Implementation and Feasibility of Artificial Muscles in Prosthesis
by Brady Russell
Abstract
Artificial Muscles (electroactive polymers, dielectric elastomers, etc.) have been looked at for potential use in prosthesis and orthotics. The purpose of this research is to look at the feasibility of using these current technologies, especially in their respective ability for load bearing capacity and power consumption. My focus was on three types of artificial muscles: Peano-HASEL actuators, McKibben actuators, and dielectric elastomers. All of these artificial muscles have shown great potential in both actuation strength and manufacturability. Both Peano-HASEL actuators and dielectric elastomers required an incredibly high voltage (from 6-10kV). Because dielectric elastomers suffer from electrical break down, and complex stacking is required to form effective actuators, I ruled both out because the feasibility of carrying materials able to employ the required voltage, as well as the danger of using the required voltage, is not only impractical but very dangerous to both the user and those around them. However, McKibben muscles can be made inexpensively and are strong with stroke length between 25-40% of original length. I used simple materials (balloon, syringe, and nylon cable tubing), and using hydraulics can lift a modest amount (4-6 lbs). Using them in conjunction with others can achieve a higher lifting capacity, and higher strength materials can yield different results. Voltage requirements are low as a muscle can be powered by a pneumatic or hydraulic chamber attached to a linear actuator or an air or hydraulic pump can be used, both of which can be powered conventionally. Given current technology, McKibben actuators provide the most logical option for prosthesis and orthotics devices.
Introduction and Predictions
The purpose of this study was to create a prosthetic device that used some form of artificial muscle (electroactive polymers, carbon nanotubes, etc.) to drive its actions. My hope was to create a prosthetic device that moved in a more energy efficient manner, and in a manner that is closer to that of an actual human appendage. A prosthetic that closely mimics human movement is better as it has been suggested that this is not only superior to the more rigid movements that can be accomplished with conventional actuators, but can also help in some psychological aspects with amputees especially those who lost their appendages as opposed to being born without them (Bar-Cohen, 2004). A prosthetic that uses artificial muscles would be beneficial as it would allow for a wider range of motion (it is not bound by the limitations of conventional actuators), greater functionality (as it would be silent as compared to mechanical actuators), and provide a better quality of life those who had lost or been born without appendages. The restrictions that are inherent in all mechanical actuators become very prevalent in complex mechanical systems (prosthetics). This is opposed to the relatively simple design of electroactive polymers, which are able to bend in more than one direction, stretch and shrink (Kim & Satoshi, 2007). The fact that dielectric elastomers (and all other artificial muscles) are also silent in operation gives them a clear advantage over loud and clunky mechanical actuators (Ireneusz, Filip, Renata, 2013).
My hypothesis states that by creating a prosthetic that functions in an almost completely lifelike manner I would be able to improve the quality of living for amputees and prove correct many of the philosophies about the efficiency of natural systems presented in the field of biomimetics.
There were some variables I had to consider. Humans are incredibly complicated beings, able to mechanically function with different degrees of ability due to a variety of factors. Humans are also built with different physiques and proportions. These limitations make it nearly impossible to create a singular device that would fit every person in need of such a device. This problem currently exists in the prosthetics industry and is usually addressed by making each device for each amputee. Not all amputations are exactly the same and because this device will be connected to the nervous system but there can be problems regarding varying degrees of nerve damage. Further, some patients may not have any difference in their mental state as a result of one of these devices, either because they were never depressed in the first place or because they were depressed from some other reason. Regardless, when conducting my study I needed to take into account the complex variable that is the human body and mind.
I am, of course, unable to fully test a device like this, (I would need a willing patient and a surgeon which is just not feasible). Along with this, the actual construction of the device will be somewhat limited by my technological capabilities. I do not possess a full lab to synthesize the most advanced electroactive polymers, nor do I have access to a full machine shop to create an incredibly complex frame that would serve as the bone structure. I desired to interview amputees. However there are very few amputees in Williamston, making my pool somewhat small.
Electroactive polymers: polymers that exhibit a change in size or shape when stimulated by an electric field. Dielectric elastomer: a class of electroactive polymers.
