Recent advances in multimodal sensing integration and decoupling strategies for tactile perception

Human skin perceives external environmental stimulus by the synergies between the subcutaneous tactile corpuscles. Soft electronics with multiple sensing capabilities by mimicking the function of human skin are of significance in health monitoring and artificial sensation. The last decade has witnessed unprecedented development and convergence between multimodal tactile sensing devices and soft bioelectronics. Despite these advances, traditional flexible electronics achieve multimodal tactile sensing for pressure, strain, temperature, and humidity by integrating monomodal sensing devices together. This strategy results in high energy consumption, limited integration, and complex manufacturing process. Various multimodal sensors and crosstalk-free sensing mechanisms have been proposed to bridge the gap between natural sensory system and artificial perceptual system. In this review, we provide a comprehensive summary of tactile sensing mechanism, integration design principles, signal-decoupling strategies, and current applications for multimodal tactile perception. Finally, we highlight the current challenges and present the future perspectives to promote the development of multimodal tactile perception.

Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence.Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.

Introduction
Human skin, as the largest somatosensory organ of human body, contains various types of tactile receptors that enable multimodal awareness of the environment, including materials recognition, temperature perception and object identification [1][2][3].Though human skins cannot quantitatively determine external complicated stimuli, qualitative tactile perception and distinction from different stimulus, such as pressures, vibrations, touches, and shear forces, can be realized with the cooperation work of various tactile receptors embedded underneath skin (including Meissner's corpuscles, Pacinian corpuscle's, Ruffini endings, and Merkel's disks) [4].Soft artificial tactile sensor, capable of emulating the somatosensory functions of human skin and converting external mechanical stimulus into processable signals, have found diverse applications in skin prosthetics, health monitoring, and human-machine interfaces [5][6][7].Although numerous monomodal sensory devices have been proposed to perceive individual elements through materials optimization and structural engineering [8][9][10][11], the intricate artificial perception of electronic skin necessitates a multisensory integration mechanism within the sensory system.This highlights the need for the development of multimodal tactile sensory device [12].
At present, tactile sensing predominantly involves the measurement of mechanical stimulus such as pressure, strain, touch, and vibration, as well as temperature and humidity.The previous literatures propose various mechanisms for the sensing of mechanical stimulus, including piezoresistive [13,14], piezocapacitive [15][16][17], piezoelectric [18][19][20], triboelectric [21][22][23][24] and optical mechanisms [25][26][27].Temperature sensing widely utilizes thermo-resistive [28][29][30] and thermoelectric [31] response.Humidity sensing can be achieved through capacitive [32], resistive [33], and moistelectric [34,35] mechanism.Despite the advancements in monomodal tactile sensors, obstacles remain when it comes to integrating multimode sensors, including complicated fabrication procedures and multi-parameter interference.Meanwhile, achieving multimodal tactile sensing through a single sensing unit remains a significant challenge.The development of multi-responsive sensors relies on the utilization of functional active materials, innovative device design and multiple response mechanisms [36].Therefore, it is essential to comprehensively review the development and prospects of multimodal sensors to inspire the next generation of soft multifunctional sensing devices.
In this review, we systematically present the advancements in tactile sensation and integration of multimodal tactile sensors, summarize their application scenarios, and discuss the associated challenges, as shown in figure 1.We provide an overview of various tactile sensing mechanisms to establish a theoretical foundation for the integration of multimodal sensors in flexible electronics.We discuss the burgeoning multimodal tactile perception from four aspects including (i) integration of monomodal sensors, (ii) multifunctional materials, (iii) multiple mechanism integration, and (iv) signal-decoupling strategies.Then, the applications of multiple tactile sensors in fields of health-monitoring and artificial intelligence are exemplified.Finally, challenges and prospects for multimodal tactile perception are highlighted.

Sensing mechanism for various tactile perception
Various sensors capable of converting pressure, strain, temperature, and humidity to computer-processable electric signals are universally utilized to mimic the tactile perception of human skin [37].Various signal transducing mechanisms involve different functional materials and structural design principles [38].In this section, we will discuss these individually.

Pressure sensors
Pressure sensors are commonly utilized tactile sensors to perceive the force stimulation in various applications, including human-machine interaction, healthcare, and artificial robots.The rapid advancement of materials and structural engineering techniques has significantly contributed to the booming progress of flexible pressure sensors.Therefore, various transduction mechanisms have been proposed to convert mechanical stimulus into processable signals.Piezoresistive, piezocapacitive, supercapacitive and optical pressure sensors are commonly utilized to measure static stimulus.Meanwhile, triboelectric, and piezoelectric mechanisms generate instantaneous signal outputs for detecting dynamic pressures.Meanwhile, despite the implementation of various technologies such as microstructure engineering (micropillars [39,40], hemispheres [41], fibers [42], porous aerogels [43,44]) and materials selection (graphene [45,46], carbon nanotubes [47], two-dimensional transition metal carbides and nitrides [14], and conductive polymers [48]) to enhance pressure sensing performance, monomodal sensors struggle to meet the demands of complex motion acquisition.It is necessary to simultaneously perceive and differentiate multidimensional spatiotemporal force stimuli such as pressure direction, static/ dynamic pressures, as well as torsion and sliding forces.
Materials structure and sensor configuration designs are effective approaches to extract multidimensional information of external pressures.The positive and negative pressures can be detected by incorporating air gaps into dielectric layers [49].The deformation under an external pressure can also be enhanced due to the air gaps within the dielectric layer, resulting in significantly improved sensitivity to negative pressure (figure 2(a)).In addition, reasonable configuration design enables the discrimination of normal and shear forces.Boutry et al mimicked the hierarchical dermisepidermis structure of human skin and proposed an pressure sensor array to measure and discriminate both normal and shear forces (figure 2(b)) [3].The biomimetic electronic skin composed of bottom electrodes with micro-hills, top  electrodes with micro-pyramids and the intermediate dielectric layer.Both the experimental and simulation results confirm the effectiveness of the biomimetic design.We encourage more attempts to replicate the structure and function of human skin, aiming to achieve the discrimination of various types of mechanical stimulus, including shear, tilted and normal forces.
Cooperation of Meissner's corpuscles, Pacinian corpuscles, and Merkel's discs allows human skin to perceive both the static and dynamic pressure [53].The integration of multiple transduction elements into a single platform enables the simultaneous detection of static and dynamic pressures.Inspired by the tactile adapting mechanoreceptors of human skin, Qiu et al proposed a dual-mode electronic skin incorporating both piezoelectric and piezoresistive sensing mechanism (figure 2(c)) [50].The synergistic effect between the top piezoelectric unit and the bottom piezoresistive unit contributes to a hybrid response mechanism, allowing the simultaneous detection of static and dynamic pressure.However, the elastic piezoresistive unit can dissipate high-frequency forces, resulting in the invalid sensing of the piezoelectric unit at high frequencies.Kong et al report a self-protective piezoelectricpiezoresistive dual-mode pressure sensor based on elastic cylinder induced flextensional transduction (figure 2(d)) [51].The cylindric design allows the transition from normal force to lateral expansion force.Meanwhile, the piezoelectric and piezoresistive units work independently without interference, opening a new possibility for self-protective piezoresistivepiezoelectric sensors with dynamic-static mechanoresponse for more complex application scenarios.In addition, integrating piezoresistive sensors with triboelectric sensors can also achieve simultaneous detection of static and dynamic stimulus [54].Overall, the signal outputs from multiple sensing modules necessitates multiple signal collection units, resulting in increased power-consumption and complex signal acquisition procedure.Meanwhile, the data acquisition speed is dependent on the switching speed between the multiple sensing modules.
To solve this problem, new sensing mechanisms have been proposed to distinguish dynamic and static pressures through a single device [55,56].A potentiometric mechanotransduction mechanism was proposed to realize the detection of both static and low-frequency dynamic pressures.The mechanism is based on a mechanically modulated electrochemical system, where the electrolyte/electrode interfaces are sensitive to various external loads and give rise to potential difference [56].Furthermore, Wu et al presented a single mode mechanoreceptor based on hybrid potentiometric and triboelectric sensing principles [57].The hybrid sensor exhibits characteristic of single-mode output, self-generated voltage, and self-adapting sensing behavior.The voltage signals from the hybrid sensor can decode various aspects of mechanical stimuli, such as frequency, loading magnitude, and releasing speed.Kim et al proposed an artificial sensory system designed to capture temporal and spatiotemporal information of dynamic tactile motions using a position-encoded spike spectrum [58].The artificial receptor consists of a mixed ion-electron conductor capable of generating a distinct potential spike, enabling the recognition of complex dynamic movements such as position, motion trace, speed, and dynamic contact area.As a result, exploring more sensing mechanisms to detect complex pressure stimuli is more promising.

