SOFT ROBOTICS – SOFT ARTIFICIAL
robotics is a fairly new emerging field which deals which robotic systems
similar to organisational structure of living organisms. Biologically inspired
robotics use sensors and materials which are very similar to that of a living
organism. One such application is soft artificial skin sensor. This paper shows
the progress in development of the artificial skin sensor.
1 The field of robotics has developed
over the years and robotics play an essential role in our daily lives. Robots
are used in fields such as medical and automation industry. Conventional robots
are made using hard and bulky materials such as aluminium and steel whereas the
development in the field of soft robotics has led to emergence of technology
and products which are similar to body structure of biological organisms such
as their skin, muscles and organs.
such sensor is the soft artificial skin sensor. This sensor is used to detect
the external strain and force applied to it. This sensor will make the robot
more responsive while performing actions such as assembly and manipulation 2.
The most common approach in designing a soft
skin sensor is use of contact sensors and embedded microfluidic channels for
detection of changes in structure of the sensors 2. Soft artificial skin
sensors which are capable of detection shear forces provide an added advantage
of giving comfort to the user of wearable devices 2.These forces are
responsible for detection of hand motion which would give an added advantage of
developing sensors for robotic arm.
skin was developed using rubber silicone for the sensor layers. Rubber silicone
has very high stretching and compressing properties 2. The skin included
three layers of silicone. the first and second layer consisted of micro
channels which are induced with a conductive fluid, EGaIn (Eutectic Gallium-Indium)
2. The third layer consisted of micro channels in a circular shape which is
responsive to external pressure 2. The second layer is placed perpendicular
to first layer to detect strain. All the layers are connected to each other
using interconnects 2.
force is applied onto the artificial skin the microfluidic channels change
form. These changes affect the cross – sectional area and lengths of the
channels due to which changes occur in electrical resistance 2. The paper 2
showed that the skin sense strain and pressure. The skin detects the strain
when the cross- sectional areas of the channels decreased and channel length
increased 2. The resistance and pressure are related and can be derived
theoretically using the equation (4) in 2.
skin was designed to be worn in 3. It is made of silicone rubber which
resembles a human hand. The glove has micro-channels which contain conductive liquids
and these channels helped in sensing changes in hand motion 3. Since the
glove is modelled after a human hand it had twenty-four degrees of freedom 3.
Two sensors were placed on each of the four long fingers, one sensor on each
joint of each finger 3. The thumb had three sensors, two on the joints on the
finger and one sensor between the thumb and the index finger 3. In this way
the skin can detect motions of the hand 3.
microfluidic channels had two types of conductive fluid flowing through them,
i.e. ionic fluid for pressure sensing and liquid metal for wiring. The ions in
the ionic fluid served as charge carriers, which helped in conductance 3. An
instrumentation circuit was used in 3 to avoid polarization of electrodes.
Due to polarization the sensors were not able to give impedance output which
was dependable 3.
gloves presented in 4 have similar construction as that of 3. In 4 the
twenty-four degrees of freedom of a human hand has been reduced to 19 degrees of
freedom by considering the joints which are responsible for movement of the
hand during the actions of handling and grasping. Sensors for detection of
pressure and strain are used in the glove 4. The changes in the shape of the
liquid – metal channels are responsible for changes in the electrical
resistance and cross – section of the channel length on which the working of
sensor is based on 4. The change in resistance is dependent on the properties
of the material, channel and elastomer properties and external forces 4. The
strain and change in resistance is given by (3) in 4.
In 5 the
sensor is made of a flexible elastomer and can measure the force in three
directions (x, y and z axis). The
conductive fluid flows through the micro-channels which is Eutectic
Gallium-Indium 5. A force transmission component is placed in the silicone
rubber which is made using 3D printer 5. Two layers of the silicone rubber
make up the sensor, the first layer consists of the force post and the second
layer consists of the micro-channels 5. The three sensors required for the
sensing the three in-plane forces placed 120 degrees to each other 5. When
external pressure is applied to the sensor in normal direction, the deformation
of channels results in changes in electrical resistance 5. These changes help
in determining of the three forces 5.
