Chain flexibility reflects the ability of polymer segments to rotate and move freely in space^[50]. Hydrogels with high chain flexibility allow polymer chains to adjust their conformations more easily^[51], thereby imparting excellent softness and deformability [Figure 2A]^[40]. For example, in poly(vinyl alcohol) (PVA) hydrogels, chain flexibility can be tuned by controlling the molecular weight and degree of polymerization^[39]. Wang et al. developed a PVA-based hydrogel for epidermal electronics that exhibits high stretchability, mechanical resilience, and stable sensing performance. The flexibility and mobility of PVA chains allow the hydrogel network to dynamically rearrange under external stress, allowing large and reversible deformation. Such behavior is particularly advantageous for flexible sensors subjected to frequent bending and cyclic deformation^[56].

Crosslinking density refers to the number of crosslinking points per unit volume, which plays a critical role in mechanical strength, swelling behavior, and stability [Figure 2B]^[53]. Increasing crosslinking density generally enhances mechanical strength by restricting chain mobility, making the hydrogel more resistant to stretching or compression^[53]. However, excessive crosslinking may lead to brittleness and reduced toughness. For instance, in polyacrylamide (PAM) hydrogels, the amount of crosslinker (e.g., N, N′-methylenebisacrylamide^[40]) can be precisely adjusted to control crosslinking density. Moderate increases in crosslinker content significantly improve tensile strength and hardness, but excessive amounts result in fragile hydrogels prone to fracture under stress^[40].

Introducing branching or side-chain architecture provides an effective strategy to regulate hydrogel mechanics and functionality. Branching increases chain entanglement, thereby enhancing mechanical strength and stability. Side chains, depending on their length, chemical nature, and spatial distribution, can impart specific functionalities^[57]. For example, Pan et al. proposed a molecular design strategy to regulate phase separation in hydrophobic association hydrogels by tuning short alkyl side-chain length [Figure 2C]^[41]. They found that enhanced hydrophobic interactions increase the size of polymer-rich domains and shift the relaxation dynamics toward higher glass-transition temperatures. Importantly, high strength and toughness are achieved when deformation conditions match the intrinsic relaxation behavior, indicating that phase separation and glass transition jointly govern mechanical performance^[41].

Self-assembly and noncovalent interactions (e.g., hydrogen bonding^[58], van der Waals forces^[59], π-π stacking^[60]) play vital roles in hydrogel structure formation and property regulation [Figure 2D]^[34]. These interactions are dynamic and reversible, enabling hydrogels to undergo reversible deformation and exhibit self-healing capability. For instance, hydrogen-bonded hydrogels can reform broken bonds after damage, restoring integrity. A typical example is PVA-borate hydrogels^[42], where hydroxyl groups on PVA chains interact with borate ions to form a three-dimensional network. Upon cutting or stretching, the damaged regions can self-repair through reformation of hydrogen bonds, recovering mechanical performance.

Double or multiple network hydrogels consist of two or more interpenetrating or interwoven networks, endowing them with superior mechanical strength and multifunctionality [Figure 2E]^[61]. Networks can be interconnected via chemical crosslinking or physical interactions, working synergistically to enhance toughness^[62,63]. A typical double-network hydrogel combines a rigid network with a flexible one: the rigid network provides strength and resistance to stretching, while the flexible network dissipates energy and prevents catastrophic fracture^[43]. This architecture yields hydrogels with high strength, flexibility, and fatigue resistance, suitable for load-bearing and wear-resistant applications.

Slide-ring networks represent an advanced topological design in hydrogel mechanics [Figure 2F]^[54]. By introducing movable ring molecules onto polymer chains, stress can be effectively redistributed during deformation^[64]. The sliding motion of rings enhances stretchability and toughness. Fully sliding structures extend this concept, allowing all network junctions to move freely, resulting in extraordinary flexibility and extensibility^[54]. Wang et al. introduced slide-ring crosslinking points into a double-network (DN) hydrogel and showed that chain sliding at these movable junctions significantly enhances mechanical adaptability^[44]. The hydrogel achieved reversible tensile deformation above 300%, an elastic modulus of approximately 300-500 kPa and maintained stable sensing signals after over 10,000 cyclic deformations. These results confirm that slide-ring dynamics play a critical role in mitigating stress concentration and improving fatigue resistance in flexible hydrogel systems.

Sacrificial networks are incorporated to improve toughness and fatigue resistance [Figure 2G]^[55]. These networks, with relatively low mechanical strength, preferentially fracture under stress, dissipating energy and protecting the primary network^[65]. By rationally designing the structure and properties of sacrificial networks, hydrogel mechanics can be finely tuned to meet diverse application requirements^[66]. He et al. proposed a DN hydrogel in which the densely crosslinked sacrificial network is built from enzyme-cleavable covalent peptide bonds^[45]. They demonstrated that, under force-coupled enzymatic catalysis, the sacrificial network can undergo reversible rupture and reformation, enabling recoverable energy dissipation rather than permanent damage. As a result, the hydrogel combines high toughness and reversibility^[45], exhibiting fracture strains above 2,000%, a maximum toughness of 15 MJ·m^-3, and a fracture energy of 9.4 kJ·m^-2, while allowing mechanical recovery and self-healing after cyclic loading.

Pore structures are fundamental characteristics of hydrogels, encompassing pore size, distribution, and connectivity^[67]. They significantly influence stress-strain transmission and fatigue life. Larger pores facilitate transport and diffusion, enhancing swelling kinetics and drug release efficiency. However, excessively large pores may weaken mechanical strength due to thin pore walls that cannot withstand high stress. Uniform pore size distribution helps distribute stress evenly, reducing local stress concentration and improving fatigue resistance. High pore connectivity facilitates rapid mass transport throughout the hydrogel, enhancing functionality^[68]. For example, in tissue engineering scaffolds, controlling pore architecture enables hydrogels to mimic the extracellular matrix by providing cell-scale void spaces and efficient nutrient transport pathways. Granular gelatin methacryloyl (GelMA) hydrogel scaffolds constructed from porous microgels exhibit hierarchical inter- and intraparticle porosity, delivering up to 170% higher void fraction and 57%-78% increase in vivo cell infiltration compared with nonporous counterparts, thereby effectively promoting cell adhesion, migration, and tissue integration. Importantly, quantitative mechanical analyses have shown that increasing pore size from ~8 to ~24 μm can redistribute stress and enhance fracture energy by approximately 2.5-fold, while simultaneously delaying fatigue crack propagation^[46]. Furthermore, the polyvinyl alcohol hydrogel prepared by the freeze casting process forms a honeycomb-like porous structure [Figure 3A and B]^[69]. After further annealing, the crystallinity of PVA hydrogels increases by ~15%-20%, which correlates with a ~2× improvement in compressive modulus and enhanced toughness. This evidence underscores the quantitative role of pore size and connectivity in enhancing fracture toughness and fatigue resistance of hydrogel networks^[69].

