Development of Bioinspired Exosuit Actuated with Hydro Muscles

and Novel Compact Robotic Flow Control Valve

Julia D’Agostino

1,2,3,

, Ellen Clarrissimeaux

2,3,*

, Shannon Moffat

1,2

, Juan D. Florez-Castillo

1,2

Felix Sanchez

1,2

, Matthew Bowers

1,2

and Marko Popovic

1,2,4,5

Robotics Engineering Program, WPI, 100 Institute Road, Worcester, Massachusetts, U.S.A.

Popovic Labs, WPI, 100 Institute Road, Worcester, Massachusetts, U.S.A.

Mechanical Engineering Department, WPI, 100 Institute Road, Worcester, Massachusetts, U.S.A.

Physics Department, WPI, 100 Institute Road, Worcester, Massachusetts, U.S.A.

Biomedical Engineering Department, WPI, 100 Institute Road, Worcester, Massachusetts, U.S.A.

Keywords: Exosuit, Wearable Robotics, Valve, Hydro Muscle, Fluid Actuator.

Abstract: The biologically inspired, wearable, exo-muscular suit has been proposed as a cost-effective, fluidly actuated

device for lower-limb physical therapy as well as for assistance with activities of daily living. The exosuit,

actuated with 12 biomimetic Hydro Muscles independently controlled with 12 5-way inexpensive, off-the-

shelf, on-off solenoid valves, has been designed, manufactured, and tested on a lightweight, biomimetic

human skeletal model. The results from testing suggested a necessity for more advanced fluid flow

management support system in the form of affordable, lightweight, and compact valves suitable for robotics

applications. To meet these metrics and fulfil the requirements of the exosuits fluid flow management system

the Compact Robotic Flow Control Valve was designed, manufactured, and tested. The CRFC Valve is lighter,

more compact, more controllable, and less expensive than any other similar valve currently on the market.

1 INTRODUCTION

Because form and function are naturally intertwined,

biologically inspired robots following the structure of

the human

body

are an attractive direction for the

advancement of a variety of biomedical applications.

Such applications include prosthetic limbs, braces,

exoskeletons, and exo-muscular suits that can be used

for physical therapy, activities of daily living (ADL),

and tasks requiring augmentation of common able-

bodied physical capabilities (Popovic, 2013; Popovic

2019).

There are a number of biologically inspired

muscles and corresponding systems (Popovic 2019).

Series Elastic Actuators (Pratt and Williamson,

1995), i.e. position controlled actuators in series with

elastic elements have been used in systems (Herr et

al, 2012; Blaya and H. Herr, 2014) with conventional

linear actuators as well as in the context of cable

driven systems (Kesner et al, 2011; Hunt et al, 2012;

Galiana et al, 2012; Mao and Agrawal, 2012; Asbeck

et al, 2013; Asbeck et al, 2014; Saint-Elme et al,

These authors contributed equally to this work.

2017). Similarly, soft and compliant fluid actuated

muscles have been actuating devices (Ueda et al,

2010; Park et al, 2014; Kurumaya et al, 2016)

utilizing McKibben Muscles and devices (McCarthy

et al, 2014; Sridar et al, 2016; Bowers et al, 2017)

employing Hydro Muscles (Sridar et al, 2016;

McCarthy et al, 2014).

McKibben muscles are currently the most popular

fluidly actuated soft artificial muscles (Popovic,

2019). Unfortunately, McKibben muscles are not

very efficient and cannot support a biologically

realistic muscle strain (Popovic, 2019; Sridar et al,

2016; Bowers et al, 2017). The Hydro Muscle,

utilized here, has excellent strain and energy

efficiency properties (Miriyev et al 2017) and can

closely mimic biological muscle dynamics (Popovic,

2019).

Many robotics researchers in academic settings

avoid hydraulically and pneumatically operated

systems in particular due to leaks, need for custom

parts, and complexities associated with a fluid

circulation system in which the entire system may be

D’Agostino, J., Clarrissimeaux, E., Moffat, S., Florez-Castillo, J., Sanchez, F., Bowers, M. and Popovic, M.

Development of Bioinspired Exosuit Actuated with Hydro Muscles and Novel Compact Robotic Flow Control Valve.

