Enabling autonomous healing in soft robotics

”In this PhD dissertation, the primary question addressed is How can soft robots autonomously perform
and intelligently control the healing action without the need for external (human) assistance?.”
Interactions between humans and animals are characterized by intimacy, wherein shared bodily prop-
erties play a pivotal role. However, closely interacting with robots, typically made from metallic compo-
nents, often evokes a sense of fear. Towards biomimetic properties, soft robots are under active research.
Crafted from soft and flexible materials, these robots are designed not only to ensure safe interactions
with their environment but also to adapt to their environment and tasks. However, while soft tissues and
materials offer resilience against shocks and impacts, they are susceptible to damage from sharp objects
and tears occurring at high elongation. Biological species tackle this issue with their healing ability. Sim-
ilarly, soft-bodied robots require this healing capability. This attribute can be integrated into soft robots
by constructing them from self-healing elastomers that leverage (thermo)-reversible chemical reactions.
The reversibility of the chemical bonds allows for healing in the bulk material.
However, merely introducing self-healing materials into soft robots cannot guarantee efficient and
functional healing. Several external interventions are required for healing at the material-based healing
level. Again, as robots move and operate in unstructured environments, external assistance and interven-
tions become more critical. The external assistance, extensively discussed in my review paper, encom-
passes the following five phases: (i) Healing should be practiced in unloaded conditions, requiring robots
to identify and analyze the damage to cease operation, necessitating the use of damage detection sensors.
(ii) Contamination and dust can affect healing efficiency; entering the damaged area will compromise
the healing process. Robots need to take actions such as cleaning (cleaning was not investigated in this
dissertation), or promptly closing the damage-induced gap. (iii) Closing the gap is essential in synthetic
self-healing materials, as they cannot grow; the damaged surface sides must be in close contact for the
chemical bonds to reform. As such, robots need a damage closure mechanism. (iv) Time is of utmost
importance, and fast healing is a significant advantage. Healing can be accelerated in self-healing materi-
als by external stimuli such as heat; sometimes serving even as a healing enabler, not just an accelerator.
Robots need the ability to heat their own bodies. (v) Finally, health monitoring and recovery assessment
should be performed. Continuing operations without proper healing would exacerbate the damage.
In all five phases mentioned, the involvement of humans would inevitably lead to both time and
monetary costs. Above that, in many cases such as (space) exploratory robots, limitations exist for human
access. In this PhD dissertation, the primary question addressed is ’How can soft robots autonomously
perform and intelligently control the healing action without the need for external (human) assistance?’.
In the field of robotics, there is potential to integrate mechatronic-based assistive mechanisms to achieve
v
vi
a self-sufficient system for healing. In this context, various multifunctional assistive healing mechanisms
have been explored to not only facilitate autonomous healing phases but also support the integration of
self-healing materials into soft robots:
• Conductive self-healing materials were customized to create a damage detection sensor skin, al-
lowing discrete localization of damage at four different zones, utilizing only one pair of electrodes.
The recovery of the sensing feature was exploited as a health monitoring criterion. This was pri-
marily intended to address phase one of the autonomous healing process, while also contributing
to phase five.
• From similar conductive materials used for damage detection and localization sensor skin, a self-
healing Joule-effect heater (reaching 90 °C with 30 V) was developed and embedded into a soft
self-healing bending actuator, serving as a stimuli provider to accelerate or enable the healing. The
heater could also function as a damage detection sensor, and also a temperature sensor. Recovery
of the resistance and heat distribution in the heater could provide information about the health of
the system as well. This was primarily intended to address phase four of the autonomous healing
process, while also contributing to phase one and five.
• Shape memory alloy (SMA) wires were primarily used in a reinforced soft bending actuator as a
damage closure agent, closing gaps of more that 2 mm in width and enable the healing. Further-
more, by utilizing the Joule-effect, the SMA wires could heat up the actuator and accelerate the
healing process, in addition to serving as reinforcement for the main robotic actuation. In a follow-
up project, with minor adjustments, the integrated SMA wires could be used as a twisting actuation
element in hybrid twisting/bending and twisting/extending soft actuators. This was primarily in-
tended to address phase three of the autonomous healing process, while also contributing to phase
four.
• In a collaborative inter-university project, adaptation in soft robots were studied. 3D-printed shape
memory polymers were tailored as adaptive fingertip elements, autonomously adapting their shape
to the objects and provide a more secure manipulation. The concept of shape memory in polymers
was discussed in the domain of damage closure as well. This has the potential to address phase
three of the autonomous healing process.
• An industrial collaboration and technology demonstration took place at Festo headquarter in Ger-
many, where self-healing materials were employed as pneumatic-based contact detection sensor
skin and variable stiffness layer-jamming based skin in an adaptive gripper. As a result, the gripper
could detect contact, delicately conform to the object’s shape, and securely manipulate it by ad-
justing its stiffness. In this setup, self-healing could restore airtightness, a critical requirement for
proper function of both the skins. It was achieved how variable stiffness can facilitate the integra-
tion of self-healing materials in robots by addressing the downside aspects. Endurance analysis was
conducted on the industrial-scale self-healing gripper. The achievements garnered attention from
Festo business developers and are currently under further consideration there. The pneumatic-based
contact detection sensor has the potential to detect damage, addressing phases one and five of the
autonomous healing process.
It is hoped that the results of this PhD dissertation will expand the applications of soft robotic systems
and encourage greater integration of healing mechanisms towards more sustainable robotics solutions.