Tactile sensing systems with tribovoltaic textile layers and methods of making and using the same
A tribovoltaic textile sensor assembly with a conducting polymer and metal layer generates direct-current signals for slip detection and normal force sensing, addressing integration and detection challenges in robotic tactile systems.
Patent Information
- Application Number
- PCT/US2026/026630
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2025-05-03
- Filing Date
- 2026-05-04
- Publication Date
- 2026-11-12
AI Technical Summary
Existing tactile sensors are inflexible, difficult to integrate into soft, curved, or stretchable robotic surfaces, and struggle to distinguish between dynamic sliding and sustained normal loading, often requiring complex geometries and external power sources.
A tactile sensor assembly with a tribovoltaic textile layer comprising a conducting polymer on a textile substrate and a metal layer, generating a direct-current electrical signal through a tribovoltaic interface during relative sliding contact, integrated with a capacitive normal-force sensor and encapsulated for protection.
The sensor provides rapid detection of sliding motion and normal force, enabling effective slip detection and robotic gripper control with flexible, compliant integration and stable operation over repeated cycles.
Smart Images

Figure US2026026630_12112026_PF_FP_ABST
Abstract
Description
Attorney Docket No.: 011520.02056TACTILE SENSING SYSTEMS WITH TRIBOVOLTAIC TEXTILE LAYERS AND METHODS OF MAKING AND USING THE SAMECross-Reference to Related Applications
[0001] This application claims the benefit of priority to U. S. Provisional Patent Application No. 63 / 799,547, filed May 3, 2025, the disclosure of which is incorporated herein by reference.Field of the Disclosure
[0002] The present disclosure relates generally to tactile sensing assemblies, tactile sensors, robotic manipulation systems, tribovoltaic textile layers, and methods of making and using the same. More particularly, some embodiments relate to textile-based tactile sensors configured to generate direct-current electrical signals in response to relative sliding contact, and to robotic gripper systems configured to use such signals for slip detection, grip adjustment, normal -force sensing, trajectory planning, and object manipulation.Background of the Disclosure
[0003] Tactile sensing is useful in robotic manipulation, automated handling, humanmachine interfaces, prosthetic devices, wearable systems, medical and healthcare devices, industrial automation, and other systems in which contact state, normal force, shear force, slip, or relative motion between a device and an object may be monitored. In robotic grasping, for example, a gripper may need to detect whether an object is slipping or beginning to move relative to a robotic finger and may need to adjust a gripping force or another manipulation parameter in response.
[0004] Human tactile perception uses different sensory mechanisms to respond to different types of mechanical stimuli. Rapidly adapting tactile responses may be associated with dynamic events such as slip, onset of contact, release of contact, vibration, or rapid changes in mechanical stimulation. Slowly adapting tactile responses may be associated with sustained normal force, pressure, stretch, or grasp state. A robotic tactile system that can provide separate or complementary information regarding dynamic sliding and sustained normal loading can therefore be useful for dexterous manipulation.Attorney Docket No.: 011520.02056
[0005] V arious tactile sensors have been developed for detecting contact force, pressure, shear, vibration, surface texture, or slip. Vision-based tactile sensors may detect motion or deformation by imaging a contact interface. However, such sensors may require cameras, optical paths, image processing, distinguishable surface features, illumination, and sufficient form factor for optical components. Other tactile sensors may use piezoresistive, capacitive, magnetic, piezoelectric, triboelectric, optical fiber, or other sensing mechanisms. Some of these sensors may be primarily responsive to normal force, may require external power at the sensing element, may have limited ability to distinguish dynamic sliding from sustained normal loading, may require complex sensor geometries for shear sensing, or may be difficult to integrate into soft, curved, compliant, stretchable, or textile-like robotic contact surfaces.