DOI: 10.5220/0008948600400049

In Proceedings of the 13th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2020) - Volume 1: BIODEVICES, pages 40-49

ISBN: 978-989-758-398-8; ISSN: 2184-4305

affected by local changes or disturbances. However,

fluid operated systems provide many advantages,

especially for wearable robotics where size and mass

matter significantly. Instead of having one strong and

heavy dedicated electric motor per actuated degree of

freedom, here only one strong and heavy electric

motor (e.g. pump) is needed for all actuated degrees

of freedom.

However, valves are necessary to operate

pneumatic and hydraulic systems. Valves that are cost

effective, lightweight, compact, can be electronically

controlled, and support a reasonable range of

pressures appropriate for wearable robotics

applications are not widely available on the market.

The Compact Robotic Flow Control (CRFC) Valve,

addressed here, was created to resolve these market

shortcomings, and to work in conjunction with the

Hydro Muscle system. When integrated, Hydro

Muscles and the CRFC Valve have the potential for

implementation in a rehabilitation robot system or a

wearable assistive exosuit that is lightweight, low-

cost, and has capabilities for fine control and

customization.

Presented here first is an overview of the design

and testing of an exosuit utilizing Hydro Muscles

(Curran et al, 2018; Moffat 2019) followed by an

explanation for the need for an advanced novel flow

control valve in Section 2. The details of the CRFC

Valve design, testing, and discussion are addressed in

Section 3.

2 EXOSUIT

The biologically inspired, wearable, exo-muscular

suit has been proposed as a cost-effective, fluidly

actuated device for lower-limb physical therapy as

well as for assistance with activities of daily living.

2.1 Exosuit: Modelling and Mechanics

The exosuit was modelled based on a lower limb

skeletal structure by 3B Scientific (Functional

Physiological Skeleton Model) with bungee cord

ligaments that allowed for lifelike degrees of

freedom.

The exo-muscular suit was modelled with 6

custom-made Hydro Muscles per each leg. This

retained anatomical integrity addressed the most

active biological muscles during regular gait cycle:

iliopsoas, tensor fasciae latae, quadriceps femoris,

gluteus maximus, hamstrings (biceps femoris and

semitendinosus), and gastrocnemius. Additional

passive spring structures were used to model the

extensor hallucis longus muscle (providing ankle

dorsiflexion), as well as the iliofemoral and

ischiofemoral ligaments.

The Hydro Muscles were placed in series with

tendons made out of Spiderwire (Spiderwire Stealth),

a thick fishing line. The series is attached to eye hooks

with threaded inserts placed at the approximated

anatomical muscle origin and insertion locations

(Hoy et al, 1990) (Fig. 1 and 2).

Figure 1: Exosuit muscle model.

Each Hydro Muscle length was uniquely

determined for the cross sectional profile of the

muscle. First, the maximal change in length of each

Hydro Muscle was obtained (assuming an always

slightly tensed inelastic tendon) based on average

biological joint angle trajectories of 15 healthy

individuals walking at a self-selected speed (Lewis

and Sahrmann, 2015). Then, by relating pressure to

Hydro Muscle length while assuming maximal

pressure of 0.69 MPa (100 psi) the unpressurized,

fully contracted Hydro Muscle length was

determined.

All 6 Hydro Muscles were composed of surgical

latex tubing and polyester Uber Hose (Uberhose153)

Development of Bioinspired Exosuit Actuated with Hydro Muscles and Novel Compact Robotic Flow Control Valve

sheathing. The tubing dimensions were 12.7 mm (1/2

in) outer diameter and 6.35 mm (1/4 in) inner

diameter. The Hydro Muscle forces were adjusted

using manual flow control valves (Elbow Pneumatic

Flow Control Valve). The valve orifice was fine-

tuned to provide for the most optimal force output of

each muscle. After initial mechanical testing, the

models gluteus maximus and quadriceps femoris

were doubled up to provide more force, i.e. each of

these two model muscles was added with an extra

Hydro Muscle connected to the same valve as the

original Hydro Muscle.

In the earlier stages of Hydro Muscle

development, an electronically controlled, a 5-way

on-off solenoid, pilot operated valve (Pneumatic

Electric Solenoid Valve) was used in series with a

manual flow control valve (Elbow Pneumatic Flow

Control Valve) to direct air flow in and out of the

Hydro Muscle. The gait was controlled through a

state machine, in which each of the six major phases

of the gait cycle was defined as a state to set each

solenoid valve to high or low, to indicate that the

corresponding muscle should be elongated or

contracted, respectively (Table 1). The six phases for

state control were obtained from eight standard

biomechanics gait phases (Fig. 3) by grouping

Loading Response, Mid Stance and Terminal Stance

into a single phase. The state transitions were deduced

by feedback supplied by Inertial Measurement Units

(IMUs), (SparkFun).

