Soft robotics represents a transformative shift in the way machines interact with the world. Unlike conventional rigid robots, soft robots leverage compliant materials — silicone elastomers, hydrogels, and textiles — to achieve movements that mimic biological organisms. Central to this paradigm are the actuators that supply the motive force. Electromechanical actuators, which convert electrical energy into mechanical motion, have traditionally been associated with stiff, high-torque applications. However, a wave of innovations is reimagining these actuators for soft, adaptive systems. Recent breakthroughs in materials science, design topology, and embedded control are enabling electromechanical actuators that are lightweight, compliant, and capable of complex, life-like motions. This article surveys the latest developments in electromechanical actuation for soft robotics, highlighting key technologies, emerging applications, and the challenges that remain on the path to practical deployment.

Recent Advances in Electromechanical Actuators

The soft robotics community has long relied on pneumatics and hydraulics for large-force, compliant actuation. These fluidic systems, however, require bulky pumps and compressors, limiting portability and miniaturization. Electromechanical actuators offer a compelling alternative: they are electrically driven, can be precisely controlled, and are easily integrated with solid-state sensors. The challenge has been to make them soft. Over the past decade, researchers have developed several classes of electromechanical actuators that retain the compliance of soft materials while providing the speed and precision of electronic control.

Dielectric Elastomer Actuators (DEAs)

DEAs are perhaps the most mature electromechanical technology for soft robotics. They consist of a thin elastomer membrane sandwiched between compliant electrodes. When a high voltage is applied, electrostatic forces compress the membrane in thickness, causing it to expand in area — a phenomenon known as the Maxwell stress effect. This expansion can be harnessed to produce linear or bending motions. DEAs offer several compelling advantages: they are lightweight, silent, and capable of large strains (over 100% in area). Their intrinsic compliance makes them natural candidates for soft grippers, artificial muscles, and wearable haptic devices.

Recent innovations have focused on improving the dielectric strength and manufacturing scalability of DEAs. For instance, researchers at the Harvard John A. Paulson School of Engineering and Applied Sciences developed a multi-layer stacking process that allows DEAs to generate higher forces without increasing footprint (see Science Robotics, 2021). Other groups have introduced self-healing elastomers that can recover from electrical breakdown, dramatically increasing actuator lifespan. Additionally, the integration of carbon nanotube and graphene electrodes has reduced resistance and improved energy efficiency. These advances are moving DEAs from lab curiosities to practical components for autonomous soft robots.

Shape-Memory Alloys (SMAs)

Shape-memory alloys, such as nitinol (nickel-titanium), undergo a reversible solid-to-solid phase transformation when heated above a transition temperature. This transformation returns the alloy to a pre-programmed shape, generating significant force and displacement. SMAs are inherently stiff in their high-temperature austenite phase, but they can be embedded in soft matrices to create hybrid actuators that blend compliance with high work density. Early SMA actuators suffered from slow cooling rates and limited efficiency, but recent developments are overcoming these hurdles.

One promising direction is the use of thin-film SMAs and micro-wires, which reduce thermal mass and accelerate cycling. Researchers at ETH Zürich demonstrated a soft robotic finger actuated by multiple SMA wires that collectively produce human-like grasping forces (see Advanced Materials Technologies, 2020). Another innovation involves embedding SMA fibers within a liquid-cooled elastomer matrix, allowing rapid active cooling and thus higher actuation frequencies. Conductive polymers and resistive heating patterns also enable localized, independent control of multiple SMA segments within a single actuator, paving the way for multi-degree-of-freedom soft robots.

Ionic Polymer-Metal Composites (IPMCs)

IPMCs consist of a perfluorinated ionomer membrane (like Nafion) plated with noble metal electrodes. When a low voltage (1–5 V) is applied, hydrated cations migrate to the cathode, causing localized swelling and bending of the membrane. IPMCs are soft, operate in wet or dry environments, and consume very little power. Their primary drawback has been low force output and slow response. Recent work at the University of California, Los Angeles has improved force generation by layering multiple IPMC sheets and optimizing electrode morphology (see Sensors and Actuators A, 2022). Additionally, hybrid IPMC-SMA actuators combine the fast response of SMAs with the low-voltage operation of IPMCs, expanding their utility in soft robotics.

Bio-Inspired Designs and Soft Actuation Principles

Nature has long been a source of inspiration for soft robotics. Muscles, for instance, are soft, distributed actuators that can modulate both force and stiffness. Electromechanical actuators are increasingly designed to mimic these capabilities through novel architectures and material gradients.

Muscle-Like Actuation with Programmable Stiffness

A key feature of biological muscle is its ability to change stiffness dynamically — a coiled bicep is far stiffer than a relaxed one. Electromechanical actuators can emulate this by employing antagonistic pairs or variable-transmission mechanisms. For example, dielectric elastomers can be arranged in a “contractile” configuration that simultaneously generates force and adjusts stiffness via voltage control. Researchers at Columbia University developed a soft actuator using twisted and coiled polymer fibers (TCPAs) that behave like natural muscles: when heated (via resistive joule heating), the fibers contract and stiffen, mimicking the force-length relationship of skeletal muscle (see Science, 2017). These TCPAs can be driven by electromechanical means (resistive heating) and offer high work densities at low cost.

