Bionic Arm With Muscle Emulation

Gugo writes "German based company FESTO has develop a bionic arm that uses muscle emulation,(video included) with a product called 'fluidic muscle.' It works like a normal animal-human muscle but moved by air inside. This new type of prosthetic offers rapid response, small size, simple assembly and ease of control. On their website they show the range of fluidic muscles with a car race simulator."

Festo pneumatic acuators for robotics - good stuff

[Animats]

Festo does good work. They're an industrial automation company, and they do demos like this for promotional purposes. Check out their videos on YouTube.

The innovation here is not "fluidic muscles". It's their piezoelectric proportional valves. It's been possible for years to do precision control of pneumatics. Twenty years ago, "Pneumatic Valves, Inc." in Palo Alto was doing control like that. But older proportional valves were big and expensive, with a voice coil actuator on the end of a spool valve. Festo has miniaturized the technology with their piezoelectric valves [patentstorm.us].

Pneumatic systems have traditionally been either force actuators or devices driven to a limit stop. Fine position control was the domain of hydraulics. This is changing. For pneumatic systems, if the valves can be brought close to the actuator, the valves are fast, position sensors are used, and the control system is well designed, the system becomes quite controllable. That's what Festo is demonstrating here.

You can also do some things with pneumatics you can't do well with electrical drive, such as create springs with variable spring constants. Muscles can be usefully modeled as spring-damper systems, where the spring constant, zero point, and damping constant are all controllable. This can be emulated with electrical actuators, but emulating a spring in software requires high-powered actuators and loses energy. Legged running work needs something like a variable spring, and pneumatics are currently the closest thing to muscles available.

Re:Now we're finally closer to the first...

[WFFS]

What, like this [tokyotimes.org]?

Re:Any idea why...

[Anonymous Coward]

This actually is a hydraulic device. There are a lot of reasons linear actuators aren't more commonly used in robotic applications: they lack the force-to-weight ratio to be effective when hooked up like a real muscle; they draw a lot of current; they're heavy; they have to be always energized. The fluidic/hydraulic or pneumatic systems allow you to displace the power generator from the power application, so you relieve the power-to-weight issue (at least for fixed installations-if you want a human to carry the hydraulic pump that runs the arm, then there are issues). The fluidic system has the further advantage of being naturally damped. Pneumatic systems are essentially springs in which the rest length of the spring is set by air pressure. This makes the very bouncy & somewhat hard to control. Either fluid or air powered systems rely on valves and flow to change lengths & forces, and this has historically been harder to control than a servo motor position.