The skeletal lumbar vertebra was connected via a

light, elastic spring to a stand made from 80/20 T-

slotted aluminum framing. A platform on the top of

that stand held all of the pneumatic valves and the

microcontroller (Arduino MEGA 2560). The Hydro

Muscles pneumatic umbilical system consisted of

6.35 mm (1/4 in) in diameter tubing with push

connects. The pressurized air was supplied from a

compressed air tank operating at 0.69 MPa (100 psi).

Finally, a 500W powered variable speed treadmill

(Exacme 6400-0108BK Treadmill) was placed

beneath.

2.2 Exosuit: Experiment

A biomimetic skeletal structure, driven by the

exosuit, walked on a treadmill with the belt moving at

a constant pace of 0.28 m/s (Fig. 2). Movements were

recorded with IMU sensors and a high-speed camera.

Figure 2: Biomimetic skeletal structure driven by the

exosuit walking on a treadmill.

2.3 Exosuit: Results

The skeletal postures are compared (Streifeneder.

Ortho lab) to biological postures (Fig. 3). The joint

angle trajectories vs. percentage gait cycle are

contrasted with biological gait data for normal (Lewis

and Sahrmann, 2015), forward leaning (Lewis and

Sahrmann, 2015), and toe (Olensek and Matjacic,

2012) walking (Fig. 4).

The stride length was 0.78 m on average. The

biomimetic skeletal structure driven by the exosuit

was able to stand upright on its own and the light

tethering forces during the gait cycle were estimated

be less than 20% of the skeletal weight

(approximately 3.3 kg) based on inverted pendulum

Table 1: Muscles’ States.

Muscles’ States Heel Strike Stance Loading-Terminal Heel-Off Pre-Swing Mid-Swing

Terminal

Swing

Iliopsoas Expanded Expanded Expanded Contracted Contracted Contracted

Tensor Fasciae Latae Contracted Expanded Expanded Expanded Contracted Contracted

Quadriceps Femoris Contracted Contracted Expanded Expanded Expanded Contracted

Gluteus Maximus Expanded Contracted Contracted Contracted Expanded Expanded

Hamstrings Expanded Expanded Expanded Contracted Contracted Expanded

Gastrocnemius Expanded Expanded Contracted Contracted Contracted Expanded

BIODEVICES 2020 - 13th International Conference on Biomedical Electronics and Devices

Figure 3: The eight gait phases: numerical simulation (top) and actual physical model in motion (bottom).

Figure 4: Joint angle trajectories for exosuit (Robot) and biological (Normal Posture, Forward Posture, Toe Walking) gait.

dynamics and estimated Center of Pressure and

Center of Mass locations.

2.4 Exosuit: Discussion

The lightly tethered biomimetic skeletal structure

driven by the exosuit was able to execute multiple

steps at a slow, steady pace and emulate close to

human-like walking trajectories.

The gait somewhat resembles a toe walking, upper

body forward leaning posture gait, similar to a

downhill walking gait. This was primarily due to lack

of active dorsiflexion and mechanical deficiency of

the skeletal model; that is the inability of the knee to

fully extend, resulting in a more bent hip in order to

provide enough foot clearance. Also, motion was not

as smooth due to the on-off, digital nature of solenoid

valves.

This preliminary system was clearly not designed to

carry the entire weight of the pneumatic system (e.g.

heavy air compressor), however, this should not

concern Lokomat-like applications. The Lokomat is a

robotic gait training system that helps people who

have suffered from various neurological or physical

conditions to regain the ability to walk. In comparison

to conventional Lokomat, which can resolve joint

level body movements, the proposed Hydro Muscle

actuated exosuit can even resolve the individual

muscle level actuation. Due to cost effectiveness, the

proposed Hydro Muscle actuated exosuit could

possibly be available for inexpensive, at-home use.

However, for more mobile wearable assistive

devices, weight, size, and controllability of the fluid

Development of Bioinspired Exosuit Actuated with Hydro Muscles and Novel Compact Robotic Flow Control Valve

circulation system must be improved. Hence a closed

(likely incompressible) fluid circulation system with

lightweight, small, and cost-effective flow control

valves is needed for an affordable, finely controlled

system. The next section proposes the use of the

Compact Robotic Flow Control (CRFC) Valve to

meet these expectations.