Review of Related Literature
Bar-Cohen, Y. (2004). Electroactive Polymer (EAP) Actuators as Artificial Muscles: Reality, Potential, and Challenges Second Edition, p. 0-20,35-36,737-740 (25 pp.)
This book offers a comprehensive look at the world of artificial muscles and their possible and preexisting applications. As I was reading I found that there are many applications for EAP that I had not previously considered, such as interfacing neuron to electronic devices. Along with this, I also found that many of my ideas are shared by the scientific community around the research of EAP such as that prosthetics made with EAP would be beneficial to the mental and psychological well-being of recipients. Along with the potential mental benefits the thought that I had about electroactive polymers being more energy efficient and less complex was reaffirmed. The article also discussed the downsides and challenges that exist in electroactive polymers such as lower actuating force. But also discussed that great strides have been made in the research and that the quality and usability of these actuators has only increased and continues to increase. This is very promising as my main concern has been that I will not find actuators at a reasonable price that are able to perform the tasks required, however I have learned that the type actuators I am looking for do in fact exist and that with some trial and practice I may even be able to produce them myself as the structure for electroactive polymers is not immensely complicated (at least not the ones I would be working with). Unlike mechanical systems electroactive polymers do not require complex machinery and precision that is required in, for instance, hydraulic and pneumatic actuators (this decrease in mechanical complexity is also more energy efficient). This book has been extremely helpful in deciding which EAP to use in my experiments as it helped me narrow down the field to a few possible candidates as well as showing me possible constraints I might have to work with.
Kellaris, N., Venkata, V. G., Smith, G. M., Mitchell, S. K., & Keplinger C. (2018). Peano-HASEL actuators: Muscle-mimetic, electrohydraulic transducers that linearly contract on activation. Science Robotics.
This article talked about a very new kind of artificial muscle known as a Peano-HASEL (hydraulically amplified self-healing electrostatic) actuator. These use an electrode and a sealed plastic pouch full of FR3 (a high breakdown strength of vegetable-based transformer oil). These actuators seem to solve many of the problems seen in other artificial muscles: they have a high breakdown capacity, are extremely fast in actuation, very cheap to manufacture, self-healing, and still relatively strong. These actuators functioned between 6 and 13 kilovolts, this is fairly normal for this type of actuator. In principle they resemble dielectric elastomer in that a voltage is used to attract two electrodes together compressing a surface, and then when the voltage is removed the muscle returns to its original shape.
This is an incredibly strong research paper, it is not only easy to understand but also conveys important information very quickly and easily. The many graphics and pictures in the paper give a clear visual of what is being talked about. There is instruction on how to create these actuators going as far as a parts list. Overall the paper is able to display and talk about data clearly and concisely.
Kaneto, K. (2016). Research Trends of Soft Actuators based on Electroactive Polymers and Conducting Polymers. Journal of Physics: Conference Series 704 012004. Retrieved from http://iopscience.iop.org/article/10.1088/1742-6596/704/1/012004/pdf
This article is largely about research trends in electroactive polymers. However, it also discusses several electroactive polymers in the real world and how to use them such as actuators that can move in a 3-dimensional plane. This article also discusses how it is an accepted design principle in the field of robotics that the way to move forward is to look at biological components and how we can mimic them as they are able to function in far greater complexity than any mechanical devices. This is done largely by bringing up already existing organic electronics such as OLEDs (organic light-emitting diodes).The article does still discussed such electroactive polymers, as Polypyrrole then some technical aspects about them. Also discusses how there is a need for more research as organic inspired electronics are the way forward as they are not bound by the limitations of mechanical actuators. This data was collected by looking at keywords in research papers and patents (the patents being in Japan).The data clearly show that after a peak in the late 2000s the research on the subject has been going down for an unknown reason.
All in all this article definitely brings up some interesting aspects and possible challenges. I found it interesting that the research trend has actually gone down though, especially due to the compelling arguments for in-depth research on this topic. This article definitely brought up some interesting possible research aspects for me especially in regards to what kinds of electroactive polymers I might use. The data collected in this paper was also very clearly presented and showed how it was collected. Because of this I had no problem comprehending this article. And the clarity also made the points without the possibility of misinterpretation.
