Julie Legrand

Detection of damage serves as the initial phase for autonomous healing or adaptation to damage in resilient robots. While signaling the occurrence of damage proves beneficial, the more critical requirement lies in localizing the damage to enable targeted actions. This paper introduces a soft, self- healing damage localization sensor capable of detecting damage at four distinct locations using only a pair of measuring points (electrodes). The sensor comprises four resistive links, forming a resistive circuit, and operates by measuring the equivalent resistance between two fixed terminals. Damage occurring to each link induces a distinct change in the value of the equivalent resistance. The system is initially characterized at the material level and subsequently at the sensor level through multiple damage trials (the sensor can restore functionality through on- demand healing, achieved by applying a temperature of 90 °C for 30 minutes). Finally, the sensor is sandwiched between self-healing layers to form a skin. Out of the 16 damage trials conducted on the sensor, in standalone and embedded configurations, 15 successful localizations were observed. Additionally, reducing the number of electrodes enhances the ease of integration of this technology into various robotic applications, such as the palm of robotic hands.

Most existing anthropomorphic robotic fingers are either too stiff to offer compliance, or too soft to provide postural stability. Yet human subjects tend to stiffen their finger when producing fingertip forces and lower their joint stiffness when grasping objects. Variable joint stiffness is therefore required to offer compliance and postural stability to the finger when interacting with its environment. We therefore propose the novel design of a robotic anthropomorphic finger capable of variable stiffness by making use of the embodied intelligence design principle through multifunctionality of the hardware parts. The ligaments of the finger are not only used to connect the phalanges together, but also to provide local variable stiffness at the finger joints through the use of miniature McKibben pneumatic artificial muscles. This novel design can therefore offer compliance at lower stiffness levels and postural stability and a higher applied force at higher stiffness levels while keeping the finger look and movement anthropomorphic and its control quite basic. The developed anthropomorphic finger with variable stiffness was tested by interacting with a flat surface. The finger presented a significantly higher stiffness (6.7 $\mathbf {\cdot }$ 10 ⁻³ $\mathbf {\pm }$ 3 $\mathbf {\cdot }$ 10 ⁻³ Nm/rad) when the stiffening system was used than when the finger was purely used in compliance mode, without stiffness adaptation (4.8 $\mathbf {\cdot }$ 10 ⁻³ $\mathbf {\pm }$ 0.7 $\mathbf {\cdot }$ 10 ⁻³ Nm/rad).

The maximum number of damage and healing cycles that can be endured by a self-healing (SH) actuator, that is, its repeatable healability, has never been assessed. One reason for this is because healability was always tested manually. Typically, an operator uses a sharp blade to manually damage the actuator and places both damaged edges back together for them to heal. This process is time consuming, which explains why only a limited amount of damage and healing cycles has been reported. Moreover, this leads to an inaccurate estimation of how repeatable the actuator's healability is, since manual damage cannot be always performed at the exact same location. Therefore, we present a method to automatically and autonomously assess the repeatable healability of a soft SH actuator. It uses a robotic system composed of a damage station and a healing station, which the actuator is automatically moved between using a robotic arm. A sensor integrated inside the actuator is used alongside a dedicated characterization algorithm to automatically indicate whether the actuator is properly damaged and healed. We present a typical use case of the system, performing and analyzing 63 damage cycles of a selfhealing Diels–Alder actuator. After 53 cycles, the actuator will never properly heal again, therefore, we consider this cycle to be the maximum repeatable healability of the tested actuator. The healability of the SH Diels–Alder actuator is, therefore, not infinite experimentally.

Most soft tactile sensors require a large number of wires to read the sensed data and complex algorithms to treat them. This makes the integration of soft tactile sensors challenging in full robotic systems. In order to simplify the integration of those soft devices, a new type of soft tactile array sensor is proposed. The sensitive keys of the sensor are made of three different self-healing Diels-Alder (DA) conductive polymers. The key positions in the sensor array are optimised in order to maximise the sensor accuracy in detecting pressed key positions and limit the number of wires required to read the sensor data. Since DA polymers are used, the keys are robustly connected to each other, ensuring high interfacial strength and conductivity while allowing reconfiguration. The reversible covalent bonds of the DA polymer allow the keys to be cut and healed back together to reform another tactile sensor tailored for another application with full recovery of the initial mechanical and electrical properties. In this work, two study cases are investigated: a keypad and a position sensor. The novel soft tactile array could reach an accuracy as low as 1.31% for the keypad and 3.8% for the position sensor compared to the developed sensor model. The keypad has a 90% success rate, whereas the position sensor has a 64% success rate.

Voxel-based robots are aggregations of soft and simple building blocks that have been extensively evolved and simulated to perform various tasks, like walking, jumping or swimming. However, real-life voxel-based robots are rather scarce because of their challenging design and assembly. With the current materials and assembling methods, the interfaces between the soft multi-material voxels are prone to failure. This work proposes to make voxels out of reversible Diels-Alder polymers, which are available in a broad range of mechanical properties. By doing so, the covalent bonds at the multi-material interface ensure strong chemical connections, while allowing for reconfiguration. A first voxel-based gripper is thus robustly assembled, then disassembled, using its pieces (voxels) for reassembling another robot, i.e. a voxel-based walking robot. This reconfigurable property allows iterative validation of the simulated voxel-based robots and fine-tuning of the simulations parameters in a sustainable and economical way. Both physical voxel-based robots show similar behaviors as their simulations with root-mean-square errors down to 10.4%.