Gradient structures are widely used to reconcile mechanical mismatches across biological interfaces^[70]. Gradient structures refer to spatial variations in physical or chemical properties within hydrogels. Such architectures enable different regions to exhibit distinct performances, meeting complex application requirements, and more importantly, they redistribute stresses under cyclic loading. By gradually varying stiffness across the interface, superficial regions experience moderated strain amplitudes while deeper regions sustain lower strain levels (< 10%), thereby alleviating stress concentration, homogenizing load transfer, and delaying fatigue crack initiation during repeated deformation cycles^[71]. For joint repair, hydrogels can be engineered with gradient mechanical properties so that the cartilage-side region presents a lower elastic modulus matching native cartilage (1 MPa) and mitigating local stress, while the bone-side region is substantially stiffer (5.8-6.4 MPa) to provide load-bearing support and interfacial stability^[72,73]. Eckstein et al. used digital light projection/microstereolithography-printed polyethylene glycol (PEG)-based hydrogel-composite micro-truss scaffolds with sub-200 μm features and high porosity (68%-81%), the study achieved a controlled cartilage-zone gradient of 0.76-1 MPa and a 6× modulus jump into the bone layer^[47]. Importantly, finite element analysis revealed that such gradients redistribute cyclic stresses: superficial regions experienced > 30% strain under compression, while deeper regions maintained < 10% strain, thereby reducing stress concentration and delaying fatigue crack initiation^[47]. In addition, Yao et al. reported a preparation method termed as electrochemical training that utilizes gradient ionic coordination and molecular locking to achieve rapid assembly of disordered hydrogels [Figure 3C]^[74]. These hydrogels exhibit multiscale anisotropic gradient structures ranging from 5 nm to 2 cm, which not only provide spatially tuned mechanical cues but also homogenize stress transfer under repeated loading, enhancing long-term fatigue resistance.

Layered or aligned fiber architectures introduce anisotropic mechanical, electrical, or transport properties into hydrogel systems. By controlling fiber orientation and hierarchical arrangement, hydrogels can achieve directional mechanical strength, conductivity, or optical performance^[75]. For example, inspired by the layered structure of the head capsule of the Asian corn borer, a chitin hydrogel with a layered structure was developed that exhibits excellent impact resistance^[76] [Figure 3D]. Conductive hydrogels can indeed be engineered by aligning conductive micro/nanofibers [e.g., cellulose nanofiber (CNF)/MXene, carbon nanotubes (CNTs)] into an ordered fiber network, which produces direction-dependent electrical pathways^[76]. Representative systems show that the conductivity parallel to the fiber alignment markedly exceeds the perpendicular direction: for example, an anisotropic CNF/MXene DN hydrogel reaches 13.08 S·m^-1 along the alignment and is explicitly used to build direction-selective strain/pressure sensors (the orthogonal direction is significantly less conductive), demonstrating how orientation programs the circuit behavior. Likewise, mechanically stretched, fiber-oriented hydrogels exhibit higher conductivity along the stretch/alignment axis than transverse to it (while simultaneously increasing strength and toughness), further confirming that anisotropy arises from ordered conductive domains^[48].

Incorporating nanofillers or nano-reinforcements provides an effective route to enhance hydrogel mechanical and functional performance. Nanomaterials possess unique physical and chemical properties, such as high surface area, superior mechanical strength, and distinctive optical or electrical characteristics^[77]. Uniform dispersion of nanomaterials (e.g., CNTs^[78], nanofibers^[79], nanosheets^[49]) within hydrogel matrices can significantly improve mechanical strength, conductivity, thermal stability, and bioactivity. Lv et al. proposed incorporating MgFe-layered double hydroxide (LDH) nanosheets into a chitosan/silk fibroin hydrogel to reinforce the polymer network and tune its gelation and mechanical properties [Figure 3E]^[49]. They showed that the nanosheets act as physical crosslinking and load-bearing units, significantly shortening gelation time (from 300 to 146 s), lowering the sol-gel transition temperature (37.4 to 32.7 °C), and enhancing the storage and compressive moduli by 12-fold and 10-fold at optimal LDH content (0.1 wt%). Based on these nanosheet-reinforced mechanical properties, the hydrogel was further demonstrated as an injectable thermos-responsive platform for effective bone regeneration.

Materials used in practical applications are frequently subjected to repeated mechanical deformation, making resistance to mechanical fatigue a key design criterion^[69]. Environmental conditions have been widely shown to exert a strong influence on material performance during cyclic loading. In high humidity environments, water molecules can penetrate polymer networks and act as plasticizers, reducing modulus while enhancing chain mobility^[80]. Although this effect may temporarily improve flexibility, prolonged exposure often disrupts crosslinking structures and accelerates crack propagation. For example, hydrogels typically exhibit good stretch recovery behavior under humid conditions but may suffer from localized dehydration and pore collapse during prolonged compressive cycling, resulting in increased energy dissipation and accumulation of residual strain^[81].

Elevated temperatures intensify thermal motion within polymer networks, potentially leading to chain disentanglement, crosslink cleavage, or phase separation, thereby reducing fatigue life^[82]. In contrast, low temperatures can embrittle soft materials by suppressing energy dissipation mechanisms, increasing susceptibility to brittle fracture^[83]. High salt environments can also significantly influence hydrogel mechanics by screening electrostatic interactions or modifying ionic crosslinking within polymer networks^[84]. However, such salt-induced stiffening is often accompanied by reduced elasticity and flexibility of the hydrogel network.

The core of material fatigue-resistant design is to construct a hierarchical protection system integrating crack suppression, energy dissipation and damage self-repair^[85,86]. While all fatigue-resistant designs aim to restrain crack expansion and reduce cumulative cyclic damage, three mainstream design strategies differ substantially in structural design principles, working mechanisms and fatigue-resistant performances, as summarized in Table 2.

Crystalline domain modulation represents a typical fatigue improvement strategy applicable to crystalline substances and crystalline polymer composites^[86]. Its core design principle focuses on precise regulation of grain size, spatial distribution, and overall crystallinity, and exploits the synergistic interaction between crystalline and amorphous phases to optimize internal stress distribution. This structural optimization strengthens intrinsic crack resistance, realizes efficient strain energy dissipation, and relieves local stress concentration. For instance, in high-molecular-weight PVA hydrogels, cracks grow through the combination of chain sliding and chain breakage. The crystalline domains reduce stress near the crack tip, and the threshold increases with the increase in crystallinity^[87]. At 18.9 wt% crystallinity, the threshold can exceed 1,000 J/m². It should be noted that brittleness dominates and fatigue resistance deteriorates beyond ~70% crystallinity^[87]. Such rigid structural regulation is limited by low ductility, which usually results in relatively low fatigue threshold and rapid crack propagation rate under large cyclic deformation, restricting its application in high-strain fatigue scenarios.