Strain sensors
Strain sensors are indispensable components in wearable electronics and tactile sensing devices.The strain-responsive mechanism could be classified into four categories, including resistive, capacitive, piezoelectric, and optical sensors.Piezoelectric strain sensors are particularly well-suited for capturing dynamic strain due to the fast charge transfer characteristics.Resistive and capacitive-type sensors have been extensively researched for static strain detection.Various conductive materials, such as carbon blacks [59], carbon nanotubes [60], two-dimensional transition metal carbides and nitrides [61] and ion-gels, are employed as active sensing elements in the design of strain sensors.Elastic soft polymers such as thermoplastic polyurethane, PDMS (polydimethylsiloxane), medical adhesive tape, and fabrics are appropriate choice for supporting substrates.Sensitivity, linearity, and sensing range are three crucial parameters to evaluate the sensing performance of strain sensors.
Considerable efforts have been devoted to achieving high sensitivity and wide linear range in uniaxial direction, including hierarchical structure design [62, 63], prestretching technology [64] and materials composition [65].However, multiaxial information of complex strain conditions has been ignored due to severe coupled signals among various directions.Regulating of the surface strain distributions through the combination of an inelastic film with a strain sensor facilitates independent sensing in the x-and y-axis directions (figure 2(e)) [52].A multidimensional strain sensor can also be constructed by utilizing multiple pre-strained conductive percolation networks, resulting in decoupled electrical responses to principal and perpendicular directional strain [66].The three-dimensional structural design of the strain sensor enables the detection of multiple strains and recognition of out-of-plane strain-direction in a single device [67].This design allows for comprehensive acquisition of mechanical feedback.
Multimodal tactile perception inevitably involved strain deformation of the sensor, which may cause stretchinginduced interference and signal crosstalk.Therefore, it is crucial to explore strain suppression technologies and realize accurate multiparameter measurements.The strain suppression methods and technologies will be discussed in detail in section 3.1.2.

Temperature sensors
Temperature, as a fundamental physical parameter, plays a critical role in tactile sensing.In recent years, thermoresistive and thermoelectric effects have been proposed for temperature sensing.Conductive composites consisting of conductive fillers and insulative polymer matrix have been extensively investigated for temperature sensing based on resistive response mechanism.Carbon-based materials, such as carbon black, graphene, carbon nanotubes, are universally employed as conductive fillers [68].Thermosensitive composite with positive and negative temperature coefficients can be designed and fabricated.For example, Someya at al proposed a flexible and printable temperature sensor utilizing composites of semicrystalline acrylate polymers and conductive graphite fillers [30].In addition, conductive nanoparticles [69] (Pt, Ni, Al), graphene [70], hydrogels [71], and ion liquid are commonly employed as resistive thermo-sensitive materials.The band gap structure and oxygen functional groups of reduced graphene oxide can be regulated by adjusting the degree of reduction, resulting in different temperature coefficients [72].Jia et al have reported a serious of thermal sensors based on ion liquids, including 1-ethyl-3-methylimidazolium acetate ([EMIm][Ac]) [73], green electrolytes [28] and 1-Ethyl-3-methylimidazolium bis(trifluoromethyl sulfonyl)imide ([EMIm][Tf 2 N]) [74], which exhibit electrical conductivity increasement with temperature increasing.
Thermoelectric materials can transform the temperature gradient into an electrical output voltage or current based on Seebeck effect.Poly (3,4-ethylenedioxythiophene): poly(styrenesulfonate) (PEDOT:PSS) [75], MXene [76], Bi 2 Te 3 [77], metal organic frameworks [78], and carbon nanotubes [31] have emerged as key materials for thermoelectric temperature sensing.Stretchable self-powered temperature sensors can be designed and fabricated with thermoelectric composites [79].Undesired mechanical interference can be eliminated by reasonable structural design including wrinkle buckling, rigid-island design and novel mechanism innovation, which will be discussed in detail in section 3.1.2.