[0006] Tribovoltaic effects can generate direct-current electrical outputs at dynamic interfaces during sliding contact. Such outputs may arise, for example, at dynamic metalsemiconductor interfaces, metal-conducting polymer interfaces, semiconductor-semiconductor interfaces, dynamic Schottky interfaces, P-N junctions, or other tribovoltaic interfaces. A tribovoltaic output may be generated by conversion of mechanical energy associated with frictional or sliding contact into electrical energy. This behavior can be useful for tactile sensing because dynamic sliding between two contact layers can provide an electrical signal indicative of slip or relative motion.
[0007] There remains a need for tactile sensing assemblies that are flexible, textilecompatible, mechanically compliant, readily integrated with robotic fingers or grippers, capable of generating slip-related direct-current signals during relative sliding contact, and capable of operating with normal-force sensing and robotic feedback control. There also remains a need for methods of making tribovoltaic textile layers, assembling such layers into tactile sensors, and using such sensors in robotic manipulation systems.Brief Summary of the Disclosure
[0008] In some embodiments, the present disclosure provides a tactile sensor assembly including a first layer and a second layer. The first layer may include a textile substrate and a tribovoltaic material disposed on fibers of the textile substrate. The second layer may be disposed adjacent to the first layer and may include a second material. The tribovoltaic material of the first layer and the second material of the second layer may be arranged for relative slidingAttorney Docket No.: 011520.02056contact with one another such that frictional contact between the tribovoltaic material and the second material generates a direct-current electrical signal.
[0009] In some embodiments, the tribovoltaic material and the second material form a tribovoltaic interface during the relative sliding contact. The tribovoltaic interface may include, for example, a dynamic Schottky interface, a P-N junction, or another interface configured to generate a direct-current electrical signal in response to sliding contact. In some embodiments, the direct-current electrical signal is generated by conversion of mechanical energy from relative sliding contact into electrical energy at the tribovoltaic interface.
[0010] In some embodiments, the tribovoltaic material includes a conducting polymer. In some embodiments, the conducting polymer includes poly(3,4-ethylenedioxythiophene) polystyrene sulfonate. In some embodiments, the textile substrate includes a stretchable textile. The stretchable textile may include spandex, polyester, nylon, cotton, or combinations thereof. The textile substrate may include a woven fabric, a non-woven fabric, a knit fabric, a braided fabric, a tufted fabric, or combinations thereof. In some embodiments, the textile substrate includes hydrophilic cotton or a superhydrophobic fiber. In some embodiments, the textile substrate includes spandex and polyester. In some embodiments, the polyester includes polyethylene terephthalate. In some embodiments, the textile substrate includes about 20% spandex and about 80% polyethylene terephthalate.
[0011] In some embodiments, the second material includes an electrically conductive material. The electrically conductive material may include a metal or metal alloy. In some embodiments, the metal or metal alloy includes titanium, aluminum, or a combination thereof. In some embodiments, the second layer includes a titanium foil.
[0012] In some embodiments, the first layer and the second layer are constrained relative to one another at an edge region while permitting relative sliding contact in an interior region. In some embodiments, the first layer and the second layer are coupled at the edge region by an adhesive. Such an arrangement may maintain the first and second layers in an assembled configuration while permitting the tribovoltaic material and the second material to undergo relative sliding contact in response to shear, slip, or other dynamic mechanical stimulation.
[0013] In some embodiments, the tactile sensor assembly further includes a normal-force sensor disposed adjacent to the second layer. The normal-force sensor may include a capacitiveAttorney Docket No.: 011520.02056sensor. In some embodiments, the capacitive sensor includes an indium-tin-oxide-coated polymer film. In some embodiments, the polymer film includes polyethylene terephthalate. The normal-force sensor may generate a normal-force signal in response to normal loading, while the direct-current electrical signal generated at the tribovoltaic interface may be used to detect dynamic sliding or slip.
[0014] In some embodiments, the tactile sensor assembly further includes an encapsulation layer disposed around at least a portion of the first layer and the second layer. The encapsulation layer may include one or more waterproof polymer sheets. In some embodiments, the encapsulation layer protects the first layer, second layer, normal-force sensor, electrical contacts, or combinations thereof from moisture, wear, dust, or environmental exposure.