Ireneusz, D., Filip, K., & Renata D. (2013). Zastosowanie Polimerów Elektroaktywnych do Budowy Protez (Prosthesis Design Driven by Electroactive Polymers). Czasopismo Techniczne Mechanika (Technical Transactions Mechanics).
This article talks about why dielectric elastomers are the most promising electroactive polymer that currently exists as they are energy efficient and resilient. First, here is a presentation of the different forms of electroactive polymers which is a list that was created by ESNAM (European Scientific Network for Artificial Muscles). Secondly, more of the technical aspects of dielectric elastomer motors are explained as well as how they can be used not only as actuators but also as measuring tools and energy harvesting components. Thirdly, the article then suggests that all of these traits could be used in conjunction to create a mechanism of superb efficiency. This is followed by an explanation and illustration showing how dielectric elastomers can be used in a stack actuator (a novel concept in which a series of electroactive polymers such as piezoelectric or in this case dielectric elastomers are used in conjunction with each other to produce linear force). Different forms of stack actuators are also discussed and the potential problems with manufacturing complexity that come with each of them. After this several concepts for devices that utilize dielectric elastomer motors are presented including an orthosis device. The article explains how the natural motions, similar to muscles, of the dielectric elastomers would provide for a superior actuator for this device as the natural motions would be better for rehabilitation of muscles This is opposed to mechanical actuators dielectric elastomer actuators would be silent and therefore not cause any inconvenience to the user besides the obvious wearing of the device.
This article definitely piqued my interest in dielectric elastomers for potential use in my prosthetic device. The manufacturing simplicity definitely made it sound more appealing as well as the low voltage cost. This article presented its information very clearly and in a way that was very easy to understand; along with this it was very thorough in presenting its evidence. This article has definitely made me seriously consider using dielectric elastomers as a potential electroactive polymer that could be used in actuation of my device.
Bar-Cohen, Y. (2012). Electroactive Polymers as Artificial Muscles – Capabilities, Potentials and Challenges. Fourth International Conference and Exposition on Robotics for Challenging Situations and Environments. Retrieved from
https://pdfs.semanticscholar.org/4966/4f1cf75bc5e17fed84c45b749fddccf17f14.pdf
The article talked at length about potential applications of electroactive polymers, especially in the field of aerospace and in prosthetics. There were mentions of several mechanisms, such as grippers and dust wipers that were powered by electroactive polymers in robots designed for space exploration. It also discussed the recent advancements in electroactive polymers as well as their strengths and weaknesses in compared to other non-mechanical actuators of the time such as electroactive ceramics. As well as electroactive materials and how they function in addition to strengths and weaknesses. The article talked extensively about both dry and wet electroactive polymers and the drawbacks of each. This included voltage requirements and temperatures functioning as well as the fact that some of them simply cannot be used in an open-air environment which would limit functionality.
This article contained very useful points about all of the available electroactive materials. I found the discussion on the carbon nanotubes electroactive polymers to be quite interesting and something I might consider using in my research. The primary drawback at the time was cost, but that has been largely corrected, as this article was published in 2000. I still found many helpful insights as well as some encouraging information in the form of examples of already existing electroactive polymer driven robotics that function much like human hands.
Kim, J. K., & Satoshi T. (2007). Electroactive Polymers for Robotic Applications Artificial Muscles and Sensors. p. 3-4, 17-21, 27-28, 32, 37-38, 128-134 (19 pp.)