3 CRFC VALVE

The Compact Robotic Flow Control (CRFC) Valve

was designed, manufactured, and tested to fulfill the

requirements of the exosuit’s fluid flow management

system.

3.1 CRFC Valve: Methods

The patent pending CRFC Valve is a simply operated

flow control mechanism, with the ability to

manipulate liquid and gas. It uses a servo motor

attached to a choking mechanism that controls an

entry and exit port for fluids (Fig. 5). The CRFC

Valve consists of a servo motor, a 3D printed servo

horn CAM mount, 3D printed casing, two tubes

incased in fabric, and two CAM-follower beads

which are connected to strings that choke off the

tubing (Figure 5 and 6).

The valve is controlled by a 0.215 Nm, 0.08

sec/60 degree @6V, ~$10 USD priced servo motor

(MG90D High Torque Metal Gear), which allows the

CRFC Valve to quickly and robustly handle over 0.69

MPa (100 PSI) of pressure. This valve replaces and

improves upon the overall performance of the

previously used valve unit addressed in Section II.

Current construction of the CRFC Valve consists

of a 3D printed motor and tube casing, and a curved

element. The curved element is used similarly to a

cam mechanism, with two spherical cam followers

(beads), the servo motor, and two tubes serving as the

flow channels, which allow for bi-directional fluid

flow (Fig. 6). Of the two tubes, at least one or both

are closed at any given point in time. One tube serves

as a fluid input for an attached system, and the other

tube is the release tube.

For one-directional flow operations, only one of

these tubes would be necessary to serve as an inlet

and outlet. On the anterior side of the valve, the two

tubes merge with a Y connector to attach to the

desired device. On the posterior side of the valve, the

tubes are separate, allowing one to be connected to a

pressurizing device, while the other releases the fluid

when the tube is opened. This is the operational

configuration of a three-way valve used to conduct

Figure 5: CRFC Valve: physical model (left) and CAD of

motor casing (bottom right) and attachment (top right).

flow and control tests; however, this valve is capable

of a few other operational configurations due to its

mechanical layout as a two-position, parallel, two-

way valve (i.e. constrained 4-way valve).

Operation of the valve involves actuating the

servo motor to tilt the curved element, on which the

beads roll along a smooth track (to minimize friction),

so that a finely controlled choking or opening of the

flow channels is achieved. String looped around each

tube is attached to each corresponding bead, which

allows the flow to be controlled by the angle of the

servo motor. The string pulls the tubing upwards

against the proximal portion of the casing, which is

curved to provide a more gradual decline in the

choking angle while preventing fluid flow.

The current design incorporates 5mm wide, 1mm

thick surgical tubing encased in kite fabric, which

prevents ballooning and the possible bursting of the

tube, as well as adding additional protection against

friction from the string.

In the resting state, the servo is at a neutral,

symmetrical position, with Tube A and Tube B both

closed, as depicted in State 2 (Fig. 6, top). This

prevents flow to pass through either tube, as they are

both constricted in this state. Rotating the servo

counterclockwise will cause the string to loosen,

allowing Tube A to open, while Tube B will remain

constricted, and thus will remain closed (State 1).

This will allow flow only through Tube A in this state.

Rotating the servo clockwise will cause Tube B to

open, but Tube A will remain closed (State 3). This

will allow flow only through Tube B in this state. The

range of operational servo angles change the size of

BIODEVICES 2020 - 13th International Conference on Biomedical Electronics and Devices

the tube inlet/outlet, thus introducing intermittent,

analog stages of operation and flow.

Figure 6: CRFC Valve: three states of the 3-port valve (top)

and principle of operation, front view (bottom). Curved

element (1, 2), CAM beads (3), Connecting strings (4, 5),

Casing base with holes (6, 7), Tubes (8).

3.1.1 Optimized Geometric Model

The valve model is dependent on the necessary

movement of the string to allow a tube to fully open,

the strain on the string on a side that is already closed,

and the desired dead-band angle for valve operation.

As shown, (Fig. 6 bottom), the curved element

rotation causes one bead to roll away from the valve’s

symmetry axis, thus reducing the distance between

the bead and the string’s anchor position above the

tube, and allowing the tube to open.

