Scott Sells
Andrew Serff
Bas Hamer


1. Abstract

The Muscles project is a proof of concept for the construction of objects that can be used as muscles for automatons, robots, and other mechanical devices. Two artificial muscles, made from rubber tubing and metallic wire mesh, expand and contract in opposition to one another. These muscles are attached to a wooden arm with a single pivoting joint. The rubber tube of one muscle is filled with compressed air, causing the tube to expand, which causes the wire mesh surrounding the tube to contract. This contraction pulls on the pivot of the arm, which rotates around its joint. This process is similar to the way real muscles in a human elbow pull the forearm up and down. The inflation of the muscles is controlled by electronic valves which allow the passage of compressed air to fill the rubber tubes. These valves are in turn controlled by a computer running LabView, and can be turned on and off at will by the user, or can be set in a repeating automated control loop.

2. Design

The design can be broken down into three parts: the physical and computational mechanisms used by the project, the materials and construction techniques used to build the project, and the design process itself.

2.1 Mechanisms

The project consists of basically three mechanisms: the muscles, the computer-controlled valves, and the computer program that controls the valves.

2.1.1 Physical Mechanisms

The muscles are the main mechanisms of the project. They do all of the physical work to rotate the arm back and forth. As one muscle flexes and pulls the arm in one direction, the other muscle is in a relaxed state. Then the relaxed muscle flexes and the flexed muscle relaxes, and the arm is pulled in the opposite direction. The muscles provide enough force to rotate the arm even when the gravitational opposing force is at its strongest.

Each muscle is hooked up to plastic tubing that connects to the computer-controlled valves. This tubing allows the flow of compressed air into and out of the muscles. When compressed air is let in, the rubber bladder inside the muscle inflates, which causes the wire mesh surrounding it to contract. A steel cable attached to the muscle is wrapped around a disc at the joint of the arm. When the muscle contracts, the cable is pulled, which in turn rotates the disc, and therefore the arm.

The computer-controlled valves are all in a single small metal block. The block contains eight valves in total, and each can be controlled individually. There is a connector for each valve, all of which are hooked up to the computer. Each valve has two controllable states: open or closed. When a valve is open, air is allowed to pass through it. When it is closed, no air can pass through. Only two valves are used in this project, one for each muscle. The other end of the block of valves is connected by plastic tubing to an air compressor, which provides the compressed air. The compressor is activated for a few minutes, which compresses the air inside it to about 60 psi, and then is deactivated. This provides enough compressed air to power the arm for 5-10 minutes.

2.1.2 Computational Mechanisms

The only computational mechanism used in the project is a program called LabView which runs on a Windows 98 computer. LabView is used to create a virtual switch that can be controlled by a user. The switch has three states: A, B, and neutral. When the switch is in position A, the left valve is turned on (causing the left muscle to flex), and the right valve is turned off (relaxing the right muscle). When the switch is in position B, the left valve is turned off, and the right valve is turned on. When in neutral, both the left and the right valves are turned off, allowing both muscles to relax and the arm rests in its natural state. There is also an option to control the valves individually.

2.2 Construction

The project was constructed in stages, using a variety of materials.

2.2.1 Materials

Wood was used to construct the main arm of the project, as well as the base on which it stands. This wood was simply left over from other projects, and has holes and other marks that have nothing to do with this project. The disc at the joint of the arm is made from 1/8-inch basswood. Originally, the project team intended to replace the wood with aluminum, which would give the project a more aesthetic appeal, and would provide more durability.

The muscles are made of rubber and metal. The inner bladder is a cut piece of a rubber bicycle inner tube. Each end of the tube is sealed with a round LEGO piece and some cold welding glue, one of which has a plastic tube running through it, to provide the compressed air. The outer mesh is made from a length of wide steel cable pulled apart. Each end of the muscle is capped by two large washers, which clamp the steel mesh. Two hooks protrude from the washers at each end, which allow thin steel cables to be attached to the muscle.

2.2.2 Construction Techniques

The entire project was made with simple construction techniques. The wood for the arm and base was already cut, and only a few holes needed to be drilled to create the joint and attach the arm to the base. The disc in the arm joint was made from three 1/8-inch discs of basswood, cut by the laser cutter. One of the three discs is slightly smaller in diameter, which is sandwiched in the middle of the other two. This provides a nice groove for the steel cable to rest in. This disc was then attached to the arm's joint.

Constructing the muscles was a matter of cutting the inner tube and wire mesh to the correct length. Holes were drilled through the large washers for the hooks that would provide both the clamping of the mesh and the attaching of the steel cables. Each mesh was then capped by the large washers, and the hooks were screwed in. The tubes were sealed and placed inside the wire meshes. The muscles were then attached to the base of the arm, and the disc at the arm's joint.

2.3 Design Process

The design for the project was inspired by a drawing from a comic book called Black Magic (fig. 1).

The drawing shows an arm with muscles attached to it. These muscles are made out of a mesh with an inflatable tube inside it. Our project took this conceptual drawing and turned it into reality. During the design process, many decisions were made on what kinds of materials to use. Our first prototype of the project consisted of a latex tube, a small wire mesh, and a surgical syringe. As force was applied to the syringe, the latex tube would expand and the metal mesh contracted.