In this book there is a large quantity of information regarding various types of electroactive polymers. Included were instructions on how to synthesize many of these electric of polymers into usable devices that are able to function in robotics, as well as the synthesization of polypyrrole and other chemicals that are used in electro active polymers. It also talked about on how to build both linearly actuating and bend actuated systems. This book had two separate sections for electronic EAP and ionic EAP and talked about the strengths and weaknesses of each. There was quite a bit of information on the electrical estimators in particular as well as ionic polymer metal composite (IPMC).The book also discussed problems with ionic EAP, namely that they were plagued with hydrolysis due the ionic reactions that occur as a result of them being used. Still in other ways they were extremely versatile being able to be utilized at low temperatures and with fairly low voltage. Electronic EAP, on the other hand faced very different problems. They require high voltage but have high response times and can operate in air without constraints as they do not require a liquid like ionic EAP. The point was brought up that as there are so many different EAP they can be tailor used for various purposes.
This book was very helpful and gave me very useful information on how to synthesize my own electroactive polymers. This book definitely turned me away from more of the ionic EAP as the problem with hydrolysis can result in a shorter life span even though they are low voltage. But the book definitely brought up some interesting points with how I might synthesize effective actuators and combined linear and bend actuators as many have done before, such as with the helical design used in one dielectric elastomer that was able to bend and actuate linearly. This book was set up in a very comprehensive and clear fashion. This would definitely be a large help in certain parts of a prosthetic.
Mohsen, S., & Kim, J. K. (2004). Ionic polymer–metal composites: III. Modeling and simulation as biomimetic sensors, actuators, transducers, and artificial muscles. Institute of Physics Publishing.
This article presented a wealth of mathematical information regarding the modeling of IPMCs. In general it talked about various ways to model these polymers, the various types of ionically actuated polymers, and the mathematical equations that go along with them. The article heavily discusses the chemistry behind these ionic polymers and certain considerations that should be taken into account with them such as thickness and type of IPMC. For the most part it is just an explanation of how the values are mathematically related in terms of reaching different equilibriums and how it will translate to the physical actuator.
If I were studying more of this kind of electroactive palmer this would be a great article. However, I am mostly focusing on the non-ionically activated side of things. The article was incredibly thorough and definitely presented its data very clearly and concisely. However this article is of little use to me as I will not be using this kind of electroactive polymer and so the article is of little use to me. This article has rolled out IPMC as a possible actuator.
Silvain, M., Zhang, Q. X., Wissler M., Löwe, C., & Kovacs, G. (2009).
A comparison between silicone and acrylic elastomers as dielectric materials in electroactive polymer actuators. Polymer International.
This article discusses ways to create a more efficient dielectric elastomer by using silicon as opposed to acrylic or other substances. An experiment was carried out which is shown in detail using this silicon dielectric EAP (there was a device which created it at different thicknesses). The article then explained the results of the experiment as well as the implications of the results. It was found that this silicon dielectric EAP was effective in the areas they wanted it to be as this is an example of a tailored EAP for a specific purpose, such as a large powerful actuator that is slow versus one that is fast. All of this was being measured against an acrylic version of the same device.
While the information in this article was presented very clearly, it did not make a lot of sense to me. Much of the terminology that was used did not make any sense to me and so some of its value was definitely lost. This article was not super helpful to me as I will most likely not be producing a dielectric elastomer of this complexity on my own as I do not own the lab equipment described in the article. Overall while the article might contain useful information it could not be comprehended by me and is of therefor little use to me. However it did show me the possible complexities possible in these actuators.
Methodology
In this study I created a baseline for technical data in regards to the actuation ability of human arms in particular motions. I constructed a rudimentary arm (when attempts to build a more complex arm failed). I tested the average actuation ability of human joints. I did this by taking a group thirteen volunteers and had them do four simple motions to test strength. They lifted the weight from their side and then proceed to bring the weight up so that there arm was extended in front of the torso (“in front of chest lifting test”). Then (in a separate test) the weight was brought all the way in the other directions so that the arm is extended in what would be the back of the torso (“backward lifting test”). The next test consisted of an exercise where the arm was fully extended so that the hand was parallel with the shoulder and then was brought across the chest (“across the chest lifting test”). There was also a standard weight curling exercise with the hand pointed at the ground. I used a dumbbell set to conduct these experiments. All of these tests were conducted by having each subject lift as much as they thought they could and letting them choose weights until the maximum was found. In subjects showing visible signs of strain, 80% of the actual value was used for calculation of the average (a 20 pound weight would be counted as a 16-pound weight). The average weight lifting ability in each test was then recorded. The exercise of moving weights to the front then to the back of the body allowed me to test flexion and extension. The test where the subject’s arm was brought up so the hand and elbow were parallel allowed me to test abduction in both directions and the bridging the weight across the body allowed me to test medial and lateral rotation of the subject. The simple curling exercise allowed me to test flexion and extension of muscles in the elbow.