[0015] In some embodiments, the tactile sensor assembly is configured to generate an open-circuit voltage or short-circuit current during relative sliding contact. In some embodiments, the direct-current electrical signal includes an open-circuit voltage signal or a short-circuit current signal. In some embodiments, a magnitude of the direct-current electrical signal varies with sliding velocity, applied force, contact area, current density, frictional power, or another parameter associated with the relative sliding contact.
[0016] In some embodiments, the tactile sensor assembly includes a processing circuit electrically coupled to the first layer and the second layer. The processing circuit may be configured to detect slip based at least in part on the direct-current electrical signal. In some embodiments, the processing circuit identifies slip when an output voltage derived from the direct-current electrical signal satisfies a threshold condition. In some embodiments, the threshold condition includes the output voltage being at least about 30 mV. In some embodiments, the processing circuit applies a moving-average filter or another filter to the direct-current electrical signal before detecting slip.
[0017] In some embodiments, the processing circuit is configured to generate a slip signal based at least in part on the direct-current electrical signal and a normal-force signal based at least in part on an output of a normal-force sensor. The slip signal may be generated in response to dynamic sliding between the first layer and the second layer. The normal-force signal may be generated in response to normal loading of the normal-force sensor.Attorney Docket No.: 011520.02056
[0018] In some embodiments, the tactile sensor assembly is configured to provide rapid detection of sliding motion. For example, in some embodiments, in a response-time test in which a voltage from the tactile sensor assembly is compared to a threshold voltage of about 3 mV. the tactile sensor assembly produces a voltage exceeding the threshold voltage within about 0.7 ms to about 40 ms after initiation of sliding motion, depending on sliding velocity. In some embodiments, the tactile sensor assembly may provide stable output after repeated sliding. For example, in some embodiments, when a poly(3,4-ethylenedioxythiophene) polystyrene sulfonate-coated fabric and titanium foil are subjected to a sliding test at a load of about 2.5 N, a sliding speed of about 100 mm / s, and a sliding distance of about 5 mm, the tactile sensor assembly is configured to generate an open-circuit voltage of at least about 30 mV after about 400,000 sliding cycles. In some embodiments, the textile substrate coated with the tribovoltaic material exhibits a resistance change of less than about 13% under about 90% tensile strain.
[0019] In some embodiments, the present disclosure provides a tactile sensor including a first layer, a second layer, a retainer, a normal -force sensor, and a processing circuit. The first layer may include a textile substrate and a conducting polymer disposed on fibers of the textile substrate. The second layer may be disposed adjacent to the first layer and may include a metal layer. The retainer may join the first layer and the second layer to one another while permitting at least a portion of the first layer and at least a portion of the second layer outside the retainer to undergo relative sliding contact with one another. The normal-force sensor may be disposed adjacent to the second layer. The processing circuit may be configured to generate a slip signal based at least in part on a direct-current electrical signal generated by relative sliding contact between the first layer and the second layer and to generate a normal-force signal based at least in part on an output of the normal-force sensor.
[0020] In some embodiments, the conducting polymer includes poly(3,4-ethylenedioxythiophene) polystyrene sulfonate. In some embodiments, the metal layer includes titanium or aluminum. In some embodiments, the metal layer includes titanium foil. In some embodiments, the normal-force sensor includes a capacitive sensor comprising an indium-tin-oxide-coated polyethylene terephthalate film. In some embodiments, the first layer and the second layer are packaged between waterproof polymer sheets. In some embodiments, the direct-current electrical signal is generated by conversion of mechanical energy from relative sliding contact into electrical energy.Attorney Docket No.: 011520.02056
[0021] In some embodiments, the present disclosure provides a robotic manipulation system including a robotic gripper, a tactile sensor mounted to at least one finger of the robotic gripper, and a controller. The tactile sensor may include a first layer comprising a textile substrate and a tribovoltaic material disposed on fibers of the textile substrate, and a second layer disposed adjacent to the first layer and comprising a second material. The first layer and the second layer may be arranged for relative sliding contact with one another such that frictional contact between the tribovoltaic material and the second material generates a direct-current electrical signal. The controller may be configured to adjust operation of the robotic gripper based at least in part on the direct-current electrical signal.