The paper 6
shows working of two sensor prototypes. The first prototype is based on paper
5, which could measure three forces. In the second prototype the conductive
metal liquid EGaIn (Eutectic Gallium-Indium) channel wiring are replaced by
copper flex circuit 6. This circuit avoids bogus sensor indications caused by
faulty mechanical trigger signals given to region of wiring.
The paper 7
elaborates on a sensor which can withstand strain. The sensor in this paper uses
two different conductive fluids which have distinguishable resistive values
7. The two conductive liquids used are ionic solution of sodium chloride and Eutectic
Gallium- Indium. For the two liquids, to be conductive an electrode was used as
an interface between the liquids. The electrode is made of silicone with micro
and nano-particle doping 7.
A low voltage
amplifier was used to design an oscillator circuit which has an advantage of
using DC supply to produce oscillatory pulses 7. When the sensor is deformed
the resistance increases and the oscillation decreases 7. Since resistance
and oscillations are dependent on each other, the relationship can be given as
(3) in 7. The input to a counter of digital processor gives the oscillator
skin demonstrated in the paper 8 is able to detect the pressure and strain
applied to it, at the same time. The notion of sensor’s that are capable of
detecting strain and pressure is used in the paper 8. These sensors consist
of micro channels with eutectic gallium- indium liquid flowing through them for
conductivity 8. The skin is designed to have three sensor layers that are
fabricated using silicone rubber 8. The silicone rubber has elastic
properties (stretchable and soft) 8. The first and second layer consists of
channels which are responsive to strain and pressure while the third layer is
responsive only to pressure. The first and the second layer are place at ninety
degrees to each other to be responsive to strain along the axis 8. This way
the sensor is able to detect the strain and pressure in x, y and z directions.
To make this circuit as a whole, the layers are connected using interconnects
3. Fabrication and Instrumentation
were followed for fabrication of artificial soft skin sensor done 2. These
steps are as follows:
Casting: The three sensor layers were developed with help of plastic
molds, which were made using a 3D printer 2. The molten silicone was poured
into the molds and left to harden at a temperature of 60 degree Celsius for
time duration more than three hours 2.
Bonding: The three hardened silicone layers were bonded using
liquid silicone 2. To avoid blockage of the interconnect holes a plastic
piece is introduced into the hole. The plastic piece is removed after
completion of the bonding process 2. To avoid liquid silicone from blocking
the micro channels, partial curing is done at 60 degrees Celsius. The process of
uniform silicone coating of silicone (i.e. spin-coating) and bonding is
repeated to get one sensor structure 2.
EGaIn Injection: EGaIn is injected into the micro channels
using one of the syringe and the other syringe is used to remove the air trapped
in the micro channels 2. After combining all the three layers, EGaIn is
injected. The duration of this process is approximately 60 seconds 2.
An electrode is used for
wiring connections 2. The holes made by the wire connections is closed using
silicone rubber 2. The micro channels are of the length 2.25 m and the
thickness of the complete prototype is approximately 3.5 mm 2.
The sensor is
acts as an input device when interfaced with computer. Whenever the sensor is
deformed due to application of pressure or strain, there are voltage drops
across the three layers of the sensor 2. The voltage drops are amplified by
the instrumentation amplifier 2. The output of the instrumentation amplifier
serves as input to a microcontroller where changes in resistance is detected
These changes are
taken by a MATLB program and a virtual model of the sensor is generated 2. The
disadvantage of the sensor prototype in 2 is that it is unable to detect
pressure and strain at the same time 2.
The glove prototype manufacturing process 3 consists of
processes similar to that of fabrication process in 2. However, the mold has
two components, a sensor bulge on the base and microchannel patterning and side
wall to decide the limit of the skin to used 3. The side wall and the base is
fixed together using screws 3.
Since the glove is designed similar to human hand,
the joints which are responsible for motion are selected and the channels are
placed on these joints 3. The molten elastomer undergoes the process of
cutting in the 3d printed mold 3. Silver threads are used between the two
conductive liquids as interfaces 3. The liquids were introduced into the skin
with the help of two syringes, one for injecting the liquids and the other for
removal of air in the channels 3.