Humidity sensors
Human respiration detection, object recognition, and environmental monitoring involve humidity monitoring that gain increasing attention in recent years.The active materials primarily utilized for humidity measurement are hydrophilic, and various properties such as resistance and dielectric behavior will be changed due to the interaction between the active material and water molecules.Therefore, materials including cellulose [80], graphene oxide [81] and transition metal carbide [82] that contains hydrophilic hydroxyl and carboxyl groups are appropriate candidates for humidity sensor.Based on the interaction between active materials and water molecules, capacitive and resistive mechanisms are the predominant strategy used in most current humidity sensors.The principle of capacitive humidity sensors is based on the changes in dielectric capacitance, while resistive humidity sensors typically rely on the changes in materials resistance during the process of water adsorption, diffusion, reaction, and expansion.Generally, active materials such as carbonized fabrics and metal-based nanofibers with high specific surface area facilitate the interactions between moisture and functional sensing groups, resulting in high sensitivity and rapid response [83,84].
However, both resistive and capacitive humidity sensors require energy input, whereas self-powered humidity sensors can operate without any power consumption.Therefore, triboelectric and piezoelectric humidity sensing devices have been developed and constructed to facilitate self-power sensing [85].Based on impedance matching effect, resistive-type humidity sensor connected with the triboelectric and piezoelectric devices in series contributes to self-powered humidity sensing [86].However, it is worth noting that this type of self-powered humidity sensor usually requires integration of a series parallel circuit and possesses a relatively large volume.Therefore, a moisture-electric mechanism was proposed to realize self-powered humidity sensing characteristics [87][88][89].
Li et al reported a MXene/cellulose/polystyrene sulfonic acid composite membrane to transfer gradient moisture into electricity due to the directional proton diffusion [88].The asymmetric swelling and expansion of the sensitive membrane give rise to bending behavior of the composite membrane.

Methodologies for multimodal sensing
Multimodal sensory systems are usually fabricated through diverse approaches, including integration of individual sensing unit, development of multifunctional materials, and integration of multiple mechanisms.The direct integration of individual elements inevitably results in structural complexity and manufacturing challenges.Meanwhile, multifunctional materials and multiple mechanism integration usually suffer from signal cross-talk and interference.Therefore, the signal decoupling method plays a crucial role in multimodal parameter acquisition, which will be discussed in the section 3.4.

Planar and multilayer configuration.
The integration of individual tactile units into a single system represents an efficient approach for achieving multi-parameter sensing functionality.In general, various types of tactile sensors can be integrated into planar or multilayer stacked configurations.For planar structure integration, simultaneous sensing of pressure and temperature can be achieved by patterned metal films [69].
To solve the problem of strain-coupling effect, the patterned metal films with and without microprotrusions were employed as the pressure and temperature sensing units, respectively.As the result, bending deformation induced resistance changes enable the pressure sensing.However, the temperature sensing unit is pressure-insensitive due to the negligible deformations.Bae et al integrated pressure and temperature sensing units with non-contacted configuration to achieve measurement of temperature/pressure in real time (figure 3(a)) [70].By novel structural design and utilization of temperature-independent material, pressure and temperature parameters are discriminated perfectly.Stretchable and conformable matrix network sensing system with multifunctional sensing capability for temperature, strain, humidity, light, magnetic field, pressure, and proximity can also be integrated (figure 3(b)) [2].The design of nodes and meandering wires allows the stable expansion of the sensing system to several orders of magnitude.Skin-inspired multilayer stacked tactile sensing systems have undergone significant development to realize multimodal sensory integration.Inspired by the fingertip patterns and interlocked microstructures between the epidermal and dermal layers, Park et al fabricated a multimodal electronic skin based on microstructured ferroelectric films composed of poly (vinylidene fluoride) and reduced graphene oxide [90].The variation in contact resistance between the microdomes allows for static pressure sensing, while the piezoelectric and pyroelectric responses enable dynamic pressure and temperature sensing.In addition, the capability of tactile perception is significant in material recognition, and the combination of thermal properties and mechanical characteristics can enhance the recognition accuracy.Gui et al proposed a multi-sensing platform by laminating temperature, pressure, and infrared light sensors into a skin like electronic device, and the three functional layer correspond to epidermis, dermis and hypodermis structure of human skin (figure 3(c)) [74].The separated layered structure of the three sensing units endows simultaneous measurement of temperature, pressure, and infrared light without signals interference.A multilayer structured multisensory system was fabricated to accurately collect the information of contact pressure and temperature including object and environment (figure 3(d)) [7].Various objects with different thermal conductivity give rise to different heat transfer, contributing to the materials recognition.The deformation of the porous material results in the changes in thermal conductivity base on piezothermic transduction, enabling the contact pressure detection.

Strain deformation suppression.
In practical wearable scenarios, undesired strain deformation or other physical inputs often introduce external disturbances and crosssensitivities to another stimulus.As a result, the tactile sensor tends to respond indiscriminately to the coupled stimulus, posing challenges in accurately perceiving specific signals.Therefore, it becomes imperative to develop effective methods to minimizing the cross-sensitivity caused by undesired strain deformation, with a particular emphasis on suppressing the influence of the tensile stress.
Strain engineering technologies, such as wrinkle buckling and rigid-island design method, are wide employed in the fabrication of strain-insensitive tactile sensors.Integrating the sensing elements on islands of high-modulus substrates embedded in the soft elastomer can suppress the strain deformation of the sensing units [91].Under tensile strain conditions, the soft substrate undergoes significant strain deformation compared to the high-modulus islands due to the substantial difference in the elastic modulus of the polymers.Meanwhile, pre-stretching strategy can also achieve tension insensitive but pressure sensitive sensing [92].Besides, technologies such as kirigami-based design [93], serpentine structures and wrinkled shar-pei like hierarchical structures [94] have been adopted to effectively suppress cross-sensitivities caused by strain deformation.However, these strategies typically require sophisticated fabrication processes and presents challenges for integrating high-density device.A circuit design strategy has been proposed to realize strain suppression without using any strain engineering method [95].The accuracy and robustness of the temperature sensor can be improved and the straindependent errors can be mitigated by employing static and dynamic differential readout approaches.Wu et al developed a novel potentiometric mechanotransduction mechanism to construct the strain-insensitive mechanical sensors [56].Since the open-circuit potential was measured, resistance variation of the elastic conductors can be neglected and the potentiometric sensors work independently from stretching deformation.However, to broaden the mechanotransduction mechanism and improve the sensing performance of the sensor, it is necessary to develop new material systems.