In sharp contrast, internal network topology engineering is the dominant design route for flexible soft materials to achieve excellent fatigue resistance^[88]. Distinct from rigid crystalline reinforcement, this strategy relies on customized network architectures including interpenetrating networks, sliding crosslinking networks, dynamically reversible crosslinked networks, and high-functionality crosslinked networks. Its unique design logic lies in realizing multi-stage energy dissipation, homogeneous stress redistribution and spontaneous microdamage rehabilitation through flexible chain motion and reversible bond dynamics. For example, a double-network elastomer can achieve fatigue thresholds as high as 500 J/m^2[89], but at the cost of substantially lower initial modulus (often < 1 MPa vs. > 10 MPa for crystalline materials). Furthermore, dynamic reversible cross-links further enable micro-damage self-repair under cyclic loading, a capability absent in purely crystalline designs. Benefiting from such flexible structural characteristics, network-regulated soft materials possess outstanding long-cycle fatigue tolerance and slow crack propagation speed, yet their fatigue stability is easily deteriorated under high-temperature and high-load service conditions.

Different from the above two strategies, the fatigue-resistant performance of multi-phase composite materials is dominated by reinforcing phase bridging behavior and interfacial load transfer efficiency^[90]. Its design philosophy centers on constructing stable filler-matrix interfacial interaction: reinforcing phases across crack gaps generate crack closure force and consume interfacial fracture energy to slow down crack growth. Further interfacial modification strategies including bonding strength optimization, gradient interfacial construction and high-energy phase introduction can effectively block crack extension paths and elevate the critical fatigue threshold of composites. In the optimized system, the fatigue threshold has exceeded 1,000 J/m^2[91]. However, unreasonable interfacial matching easily causes interfacial debonding failure, which becomes the main factor restricting the further promotion of comprehensive fatigue performance.

In summary, the three strategies show obvious differences in key fatigue-related quantitative indicators: crystalline regulation achieves high static mechanical strength but poor cyclic fatigue durability; network topology design balances large deformation adaptability and long-cycle stability; composite reinforcement realizes performance complementarity at the expense of increased structural complexity. Recently, emerging advanced material systems integrate the merits of the above design concepts to obtain optimized fatigue resistance under complex service environments, such as dual-network hydrogels, nanofilled elastomers and bionic layered structural materials. Specifically, bidirectional freeze-casting induced layered microstructure provides a feasible route to fabricate isotropic fatigue-resistant materials^[76,92,93]. Hydrogels with ordered 2D layered structures and finely tuned nanocrystalline domains achieve greatly improved comprehensive mechanical properties^[94]. The introduction of rigid nanofillers such as cellulose nanocrystals and MXene can further suppress crack initiation and propagation, endowing materials with reliable fatigue-resistant capability in humid service environments [Figure 4A]^[95,96]. Moreover, gradient crosslinking design optimizes internal stress field distribution and alleviates local stress accumulation, which is an effective auxiliary means to further enhance the cyclic structural resilience of fatigue-resistant materials.

Moisture exchange is a central factor governing the environmental stability of soft materials, particularly hydrophilic systems such as hydrogels and hygroscopic polymers. Repeated drying and wetting cycles often induce irreversible structural damage. Under low humidity conditions, water evaporation leads to volumetric shrinkage, pore collapse, and network densification, resulting in increased Young modulus at the expense of ductility^[97]. This drying-induced contraction is commonly accompanied by interfacial delamination and surface cracking, severely limiting material reusability.

More critically, repeated dry-wet cycling causes cumulative structural degradation. Although rehydration may partially restore volume, incomplete network reconstruction frequently leads to anisotropic swelling, microcrack propagation, and blockage of ion transport pathways. Rapid evaporation can further generate capillary forces that induce macroscopic wrinkling and delamination, undermining structural integrity. Strategies such as incorporating alcohol have been reported to improve water retention in hydrogels [Figure 4B]^[98]. Nevertheless, complete prevention of water loss remains challenging.

To address these issues, several approaches have been proposed, including construction of amphiphilic networks using block copolymers or organic inorganic hybrids to regulate water diffusion kinetics^[99,100], integration of self-healing functionalities through dynamic covalent bonds (e.g., imine or boronate ester linkages) or noncovalent interactions such as hydrogen bonding and metal coordination. to enable post damage repair [Figure 4C and D]^[101-104], and application of encapsulation strategies such as polydimethylsiloxane (PDMS) coatings or microencapsulation to physically shield active layers from environmental exposure^[105,106].

Beyond short term cyclic loading, materials deployed in practical systems must maintain mechanical performance under prolonged static or quasi-static deformation. Most polymer materials, however, exhibit creep and stress relaxation^[107], manifested as progressive deformation under constant load or gradual decay of stress over time, posing challenges for applications requiring dimensional stability or long-term sealing.

Environmental factors further exacerbate these phenomena. Elevated temperatures accelerate molecular chain slippage and enhance stress relaxation rates^[108], while humidity fluctuations alter free volume and dielectric properties, disrupting internal energy dissipation mechanisms^[109]. To improve long term reliability, modern material design increasingly emphasizes multiscale stable network architecture. Crystallizable elastomers, for example, employ reversible crystalline domains as physical crosslinks to preserve modulus even at elevated temperatures^[110]. Shape memory polymers provide the ability to recover original shape upon external stimulation, extending effective service life^[111]. In addition, environmentally responsive crosslinking modulation, such as photo triggered or pH sensitive networks, enables mechanical reinforcement and opens new opportunities for adaptive material systems [Figure 4E and F]^[112,113].

Ionic conductive hydrogels conduct electricity through the directional migration of mobile ions within their hydrated polymer networks. Owing to their high water content, ions such as H⁺, OH^-, Na⁺, and Cl^- are readily generated by water dissociation or polymer ionization^[124]. Under an external electric field, these ions migrate along concentration or potential gradients, producing ionic current. For example, in sodium polyacrylate-based hydrogels, partial dissociation of sodium carboxylate groups releases Na⁺, and carboxylate counterions, which act as charge carriers within the three-dimensional network^[125]. Ionic conductivity is primarily governed by ion concentration, network architecture, and temperature. Higher ion density increases the number of charge carriers, while a loose and porous network provides low-resistance pathways for ion diffusion. Elevated temperature enhances ionic mobility by accelerating segmental motion and diffusion, although excessive heating may destabilize the hydrated structure and compromise conductivity^[126]. Importantly, because ionic transport depends on hydrated channels, microcracks or dehydration induced by mechanical fatigue can interrupt ion pathways, linking conductivity decay directly to structural durability.

Electronically conductive hydrogels rely on electron transport through embedded conductive phases within an otherwise insulating polymer matrix. This is commonly achieved by introducing conductive fillers such as CNTs^[78], graphene^[127], or metallic nanoparticles^[128], that form percolated networks or continuous conductive pathways. When a sufficient filler concentration is reached, electrons can hop between adjacent fillers or travel along interconnected networks under an electric field. For instance, graphene sheets dispersed in a PVA hydrogel can overlap to form conductive channels, enabling efficient electron transport^[127]. The conductivity of electronically conductive hydrogels depends strongly on filler type, loading level, dispersion quality, and interfacial interactions with the polymer network. Well-dispersed fillers with strong interfacial bonding facilitate continuous conductive pathways, whereas aggregation or poor compatibility disrupts electron transport. In addition, the hydrogel’s network flexibility and porosity influence the stability of conductive networks during deformation. Large strains or repeated fatigue cycles may fracture conductive bridges, showing how mechanical resilience and electrical performance are inherently coupled.