The glove is interfaced with an instrumentation circuit
through which the electrical changes in the glove are measured and the sensor
response is observed. The circuit consisted of chip which had a microcontroller,
digital and analog components which are programmable and amplifiers 3. A
multiplexer was used so that the chip could be individually connected to any of
the sensors present on the joints 3. The response of the sensors is detected
and read by a MATLAB program 3.
fabrication process in 4 is same as that of fabrication in 2. The process
involves four steps, “1. Functional component embedding, 2. Silicone casting,
3. Layer bonding, 4. Conductive liquid injection,” 4. Three methods are
considered in 4 for obtaining the desired bond in the elastomer layers of the
glove of which mechanical method has been used 4. The sensors are casted into
a glove form using process of encapsulation 4. Encapsulation provides a
layout for the sensors to be placed in the glove form 4. The glove was tested
and was able to detect both pressure and hand motion at the same time 4.
used in 56 has been fabricated using set similar to that of 3 and 4.
However, the instrumentation setup differs. Firstly, a known load of “6-axis
force and torque sensor”, 56 has been applied to the sensor. Further the
sensor is connected to a PC so that the data can be read by MATLAB 5. The
sensor in 5 has a maximum capacity load of 13.3 N and range of 9.5 mm while
in 6 the sensor has a load capacity of around 44.1 N and range of 13 mm. The
forces are applied to the sensor in in plane force directions (x, y and z
directions) 56. For prevention of the two layers of sensor sliding away
when forces were applied onto it, sandpaper was glued to the surface 56.
The sensor in 7 was fabricated using
CO2 laser cutting. Three molds were prepared for shaping
of interfaces, determination of micro-channels and a final layer 7. The width
of the channels can be determined by keeping a check on power and laser speed
7. The elastomer consists of conductive particles which were introduced by
the process of mixing. The foundation of surface between the two conductive
liquids is “platinum cure silicone”, 7. For successful bonding of the
elastomer and Ecoflex, process of spin coating was used while elastomer was
placed in an oven 7. The last step was injection of the conductive liquids
using two syringes into the channels 7. After this step Ecoflex is used to
close holes made when wires were placed into the micro-channels 7. These
wires are used to connect the sensor to the PCB 7.
the design for its response and conductivity, it was observed that there was water
evaporation which resulted in slight change in resistance of the conductive
solutions 7. Further more strain and pressure were applied for testing the
response of micro-channels which did not cause any noticeable change in the
resistance of the sensor 7.
process of sensor in 8 is similar to fabrication process in 5,6. The
process follows the steps of casting, bonding and conductive liquid (EGaIn)
injection 8. Three sensor layers are cast using 3D printed molds. The process
of bonding takes place with the help of spin coating of silicone. The final
steps involve insertion of the EGaIn into the micro-channels 8. The sensor
layers are connected to each other using interconnects 8.
instrumentation used to connect the sensor to computer involves usage of
amplifiers 8. These amplifiers detect the voltage difference between the
three sensor layers 8. The detected values are given as inputs to the
microcontroller. The robustness of the sensor was measured by applying strain
and pressure in x, y and z directions 8. The response was distinguishable in
all the three cases; therefore, the sensor can detect strain, pressure along
with the different types trigger source 8.
fabrication process introduced in 9 can used to print intricate micro
channels. This construction of the sensors involves two sections, first section
being, fabrication of micro fluidic channels and the second being construction
of electrical components for sensing element 9. The fabrication of the
complete involves five steps. The steps are as follows:
Printing of the mold using inkjet:
Using inkjet printing for determination of the micro-channels is much
more environment friendly and cost effective than usual soft photolithography
process 9. The glass on which the mold
is printed is treated with an acetone solution and UV– ozone so that the ink
drops can adhere to the glass better 9. The holes present for the movement of
the fluids were reserved by placing a 1.5 mm holder 9.
The channels were patterned using a silicone elastomer kit 9. The kit
consists of two mixing agents which were mixed in a ratio of 10:1and was filled
into a container until the liquid overflowed 9. The glass used in the
previous step is used to cover the container 9. The pattern is obtained on
the PDMS sheet after the container cools down 9.