Multifunctional materials for multimodal sensing
The integration of multiple individual functional units usually involves sophisticated fabrication processes and signal processing algorithms.Multifunctional materials, capable of responding to diverse stimulus simultaneously, have been widely employed to fabricate multimodal tactile sensors.Typically, multimodal sensory systems with homogeneous output signals and heterogeneous output signals will be introduced below.
Various functional materials are adopted to fabricate multiresponsive tactile sensors with homogeneous output signals, including carbon black, graphene, carbon nanofiber, hydrogels, and conductive composites.For instance, Zhang et al reported silk-nanofiber-derived carbon fiber membrane as a multimodal sensing material for pressure and temperature detection (figure 4(a)).The conductive carbon fibers are capable of transducing temperature and strain signals into resistance outputs through thermosensitive and piezoresistive mechanisms [96].In addition, piezo-sensitive and thermosensitive composites can be utilized as the active materials to fabricate multimodal sensors [37].Moisture absorptioninduced changes in ionically conductive pathways enable humidity sensing.Hydrogel-based polymer matrix filled with conductive components or electrolytes have been widely developed as resistive tactile sensor.For example, Liu et al fabricated a temperature-strain sensitive hydrogel using direct ink writing technology (figure 4(b)) [97].The hydrogel transfers the tensile strain into resistance signals due to the changes in conductive paths within the hydrogel.Additionally, it responds to temperature through the thermal-induced tunneling effect within the hydrogel.Composite aerogels consisting of conductive fillers and semicrystalline polymer allow for bimodal sensing of temperature and pressure (figure 4(c)) [98].The semicrystalline polymer, as an intercalation material, exhibits shrinkable behavior due to melt flow-triggered volume change, resulting in thermoresponsive conductive channels and pyroresistive properties.
The aforementioned-functional materials are specifically sensitive to certain stimuli and produce corresponding electrical signals for tactile sensing purposes.Except for resistance-responsive mechanisms, self-powered multimodal tactile sensors without energy consumption are gaining increasing attention.Integrating of thermoelectric and piezoelectric materials allow for temperature-pressure sensing [100,101].Thermoelectric and piezoelectric components enable self-powered sensing capabilities via transducing temperature and pressure stimulus into voltage signals, respectively.Multifunctional ionic hydrogel with gradient structures can directly and accurately perceive temperature, pressure, and humidity stimulus (figure 4(d)) [99].The gradient distribution of charged groups and diffusion of counterions contribute to self-induced potential that is influenced by factors such as thickness, temperature, and humidity.The temperature, pressure and humidity stimulus are transformed into voltage signals via mechano-electric and thermo-electric effects.Unfortunately, the abovementioned multimodal tactile sensors based on multifunctional materials tend to output homogeneous signals, resulting in significant signal interference and crosstalk.Differentiated and heterogeneous outputsbased sensing system based on multiple mechanisms integration was urgently exploited.

Mechanism integration for multimodal sensing
For multimodal tactile sensing, multiple mechanism integration allows for multimodal sensing without interference and crosstalk.Up to now, piezoresistive-thermoelectric, piezoelectric-thermoresistive, and piezocapacitivethermoresistive mechanisms have been developed for simultaneous detection of pressure and temperature.Piezoelectric, triboelectric and piezoresisitive mechanisms are widely employed to differentiate dynamic and static mechanical stimulus.Herein, we discuss recent advances on mechanism integration for multimodal sensing.

Mechanisms integration for multiple sensing of pres-
sure and temperature.The thermoelectric and piezoresistive effects can transduce the temperature and pressure stimulus into independent voltage and resistance outputs.For example, Zhu et al reported a microstructure-frame-supported thermoelectric PEDOT:PSS material that exhibits independent thermoelectric and piezoresistive effects (figure 5(a)) [102].The temperature stimulus is transduced into thermovoltage signal, while the pressure stimuli is transduced into current signals.Based on their porous characteristic, the thermoelectric fabrics can be used to fabricate large-area pressuretemperature e-skin [103].The choice of thermoelectric materials and structural design of the piezoresistive frameworks play an important role to achieve independent bimodal sensing.Solvent-treatment of thermoelectric PEDOT:PSS aerogels contributes to the decoupled readout of dual parameters by regulating the mechanism of charge carrier transport [75].The pressure and temperature stimulus can be decoupled by analyzing the linear slope and voltage shift in the I-V curves.
Structural design serves as an effective strategy to eliminate signal crosstalk and interferences.The active materials of dual-mode pressure and temperature sensor can be designed by decorating predesigned elastic and porous substrates with conductive or thermoelectric materials [76].Microscopic firstprinciples calculations confirm that pressure stimuli exhibit negligible effects on energy band of carriers, which explains the independent dual mode sensing performance.Meanwhile, the mutually non-interfering mechanism can also be achieved by in situ growing the conductive metal-organic framework onto microstructured mixed cellulose substrates [78].
Coupled with thermoresistive mechanism, the piezoelectric and triboelectric effects have been widely utilized to measure dynamic mechanical stimulus and temperature.Rao et al developed a tactile electronic skin capable of detecting and distinguishing pressure and temperature stimuli.This was achieved using a single-electrode-mode triboelectric nanogenerator with thermoresistive electrode [108].The triboelectric layer with pyramidal microstructures responds to dynamic pressures, while thermoresistive electrode exhibits a respond to temperature with negative thermal coefficients.In comparison to the large internal impendence of the triboelectric device, the resistance variation of the thermoresistive electrode has negligible impact on triboelectric voltages, explaining the non-interference output signals of pressure and temperature stimuli [109].Alam et al fabricated an multifunctional sensor using electrospun nanofibers, which incorporates triboelectric, piezoresistive, and thermoresistive sensing effects [110].The triboelectric layer converts dynamic stimuli into voltage signals, while the thermoresistive nanofiber layer serves as thermoresistive and piezoresistive material to measure temperature and static pressure.
The piezocapacitive and thermoresistive mechanisms can convert pressure and temperature stimuli into capacitance and resistance signals, respectively.Therefore, integration of piezocapacitive and thermoresistive mechanisms allow for the simultaneous detection of pressure and temperature.For example, Kim et al developed a three-dimensional microstructure-based electronic skin sensor using a dropletbased microfluidic-assisted emulsion self-assembly process [111].The microstructured elastomer functionalized with multiwalled carbon nanotubes exhibited high sensitivity to pressure stimuli through piezoresistive response.Since the capacitance output is insensitive to temperature stimuli, pressure and temperature can be differentiated by comparing resistance and capacitance signals from the sensors.Meanwhile, textilebased multimodal sensor capable of discerning pressure, proximity, and temperature can also be fabricated (figure 5(b)) [104].The thermoresistive effects due to thermal energyinduced charge carrier hopping facilitate the signal transformation from temperature to resistance.The signal variation in capacitance was monitored for pressure and proximity sensing.While temperature could introduce some crosstalk to the capacitance signal, the decoupling of bimodal parameters was achieved through capacitance baseline calibration to distinguish small pressure differences.
The optical sensing mechanism can transduce applied stimuli into light signals.Coupled with the optical sensing mechanism, multiple parameters can be encoded into light and electric signals, allowing for accurate discrimination of various stimuli and minimal signal interference.For instance, Wang et al reported a flexible hybrid optoelectronic, resistive and capacitive multimodal sensor capable of detecting proximity, pressure, and temperature [112].The sensor consists of a light waveguide and conductive interdigital electrode on a regenerated silk fibroin substrate.It encodes proximity into capacitance signals based on the fringe field effect, pressure into light intensity by measuring deformation-induced optical loss, and temperature into resistance due to the thermoresistive effect of PEDOT:PSS.Due to the independent optical and electrical sensing effects of the device, non-interference signals are achieved without requiring complex decoupling algorithm.Thermochromic mechanism can be utilized to fabricate temperature sensor via color change [113].Coupled with capacitive and triboelectric configuration, multimodal devices are constructed to detect strain and pressure.