[0022] In some embodiments, the controller is configured to increase a gripping force of the robotic gripper in response to detected slip indicated by the direct-current electrical signal. In some embodiments, the controller is configured to modulate a servo motor to adjust a gripping force of the robotic gripper. In some embodiments, the controller is configured to compute a magnitude of an error based on a reference no-slip signal and an output signal derived from the direct-current electrical signal and to change a servo angle when the magnitude of the error satisfies a threshold condition. In some embodiments, the controller changes the servo angle by a predetermined angular increment in response to detected slip.
[0023] In some embodiments, the controller is configured to reduce linear displacement of an object held by the robotic gripper in response to detected slip. In some embodiments, the controller is configured to reduce rotational displacement of an object held by the robotic gripper in response to detected slip. In some embodiments, the robotic gripper includes a compliant gripper configured to adjust gripping force independently of grasp width. In some embodiments, the compliant gripper includes a variable-stiffness actuator. The variable-stiffness actuator may include one or more permanent magnets arranged to define an adjustable air gap, and the controller may adjust the adjustable air gap in response to detected slip.
[0024] In some embodiments, the controller is configured to use detected slip as an input to trajectory planning for the robotic gripper. In some embodiments, the controller is configured to provide detected slip as an input to a control policy for robotic manipulation. In some embodiments, the control policy is configured to adjust a grasping parameter based at least in part on the detected slip.Attorney Docket No.: 011520.02056
[0025] In some embodiments, a processor, controller, processing circuit, or computing device described herein may include one or more modules or components. Each module or component may include hardware, software, firmware, or a combination thereof. For example, a module or component may include a field-programmable gate array (FPGA), an applicationspecific integrated circuit (ASIC), a digital signal processor (DSP), a microcontroller, a central processing unit (CPU), an accelerated processing unit (APU), a graphics processing unit (GPU), a programmable logic device, a software module comprising instructions stored in memory and executed by one or more processors, a virtual machine, or combinations thereof. The memory may store instructions that, when executed, cause the processor, controller, processing circuit, or computing device to perform one or more operations described herein.
[0026] In some embodiments, the present disclosure provides a method of sensing slip. The method may include contacting a first layer with a second layer, the first layer comprising a textile substrate coated with a tribovoltaic material and the second layer comprising a second material; permitting relative sliding contact between the tribovoltaic material of the first layer and the second material of the second layer; generating a direct-current electrical signal at an interface between the tribovoltaic material and the second material in response to the relative sliding contact; and detecting slip based at least in part on the direct-current electrical signal.
[0027] In some embodiments, generating the direct-current electrical signal includes generating the direct-current electrical signal at a dynamic Schottky interface. In some embodiments, generating the direct-current electrical signal includes generating the direct-current electrical signal at a P-N junction. In some embodiments, generating the direct-current electrical signal includes converting mechanical energy from the relative sliding contact into electrical energy.
[0028] In some embodiments, the method further includes sensing a normal force using a capacitive sensor disposed adjacent to the second layer. In some embodiments, the method further includes distinguishing a shear-related slip stimulus from a normal-force stimulus based at least in part on the direct-current electrical signal and an output of the capacitive sensor. In some embodiments, detecting slip includes comparing an output voltage derived from the direct-current electrical signal to a threshold. In some embodiments, the threshold is about 30 mV. In some embodiments, the method further includes filtering the direct-current electrical signal before detecting slip.Attorney Docket No.: 011520.02056
[0029] In some embodiments, the method further includes controlling a robotic gripper based at least in part on the detected slip. Controlling the robotic gripper may include increasing a gripping force of the robotic gripper, changing a servo angle based on a magnitude of an error between a reference no-slip signal and an output signal derived from the direct-current electrical signal, changing a stiffness state of a compliant gripper by adjusting an air gap between permanent magnets, changing a trajectory-planning input, changing a control-policy input, or combinations thereof.
