Air muscles vs compliant mechanisms

[Agatha_012]

Hello, I am currently working on developing a soft tentacle arm robot, that can work in close proximity to humans. I initially intended it to be entirely made out of silicone. However, it seems it's not possible to make a silicone robot both soft and strong at the same time.
So, I came up with 2 designs, but can't choose.
1. The tentacle is divided into 8-10 sections horizontally; each section having 4 air muscles, that will be individually controlled through air vessels and a solenoid valve mechanism.
2. The tentacle is again divided into 8-10 sections horizontally. However, this time, it will be controlled with a compliant mechanism mostly inspired by a compliant gripper. Here, I am planning to incorporate a solenoid valve mechanism into each section. Using electricity, I am planning to control the direction of the bending.
Would appreciate any thoughts and discussions on this.
Note: This is not the final design yet. Both versions need a lot of improvement. The question is, which one is worth investing more time in?

 

[Compositepro]

I'm not sure how your structure is supposed to work. I assume you are using air bladders with a braided reinforcement. In addition to muscles you require bones and joints to create a mechanism. Don't forget about torsional stability.

 

[rb1957]

muscles don't work in isolation ... they need a skeleton, some rigid components to carry compression loads.

another day in paradise, or is paradise one day closer ?

 

[Agatha_012]

I plan the main "skeleton" to be a slightly stiffer air vessel. They can bend into any shape without permanent deformation.

 

[MintJulep]

The question is, which one is worth investing more time in?

I'm sorry, but the direct and honest answer is neither.

The skeleton needs to carry loads and constrain motion to the desired degrees of freedom and provide a path to "ground" to react loads against.

Something that "can bend into any shape" does not provide any constraints and cannot react anything.

Your approach will simply result in uncontrolled and uncontrollable writhing.

I suspect that this is a student course project. Possibly with the goal of teaching how the process of design should work.

Start with quantifiable requirements. Then you can begin imagining designs to fulfill them.

Please take this to the student formum forum1630

 

[Agatha_012]

Got it, thanks ✓

 

[rb1957]

yeah, not seeing it.

something that can bend into any shape ...
I see a structure of rigid elements with "tendons" (cables running along the shape) to control the shape. Each rigid element (I see many short elements, but you could trial with 3 of 4) has a bearing at both ends and these can "somehow" (electronically controlled ?) be locked or moveable as needed. I see some "AI" controlling the end of the structure, controlling the shape.

another day in paradise, or is paradise one day closer ?

 

[BrianPetersen]

You buy a robot designed for collaborative operation, and then set it up properly to do what you want it to do. It already has the capability of speed, force, and power monitoring/limiting built in, you just need to set it up.

Example:

https://www.universal-robots.com/products/

 

[Agatha_012]

I'm sorry, but the direct and honest answer is neither.

The skeleton needs to carry loads and constrain motion to the desired degrees of freedom and provide a path to "ground" to react loads against.

Hmm, I was suspecting that could be the case but thought that if the mechanism is strong, then it will be able to hold itself without a need for a strong skeleton. Do you have any suggestions on what to research to find better methods?

 

[Agatha_012]

You buy a robot designed for collaborative operation, and then set it up properly to do what you want it to do. It already has the capability of speed, force, and power monitoring/limiting built-in, you just need to set it up.

But I am learning to design and innovate. I can not do that with an already designed and built robot..?

 

[EdStainless]

What kind of placement accuracy will you need?
A fully compliant system will likely repeat within a few inches.
The 'gentle' systems that I have seen still have a solid frame with joints under them, but the joints are designed so that they are not locked.
This gives you better control and your overall stiffness is set by the pressures that you use.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, consulting work welcomed

 

[Agatha_012]

What kind of placement accuracy will you need?

It needs to be able to take a ball and put it inside a box. That is the goal I set for a minimum viable product. So a decent level of accuracy and stability is needed.

The 'gentle' systems that I have seen still have a solid frame with joints under them, but the joints are designed so that they are not locked.

Could you give a sample of the systems you have seen? Perhaps a name? So that I could analyze them closer.