Biohybrid Actuators

Another frontier is the integration of living cells or tissues with electromechanical components. Biohybrid actuators use contractile cardiac muscle cells or skeletal myotubes to generate motion, with electrical stimulation provided by microelectrode arrays. While still in early research, these actuators promise unparalleled efficiency and self-healing. For instance, a team at Stanford University built a soft robot powered by rat heart muscle cells that “swam” using electrical pacing (see Nature, 2023). Although not purely electromechanical, biohybrid systems rely on electronic scaffolds for control and energy delivery, blurring the line between biological and synthetic actuation.

Control and Integration

Soft robots are inherently underactuated and nonlinear, making control a formidable challenge. The latest advances in electromechanical actuators are inseparable from progress in sensors, embedded systems, and machine learning. Without precise feedback and adaptive algorithms, the compliance of soft materials leads to unpredictable behavior.

Embedded Sensing and Closed-Loop Control

Modern soft robots incorporate stretchable sensors—capacitive, piezoresistive, or magnetic—that measure strain, curvature, and contact forces. These sensors feed data to microcontrollers that adjust actuator inputs in real time. For DEA-based systems, proprietary driver electronics now enable high-voltage waveforms with sub-millisecond resolution, allowing precise oscillation for crawling or swimming gaits. Researchers at University of Southern Denmark developed a soft robotic hand with embedded DEA sensors and actuators, achieving closed-loop grasping at bandwidths exceeding 10 Hz (see ICRA 2023).

Machine Learning for Actuator Coordination

Reinforcement learning and neural network models are increasingly used to generate control policies for soft robots with many degrees of freedom. Because electromechanical actuators can be independently addressed, learning-based approaches can discover efficient movement patterns that would be impossible to engineer manually. For example, a team from MIT CSAIL trained a soft quadruped robot (powered by SMA coils) to walk and trot using deep reinforcement learning, compensating for material fatigue and environmental variation (see Nature Communications, 2022). This synergy between compliant hardware and intelligent control is a defining trend in the field.

Applications and Impact

The innovations in electromechanical actuation are unlocking applications that were previously impossible with rigid robots. Three areas stand out.

Medical Robotics

Soft electromechanical actuators are ideal for endoscopic tools, wearable assistive devices, and implantable pumps. Their compliance reduces tissue damage, and their electrical control enables precise, low-latency operation. For example, a soft robotic glove using DEA sensors and actuators can provide therapeutic hand motion for stroke rehabilitation, adapting to each patient’s unique finger anatomy. A recent clinical trial at Shirley Ryan AbilityLab demonstrated significant improvement in hand function after six weeks of training with such a device.

Manufacturing and Handling

Soft grippers equipped with electromechanical actuators can handle fragile objects—fruit, glass, electronic components—without damage. Unlike pneumatic grippers, these electric counterparts require no external air supply, making them suitable for mobile autonomous robots in warehouse settings. Companies like Soft Robotics Inc. are already commercializing FDA-approved electric grippers for food handling, using a combination of SMAs and dielectric elastomers to achieve energy efficiency and hygiene.

Environmental Exploration

Soft robots can traverse loose sand, water, and uneven terrain that would stall rigid machines. Electromechanical actuators enable untethered operation, essential for fieldwork. Researchers at Carnegie Mellon University designed a soft snake robot driven by modular DEA segments that can crawl through narrow pipes and over rocks, powered by a battery and high-voltage inverter worn on its back. Such robots could inspect underground infrastructure or assist in search-and-rescue operations.

Challenges and Future Directions

Despite impressive progress, several hurdles remain before electromechanical actuators become ubiquitous in soft robotics.

Energy Efficiency and Power Supply

Many electromechanical actuators, particularly DEAs, require kilovolt-level driving voltages. While the currents are tiny, generating high voltage efficiently from compact batteries is an ongoing engineering challenge. Low-voltage alternatives (like IPMCs and electrostrictive polymers) suffer from low force and strain. Research into piezoelectric harvesters and hybrid energy storage (battery + capacitor) aims to improve system-level efficiency. A recent breakthrough using silicon-based micro-DEA arrays can operate at 10 V by thinning the elastomer to sub-micrometer dimensions, but mass production remains elusive.

Durability and Self-Healing

Soft actuators experience mechanical and electrical fatigue: dielectric breakdown in DEAs, thermal cycling in SMAs, and delamination in IPMCs. Self-healing materials (elastomers with reversible bonds) and redundant electrode pathways can mitigate damage. However, long-term reliability under realistic loading (thousands to millions of cycles) has not been demonstrated for most systems. Future work must standardize testing protocols and develop accelerated life-testing methods.

Autonomous Operation

Fully untethered soft robots require onboard power, control electronics, and computation — all of which add weight and stiffness. Lightweight printed circuit boards (flex PCBs) and high-density batteries are being integrated directly into actuator structures. The RoboSoft consortium has set a roadmap for 2030 that envisions a fully autonomous soft drone capable of perching and morphing, powered entirely by electromechanical actuation and sensing.

Conclusion

Electromechanical actuators are no longer merely rigid motors; they have become versatile, compliant building blocks for the next generation of soft robots. Through innovations in dielectric elastomers, shape-memory alloys, and biohybrid systems, researchers are achieving force, speed, and control that rival natural muscle. Combined with embedded sensing and intelligent algorithms, these actuators enable soft robots to safely interact with humans, adapt to unpredictable environments, and perform tasks that were once the exclusive domain of living organisms. The road ahead is steep — efficiency, durability, and autonomy must be tackled — but the trajectory is clear: electromechanical actuation will be a cornerstone of soft robotics for years to come.