We soon came to realize that the latex tube was not going to work for our project. The small latex tube we were using had two stable states: relaxed, and fully inflated. It did not fill smoothly with air like we needed. Therefore, other types of tubes were tested. Among them were: a party balloon, thicker latex tubes, a small rubber tube, a condom, and a bike inner tube. The balloon and the small rubber tube both exploded without much force. The condom didn't explode but was easily punctured by broken wires in the metal mesh. The other latex tube suffered from a different flaw. When you put pressure on the tube it would not inflate, until it would reach its pressure limit and would start to swell at one point. After putting more pressure on it the area that was swelled would expand along the tube, while the rest of the tube would stay the original diameter. This left us with the bike inner tube. This tube turned out to work just like we wanted. It's walls are strong, and it would inflate smoothly. After choosing a tube, we needed to get a larger wire mesh as this tube was larger than the small latex tube we were using with the prototype. The wire mesh was easily obtainable from a hardware store, where it was sold as the wire shielding.

Once we obtained each of these items, we needed to figure out how to get the air into the tube. We already had an air compressor and some equipment that was used for air pressure. The air pressure equipment came from and old CD manufacturing machine, which had some air pressure controlled arms. Now all we needed to do was make an airtight seal in the tube, and figure out how to combine the air pressure equipment so that it would work.

We found that a few round LEGO pieces fit perfectly into the tube as plugs, and with a little drilling, the plastic tube would also fit through the LEGO piece and go into the tube. We placed one of these LEGO pieces into each side of the rubber tube. One was sealed with glue, and the other was sealed with the plastic tube running through it. We then used a bit of heat shrink tube to complete the seal. After we had one of these completed, we placed it into the metal mesh housing.

The metal mesh had two washers on each end of it. The mesh was frayed and bent outwards to make a flat, circular surface. We then place a washer on each side and bolted them together. The bolt used had hooks on the ends that would be used later as a connection to the arm. Now with the first muscle complete with the rubber tube inside, we needed to test it. We hung the muscle from a suspended rod and hung some weights on the other end. We then hooked the air compressor up to the muscle. After filling the compressor, we released a controlled 15PSI into the tube. This was done using a reduction valve. The muscle worked perfectly. We gave the muscle a little work out, and then we let it rest until we finished the next one.

After we had both of the muscles put together, we attached them to the arm on the disc of the pivot point. We decided to point the arm straight up and have the muscles pull from each side. At first, we used a weak nylon string to attach the muscles to the arm. This was quickly changed to a strong steel cable that was flexible, but would not stretch. We now had a working arm and it was time to test it with the computer.

Now the computer part needed to start working. This took a while because the first machine that was set up to work in the project stopped working. The hard disk decided to die on us. After bringing in a new machine we found that the card we were using in the computer to control the valves wasn't compatible with Windows 2000. After this we got the third computer set up, which worked. The card that is being used in the computer has a fixed location in physical memory and takes a 4-bit argument. The first 3 bits represent the address of the relay on the card and the last one is whether you want it to be on or off. The card can control 8 valves in all, but only 2 are used for this project.

The control program of our project is fairly straightforward. The program shows a graphical display with which you can control the valves and let certain sequences run. We hooked the muscles to the computer-controlled valves and then used the LabView to turn each valve on or off. It worked perfectly. We even had a setting where it would cycle though turning each muscle on and off. Therefore, we could let it run by itself for a demo.

After completing the design process, we were ready to demo the project. We encountered a few design problems, but nothing that set us back too much. The experimentation helped smooth out many of the bumps we encountered. All in all, it was a very smooth process.

3. Evaluation / Education

The main purpose of this project was to demonstrate how muscles work. We decided to simulate an arm to accomplish this task. We intended our project to be a museum exhibit. It could be on display with a computer to run the simulation. A user would press a button and the computer would start alternating the the muscles back and forth. This would show the muscles contracting and expanding. The user should notice how the muscle shrinks length-wise when it expands width wise. This is much like the human bicep. The audience this exhibit would be for is probably elementary school children and maybe children a little older. It would, of course, be interesting to all audiences. It is pleasing to look at and it is even pleasant to listen to.

Our final project shows very well how an individual muscle works, but it does not show how a complete arm works nearly as good. As is, this project is a very good proof of concept that these different materials used together can be used to simulate muscles. It may be interesting to some to discover that a metal mesh, much like a Chinese finger trap, could be used to create a muscle.

The project, as an experience, could be compared to such objects as an old re-breather or iron lung. The sounds produced are very similar. The motion of the project is similar to a metronome, with a slight delay before each tick. The idea of the project stems from futuristic robot designs, and therefore the project could be the early predecessor to something much more advanced and functional.

If the project evokes any sort of affective response from its viewers, it is most likely fear. Due to the large, noisy air compressor (which has the potential, though small, to make things explode), the force of the arm movement, and the monstrous computer controlling it, most people tend to stand back away from the project when it is in operation. This was not intentional. We intended for it to evoke more curiosity and aesthetic pleasure (it really is pleasant to listen to). Those feelings are indeed present in viewers, but the uncertainty of safety tends to keep them back.