Hybrid conductive hydrogels combine ionic and electronic conduction mechanisms by simultaneously incorporating mobile ions and electronically conductive components within a single hydrogel matrix. In such systems, ionic species migrate through hydrated polymer domains, while electrons are transported through interconnected conductive networks^[129]. For example, sodium polyacrylate can provide ionic conduction via dissociated Na⁺ ions, whereas polyaniline (PANI) embedded in the same matrix enables electronic conduction through percolated networks^[130]. The coexistence of these two charge transport pathways results in synergistic conductive behavior. The relative contribution of ionic and electronic conduction can be tuned by adjusting the composition ratio, dispersion state, and interactions between the ionic and electronic components. Environmental factors such as temperature and humidity further modulate performance by influencing ion mobility and the integrity of electronic networks. While dual pathways provide redundancy, mechanical damage such as crack propagation or network collapse can simultaneously impair both ionic and electronic channels, underscoring the need for architectures that balance conductivity with fatigue resistance^[131].

In summary, fatigue performance and environmental stability of soft materials in humid, thermal, and saline environments are governed by the coupled effects of moisture migration, ionic interactions, and thermally activated processes. Although substantial progress has been achieved in enhancing fatigue resistance and environmental adaptability, several challenges remain. These include achieving simultaneous optimization of conductivity, flexibility, and durability under high humidity and high salt conditions, establishing standardized metrics for quantitative evaluation of structural reversibility during dry-wet cycling, and developing in situ monitoring techniques capable of tracking aging mechanisms during long term operation.

Future research should focus on intelligent and adaptive material systems supported by in situ characterization tools such as Raman spectroscopy and atomic force microscope-infrared (AFM IR) spectroscopy to elucidate microstructural evolution during service^[132,133]. Integration of machine learning models to predict lifetime performance based on coupled environmental and mechanical inputs will further enable rational design and engineering scale deployment of durable, high-performance soft functional materials.

The long-term performance of soft hydrogels in bio-integrated devices depends strongly on interfacial adhesion to skin, tissues, or soft robotic structures^[2]. Hydrogels are highly hydrated networks with low modulus and high permeability, yet they must interface with biological substrates that are persistently wet, dynamic, and chemically heterogeneous^[129]. Without proper interface matching, even advanced bulk materials cannot maintain stable system functions during real operation^[176].

A major challenge arises from insufficient adhesion in fluid-dominated environments such as sweat, gastric acid, and tears^[177]. Mechanical mismatch between hydrogel layers and viscoelastic tissues further accelerates delamination under cyclic motion or long-term wear^[178]. Many commercial packaging films are either fully impermeable, causing irritation, or contain reactive glues that raise safety concerns^[179]. The objective of interface design is therefore to construct bonding zones that are robust, reversible, and biocompatible, enabling conformal integration of hydrogels under physiological conditions.

Physical strategies are inspired by the way organisms maintain attachment on naturally moist boundaries^[180]. Hierarchical micro and nanostructures, exemplified by gecko-like fibrillar arrays^[181] and crab claw-resembling interlocking textures^[182], create multipoint contact and efficient energy dissipation to help soft materials adapt to irregular tissue surfaces. For example, Hu et al. fabricated bioinspired hierarchical architectures that significantly improved contact-sensible adhesion performance under dynamic loading [Figure 7A]^[183]. These architectures convert unavoidable fluid layers such as sweat^[94], gastric juice^[184], or tears^[185] into supportive liquid bridges, increasing peeling resistance while alleviating mechanical mismatch. Consistent with this concept, Yuk et al. reported a tough hydrogel adhesive that leveraged interfacial water rather than excluding it, achieving durable adhesion on wet biological tissues and implanted devices [Figure 7B]^[186]. Li et al. engineered a polyphenol-based adhesive hydrogel capable of forming dynamic hydrogen bonding and metal coordination with tissue surfaces, enabling repeatable wet adhesion and on-demand debonding [Figure 7C]^[187]. In clinical practice, patterned films^[188] and porous tapes^[189] based on this philosophy have been explored to improve the comfort and removal safety of wound dressings and wearable sensors.

Chemical inspirations provide a complementary molecular route^[190]. Networks rich in hydrogen bonds and aromatic groups interact with interfacial proteins and mucin glycoproteins, resembling transient biological bonding^[191,192]. Polyphenols illustrate how metal coordination and dynamic covalent bonds can reinforce adhesion without sacrificing tolerance, allowing application-oriented balance between residence and safe removal^[193]. Such adhesion must be defined by therapeutic outcomes rather than maximal strength at all costs.

Many conventional hydrogels rely on ester- or amide-based cross-linking pathways that can be vulnerable to hydrolysis in strongly acidic or high-salt media. Cleavage of cross-linking nodes within the backbone leads to decline in modulus, excessive swelling, and eventual structural failure. For example, Sun et al. showed that ionically crosslinked hydrogels rapidly lost mechanical integrity and adhesion under high ionic strength conditions due to charge screening and network dissociation^[62]. Protonation and ionic screening at tissue interfaces further weaken electrostatic attraction and hydrogen-bonding, giving rise to short-lived adhesion and delamination^[194]. These phenomena are particularly problematic in the gastrointestinal tract, sweat-rich chronic wounds, and the ocular surface dominated by tear fluids^[195]. Addressing this challenge, Yang et al. developed a mucus-inspired hydrogel with protonation-driven adhesion that remained structurally stable and strongly adhesive under highly acidic conditions (pH ≈ 2), enabling effective esophageal tissue repair in vivo [Figure 8A]^[196]. Therefore, reliable performance requires dual or dynamically reinforced networks incorporating chemically inert motifs, so that resistance to chemical attack and maintenance of interfacial coupling can be achieved simultaneously.

Interface hydration must be preserved during exposure to air, as dehydration alters ionic mobility, interfacial impedance, and consequently sensing accuracy^[197]. Along this line, Yeon et al. demonstrated that sweat pore-inspired perforated e-skins effectively mitigate sweat accumulation at the skin-device interface, thereby preserving conformal contact and long-term sensing stability during daily activities [Figure 8B]^[198]. Two complementary strategies are commonly used. Breathable yet waterproof polymeric membranes serve as physical barriers to evaporation while permitting gas and metabolite diffusion^[199]. Hydrophobic or mucus-mimicking surface coatings provide additional regulation of interfacial water activity, enabling consistent permeability and signal stability under long-term wearing^[200]. The combination of encapsulation engineering and biomimetic skins has emerged as a practical design principle for translating soft hydrogels into continuous monitoring platforms.

Bio-interfaces are evolving from single adhesive layers to hubs for electrical and therapeutic exchange^[201]. Conductive zones must support reliable percolation for sensors, while elastic networks demand low filler loading to preserve flexibility. Integration introduces challenges including sensing crosstalk and mismatch with flexible electronics or clinical workflows^[202]. Therefore, molecular composition, microstructure, and encapsulation should be co-optimized to ensure predictable transport and stable system compatibility.

Progress toward clinical translation will rely on three dimensions. Programmable stimuli-responsive interfaces with reversible adhesion will address dynamic wet boundaries. AI-assisted prediction of interfacial lifetime could reduce empirical trial in material selection. Standardized testing protocols and long-term durability assessments will enable cross-laboratory comparison and regulatory evaluation. Convergence of these routes will determine how bioinspired soft interfaces mature into reliable wearable and implantable technologies.