The increasing interest in soft robotics has led to new designs that exploit the combination of multiple materials, increasing robustness and enhancing performance. However, the combination of multiple non-linear materials makes modelling and eventual control of these highly flexible systems challenging. This article presents a methodology to model multi-material soft pneumatic actuators using finite element analysis (FEA), based on (hyper)elastic constitutive laws fitted on experimental material characterisation. Modelling in SOFA, a FEA software, allows to simulate and control in real-time soft robotic structures. One of the novelties presented in this paper is the development of a new user-friendly technique for the mesh partitioning in SOFA, using MATLAB algorithms, that allow the creation of uniform and more refined meshes and a mesh domain partitioning that can be adapted for any geometry. As a case study, a cylindrical multi-material soft pneumatic actuator is considered. It is composed of an internal chamber, which is constituted of an autonomous self-healing hydrogel, modelled as a hyperelastic material, and an external elastic reinforcement, made of thermoplastic polyether-polyurethane elastomer (TPPU), approached as a linear elastic material. The simulation of the combination of a hyperelastic and a linear elastic material in a single design is another contribution of this work to the scientific literature of SOFA simulations. Finally, the multi-material model obtained with the new mesh partitioning technique is simulated in quasi-static conditions and is experimentally validated, demonstrating an accurate fit between simulation and reality.

The current state of the art in compliant actuation has already good performance, but this is still insufficient to provide a decent autonomy for the next generation of robots. In this paper, a next step is taken to improve the efficiency of actuators by tackling and enhancing the Series-Parallel Elastic Constant Torque Actuation (SPECTA) concept, which has previously been analyzed in simulations. In this work, the efficiency is increased further by decoupling the springs and their driving parts through the use of locking mechanisms, such that the motors are not always loaded and the springs can easily store energy from both input or output. Simulations have been performed to confirm this and they also showed that, in the SPECTA concept, it is always better to use high-speed motors instead of high-torque motors, even with non-efficient gearing. In this paper, the SPECTA concept is also validated experimentally with the use of a newly built test setup. In light of the obtained results, showing an increase in efficiency for almost all working points, it can be stated that SPECTA is a promising new actuation technology that allows for an increase in energy recuperation, efficiency, and autonomy.

Self-healing polymers can address the damage susceptibility in soft robotics. However, in most cases, their healing requires a heat stimulus, provided by an external device. This letter presents a self-healing soft actuator with an integrated healable flexible heater, functioning as the stimuli-providing system. The actuator is constructed out of thermoreversible elastomers that are crosslinked by the Diels-Alder (DA) reaction, which provides the healing ability. The heater is manufactured from a DA-based composite network filled with 20 wt% carbon black to provide electrically conductive properties for resistive Joule heating. The flexibility of the heater does not compromise the actuator performance upon integration and the self-healing properties of both heater and actuator allow for damage repair. This includes very large damages, as both heater and actuator can recover (near 100%) from being cut completely in two pieces, using Joule heating at 90°C with a bias voltage of about 30 V. The embedded heater avoids the need for external intervention in the healing process, and provides healing quality assessment and a healing on-demand mechanism, paving the way for an optimum healing solution of damage resilient soft robots that require heat as a healing stimulus.

In endoscopic maxillary sinus surgery, the maxillary sinus is accessed through the nasal cavity which constitutes a narrow and tortuous pathway. However, surgeons still use rigid endoscopes and rigid, straight or pre-bent instruments for this procedure. Resection of the uncinate process and creation of a medial antrostomy is warranted to access the pathology inside the maxillary sinus and depending on the location of the pathology (lateral, inferior or anterior wall), additional resection of healthy tissue and/or functional structures like the lacrimal duct and/or inferior turbinate is necessary to gain optimal access. In order to avoid this additional resection, a functional single-handed, steerable endoscope for endoscopic maxillary sinus surgery has been designed and built. This endoscope is, to our knowledge, the most slender active steerable endoscope ever reported for maxillary sinus surgery. The performance of the endoscope was validated by two surgeons on a cadaver. An increased field of view was found in comparison to currently used endoscopes. As a direct consequence, a reduced need for resection of healthy tissue was confirmed.

This work gives an overview of all types of constant torque mechanisms (CTMs) based on both function and structure. Based on their architecture, they can be divided into five distinct categories, which all have their specific behavior. It is also shown that these CTMs can be divided into two application types, namely, power assistance or torque stabilization. Constant force mechanisms (CFMs) are more prevalent in the literature; therefore, some discussion is proposed on how CFMs can be transformed into CTMs. It is also shown that some of these CTMs have a very high specific energy, which makes them potentially interesting for use as energy storage/power providers in novel fields like robotics.