What Is Soft Robotics?

Soft robotics is a subfield of robotics concerned with the design, fabrication, and control of robotic systems built from compliant, deformable materials rather than rigid links and joints. Where conventional robotic arms rely on stiff structural members and discrete revolute or prismatic joints to transmit force, soft robots distribute deformation continuously through elastomeric bodies, producing motion that more closely resembles the musculature and appendages of biological organisms. The field draws on materials science, mechanical engineering, and control theory, and has grown rapidly since the early 2000s as advances in elastomer processing and additive manufacturing made complex compliant structures practical to fabricate.

The motivation for soft robots comes from limitations of rigid systems in unstructured environments. A rigid manipulator that contacts an unexpected obstacle can damage the object or itself; a soft gripper can passively conform to the geometry of an irregular object and grasp it safely without precise force control. This compliance also makes soft robots intrinsically safer for human-robot contact, which is why the field has strong connections to collaborative robotics and medical applications.

Actuation Mechanisms

Pneumatic actuation dominates soft robotics because pressurized fluid provides a high power-to-mass ratio and is compatible with the extensible elastomers used to construct soft bodies. A soft pneumatic actuator is typically a molded elastomeric chamber whose walls are geometrically constrained to bend, extend, or twist when pressurized, depending on the wall thickness distribution and fiber reinforcement. Fiber-reinforced actuators embed inextensible fibers in the elastomer matrix to direct deformation along a prescribed trajectory; angle-wound fibers produce twisting, while longitudinal fibers produce elongation. An alternative approach described in research on integrated fabrication of pneumatic soft robot actuators uses water-soluble cores printed by fused filament fabrication to cast hollow channels in a single molding step, reducing fabrication complexity significantly. Beyond pneumatics, soft actuation approaches include hydraulics, shape-memory alloys, dielectric elastomer actuators driven by high voltage fields, and hydrogel-based systems that swell or contract in response to chemical or thermal stimuli.

Materials and Fabrication

Silicone elastomers, particularly platinum-cured polydimethylsiloxane variants such as Dragon Skin and Ecoflex, are the most widely used structural materials in soft robotics because they can undergo strains exceeding 500 percent without tearing, are biocompatible, and are processable at room temperature in laboratory settings. Polyurethane elastomers offer higher tear resistance and are preferred in applications demanding durability. Mold casting is the primary fabrication method for laboratory prototypes, but additive manufacturing using direct ink writing or multi-material fused deposition is increasingly used to produce complex internal geometries that would be impossible to demold. The IEEE Xplore publication on characterization of silicone rubber-based soft pneumatic actuators established foundational characterization methods for these materials that inform model-based design.

Sensing and Control

Controlling a soft robot presents challenges absent from rigid robotics. Because soft bodies have theoretically infinite degrees of freedom and their deformation depends on contact forces that vary during operation, classical rigid-body models do not apply. Proprioceptive sensing for soft robots uses embedded strain sensors made from conductive elastomers, optical fibers, or liquid metal channels that change resistance as the body deforms. Exteroceptive sensing draws on vision-based methods that track body shape from camera images, avoiding the need for embedded sensors that can alter the mechanical properties of the structure. Model-based control approaches use finite element simulations or reduced-order models derived from them; data-driven methods including reinforcement learning are also actively studied. The ScienceDirect review of soft robotic systems for unstructured environments surveys sensing and control strategies used in field-deployed soft robots.

Applications

Soft robotics has applications in a wide range of domains, including:

• Surgical robotics and endoscopic tools that navigate tortuous anatomical pathways
• Agricultural harvesting robots that grip soft produce without bruising
• Search-and-rescue systems that squeeze through confined spaces in collapsed structures
• Wearable rehabilitation exosuits for stroke recovery and mobility assistance
• Underwater exploration vehicles inspired by cephalopod locomotion