[0030] In some embodiments, the method further includes determining an estimated value of at least one of a contact force, sliding velocity, contact area, or current density using a relationship between the direct-current electrical signal and the at least one of the contact force, sliding velocity, contact area, or current density. In some embodiments, the relationship is based at least in part on a Hertzian contact model.
[0031] In some embodiments, the present disclosure provides a tribovoltaic textile layer for a tactile sensor. The tribovoltaic textile layer may include a textile substrate comprising a plurality of fibers and a conducting polymer coating disposed on the plurality of fibers. The conducting polymer coating may be configured to form, during sliding contact with a second material, a tribovoltaic interface that generates a direct-current electrical signal. In some embodiments, the conducting polymer coating includes poly(3,4-ethylenedioxythiophene) polystyrene sulfonate. In some embodiments, the textile substrate includes a stretchable fabric. In some embodiments, the stretchable fabric includes spandex and polyethylene terephthalate. In some embodiments, the conducting polymer coating wraps around individual fibers of the textile substrate. In some embodiments, the conducting polymer coating forms conductive interconnections between adjacent fibers of the textile substrate. In some embodiments, the conducting polymer coating includes three coating layers.
[0032] In some embodiments, the present disclosure provides a method of making a tribovoltaic textile layer. The method may include preparing a coating composition comprising a conducting polymer, applying the coating composition to a textile substrate, and drying or curing the textile substrate after applying the coating composition, thereby forming a conducting-polymer-coated textile substrate configured to generate a direct-current electrical signal during sliding contact with a second material.Attorney Docket No.: 011520.02056
[0033] In some embodiments, the conducting polymer includes poly(3,4-ethylenedioxythiophene) polystyrene sulfonate. In some embodiments, preparing the coating composition includes combining the conducting polymer with a conductivity -enhancing additive comprising ethylene glycol, dimethyl sulfoxide, or a combination thereof. In some embodiments, preparing the coating composition includes combining poly(3,4-ethylenedioxythiophene) polystyrene sulfonate, ethanol, isopropanol, and ethylene glycol in a weight ratio of about 100:50:30:2.05. In some embodiments, applying the coating composition includes dipping the textile substrate in the coating composition. In some embodiments, the method further includes placing the textile substrate under vacuum after applying the coating composition. In some embodiments, drying or curing the textile substrate includes heating the textile substrate. In some embodiments, the method further includes repeating applying the coating composition and drying or curing the textile substrate to form three coating layers.
[0034] In some embodiments, the method further includes assembling the conducting-polymer-coated textile substrate with a metal counter-layer by coupling the conducting-polymer-coated textile substrate and the metal counter-layer at an edge region while permitting relative sliding contact between the conducting-polymer-coated textile substrate and the metal counterlayer. In some embodiments, the method further includes placing the conducting-polymer-coated textile substrate and the metal counter-layer adjacent to a capacitive normal-force sensor. In some embodiments, the method further includes encapsulating the conducting-polymer-coated textile substrate, the metal counter-layer, and the capacitive normal-force sensor in one or more waterproof polymer sheets.Description of the Drawings
[0035] For a fuller understanding of the nature and objects of the disclosure, reference should be made to the following detailed description taken in conjunction with the accompanying drawings.