By incorporating photothermal nanomaterials (such as Mxene, polydopamine, and gold nanorods), hydrogels can efficiently convert light energy of specific wavelengths (primarily near-infrared) into heat^[203], thereby inducing localized thermally triggered volume phase transitions, which in turn generate programmed bending, curling, or grasping motions^[204,205]. Inspired by the phototropic behavior of sunflowers, Qin et al. developed a bionic sunflower using a light-responsive soft hydrogel that autonomously bends towards a light source, enabling a photocatalytic disk to maintain near-vertical alignment for highly efficient H₂O₂ production by maximizing light energy capture^[206]. Thermoresponsive hydrogels, exemplified by poly(N-isopropylacrylamide) (PNIPAM, also known as PNIPAAm), undergo volume transitions driven by the temperature-dependent balance between hydrophilic and hydrophobic interactions of polymer chains^[207,208]. Chung et al. discovered that incorporating NaClO₄ during polymerization creates phase-separated, porous structures in PNIPAAm hydrogels, enabling ultrafast and large volume shrinkage upon heating. This significantly enhanced thermoresponsiveness, which is critical for applications like soft actuators and drug delivery systems, overcomes the slow deformation limitation of conventional hydrogels [Figure 9A]^[209]. Furthermore, aligning superparamagnetic nanoparticles (e.g., Fe₃O₄) within hydrogel networks enables heat generation under alternating magnetic fields or direct actuation under gradient magnetic fields^[210,211]. Preprogramming the spatial distribution of magnetic particles allows complex three-dimensional deformations and directional motion^[212], offering great potential in targeted drug delivery and wireless control of microrobots. Additionally, constructing ionic concentration gradients or incorporating electroactive polymers within hydrogels allows electric fields to induce ion migration, local pH changes, or redox reactions of polymer chains, resulting in asymmetric swelling or shrinkage. This mechanism provides rapid responsiveness, making it suitable for artificial muscles and microfluidic valves^[213].

Chemical responsiveness enables hydrogels to deeply interact with biological microenvironments^[214,215]. Networks containing weak acidic/basic groups (e.g., carboxyl groups in polyacrylic acid (PAA) or amino groups in chitosan) exhibit significant swelling changes in response to environmental pH^[216,217]. For example, Zhang et al. developed a polyampholyte hydrogel capable of spontaneous shape-memory cycles and macroscopic motion in response to alternating acidic/alkaline solutions, demonstrating its potential as a soft actuator [Figure 9B]^[33,218]. Recently, Yao et al. developed a solvent-adaptive hydrogel with a layered, constrained honeycomb structure [Figure 9C]^[219].

Volume change represents the most fundamental stimulus-responsive behaviour of hydrogels, primarily driven by abrupt variations in osmotic pressure within the network or significant alterations in polymer-solvent interactions^[220]. To transform uniform volumetric changes into directional, programmable deformations (such as bending, twisting, or curling), asymmetry or gradient structures must be introduced within the hydrogel to generate internal stress differences under stimulation^[221,222]. For example, Wu et al. researched a novel soft hydrogel system where in-situ proton transfer enables programmable shape-morphing under light stimulation, independent of external water exchange^[223]. Patterned printing of hydrogels with distinct responsiveness enables programmed two-dimensional to three-dimensional deformations^[224]. Chen et al. developed a multiprogrammable anisotropic soft hydrogel by embedding magnetically oriented ferromagnetic nanoplates, which exhibits anisotropic volume response with distinct swelling ratios parallel and perpendicular to the alignment direction. This deliberate structural ordering enables the programmable integration of optical, mechanical, and magnetic properties for advanced soft actuators^[225]. A typical example involves aligning magnetic nanoparticles along a single direction and fixing them within the network, so that the hydrogel bends predominantly in the vertical direction under magnetic or photothermal actuation^[225]. Such preprogrammed anisotropy provides a powerful tool for achieving complex and predictable deformations.

Building on the concept of volume-dependent transformations, hydrogels can also demonstrate color responses through structural coloration mechanisms. The color response of hydrogels (structural coloration) does not originate from chemical dyes but rather from interference, diffraction, or scattering of light by their internal periodic nanostructures^[226,227]. This coloration dynamically adjusts with changes in nanoscale spacing. For example, Chen et al. developed a photochromic organohydrogel that exhibits distinct color changes under UV light due to the reversible molecular conformational switching of naphthalene (Np) groups, enabling dynamic optical properties [Figure 9D]^[228]. When stimuli induce uniform swelling or shrinkage of the hydrogel^[229], the lattice constant (interparticle spacing) changes accordingly, resulting in shifts in the wavelength of Bragg-diffracted light. Macroscopically, this manifests as continuous and reversible color changes. The quantitative relationship between color and stimulus intensity makes such hydrogels ideal candidates for visual sensing applications^[229].

Expanding beyond optical responses, hydrogels can also exhibit photoelectric effects for energy conversion and sensing. Photoelectric response refers to the ability of materials to generate or modulate electrical signals (current, voltage, or resistance) under illumination^[230]. Achieving efficient photoelectric response in hydrogels requires constructing a complete energy conversion pathway, from light harvesting to charge generation and subsequent charge transport^[231]. Wu et al. developed a self-powered photoelectric sensor based on soft hydrogel diodes doped with photoacid, where the photoelectric response mechanism involves light-induced proton release from the photoacid and subsequent ion migration driven by the built-in electric field of the hydrogel PN junction^[232]. The integrating photoelectric units into hydrogel networks could enable self-powered, soft bio-interfaces for wearable sensors or devices.

The central mission of flexible electronics and e-skin is to endow machines or prosthetics with sensory capabilities comparable to, or even surpassing, those of natural skin^[251]. High-precision and reliable detection of diverse mechanical signals, including stretching, compression, bending, and their derivatives such as pressure and strain, is fundamental to achieving tactile perception, physiological monitoring, and motion feedback^[252]. Owing to their soft mechanical properties matching biological tissues, tunable conductivity, and outstanding fracture toughness, soft hydrogels have emerged as ideal materials for constructing the next generation of durable flexible mechanical sensors^[253,254]. Based on their working mechanisms, these sensors can be categorized into force/pressure sensors, strain sensors, and multifunctional integrated sensors, each designed to accurately detect tension, compression, bending, and pressure signals^[255-258].