Mechanisms integration for multiple sensing of strain and temperature.
The development of stretchable strain sensor is crucial in wearable electronics, and the combination of strain-temperature stimuli measurement has been reported recently.For example, Li et al developed a strain-temperature dual-parameter sensor by printing a stretchable conductive and thermoelectric nanocomposite [105].Stretchable conductive material, based on stretching-induced crack-propagation effect, exhibits resistive response behavior.Thermoelectric nanocomposite with high Seebeck coefficient was selected as temperature sensing material.During the sensing process of strain and temperature, resistance and voltage signals can be collected without crosstalk (figure 5(c)).A multifunctional and stretchable thermoelectric fabrics has also been fabricated (figure 5(d)) [77].The thermoelectric fabric utilizes the Seebeck effect to generate output voltages for temperature measurement.Meanwhile, under lateral strain, the fabric exhibits a decrease in resistance due to the reduced band gap of thermoelectric material.
Chromotropic ionic films have been exploited to visualize strain stimuli through colorimetric changes.By integrating mechanochromism and thermoresistive effect, these films exhibit multi-responsive behavior to temperature and mechanical stimuli.Zhang et al reported a stretchable chromotropic sensor to discriminate strain, temperature and pressure stimulus based on photonic crystal particle arrays embedded in a polymer matrix (figure 5(e)) [106].The highly ordered photonic crystal arrays exhibited strain-sensitive properties, resulting structural color changes across the full visible spectrum from violet to red.Additionally, the thermoresistive principle of the ionic hydrogel enables the temperature sensing with high sensitivity.Moreover, the pressure sensing performance was enabled by a triboelectric mechanism.The chromotropic mechanism could potentially contribute to multimodal sensing in a visual and energy-efficient manner.

Mechanisms integration for multiple sensing of humid-
ity and other stimuli.Ambient humidity might influence the functionality of the pressure, strain, and temperature sensors.Therefore, the decoupling of humidity from other physical stimulus is necessary to achieve multimodal tactile sensing, and in recent years, the corresponding mechanisms integration strategies have also been proposed.For example, Yang et al reported a single-component multimodal sensor based on graphene oxide, which enables the simultaneous monitoring of changes in temperature, humidity, pressure, and light intensity [87].The moist-electric mechanism can generate voltage signals for detecting humidity.However, temperature, pressure, and ultraviolet light stimuli can introduce cross-talk to the moist-electric potential.To address this issue, a machine learning decoupling method is introduced to enable simultaneous multi-metric monitoring.In addition, the ionic-electronic aerogels could be utilized to develop a pressure-temperaturehumidity sensing device (figure 5(f)) [107].Thermoelectric and piezoresistive mechanisms contribute to transducing the pressure and temperature stimulus into voltage and resistance signals.Humidity gives rise to the ions motivation within the aerogel and increased the thermoelectric potential, resulting in the increasing of Seebeck coefficient.Therefore, the contribution of ions to the Seebeck coefficient is calculated to assess the humidity level.It is still challenging to achieve multimodal sensing by measuring fewer electrical parameters.
Colorimetric sensing mechanism is employed in chemical and physical multimodal sensors, which provides additional opportunities for sophisticated sensory and chemical discrimination.Khatib et al proposed an electronic and optoelectronic mechanism integration to achieve the detection and discrimination of temperature, relative humidity and volatile organic compounds [114].The hierarchical nano-bilayer composed of graphene oxide and colorimetric dyes provides two sensing outputs.The colorimetric outputs of colorimetric layer reflect the levels of volatile organic compounds, while resistive output of the graphene-dye bilayer reflect the variations in relative humidity and temperature.Selective discrimination and quantification of multiple stimulus can be achieved based on the colorimetric and electrical response.