Figure 7: Tube choking - simplified elliptical model.

The total change in length needed for the tube to fully

open is obtained from the elliptical model (Fig. 7) as:

∆  2 





3













3 



3









(1)

The general valve geometric model (Fig. 8) for the

initial resting state (with the servo at a neutral, zero-

degree angle) can be used to relate string slackening,

for the tube that is opening, and string strain, for the

tube that is closed, to the servo angle for a set of

specified valve parameters. This general

configuration includes two separate, symmetric,

curvature radii (of which only one R is shown in (Fig.

8)), and a dead band angle, which prevents the bead

from rolling until points R, B, and A are collinear.

Figure 8: The general valve geometric model.

The model’s parameters were optimized such that:

(1) the valve volume is minimized and scaled to the

appropriate operation conditions. (2) The slackening

for the fully open condition is attained within a

rotation range. (3) The amount of strain on the closed

side is minimized. (4) The dead band is sufficient to

account for servo positioning errors, to prevent

undesirable flow, while not being so large as to

substantially affect simple control. (5) The servo

motor can have sufficient torque as to easily close the

tube for the desired operational pressure.

The optimized geometric configuration has

coinciding R’s (Fig. 8) positioned on the valve

symmetry axis, i.e. 



0



, AB axis (



) parallel to

the symmetry axis with 



 135



,



0



. The

values for 



,



,



are dependent on the dimension of

the tube, overall valve, and moment that servo motor

can produce. As this model can be scaled, the

optimization procedure can be easily reproduced with

Development of Bioinspired Exosuit Actuated with Hydro Muscles and Novel Compact Robotic Flow Control Valve

different tube diameters, fluid pressures, and desired

valve dimensions for given servo.

The current CRFC valve has a curvature radius of

20 mm and a total operational angle span of 42

degrees that can be used to finely control flow. Its

dimensions are 6  5  2, though it only

occupies 2/3 of that volume due to its L-like shape,

and it has a total mass of only 28 grams. This

compares favorably to the 12  10  2 ,

276 gram, on-off flow control modules addressed in

Section II.

3.2 CRFC Valve: Experiments

1) Response Time: To test system response time, end-

stops were placed at the CRFC Valve’s maximal

operational angle and neutral position. Contact with

these end-stops triggered or stopped an internal timer

for the channel open and channel close movement at

100 psi of fluid pressure. Ten tests were recorded for

both air and water.

2) Flow Rate: The flow rate across a range of servo

angles was determined by taking the steady state flow

rate readings at 6 degree increments from 0 (fully

closed) to 42 degrees (fully open). The test setup for

air consisted of a compressed air reservoir connected

to a valve inlet and a digital anemometer at the outlet.

Similarly, the test setup for water consisted of a 12V

diaphragm pump connected to a valve inlet and a

digital paddle-wheel flow meter at the outlet.

3) Hydro Muscle speed: To address the servo angle in

relation to the Hydro Muscle elongation speed, the

rate of elongation was collected with the curved valve

attachment being rotated at various degrees and

timing the full elongation of a 10.4cm Hydro in

contracted state.

4) Controllability: The controllability of the CRFC

Valve using both air and water were evaluated with

the test setup seen in Figure 9.

A simple rigid element (‘leg’) was attached with a pin

joint to a fixed base, and actuated by a Hydro

Muscle. A desired ‘leg’ angular trajectory was

specified in the form of absolute value of the sine

function over period of 2 seconds. A simple

proportional, dead-band adjusted controller was

developed for control of the servo motor. The ‘leg’

angular displacement values were provided by a

potentiometer. The same test was performed with 5-

way pneumatic solenoid valve, which utilized a

custom, pseudo-analog, PWM loop with a cycle time

of 5 ms and a tuned P-controller.

For the air test, an air compressor maintained a

constant pressure of 0.69 MPa (100 psi), and the

exhaust was vented into the ambient space. For the

Figure 9: Test setup for controllability of CRFC Valve with

3D printed rigid element.

the water test, a 12V pump with an accumulator

maintained steady fluid pressure in closed loop

hydraulic system.

3.3 CRFC Valve: Results

1) Response Time: The average CRFC Valve

response times for both air and water at 0.69 MPa

(100 PSI) are shown in Table 2. According to

information obtained directly from the manufacturer,

the specified solenoid valve has a response time

between 1 and 2 seconds to fully transition between

closed and open states.