Force and pressure sensors are primarily employed to detect normal forces or pressures perpendicular to the device surface^[255,256]. Wang et al. developed a multimodal hydrogel-based sensing system that integrates EMG and force myography (FMG) signals, leveraging a macroporous Foam-PAM hydrogel pressure sensor to detect subtle muscle forces for active rehabilitation in human-machine interfaces [Figure 9E]^[259]. In addition, Rahman et al. discovered that incorporating zeolitic imidazolate framework-8 (ZIF-8) nanoparticles into a poly(acrylamide)-co-hydroxyethyl acrylate (PAAm-co-HEA) hydrogel significantly enhances its force-responsive properties, enabling the fabrication of a highly stretchable triboelectric nanogenerator (TENG) that effectively converts biomechanical forces into electrical signals for self-powered sensing and energy harvesting^[260]. Strain sensors are used to monitor stretching, compression, or bending of surfaces, with outputs typically correlated to strain (ε)^[261]. Tensile strain detection is predominantly piezoresistive^[262], when hydrogels are stretched, internal conductive pathways (such as ionic channels or networks of conductive nanofillers) are geometrically elongated and cross-sectional areas reduced, leading to increased resistance (positive strain effect). Soft hydrogels can withstand reversible tensile strains, thereby offering a broad detection range^[254]. Their high toughness ensures that even under extreme stretching or localized damage, conductive networks remain continuous, allowing signals to recover. Compression sensing mechanisms overlap with those of normal pressure detection but emphasize volumetric deformation. Porous or foam-like hydrogels are particularly advantageous^[263,264], upon compression, pore closure densifies conductive pathways, reducing resistance, while capacitance characteristics also change. Bending can be regarded as a composite state of tensile strain on one side and compressive strain on the other^[261]. Sensors are typically attached to bendable surfaces, with resistance changes reflecting curvature. Critical factors include interfacial adhesion^[265,266], ensuring hydrogels deform cooperatively with substrates without delamination.

Key performance indicators for sensors include sensitivity, detection range, response time, and stability^[267-269]. Sensitivity quantifies the ability of sensors to convert mechanical stimuli (e.g., strain or pressure) into electrical signals (e.g., resistance or capacitance changes), determining resolution and amplification of small stimuli^[267,269]. The high sensitivity of soft hydrogel sensors arises from the synergistic interplay between microstructural features and macroscopic mechanical properties^[270]. Detection range is defined by the minimum (lower limit) and maximum (upper limit) stimuli that elicit effective responses. The broad elasticity and structural integrity of soft hydrogels provide a physical basis for extending detection ranges. For instance, a highly cracked hydrogel design that achieves multi-sensing with high sensitivity and a large detection range for soft machines by utilizing the opening and closing of pre-cut cracks to significantly alter current flow^[271]. Response time characterizes the ability of sensors to track dynamic or transient mechanical stimuli^[272], which is crucial for monitoring vibrations, speech, high-speed impacts, or physiological pulses. Benefiting from low hysteresis and porous architectures^[263], soft hydrogels typically exhibit superior performance. Stability reflects the ability of sensors to maintain functionality under repeated use and environmental fluctuations, representing a core advantage of soft hydrogels^[269]. Their energy dissipation mechanisms, fatigue resistance, and synthetic strategies conferring environmental robustness collectively determine sensor lifespan and reliability. Despite their excellent performance in mechanical sensing, soft hydrogel sensors still face multiple challenges in practical applications^[83,273,274]. Conventional hydrogels undergo significant water-induced swelling in aqueous environments, leading to structural deformation, mechanical degradation, and signal instability. This limits their reliable use in complex conditions such as underwater sensing or high-humidity environments. In contrast to swelling, hydrogels are also prone to rapid moisture loss in ambient or dry conditions, leading to dehydration within hours. This compromises their mechanical integrity, ionic conductivity, and sensing functionality. Practical applications demand a synergy of features, such as mechanical robustness, high conductivity, self-adhesion, self-healing, biocompatibility, and antimicrobial properties, which are challenging to incorporate into a single hydrogel system.

One of the primary objectives of flexible sensors based on soft yet robust hydrogels is the construction of biomimetic, multifunctional e-skin^[275,276]. An ideal e-skin system must not only transcend the detection of single mechanical signals but also achieve multimodal environmental perception^[277]. Moreover, as an interface in long-term, intimate contact with biological systems, it must exhibit excellent wearing comfort, conformability, and biocompatibility. Owing to their tunable physicochemical properties, mechanical compatibility with tissues, and inherently hydrophilic interfaces, soft hydrogels represent an ideal material platform for this purpose. Recent advances can be reviewed from three perspectives, multimodal sensing integration, conformability, and biocompatibility.

E-skin must emulate the ability of natural skin to perceive multidimensional information, integrating synchronous or selective responses to stimuli such as temperature, humidity, chemical composition, and tactile signals. The design versatility of soft hydrogels provides unique advantages. Temperature sensing can exploit the temperature dependence of ionic conductivity in hydrogels (following the Arrhenius relationship), enabling a single conductive hydrogel to achieve self-sensing of temperature while monitoring mechanical signals. The working mechanism introduced by Ren et al. involves a temperature-sensing schematic where variations in temperature directly influence electrical resistance [Figure 9F]^[278]. This alteration in resistance is subsequently detected by an external circuit, facilitating temperature measurement. This well-defined linear dependence enables precise discrimination between different temperature levels. Humidity sensing relies on the hygroscopic nature of hydrogels, with electrical properties (e.g., ionic conductivity and dielectric constant) closely correlated to ambient humidity^[279]. Ding et al. developed a groundbreaking self-powered, flexible chemosensor by engineering a smart hydrogel incorporated with zinc and lithium ions, enabling it to reversibly switch between water-rich and water-deficient states^[280]. This single device achieves crosstalk-free detection of both oxygen and humidity with remarkable sensitivity (up to 4,170.5 %/% for O₂ and 380.2 %/% RH for humidity) by leveraging a metal-air battery structure where the target molecules selectively act as limiting reactants or catalysts in the oxygen reduction reaction depending on the hydrogel’s state. Chemical sensing can be achieved by immobilizing specific recognition elements (e.g., glucose oxidase, aptamers, or molecularly imprinted polymers) within hydrogel networks. Inspired by human skin, Yun et al. developed a novel hydrogel-based sensing platform for two-dimensional imaging of external stimuli like chemicals^[281]. This design enables high sensitivity 2D imaging through electrochemiluminescence, representing a significant advance toward biocompatible e-skins that leverage chemical reactions for multi-stimuli detection. Tactile sensing (pressure, strain, shear force) forms the foundation of multimodal perception, yet signal decoupling remains a central challenge. Tao et al. developed a self-powered tactile hydrogel sensor featuring a micro-pyramid-patterned double-network ionic organohydrogel, which detects subtle pressure changes through triboelectric signals without an external power supply^[282]. This sensor exhibits remarkable sensitivity (45.97 mV·Pa^-1), a fast response time (~20 ms), and operates reliably across a broad temperature range (-20 to 60 °C). Furthermore, its successful integration into a system that controls electronic devices and a robotic hand by mimicking finger gestures demonstrates significant potential for wearable electronics and human-machine interfaces. Comfort and skin-fitting characteristics are essential requirements for e-skin. Soft hydrogels, through precise tuning of mechanical properties and interfacial characteristics, provide key solutions to this requirement^[283]. Chen et al. developed a wet-adaptive electronic skin (WADE-skin), highlighting that maintaining biological comfort at the device-skin interface in wet environments is crucial for the long-term, imperceptive wearability of epidermal electronics^[284]. Zhang et al. developed a porous thermoplastic polyurethane (TPU)/carbon black hydrogel sensor via water vapor-induced phase separation, which exhibited high water vapor transmission rates (≈ 3,800-4,100 g·m^-2·day), enabling comfortable wear while maintaining multimodal sensing capabilities^[285]. Furthermore, Zhou et al. researched a strain sensor with an ultra-wide sensing range by forming a porous conductive network in a CNT/TPU composite through salt leaching, enhancing both mechanical and breathable properties^[286].