Discrimination of complex mechanical stimulus.
Dexterous manipulation in artificial robotics requires intricate mechanical sensing to detect normal and shear forces, ensuring slip detection and facilitating interaction with fragile objects.Discriminating various types of mechanical stimulus, such as pressure, strain, torsion, and sliding, remains a significant challenge due to their similar electrical outputs.Microstructural engineering strategies offer a feasible approach to achieve sensing and discrimination of diverse mechanical stimulus.Ge et al developed an electronic fabric consisting of an intertwined fibrous shell with piezoresistive rubber and a helical core stretchable silver electrode [115].The coaxial structure and fibrous architecture of the electrotonic fabric enable multiple force sensitivities.Different types of mechanical stimuli can be distinguished by examining the variations in electrical resistance of silver electrodes under different conditions of pressing, stretching, and flexing deformations.Negligible changes in resistance were observed under vertical loading force as there was no lateral straining involved.During stretching, both silver electrodes underwent the same tensile strain and rate of resistance change.During flexion, only the bent sensor electrode underwent tensile stress and exhibited changes in resistance.As a result, the different resistance responses of the silver electrodes allow for the differentiation of press, stretch, and flexion deformation.Interlocked microdome arrays allow for the detection of directional mechanical stimuli along three different axes [116].The interlocked microstructures facilitate stress concentration at the contact spots, thereby enhancing sensitivity of the piezoresistive response to various tactile stimuli.Therefore, we appeal more microstructural engineering strategies to explore the differentiation between normal and shear forces.
Anisotropic structure design also allows for multidimensional mechanical sensing.The anisotropic structures usually exhibit distinct resistance changes in various loading directions, which enables the differentiation of multidimensional mechanical stress.For example, Tang et al developed an all-in-one flexible anisotropic sensor with MXene coated piezoelectric fibers [117].The aligned piezoelectric nanofiber mats were fabricated via electrospinning with a high-speed collector, and MXene was coated on the surface of the nanofiber mats to create anisotropic conductive path.The nanofiber mats exhibited substantial disparity in elastic modulus between x-axis direction and y-axis direction.This discrepancy resulted in distinct deformation capabilities and varied electrical resistance changes in response to lateral strains along x and y-axes.Specifically, the oriented piezoelectric nanofiber mats allowed for capability of detecting mechanical force in z-axis direction via voltage output.As the result, the anisotropic structures effectively enable the distinction of multi-directional mechanical stimuli.The difference in structural change caused by different mechanical inputs was also utilized to differentiate between pressure and strain forces.Pressure-insensitive strain sensors can be fabricated by designing porous conductive composite [118].The vertical pressure results in close and denseness of the pores, while causes negligible influence on the percolation network of the composite, allowing for pressure-insensitive strain sensing.However, tensile strain can cause the propagation of microcracks and reduction in the percolation paths, resulting in significant changes in electrical resistance.The differentiated structural changes induced by compressive and tensile strain can also be utilized to discriminate between shear stress and normal pressure.
As discussed in part 2.1 (figure 2(b)), microstructured engineering of sensing layers is an effective strategy to discriminate shear stress and normal pressures [3].Bao et al incorporate microstructured features with an array of capacitors which is capable of measuring and discriminating both normal and tangential forces, but with relatively small linear dynamic range.Three-dimensional architecture design enables separate, simultaneous measurements of multiple, such as normal force, shear force, and bending.For instance, Won et al constructed a three-dimensional table-like configuration by integrating four piezoresistive elements onto a patterned flexible film [119].Each piezoresistive element within this three-dimensional configuration generates separate and independent electrical resistance outputs, allowing for the extraction of detailed information regarding complex mechanical stimulus from the sensing properties of the four piezoresistive elements.Normal force, uniaxial shear force, and convex bending force induce distinct tensile strain and fractional changes in resistance within the four piezoresistive elements, achieving the determination for the three loading conditions.Inspired by suspension bridge, structural design concept based on piezoelectric nano/microwire with polymer protrusions has been proposed to transform vertical force to lateral force, allowing for dual-modal sensing of vertical force and lateral strain sensing [120].Microstructural engineering strategies and multiple mechanisms integration tend to increase the thickness of the whole integrated multimodal devices and adverse to fabricate conformal multimodal sensing system.Besides the decoupling of various stimulus, the coupling of the same stimuli also occurs especially for the high-density sensing array due to in-plane or out-of-plane deformation.The configuration optimization, signal modulation and intelligent algorithms deserves more attention to increase both the sensing accuracy and array density.

Signal decoupling strategies for multimodal sensing
Considering the sensitivity and responsiveness of multimodal sensors to various input signals, it is crucial to develop signal decoupling mechanisms to detect the target signal without cross-sensitivity.Multimodal sensors with signal decoupling mechanisms can simplify the complexity of signal processing.Time dependent variation of signals has been utilized to distinguish temperature and pressure stimulus.For instance, Park et al developed a time-dependent strategy to decouple temperature and pressure inputs, which exhibits positive and negative resistance variation in response to different temperature [90].When the object with various temperatures and pressures were loaded onto the sensor, the equilibrium resistance reflected the pressure level while the difference between instantaneous and equilibrated resistance values represented temperature level.As the result, time-dependent signal decoupling of pressure and temperature was realized.Signal decoupling-assisted piezoresistive and pyroresistive mechanism integration facilitate the dual-modal sensing of temperature and pressure [121].The pressure was determined by analyzing the electrical resistance variation via the piezoresistive effect.The temperature was evaluated by the recovery time of signal, which is influenced by the release of thermal energy and the movement of the dipoles.Meanwhile, multimodal sensing based on  [123].Copyright (2023) American Chemical Society.(c) Frequency-dependent ion relaxation dynamics for strain and temperature sensing.From [124].Reprinted with permission from AAAS.(d) Resistance and capacitance decoupling method to discriminate pressure, strain, and temperature.Reprinted with permission from [125].Copyright (2022) American Chemical Society.triboelectric-piezoelectric-pyroelectric principles also depend on the signal decoupling strategy (figure 6(a)) [122].The lower polarized film is utilized to evaluate temperature stimulus based on the pyroelectric effect, while the capability of pressure sensing is evaluated via the coupling effect of triboelectric and piezoelectric effects.Despite yielding the same type of output signals, pressure and temperature stimuli can be differentiated by considering the contrasting response time difference of the coupled bimodal voltage.
Frequency-based decoupling approach provides a new route for decoding of multitype response in ionic materialbased system.Ma et al developed a bimodal temperature/ pressure sensor utilizing a chitosan film as the active layer [126].Both pressure and temperature are measured using capacitance signals.The response to pressure is determined by the variations in the contact surface caused by external loads, whereas the response to temperature depends on the temperature-sensitive proton migration and the dielectric constant of chitosan.Suh et al developed an ion-gel based tactile sensor composed of spherical cap-shaped top electrode and four flat bottom electrodes (figure 6(b)) [123].By modulating the sensitivities of ion dynamics with frequency, shear forces and normal forces are differentiated by measuring impedance values at two predetermined frequencies.Furthermore, directional information of shear forces is obtained by calculating impedance changes from the four bottom electrodes at high frequency.Meanwhile, they proposed a self-decoupling artificial ion receptor for measuring temperature and strain, utilizing frequency-dependent ion relaxation dynamics (figure 6(c)) [124].The charge relaxation time was chosen as a temperature-sensitive intrinsic variable to eliminate the interference of strain or shear deformation.Additionally, normalized capacitance was utilized as a temperature-insensitive extrinsic variable for evaluating strain.Consequently, temperature and strain parameters are accurately obtained by measuring the charge relaxation time and normalized capacitance at two specified frequencies.However, simultaneous measuring of charge relaxation time and capacitance limits the sensing rate and needs specially designed circuits.
Signals decoupling method based on the trends in responses can also be utilized to distinguish multiple stimuli.Lo et al developed a sensor based on a sponge material capable of detecting pressure, strain, and temperature stimuli through capacitive and resistance variations (figure 6(d)) [125].The sensor was constructed by assembling PEDOT:PSS onto a porous PDMS foam, enabling it to selectively respond to the three stimuli with distinct response trends.External pressures lead to a decrease in resistance and an increase in capacitance, tensile strain causes an increase in resistance and a decrease in capacitance, and elevated temperature causes a decrease in resistance with negligible changes in capacitance.The decoupling of the three stimuli is achieved by simultaneously measuring resistance and capacitance.In addition to the abovementioned signal decoupling methods, machine-learning technology has been extensively utilized to decipher signal cross-talk and accurately acquire tactile stimuli.Yang et al developed a machine-learning assisted method to accurately identify temperature, humidity, pressure, and light intensity [87].

