into a sort of binding and the binding is coupled to the shoe via the two rods. There are bearings behind the binding that allows me to rotate my foot forward and back, and the linear bearings slide along the rods so I can lift my foot up and down.
That’s basically how my current design works. Right now I’m working through some kinks and updating the design to make it easier to put the thing together. Once I’m happy with one shoe I’ll make another and do more testing. If you guys are interested in hearing about the algorithm I’m going to use to automatically control the shoes let me know. The algorithm will use the output from the hall effect sensors and accelerometers. If you guys have any feedback for me I’d love to hear it.
This design is open source. I’m going to be uploading the design file to github later this weekend. If anyone wants to collaborate please let me know. If people find this post interesting I’ll keep posting weekly updates. I also have a YouTube channel with a build log series for this project (link to the channel is in my profile).