Biocompatibility is a paramount requirement for hydrogel-based e-skin to ensure safe, long-term wearability and prevent skin irritation or inflammatory responses, which are common issues with impermeable materials. Ma et al. developed a highly permeable and superelastic liquid-metal fibre mat (LMFM) that demonstrated excellent biocompatibility, as confirmed by in vitro cell viability tests showing over 95% cell survival and in vivo tests on rabbit and human skin showing no significant irritation, making it ideal for monolithic, multi-layered e-skin devices^[287]. Additionally, Yan et al. developed conductive cellulose-based bio-nanosheet hydrogels by using polydopamine-reduced graphene oxide (GO) as a template for cellulose assembly. This approach resulted in a hydrogel that exhibits remarkable stability under physiological conditions while maintaining excellent electrical functionality and the ability to support cell growth, showcasing its high potential for creating long-term, bio-integrated electronic devices^[288].

Soft actuators can be categorized based on stimulus type, including electrostatic, electroosmotic, thermal, photoinduced, moisture-responsive, and ion-driven systems.

Electrostatic actuation, exemplified by dielectric elastomer actuators (DEAs), relies on Maxwell stress generated by electric fields across dielectric layers between compliant electrodes^[293]. Voltage application compresses the thickness and expands the in-plane area, enabling large strains^[294]. These actuators offer high energy density, fast response, and silent operation, making them ideal for artificial muscles and dynamic morphing structures [Figure 10A]^[295], but often require high driving voltages, raising challenges for portability and safety, and may suffer from electrical breakdown and instability.

Electroosmotic actuation exploits the movement of ions and solvent in porous or hydrogel matrices under electric fields [Figure 10B]^[296]. It operates at low voltage and is suitable for micro-robot and microfluidic applications but can be limited by diffusion-constrained kinetics and modest force output. Thermally induced actuation relies on thermal expansion mismatch or phase transitions in materials such as shape memory polymers^[297], liquid crystal elastomers^[298], and bimorph composites^[299]. Liquid crystal elastomers can contract anisotropically along mesogen alignment above phase transition temperatures, resembling muscle behavior [Figure 10C]^[300]. Thermal approaches provide large stroke and force but can be slow and energy-intensive. Photothermal agents such as CNTs, graphene, and gold nanoparticles can enable remote and spatially controlled heating under light^[301]. Photoinduced actuation enables wireless control by converting light into mechanical work. Direct systems use molecular photoswitches such as azobenzene through trans cis isomerization^[302], while indirect systems combine light absorbers with thermoresponsive matrices to achieve noncontact actuation^[303]. These strategies are attractive for tether-free biomedical operation. Moisture-responsive actuation relies on differential swelling of hygroscopic materials such as cellulose under humidity gradients^[304]. Biomimetic designs inspired by seed dispersal mechanisms can generate coiling, twisting, crawling, and other motions^[305,306]. Such actuators can operate autonomously but may face slow response, limited reversibility in dry conditions, and durability concerns. Ion-driven actuation includes ionic polymer metal composites and conductive polymer actuators^[307]. In ionic polymer metal composites, cation migration under low voltage bias causes hydration gradients and bending. These systems offer flexibility and bidirectional control^[308] but may be limited by the solvent in electrolyte evaporation, electrode delamination, and finite lifetime.

Structural architecture strongly influences motion richness and controllability. Advances in fabrication, including multi-material three-dimensional printing, soft lithography, and kerygma and origami-inspired patterning, have enabled sophisticated geometries. Basic kinematic modes include bending, contraction, and elongation, and twisting and helical motion. Bending can be achieved through asymmetric material distribution or gradient crosslinking and is widely used in grippers and locomotive units^[309]. Contractile actuators based on pneumatic networks, DEAs, or liquid crystal elastomers generate linear motion for propulsion^{[294,298,309]}. Twisting and helical motion can be produced through chiral winding or anisotropic alignment, enabling rotary propulsion in swimming microrobots^[310,311]. Programmable motion can be achieved by patterning actuation domains for sequential or selective activation. A dual stimuli responsive hydrogel combining light and humidity sensitivity demonstrated autonomous sequence execution through environmental feedback^[312]. Electrode arrays integrated into DEA membranes can generate traveling wave deformation^[313]. Biomimetic structural designs include octopus inspired grippers with tapered chambers for adaptive grasping [Figure 10D]^[314] and vine tendril inspired elongating robots using localized swelling for tip extension [Figure 10E]^[315]. Layered architectures integrating sensing layers, actuator layers, and reinforcement elements enable closed loop systems with proprioception and adaptive response^[316]. Embedded sensors based on stretchable conductors such as liquid metals and conductive hydrogels can monitor strain and pressure for feedback control^[317,318]. A soft gripper equipped with resistive strain sensors, for example, can adjust gripping force based on object compliance, preventing damage^[319].

One particularly promising application of hydrogel actuators is in relatively complex multi-arm robots or grippers that can operate in a variety of environments^[320-322]. These hydrogel actuators can grasp objects when exposed to certain stimuli and release them. Such responsive actuators are now extensively used as intelligent components in robotic systems, flexible sensing devices, and mobile medical or surgical biopsy instruments, largely due to their ability to be programmed into specific shapes. Although this line of application is still emerging, soft grippers that can grip and release objects are among the most promising developments. For example, to expand the range of robotic hand applications, by incorporating GO and polypyrrole (PPy), Dong et al. developed a programmable, patterned GO/PPy dual-responsive hydrogel actuator that responded efficiently to both temperature and infrared irradiation^[323]. Remarkably, the gripper can lift objects 38 times its own weight, mimicking the way an eagle catches its prey. This research offers an effective approach for developing hydrogel actuators and paves the way for potentials in intelligent actuator technology.

A similar humidity-responsive hydrogel actuator was developed by Yao et al., who employed an Eu³⁺ ionic polyethyleneimine–acrylic acid copolymer (PEI-co-PAAC) in combination with poly(ethylene glycol) diacrylate (PEGDA)^[324]. This actuator exhibited sensitivity to humidity and was capable of object detection and identification under ultraviolet light. Leveraging these properties, they designed a hexagonal hydrogel inspired by jellyfish camouflage, enabling the gripping and release of objects underwater. Zhao et al. reported another 4D printing innovation: an asymmetric bilayer-structured hydrogel actuator^[325]. When gradually immersed in heated water, the actuator would catch and securely wrap a rubber block attached to a rope as immersion time increased. The wrapped object could then be moved by pulling the rope. Additionally, the actuator’s bending and releasing mechanism was adapted into bionic plant flower and hydrogel manipulator designs, making it easy to open/close or grasp/release target objects in response to thermal stimulation. Building on these strategies, researchers have recently developed even more advanced hydrogel actuators. These new designs function as intelligent fixtures, offering solutions that overcome the limitations of traditional fixtures, especially in challenging environments.