Multimodal sensors for health monitoring
Recent progress in flexible technologies for measuring multiple physical or physiological signals of human body boosts the application scenarios in the fields of wearable healthcare systems.For example, Kwon et al developed a wearable and implantable system for monitoring blood pressure, flow velocity and temperature in a continuous mode of operation (figure 7(a)) [127].The main components of the subsystem include a sensing module that is inserted into blood vessels and a battery-free electronic module that harnesses energy through a receiver coil and transmits data.The biosensing module is equipped with 3D structured strain gauges for measuring forward and backward flows, as well as pressure and temperature sensors for measuring the pressure and temperature of the blood in and around the heart.By incorporating a bimodal sensing element, an integrated wireless control and communication module, a power management circuit, and a mobile application, the system successfully achieved realtime display of blood flow, pressure, and temperature data.This comprehensive system allows for the evaluation of cardiovascular status and the diagnosis of valvular heart disease with clinical-grade accuracy.Cho et al proposed a system that combines skin-mountable multimodal sensors and a movable platform to monitor pressure, temperature, and hydration at the skin interfaces of patients seated in wheelchairs (figure 7(b)) [128].The wireless sensing platform is comprised of an NFC system, electronic components, an instrumentation amplifier, and a multimodal sensing circuit that includes serpentine interconnects, a pressure sensor activated by cracks, and a temperature sensor based on a negative temperature coefficient thermistor.The developed system enables continuous assessment of various parameters such as pressure at specific mounting locations, temperature, and skin hydration levels for wheelchair users.Additionally, Li et al designed and produced a glove equipped with multimodal sensors to accurately measure and quantify the severity of Parkinson's disease symptoms in patients [129].The multimodal sensing system consists of a bend sensor module to capture finger bending information, a pressure sensor module to evaluate hand muscle strength, and an inertial measurement unit to record acceleration signals caused by hand tremors.Through reliability analysis, it was confirmed that the stability assessments made by the glove system strongly correlate with assessments made by doctors.This indicates that the glove system is reliable for diagnosing Parkinson's disease.
Modulus of biological tissue is a ubiquitous mechanical property, which are valuable in clinical diagnosis and humanmachine interactions.Cui et al fabricated bimodal sensors by adding a protrusion on a piezoresistive pressure sensor, endowing the resistance signals with combined information of pressure and the softness of samples [131].A thin, conformal, and stretchable tactile glove was fabricated to digitalize the spatial tactile information of force and softness.Meanwhile, Beker et al developed a thin sensor that consists of a strain sensor coupled to a pressure sensor, which can identify compliance of touched materials [132].The compliance sensor was achieved through the integration of a bilayer structure.The first strain-sensitive membrane detected surface deformation, while the second pressure-sensing layer detected applied pressures.In addition, the compliance sensing array was integrated into robotic systems to achieve high-spatial resolution during the grasping of objects made of multiple materials with varying degrees of compliance.
To ensure long-term use, the multimodal sensing platform is designed with multiple functionalities to enhance durability and comfort during wearing applications.For instance, Hao et al developed a bioinspired electronic skin sensor that features a multi-layered architecture, which was fabricated using an electrospinning and spray-coating procedure [133].By incorporating self-protection, self-adhesiveness, and breathability properties, the electronic skin sensor was able to attach for long-term strain and electrophysiological sensing adaptively and comfortably.Peng et al proposed an all-in-one electronic textile that is both superhydrophobic and antibacterial, relying on the integration of antibacterial Ag/Zn nanoparticles [134].The fabricated sensor demonstrated precise detection of underwater motion and real-time monitoring of electrophysiological signals, showcasing its remarkable potential in the field of next-generation comfortable skin electronics.

Multimodal sensors for artificial sensation
Human-machine interaction refers to the interconnected relationship between users and equipment, demonstrating Applications of the multimodal tactile sensory system.(a) A wearable and implantable system for blood pressure, flow velocity and temperature detection.Reproduced from [127], with permission from Springer Nature.(b) Movable platform to measure pressure, temperature, and hydration of skin for patients seated in wheelchairs.Reproduced from [128].CC BY 4.0.(c) A multimodal sensor based robotic systems of remotely controlling robots.From [5].Reprinted with permission from AAAS.(d) A soft glove consisting of sensing and intelligent feedback systems.Reprinted with permission from [130].Copyright (2022) American Chemical Society.extensive potential applications in industries such as rehabilitation and virtual reality.Feng et al a flexible humanmachine interaction platform that utilizes a bidirectional resistive sensor.The sensing module consists of a micro-cilia array on the surface, which allows for the detection of inward bending, outward bending, and normal pressures [135].By converting the user's motions into controlled commands, the bidirectional sensing interface enables effective robot control or disability care.The end-effector of the robots and sensing arrays provide feedback as part of the process.Kim et al utilizes three different sensing mechanisms, including optoelectronics, microfluidics, and piezoresistivity, to effectively discriminate stretching, bending, or compression (figure 7(c)) [5].The sensor was implemented to construct robotic systems of remotely controlling robots.Zhang et al presented a multimodal proximity/pressure/temperature sensor that was incorporated into a doll.This sensor enabled secure and engaging interactions between the doll and the user, ensuring a safe and immersive experience [112].
Artificial sensory systems capable of mimicking the somatosensory functions of human skin is crucial for the progress of artificial prosthetics, soft robots, and electronic skin applications.Cho et al designed a stretchable multimodal device that combines piezoelectricity, triboelectricity, and piezoresistivity.This device is capable of identifying and distinguishing multiple stimuli, such as pressure, tensile strain, or vibration [136].The multimodal sensor was integrated onto a robot prosthesis to enable applications such as materials identification and texture recognition.This integration enables artificial prosthetics to provide human-like sensations.Zhu et al proposed a modular soft glove capable of providing both sensing and feedback functions from the same unit (figure 7(d)) [130].An intelligent feedback system was introduced, wherein machine learning was employed to assist multimodal sensing and tunable feedback.By integrating a triboelectric and strain sensing system, the system was able to provide comprehensive sensations of robotic fingers and palm motion.Additionally, temperature feedback, accomplished through a resistive heater, ensured a realistic perception of the operating environment.This augmented tactile haptic feedback expands the dimensions of perception beyond simple point-to-point stimulation, offering more meaningful and object-specific information.