2) Flow Rate: The flow rate vs servo angle (Fig. 10).

3) Hydro Muscle Speed: The result of test relating

servo angle to the Hydro Muscle elongation speed

exhibits 



-value of 0.967 (Fig. 11).

4) Controllability: The results of the controllability

experiments (Fig. 12).

Table 2: CRFC Valve response time for full state transition.

Response Open to Closed Closed to Open

Water 75 ms 70 ms

Air 70 ms 65 ms

BIODEVICES 2020 - 13th International Conference on Biomedical Electronics and Devices

Figure 10: Flow vs. angle; water (bottom) and air (top).

Figure 11: Hydro Muscle speed vs. servo angle.

Figure 12: Controllability tests: water CRFC Valve (top),

air CRFC Valve (middle), air solenoid valve (bottom).

3.4 CRFC Valve: Discussion

The exosuit, addressed in Section II utilized heavy

12cm by 10cm by 2cm on-off flow control modules

[26,27] that had a pre-set orifice size. When the

exosuit was first constructed, these were the most

practical, affordable solution on the market. To

Development of Bioinspired Exosuit Actuated with Hydro Muscles and Novel Compact Robotic Flow Control Valve

address already specified limitations of such flow

management system, the CRFC Valve was created.

This new valve is small, lightweight, cost-effective,

and can operate using both air and liquid. The new

CRFC valve takes up only 1/6th of the original valve’s

volume, and has a mass of only 28 grams (i.e. 1/10th

of the original weight). Additionally, the CRFC Valve

exhibits a relatively fast response time with very little

difference between water and air mediums.

Based on reviews of commercially available valves

(ASCO, Engineering Solutions), the quick response

times for fully opening and fully closing of 4-8 mm

inner diameter, commercially available, pilot solenoid

valves operating at ~100 PSI air pressure range from

10ms to 20ms and 20ms to 80ms respectively. In the

case of liquids these ranges are typically 15ms to 30ms,

and 30ms to 120ms respectively. The valves with these

parameters are typically valued at greater than $100

USD.

In comparison, the CRFC valve costs about $10

USD to produce. The full closing and opening times

for the same conditions are approximately 65ms for air

and 70ms for water. Additionally, the CRFC Valve

allows for continuous fine control of flow. The CRFC

Valve’s speed is well suited for wearable robotic

actuation systems, as it takes about 250ms for a skeletal

biological muscle to develop a peak force.

The CRFC flow is reasonably large with 

2.5/ and 210



/ for water and air

respectively. The flow can be increased by using

different tube dimensions. The flow results exhibit a 6



deadband angle, which addresses potential servo

inaccuracies and introduces control delays. Before it

saturates, the flow is roughly proportional to valve

angle.

The result of the test relating servo angle to the

Hydro Muscle elongation speed exhibits an 



-value

of 0.967. There is a strong linear relationship between

flow rate and servo angle. This largely linear behavior

is a characteristic of an optimized valve design. This

linear control of the flow allows for better control of

the Hydro Muscle than with the original on-off

solenoid valve.

The CRFC Valve operating with air has a more

precise and accurate tracking than the original 5-way

on-off solenoid valve. There was significantly less

oscillation in the CRFC Valve tests when compared to

the original solenoid valve due to the CRFC Valve’s

mechanism preventing sharp, jerky movements. While

both valves had oscillations, the CRFC Valve’s were

mainly due to sensor noise, while the solenoid valves

oscillations were due to the 2-state nature of the valve.

However, with improved sensors and a revised control

scheme, the response of the CRFC Valve could be

drastically improved.

From the response tests of the CRFC Valve, it is

clear that the choice of fluid impacts the response of

the system, however, the valve is still able to follow the

desired trajectories in a smooth and controlled manner.

The main issue in the system operating with water was

the inexpensive diaphragm pump being used. The

pump would frequently turn on and off, hence creating

an oscillatory behavior throughout the system; it is

likely that with improved system components the

results would be more indicative of the valves

capabilities. Additionally, due to the CRFC Valve’s

significantly reduced size and weight, the exosuit has

improved portability.

Due to the precise control over the fluids in the

actuators, the design has potential to create highly

controllable robotic systems.