Thermoresponsive hydrogels are highly promising for use in soft actuators because they can reversibly change shape in response to temperature fluctuations^[326,327]. Recent advance includes the development of tough, rapidly responding thermoresponsive DN hydrogels engineered for soft actuators by Zhang et al. The resulting hydrogels exhibit excellent mechanical performance, with an ultimate compressive stress of approximately 8 MPa^[328]. They also display rapid actuation, reaching about 30% linear contraction and 28% radial contraction within 2 min. Furthermore, tubular soft actuators can serve as fluidic temperature sensors, switching the direction of fluid flow due to temperature variations.

The integration of intelligent control and feedback systems marks a major advance forward in the development of hydrogel-based soft robots. By embedding sensors and control units directly into the soft robotic structure, it becomes possible to continuously monitor key parameters such as fatigue, deformation, and actuation response in real time^[329,330]. This closed-loop feedback allows for adaptive adjustments to actuator performance, ensuring reliable operation even in complex and changing environments. These smart systems not only improve the durability and functional reliability of hydrogel actuators but also broaden their potential for advanced applications in robotics, flexible sensing, and minimally invasive medical devices. Ultimately, combining responsive hydrogel materials with advanced control strategies sets the stage for the next generation of autonomous, adaptive, and multifunctional soft robotic systems^[331,332]. For example, Huang et al. developed a foldable inductive sensor that could be integrated with an origami-inspired actuator^[333]. With an increasing number of folds, the coil inductance decreased, leading to stable, sensitive, and highly repeatable performance, providing timely and reliable proprioceptive feedback for the gripper^[333]. The final smart jellyfish design featured both buoyancy control and grasping capabilities under the water. In another example, Jin et al. developed a gas-driven flexible caterpillar robot equipped with advanced sensing capabilities^[334]. The tactile sensors achieve a detection limit as low as 0.05 kPa along with an ultrafast response time of just 0.03 s. Meanwhile, the resistive strain sensors provide a sensitivity of 2.94 and can stretch up to 180%, allowing them to fully accommodate the robot’s body bending.

Quadrupedal soft robots also demonstrate unique advantages, such as rapid crawling speeds and versatile movement patterns, making them well-suited for tasks like navigation, obstacle avoidance, and exploration. For example, Xu et al. introduced a multifunctional, controlled-buckling sensor for monitoring the locomotion of a quadruped soft robot^[335]. The robot’s body was constructed to be independently actuated along the grid in two directions, enabling movement within the first quadrant of a two-dimensional surface. In another work, Tang et al. advanced the field further by using 3D printing to create magnetic hydrogel soft structures capable of complex shape changes^[336]. By incorporating a hard-magnetic filler into the hydrogel matrix and employing nano-colloids to tune the rheology of the precursor, they were able to directly print intricate magnetic hydrogel shapes. These printed components were assembled into a 3D-printed magnetic hydrogel lotus, which is fully bloomed without a magnetic field. Under a uniform magnetic field, the lotus spreads flat; an upward field causes it to close completely, while a downward field induces opening.

Soft hydrogels are pivotal for wearable biosensors, offering skin-like compliance, high signal fidelity, and multimodal sensing capabilities by seamlessly integrating with biological tissues. In recent years, research on wearable soft hydrogel biosensors has shifted from merely enhancing mechanical performance to achieving more stable, precise, and intelligent health signal acquisition. Pan et al. developed a MXene-functionalized PEDOT:PSS conductive hydrogel for non-invasive sweat glucose monitoring, demonstrating high conductivity and stability enabled by the material’s tailored porous network [Figure 11A]^[337]. Similarly, Zhao et al. engineered a dual-network hydrogel architecture with allylated hydroxyethyl cellulose with poly(3,4-ethylenedioxythiophene):poly(sodium-p-styrenesulfonate) and PAM, achieving remarkable pressure sensitivity and fast response for human motion tracking without external power^[338]. Additionally, Wang et al. designed a rapid-polymerization hydrogel exhibiting autonomous self-healing and stable adhesion, allowing reliable operation under mechanical deformation and varying environmental conditions^[339]. Furthermore, Shan et al. created an injectable hydrogel with integrated bacteria theranostic and motion monitoring functions, enabling real-time detection of joint movements and antibacterial protection for wound healing applications [Figure 11B]^[340]. These advances highlight how soft hydrogels, through rational multiscale design, are evolving from passive conductors to intelligent, multifunctional interfaces for continuous health monitoring. Despite these advances, soft hydrogel-based wearable biosensors are hindered by inherent material vulnerabilities, including environmental instability and signal interference from calibration drift and biofouling^[341,342]. These limitations collectively challenge their long-term reliability and commercial viability for continuous health monitoring.

Multimodal biosensing integrates diverse physiological signals to provide comprehensive health monitoring, enabling real-time assessment of both biochemical and biophysical parameters for personalized healthcare. Soft hydrogels, with their intrinsic flexibility, biocompatibility, and tunable conductivity, serve as ideal platforms for developing wearable sensors that can conform to biological tissues and facilitate simultaneous data acquisition from multiple sources. For instance, Imani et al. developed a hybrid chemical-electrophysiological sensing system that concurrently monitors sweat lactate and electrocardiogram (ECG) signals, utilizing hydrogel-based interfaces to ensure skin compatibility and stable signal transmission during physical activities^[343]. Lin et al. developed a non-invasive sweat glucose sensor that concurrently monitors ECG signals, demonstrating the integration of chemical and electrophysiological sensing in a single wearable platform. This hybrid system allows for the correlation of metabolic activity with cardiac function, providing a more holistic health assessment^[344]. Furthermore, Liao et al. engineered a programmable microfluidic-assisted hydrogel patch with a silver-lignin-PAM composite for monitoring physiological signals like heart rate and pulse. This approach highlights the use of soft hydrogels in creating structured, multifunctional sensing interfaces through advanced fabrication techniques [Figure 11C]^[345]. However, the broader deployment of soft hydrogel-based multimodal sensing systems is currently challenged by achieving environmental stability and integrated multimodal sensing for reliable clinical translation.

Soft hydrogels serve as an ideal platform for AI-driven health monitoring by providing high-quality, multimodal physiological data through their biocompatible and conformable interfaces. The integration of deep learning algorithms with these sensors addresses challenges in processing complex, noisy data, enabling intelligent diagnosis and prediction. Guo et al. developed a deep learning-assisted triboelectric hydrogel sensor that achieved 100% recognition accuracy in monitoring infant movements by analyzing signals from various activities. This system enables real-time safety monitoring through wireless data transmission to a mobile device [Figure 11D]^[346]. Li et al. engineered a hydrogel-based optical waveguide stretchable sensor integrated with a convolutional neural network to differentiate between normal and abnormal gait patterns for remote patient monitoring. This smart tele-healthcare system facilitates multimodal human-machine interaction, serving as a communication tool for patients with speech impairments [Figure 11E]^[347]. However, a key limitation in current systems is the need for extensive, individualized data pre-training to overcome signal variability and achieve robust model generalization for reliable clinical insights.