Future perspectives
The demand on intelligent robotics, health-monitoring, and human-machine interfaces has sparked rapid advancement in research on the multimodal sensors.Numerous multifunctional materials have been explored, and investigations into multiple mechanism integration and decoupling strategies have yielded valuable insights.These promising findings have significantly advanced the field of human skin-inspired multiple sensory systems.However, there are still significant challenges associated with truly multiple sensing without crosstalk that require urgent attention.Therefore, further fundamental research is necessary to fully unlock the potential of these multiple tactile sensing system.
(1) Materials: As the foundation and core of tactile sensing devices, the innovation of materials should be prioritized.Multifunctional materials have been fabricated to achieve multimodal sensation, however, most present active materials sensitive to multiple stimuli result in accompanying interference, which influences the accuracy of the signal outputs for each sensing mode.Self-decoupling materials have the capability to eliminate signal interference inherently by employing innovative sensing mechanisms.
Coupled with frequency-dependent ion relaxation dynamics, ionic-based materials are a potential choice for developing self-decoupling sensing systems [124].Meanwhile, exploration of biocompatible, skin-conformal, antibacterial, and non-toxic materials are highly demanded to fit the human skin and achieve long-term utilization.
In this regard, textile, silk, and cellulose-based materials hold the advantages of breathable and sustainable properties.Ultrathin skin-conformal sensors can acquire precise subtle deformation of human skin, and largescale, cost-effective processing-transfer printing technologies are urgently needed to meet the requirement for high-precise detection.(2) Structural design and integration: Most existing multimodal sensors only offer a limited stimulus-response functions.To advance the detection of information from the full-range of the whole human body, further research on structural design and compatible integration is required.Meanwhile, the elimination of interferences caused by multiple stimulus requires increased efforts on structural design and mechanism integration.Island-bridge structure, serpentine, anisotropy structures are developed to suppress strain deformation interference, and the structural engineering usually involves sophisticated processing techniques.It is required to develop low-cost manufacturing or patterning procedures, such as inject printing and 3D printing techniques.However, the trade-off between structural complexity and modality amount should be prioritized.Exploring novel mechanisms with an exclusive response to multiple stimulus can reduce fabrication complexity and signal acquisition difficulties.(3) Intelligence: Integrating multimodal tactile sensors with intelligence algorithms to develop a truly multimodal tactile perception system is an effective approach to mimic the tactile capacities of human skin.More machine learning algorithms, including but not limited to convolutional neural networks, artificial neural networks, and support vector machines, should be investigated to processes and analyze the data collected from the multimodal sensors.Meanwhile, more approaches to decouple and fuse multimodal data might facilitate the intelligent feedback of the multimodal tactile perception system.Signal processing of multimodal sensing platform and feedback control algorithms are necessary to achieve human-machine interaction and intelligent robots.
The integration of decoupled multi-sensing system and intelligent computing parts may eliminate the signal interference of multi-parameter inputs.In addition, present wearable multimodal sensing platform is limited by additional wiring for signal transition and energy supply, therefore, wireless communication and energy-supply system deserve more research attention.

Figure 1 .
Figure 1.Overview of multimodal tactile perception including tactile sensing mechanisms, multimodal integration, applications, and challenges.

Figure 2 .
Figure 2. Strategies to detecting complex pressure and strain stimulus.(a) Air-gap strategy to distinguishing positive and negative pressures.[49] John Wiley & Sons.[© 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim].(b) Hierarchically pattern strategy to detecting direction of pressures.From [3].Reprinted with permission from AAAS.(c) Piezoelectric-piezoresistive mechanism for detection of static and dynamic stimulus.Reprinted from [50], Copyright (2020), with permission from Elsevier.(d) Poisson's effect-based mechanism for detection of dynamic and static pressures.Reprinted from [51], Copyright (2022), with permission from Elsevier.(e) Crack-propagation mechanism for detection of directions of strain.Reprinted with permission from [52].Copyright (2018) American Chemical Society.

Figure 4 .
Figure 4. Multifunctional materials for multimodal sensing.(a) Silk derived multimodal sensor for simultaneous sensing of temperature and pressure.Reprinted with permission from [96].Copyright (2017) American Chemical Society.(b) Multifunctional hydrogel for temperature and strain perception.Reproduced from [97].CC BY 4.0.(c) Composite aerogels consisting of conductive fillers and semicrystalline polymer for bimodal sensing of temperature and pressure.Reprinted with permission from [98].Copyright (2022) American Chemical Society.(d) Gradient polyelectrolyte membranes for pressure, temperature, and humidity sensing.Reprinted with permission from [99].Copyright (2022) American Chemical Society.

Figure 6 .
Figure 6.Signal decoupling strategies for multimodal sensing.(a) Time dependent decoupling strategy for multimodal sensing of pressure and temperature.Reprinted from [122], Copyright (2019), with permission from Elsevier.(b) Frequency-based decoupling strategy for shear force and normal force.Reprinted with permission from[123].Copyright (2023) American Chemical Society.(c) Frequency-dependent ion relaxation dynamics for strain and temperature sensing.From[124].Reprinted with permission from AAAS.(d) Resistance and capacitance decoupling method to discriminate pressure, strain, and temperature.Reprinted with permission from[125].Copyright (2022) American Chemical Society.

Figure 7 .
Figure 7. Applications of the multimodal tactile sensory system.(a) A wearable and implantable system for blood pressure, flow velocity and temperature detection.Reproduced from[127], with permission from Springer Nature.(b) Movable platform to measure pressure, temperature, and hydration of skin for patients seated in wheelchairs.Reproduced from[128].CC BY 4.0.(c) A multimodal sensor based robotic systems of remotely controlling robots.From[5].Reprinted with permission from AAAS.(d) A soft glove consisting of sensing and intelligent feedback systems.Reprinted with permission from[130].Copyright (2022) American Chemical Society.