4 CONCLUSIONS

The bioinspired exosuit strives to be a cost-effective,

fluidly actuated, wearable robotic device that can be

used for physical therapy and/or assistance with

activities of daily living. The bioinspired exosuit has

been designed, manufactured, tested on a lightweight

biomimetic human skeletal model. This initial study

motivated a need for better flow management system,

which inspired the creation of the CRFC Valve.

The CRFC Valve is lighter, more compact, more

controllable, and less expensive than any other similar

valve currently on the market. Now that the CRFC

Valve has been developed, future work will be focused

on constructing a wearable exosuit with the CRFC

Valves. Other applications for the CRFC Valve will

also be explored. The synergy of the CRFC Valve with

the cost-effectiveness, energy efficiency, and excellent

strain properties of Hydro Muscle opens a door into a

new age of very interesting, useful, and

accessible/affordable fluid operated wearable robotics

solutions.

REFERENCES

Arduino MEGA 2560. Arduino. Microcontroller.

Asbeck, A. T., Dyer, R., Larusson, A., and Walsh, C. J.,

“Biologically inspired soft exosuit,” 2013 IEEE

International Conference on Rehabilitation Robotics

(ICORR), pp.1-8, 2013.

Asbeck, A.T., De Rossi, S.M., Galiana, I., Ding, Y. and

Walsh, C.J., 2014. Stronger, smarter, softer: next-

BIODEVICES 2020 - 13th International Conference on Biomedical Electronics and Devices

generation wearable robots. IEEE Robotics &

Automation Magazine, 21(4), pp.22-33.

ASCO, Engineering Solutions, Solenoid valves.

https://www.asco.com/ASCO%20Asset%20Library/asco-

solenoid-valves-engineering-information.pdf

Blaya, J.A., and Herr, H.. “Adaptive control of a variable-

impedance ankle-foot orthosis to assist drop-foot gait,”

IEEE Transactions on neural systems and rehabilitation

engineering, 12(1), pp.24-31, 2014.

Bowers, M., Harmalkar, C., Agrawal, A., Kashyap, A., Tai,

J., and Popovic, M.B., Design and test of biologically

inspired multi-ﬁber Hydro Muscle actuated ankle,

Proceedings of 2017 IEEE International Workshop on

Advanced Robotics and its Social Impacts, March 8-10,

2017, University of Texas at Austin, Austin, TX, USA.

Curran, A., Colpritt, K., Moffat, S., and Sullivan, M.,

“Humanoid Walking Robot,” Major Qualifying Project,

Worcester Polytechnic Insitute, 2018.

Elbow Pneumatic Flow Control Valve. Utah Pneumatic. 1/4

OD 1/8 NPT Push to connect valve.

Exacme 6400-0108BK Treadmill. Exacme. Combo 500W

folding electric motorized treadmill.

Functional Physiological Skeleton Model - Frank - Hanging

Stand, Anatomical Models - Human Skeleton Models.

[Online]. https://www.a3bs.com/functional-

physiological-skeletonmodel-frank-hanging-

stand,p_164_20.html [Accessed: 15-Oct-2017].

Galiana, I., Hammond, F.L, Howe, R.D., and Popovic, M.B.,

“Wearable Soft Robotic Device for Post-Stroke Shoulder

Rehabilitation: Identifying Misalignments” 2012

IEEE/RSJ International Conference on Intelligent

Robots and Systems, October 7-12, 2012. Portugal.

Herr, H., Blaya, J.A., and Pratt, G.A., Massachusetts Institute

of Technology, 2012. Active Ankle foot orthosis. U.S.

Patent 8,287,477

Hoy, M.G., Zajac, F. E., and Gordon, M. E., “A

musculoskeletal model of the human lower extremity: the

effect of muscle, tendon, and moment arm on the

moment-angle relationship of musculotendon actuators

at the hip, knee, and ankle”. Journal of Biomechanics,

23(2), 1990, pp. 157-169.

Hunt, T., Berthelette, C., Iannacchione, G. S., Koehler, S.,

and Popovic, M. B., “Soft Robotics Variable Stiffness

Exo-Musculature, OneTo-Many Concept, and Advanced

Clutches.” In IEEE ICRA 2012 WORKSHOP: Variable

Stiffness Actuators moving the Robots of Tomorrow, St.

Paul, Minnesota, May, vol. 14. 2012.

