spherical robot climbing out of a hole

[mcpriess]

I'm a bit confused by your description- does your robot produce torque by reacting against gyroscopes which are allowed to precess (like in a control moment gyro), or does it produce torque by braking flywheels which you have spun up? Either way, mad props for overcoming whats been a very fundamental limitation of spherical robots. Please tell me you're going to publish on this!

[highgearengineer]

@mcpriess It uses the flywheels like control moment gyroscopes to produce torque. The difficulty is now controlling the dynamics of the sphere with the added gyroscopic effects.

[mcpriess]

@highgearengineer So are the gyros only capable of providing high reaction torque for a limited time then? I would assume that as you reacted against one of them, it would precess until it's flywheel axis was parallel to the axis of the applied torque.

[highgearengineer]

@mcpriess That is correct. Unfortunately, there is no way to get around the limits of the Conservation of Angular Momentum.

[mcpriess]

@highgearengineer Our lab has been doing some work on a novel gyro-stabilized platform, and as I'm sure you've found out, 3D rigid body dynamics involving gyros is no fun at all. I look forward to seeing where you go with your new prototype!

[basheersubei]

more videos please!!! Very nice!

[highgearengineer]

This machine is just a proof of concept prototype. It doesnt yet implement all of the features I described below. It is entirely remote controlled at the moment. (I only have two thumbs so its hard to control). No sensors or microcontroller. It uses RC car motors and ESCs with a helicopter radio.  I machined all of the internal mechanical parts. I'm building a new prototype with all the sensors and electronics and 5-10x the performance.

[Lasrin]

Sweet! Schematics? Diagrams? Homebrew or off-the-shelf circuits? Where's the pron? There are techblogs that are ready to spam this. I saw the article in Pop.Mech. Nice!

[wdfowty]

that's pretty sweet. you need to add jesus mode and have it roll on water!

[Lasrin]

Very nice! What about turning?

[highgearengineer]

a combination of shifting mass inside (moving a pendulum) and manipulating flywheels allows the robot to rotate about all three (x-y-z) axes.

[wer6399]

@highgearengineer big words make my brain hurt. . .

[Excalib12e]

@highgearengineer

How does it rotate about the Z-axis or Y-axis..I take that to mean spinning on the spot? It seems that it would support full 360 degree rotation in only one axis. In this case X

[highgearengineer]

@Excalib12e It is not a truly omnidirectional sphere in that it cannot rotate continuously about any arbitrary axis. It does have a primary rolling axis (X as you said). The ball can tilt side to side about the Y axis using the pendulum inside.  Using a "simple" configuration of control moment gyroscopes, additional torque can be generated about the X-axis, but it can also generate torque about the Z-axis to spin in place. A more complicated CMG mechanism can generate torque about Y as well.