[0036] Fig. 1. Comparison of human and the Bionic Textile Sensing (BTS) integrated robotic sensory' and actuation systems: (a) For humans, different types of mechanoreceptors in the skin detect tactile stimuli, transmit these signals to the brain for processing, followed by muscle actuation, (b) For robots, the proposed BTS system detects inputs, processes data through computers and processors with algorithms, and actuates motors to perform actions, (c) Schematic diagrams show the hypothetical scenario of touching experiments of a human finger withAttorney Docket No.: 011520.02056different steps, (d) Schematic diagrams show the analogous experimental setup and the steps for the touching experiment of BTS system based on dynamic DC generation mechanism with a modeled finger, (e) shows the afferent response signals generated in different mechanoreceptors, in which Rapidly Adapting (RA) type of mechanoreceptors (MR) being responsible for the identification of objects slipping from the human hand and the Slow Adapting (SA) MRs being responsible for the perception of object grasp. The signal further reaches the brain for processing and then to muscle for actuation, (f) The variation of force (F), capacitance (C) and open-circuit voltage (Toe) response at different steps in (d). BTS is capable of detecting slip due to the dynamic sliding between the PEDOT: PSS coated fabric and the titanium (Ti). Electronic excitation of electron-hole pairs is induced at the dynamic Schottky interface (Inset). Further, the signals are transmitted to the processor and to the servo for actuation.
[0037] Fig. 2 Fabrication and characterization of PEDOT: PSS-coated E-textile as fundamental element of BTS for dynamic DC generator-based self-powered sensing: (a) Schematic illustration of the process for fabricating PEDOT: PSS coated fabric on a spandex substrate, (b) Optical images showing the spandex fabric before and after 3 coatings, (c) Layer-by-layer structure of the BTS sensor prototype, (d) SEM images of fabric surface morphology with increasing number of PEDOT: PSS coating showing the microstructural changes, (e) SEM images with higher magnification, highlighting the coating quality' on the fibers, (f) Elemental EDS mapping of sulfur distribution on clean spandex fabric substrate without PEDOT: PSS coating and (g) after 3 coatings, showing uniformed distribution of the PEDOT: PSS ink on the fabric, (h) SEM image showing the individual strand of fiber from the uncoated fabric, (i) SEM image showing the fabric thread with successful wrapping of the PEDOT: PSS ink coating around the fiber and interconnects of the same ink between tw o adjacent strands, (j) Tensile test data displaying the change in resistance under mechanical deformation showing less than 13% of \ / ? / / ?o under the strain (s) over 90%. (k) Experimental setup for tensile testing of fabric using a mechanical tester. (1) I-V characteristics of the coated fabric under different normal forces in static conditions, (m) Statistical analysis of the open-circuit voltage (Foe) output in a Ti / PEDOT: PSS coated fabric dynamic DC generator system in one-million cyclic testing, (n) Schematic diagram of the experimental setup for evaluating the power output performance during sliding motions.
[0038] Fig. 3 Slip Detection and Mitigation with BTS-Integrated Robotic Gripper System: (a) Schematic of the algorithm used for the BTS system to detect and mitigate slip, (b)Attorney Docket No.: 011520.02056Schematic of the gripper's operation without feedback, illustrating continuous object movement: (1) Object positioned between robotic fingers; (2) Fingers engage the object, applying a grasping force ( / 'grip): (3) Mechanical tester pulls the object upwards with a force (Fpull) of 1.5 N; (4) Object continues to move as the maximum F_pull of 4.5 N is not reached; (5) Gripper fails to stop the object due to lack of feedback, resulting in a final displacement (D) of 30 mm. (c) Data plots for the gripper’s operation cycle stages, showing synchronized final displacement of the object ( / J). moving velocity of the object (v), BTS sensor feedback (Vout), change of grip force ( Fgrip), and change of the pulling force (ΔF_pull) exerted by the mechanical tester, (d) Photos of the gripper manipulating the object without feedback, showing continuous movement and final D of 30 mm. (e) Schematic of gripper operation with feedback activated, reducing displacement to D= 4.1 mm: (1) Object positioned between fingers; (2) Fingers engage the object; (3) Displacement occurs, but feedback is activated; (4) Gripper exerts higher force as feedback is initiated by the V_out, stopping object movement upon reaching maximum force; (5) Final displacement of 4.1 mm. (f) Data plots for the 4.1 mm displacement cycle, showing force, displacement, and system response, (g) Photos of the gripper demonstrating 4.1 mm displacement, (h) Schematic of the gripper’s initial position and response to added weights without feedback for rotational control: (1) Initial position; (2) Grasping; (3) Weights added, initiating object rotation; (4) Rotation without feedback, (i) Data plots showing system response without feedback, (j) Photo of the gripper rotating the object without feedback, resulting in a larger angle (6*=70°). (k) Schematic of initial position and response with feedback for rotational control: (1) Initial position; (2) Grasping; (3) Weights added to initiate rotation; (4) Rotation with feedback. (1) Data plots with feedback activated, demonstrating precise control, (m) Photos showing gripper rotation with feedback, achieving a smaller angle (9=7°).