Kesner, S.B., Jentoft, L., Hammond, F.L., Howe, R.D., and

Popovic. M.B., “Design Considerations for an Active

Soft Orthotic System for Shoulder Rehabilitation” 33rd

Annual International IEEE EMBS Conference, August

30 - September 02, 2011, Boston, USA.

Kurumaya, S., Suzumori, K., Nabae, H., and Wakimoto, S.,

“Musculoskeletal lower-limb robot driven by

multiﬁlament muscles”. Robomech Journal, 3(1), p.18,

2016.

Lewis, C. L., and Sahrmann, S. A., “Effect of Posture on Hip

Angles and Moments during Gait,” Manual Therapy

(20), pp. 176-182, 2015.

Lokomat, Hocoma. Zurich, Switzerland.

Mao, Y., and Agrawal, S.K., “Design of a cable-driven arm

exoskeleton (CAREX) for neural rehabilitation”. IEEE

Transactions on Robotics, 28(4), pp.922-931, 2012.

McCarthy, G., Effraimidis, D., Jennings, B., Corso, N., Onal

C., and Popovic, M.B., “Hydraulically Actuated Muscle

(HAM) Exo-Musculature” in “Robot Makers: The future

of digital rapid design and fabrication of robots” (RoMa)

Workshop, the 2014 Robotics: Science and Systems

Conference, Berkeley, CA, July 12, 2014.

MG90D HIgh Torque Metal Gear. Adafruit. Micro servo.

Miriyev, A., Stack, K., and Lipson, H., “Soft material for soft

actuators.” Nature communications 8.1 (2017): 596.

Moffat, Shannon Marija. "Biologically Inspired Legs and

Novel Flow Control Valve Toward a New Approach for

Accessible Wearable Robotics." Master Thesis,

Worcester Polytechnic Institute (2019).

Olensek, A., and Matjacic, Z., “Adjusting kinematics and

kinetics in a feedback-controlled toe walking model,”

Journal of NeuroEngineering and Rehabilitation (9), pp.

1-12, 2012.

Park, Y.L., Chen, B.R., Pérez-Arancibia, N.O., Young, D.,

Stirling, L., Wood, R.J., Goldfield, E.C. and Nagpal, R.,

2014. Design and control of a bio-inspired soft wearable

robotic device for ankle–foot rehabilitation.

Bioinspiration & biomimetics, 9(1), p.016007.

Pneumatic Electric Solenoid Valve. U.S. Solid. 1/4 5 way 2

position DC 24 V.

Popovic, M. B. 2013. Biomechanics and Robotics, Pan

Stanford Publishing Pte. Ltd., 1

edition.

Popovic, M. B. 2019. Biomechatronics, Academic Press,

Elsevier, 1

edition.

Pratt, G.A. and Williamson, M.M., 1995, August. Series

elastic actuators. In Intelligent Robots and Systems

95.'Human Robot Interaction and Cooperative Robots',

Proceedings. 1995 IEEE/RSJ International Conference

on (Vol. 1, pp. 399-406).

Saint-Elme, E., Larrier, M.A., Kracinovich, C., Renshaw, D.,

Troy, K., and Popovic, M.B., “Design of a Biologically

Accurate Prosthetic Hand”, IEEE RAS International

Symposium on Wearable & Rehabilitation Robotics

Houston, TX November 5-8, 2017.

SparkFun Triple Axis Accelerometer and Gyro Breakout

MPU-6050. SparkFun. Inertial measurement unit.

Spiderwire Stealth SCS50G-200. Spiderwire. 200yd, 50lb,

Braided ﬁshing line.

Sridar, S., Majeika, C. J., Schaffer, P., Bowers, M., Ueda, S.,

Barth, A. J., Sorrells, J. L., Wu, J. T., Hunt, T. R., and

Popovic, M., “Hydro Muscle - a novel soft ﬂuidic

actuator”, IEEE International Conference on Robotics

and Automation (ICRA), pp4104-4021, 2016.

Streifeneder. “Ortho.lab - Software and systems for video-

based motion analysis.”

Uberhose153. Uberhose. Watering hose.

Ueda, J., Ming, D., Krishnamoorthy, V., Shinohara, M., and

Ogasawara, T., “Individual muscle control using an

exoskeleton robot for muscle function

testing”. IEEE

Transactions on Neural Systems and Rehabilitation

Engineering, 18(4), pp.339-350, 2010.

Development of Bioinspired Exosuit Actuated with Hydro Muscles and Novel Compact Robotic Flow Control Valve