In regards to the construction of the arm, a design was drafted based on a human skeletal arm and necessary motions the arm needed to perform. From there, the initial design was drafted using paper and then on CAD. Unfortunately, the design was not able to be completed as I was unable to properly mill out my design with the limited tooling I had available. After this failure, a rudimentary arm was set up that did the curling motion, but none of the shoulder motions. While I had the correct circuit and voltage the electrodes I was using were too rigid. Unfortunately, I did not have the facilities to construct proper electrodes. Because of this I moved to McKibben actuators but was still able to collect pre-existing data from published articles on the actuation force of Peano-HASEL actuators as well as dielectric elastomers. I moved on to gathering data on other already existing McKibben muscles, as well as testing my own McKibben muscles made from supplies of easy replication, for not only manufacturers but private citizens. Actuation capacity tests were conducted and the data was collected and recorded.
Results
Through experimentation, modeling, and data collecting, requirements about such things as number of muscles necessary to operate a fully functioning prosthetic arm and that arm’s ability to lift specified amounts of weight, were determined, as shown in Figure 1. The amount of weight needed to be lifted was determined to be between about 10-15 kg. The number of muscles required to make a fully functioning arm is eight, as is shown in Figure 2. Additionally it was shown that McKibben muscles can the lift approximately 2.7 kg at 12 volts, as shown in Figure 3.
Figure 1
The average scores of four different lifting different tests mentioned earlier in the paper. If a person showed visible straining while completing the test the maximum weight they were able to lift was numerically decreased by 20%. Data suggest while different muscles human are stronger than others they are still within a relatively close range (about 5kg).
Figure 2
The CAD file shows the proposed arm. The colored lines show where proposed muscles would be placed. Seven are visible but eight would exist (the yellow would be mirrored on the other side of the arm). The green muscles are for a curling motion. The purple and blue muscles are for lifting the arm up and down at the elbow. The yellow muscles are for swinging the arm at the shoulder. The red and orange muscles are for rotating the elbow. In total there are four pivot areas, three in the shoulder in the shoulder. One is a bearing that would be located in the cylinder that connects the two halves of the universal axle, and the two found in a normal universal axle. The remaining pivot area would be located in the elbow.
Figure 3
| Types of Actuator | McKibben |
| Voltage Requirements | 12V |
| Load Capacity | 2.7 kg |
The data shows that McKibben muscles can lift almost 3 kg with only 12V. The load capacity for my McKibben muscles is shown as is the voltage required to operate based off a 12 volt linear actuator or hydraulic pump (depending on possible design as described earlier).
Discussion and Conclusions
My original hypothesis, that it would be psychologically beneficial to use artificial muscles in prosthetics, was a bit grandiose. However, the data does show that using artificial muscles is feasible when using McKibben muscles. According to the data it would take approximately 4-6 of my own McKibben muscles to make a functioning arm. This backs up at least part of my original hypothesis, that it is possible to create a life-like prosthetic using artificial muscle. The McKibben muscles I created only cost a few cents to make and the actual arm itself could be milled out of aluminum or titanium for a comparatively small cost (to current prosthetics). The primary cost would be fitting and customizing the prosthetic to each particular user.
Unfortunately, aside from the more lifelike movement, none of the benefits such as silence as discussed in “Prosthesis Design Driven by Electroactive Polymers” would exist because of the need for either linear actuators or a pump (Ireneusz, Filip, Renata, 2013). However, other research into the field of McKibben actuators has yielded promising results. At the Tokyo Institute of Technology McKibben muscles have been developed and produced that are thin, lightweight and even sold in bundles like wire (University-Born, 2016). Along with this other more professionally produced McKibben muscles that can lift up to 15 kg at an inner diameter (of the muscle) of 0.5 cm² and this is scalable (University-Born, 2016).