Printing of isolation and metallization characteristics using
Using silver ink, the metallization characteristics were printed onto a
polyethylene terephthalate (PET) sheets 9. These sheets can be replaced by
paper, silicon, LCP (liquid crystal polymer) and glass 9. To obtain an ideal
resistance of sheet, silver patterns were printed four times 9. The isolation
layer is made of the material SU-8 due to its properties of high chemical
Van Der Waals forces are considered in sealing the PDMS with a smooth
surface 9. The PDMS and PET are bonded together tightly in such a way that
there is no presence of air bubbles 9. This prevents conductive fluid
drainage due to high pressure 9. The ability of peeling of the microfluidic
layer of the sensor, helps in replacement if needed, and also helps in making
changes to the sensitivity 9.
The two basic properties of a wearable sensor are
sensitivity and flexibility 9. For measurement of the sensitivity factor, a
considerable number of liquids were used with minimum ground interference 9.
It was observed that the variations in attenuation and bandwidth are due to the
dielectric losses due of the different liquids 9. Very small amount of liquid
is required for the sensitivity measurement due to the small size of the
channels 9. To check the flexibility of the sensor, it was folded to the size
of four cylinders with different measurements 9. It was concluded that the
prototype is sensitive as well as flexible 9.
All the papers
have demonstrated use of biologically similar materials to implement the
sensors used in the porotypes. These sensors can find application in areas of
humanoid, haptics, smart skins, health care and machine- human interactions.
The manufacturing process of these sensors may include 3D printing, Ink-jetting
or soft lithography. The sensor prototypes have been designed using a very
elastic material like silicone rubber through which microfluidic channels run.
These channels have conductive liquid flowing through them which can be EGaIn
or a mixture of ionic solution and EGaIn. When these channels undergo
deformation, there is change in electrical resistance of the conductive liquid.
The change in resistance is measured and transferred to the interfaced
processor. These readings are read by MATLAB and simulation is displayed. These
prototypes can be designed to be adapt to different users. The future scope of
soft robotics includes the ability to perform medical procedure’s such as
implantation’s and surgeries, wearable robotic prosthetics , tissue budding and
1 Iida F, Laschi C. Soft robotics:
challenges and perspectives. Procedia Computer Science 2011; 7: 99-102.
2 Y.-L. Park, B.Chen,
and R. J. Wood., “Design and fabrication of soft artificial skin using embedded
microchannels and liquid conductors,” IEEE Sens J., vol. 12, no. 8, pp.
3 Jean-Baptiste Chossat, Yiwei Tao, Vincent Duchaine,
and Yong-Lae Park., “Wearable Soft Artificial Skin for Hand Motion detection
with embedded Microfluidic Strain Sensing,” 2015 IEEE International Conference
on Robotics and Automation (ICRA).
4 F. L. Hammond, Y. Menguc,
and R. J. Wood, “Toward a modular soft sensor-embedded glove for human hand
motion and tactile pressure measurement,” in Proc. IEEE/RSJ Int.Conf. Intell. Rob. Syst.,
(Chicago, IL), September 2014.
Vogt, Y. L. Park and R. J. Wood, “A soft multi-axis force
sensor,” 2012 IEEE Sensors, Taipei, 2012, pp. 1-4.
6 D.Vogt,Y.-L. Park, and
R.J. Wood, “Design and Characterization of a soft multi-axis force sensor using
embedded microfluidic channels,” IEEE Sens J., vol. 13, no. 10 pp. 4056-4064,
7 J.-B. Chossat,
Y.-L. Park, R. J. Wood, and V. Duchaine, “A soft strain
sensor based on
ionic and metal liquids,” IEEE Sens J., vol. 13, no. 9,
pp. 3405–3414, 2013.
8 Y. L. Park, B. r. Chen and
R. J. Wood, “Soft artificial skin with multi-modal sensing capability
using embedded liquid conductors,” 2011 IEEE SENSORS Proceedings, Limerick, 2011, pp. 81-84.
9 W. Su, B. S. Cook and M.
M. Tentzeris, “Additively Manufactured Microfluidics-Based
“Peel-and-Replace” RF Sensors for Wearable Applications,” in IEEE Transactions on Microwave Theory and
Techniques, vol. 64, no. 6, pp. 1928-1936, June 2016.