Figure 4
| Types of Actuator | Peano-HASEL (Kellaris et al 2018) | Dielectric Elastomers (Plante, Devita, & Dubowsky 2007) | McKibben |
| Voltage Requirements | 6-13kV | 8.6 -9.4 kV | 12V |
| Load Capacity | 1 kg | 0.61 kg | 2.7 kg |
It is simply not feasible to use things like Peano-HASEL or dielectric elastomer actuators in prosthetic devices at present due to incredibly high voltage requirements, Peano-HASEL and dielectric elastomer actuators simply have too high of a voltage requirement, between 6-13kV, to be practical as is illustrated in Figure 4. McKibben muscles unfortunately produce noise, but are able to do the same thing (actuate) as these other muscles while still remaining practical to use. Current battery technology is simply not able to accommodate such high voltages for long periods of time as would be required of these prosthetic devices. As battery technology increases and as more efficient artificial muscles are developed, the topic should be revisited to find new possible solutions to current problems. This is especially true in the field of power consumption and requirements. Other electrodes can be used in Peano-HASEL or dielectric elastomer actuators; these should be explored and developed further to find more efficient electrodes. Peano-HASEL and dielectric elastomers have been used in a stack configuration to be stronger as have other artificial muscles and development in new and innovative configurations have proven fruitful. This topic has great potential to improve strength through mechanical means. More research needs to be done to create entirely new artificial muscles, such as the Peano-HASEL actuator.
Work Cited
(2016). University-Born Venture Established for Thin, Flexible Artificial Muscle Developed and going on sale as a key device for nursing care power suits and humanoid robots. Tokyo Institute of Technology. Retrieved from https://www.titech.ac.jp/english/news/2016/035340.html
Bar-Cohen, Y. (2004). Electroactive Polymer (EAP) Actuators as Artificial Muscles:Reality, Potential, and Challenges Second Edition, p. 0-20,35-36,737-740 (25 pp.)
Bar-Cohen, Y. (2012). Electroactive Polymers as Artificial Muscles – Capabilities, Potentials and Challenges. Fourth International Conference and Exposition on Robotics for Challenging Situations and Environments. Retrieved from
https://pdfs.semanticscholar.org/4966/4f1cf75bc5e17fed84c45b749fddccf17f14.pdf
Ireneusz, D., Filip, K., & Renata D. (2013). Zastosowanie Polimerów Elektroaktywnych do Budowy Protez (Prosthesis Design Driven by Electroactive Polymers). Czasopismo Techniczne Mechanika (Technical Transactions Mechanics).
Kaneto, K. (2016). Research Trends of Soft Actuators based on Electroactive Polymers and Conducting Polymers. Journal of Physics: Conference Series 704 012004. Retrieved from http://iopscience.iop.org/article/10.1088/1742-6596/704/1/012004/pdf
Kellaris, N., Venkata, V. G., Smith, G. M., Mitchell, S. K., & Keplinger C. (2018). Peano-HASEL actuators: Muscle-mimetic, electrohydraulic transducers that linearly contract on activation. Science Robotics.
Kim, J. K., & Satoshi T. (2007). Electroactive Polymers for Robotic Applications Artificial Muscles and Sensors. p. 3-4, 17-21, 27-28, 32, 37-38, 128-134 (19 pp.)
Mohsen, S., & Kim, J. K. (2004). Ionic polymer–metal composites: III. Modeling and simulation as biomimetic sensors, actuators, transducers, and artificial muscles. Institute of Physics Publishing.
Plante, J., Devita, L. M., & Dubowsky, S. (2007). A Road to Practical Dielectric Elastomer Actuators Based Robotics and Mechatronics: Discrete Actuation. ResearchGate.
Silvain, M., Zhang, Q. X., Wissler M., Löwe, C., & Kovacs, G. (2009). A comparison between silicone and acrylic elastomers as dielectric materials in electroactive polymer actuators. Polymer International.