[0039] Fig. 4 Investigation of Parametrized Slip Detection and Mitigation. Moving distance of the object when the gripper grasp force (Fgrip) reaches 3 N (a) at different gripper actuation velocities (vgripper), while the pulling velocity of the mechanical tester is kept constant (vpull= 18.33 mm / s). D decreases as vgripperincreases due to the faster actuation of the gripper in response to the detected slip; (b) At different pulling velocity of the mechanical tester (vpull), while the gripper actuation velocity is kept constant (vgripper= 126.75 mm / s). D decreases as vpullincreases due to the enhanced frictional power input, resulting in a faster rising of BTS signal and thus an earlier actuation of the gripper, (c) Final rotation angle (9) of the object at different gripper actuation velocity (vgripper). The weight of the added object is the same for all the fiveAttorney Docket No.: 011520.02056experiments (F= mg= 0.26 N). (d) Graph showing the response time of BTS system with different velocities and the colored boundaries indicating the response time for different types of human MRs. (e) Image of the BTS sensor integrated compliant gripper system on a robotic arm. (f) The detection time delay (Ata) vs. Fgrip. It is evident that the detection of slip via BTS sensor is faster than GelSight. (g) The 95% confidence interval of the detection delay for BTS and GelSight are 28.7 - 39.8 ms and 89 -103 ms respectively. It is evident that the detection of slip via BTS sensor is faster than via GelSight (t-test statistic: -13.42 and p-value: 4.7e-32). (h) Schematic of the step-by-step procedure of the expenment conduction where, Step #1: No grasp; Step #2: Engaged (minimum grasp force Fgripapplied to prevent container dropping); Step #3-9: Dropping of the 1st, 2nd, 3rd, 4th, 5th, 6th, 7th bead. Stainless beads with a mass of 17.0 g with 16 mm diameter are continuously dropped into a 3D printed container, (i) Video captures of the falling bead experiment with BTS control feedback at different steps, with their corresponding schematics depicted in (j) Force torque sensor signal, (k) BTS sensor signal, (1) GelSight score, and (m) Fgripas a function of time throughout the experiment. The increasing Fgripfrom the gripper at each bead-dropping event is triggered by the BTS sensor signal to mitigate the slip of the container.
[0040] Fig. 5 Quantitative analysis of force- and velocity- dependent DC output: (a) Short-circuit current ( / sc) generated by the PEDOT: PSS coated fabric on the modeled finger and Ti placed on the linear actuator driven at different velocities ranging from 0.75 to 116 mm / s. The inset shows the experimental setup with a force gauge and linear motor, (b) / sc generated under different applied forces (F) ranging from 1 to 5 N. (c) Schematic of the Hertz model with the contact area A under F. (d) Plot showing the / sc vs. F. The solid lines correspond to the data fitting based on Eq. 6. (e) plot shows the derived current density (Jsc) increasing with v due to the enhanced frictional energy input and the intensified electron-hole pair excitation with increasing v.
[0041] Fig. 6. Capacitance measured for ITO-PET normal force sensor under different force (a) 0.5 N. (b) 1.0 N, (c) 1.5 N, (d) 2.0 N (e) 2.5 N and (1) 3.0 N.
[0042] Fig. 7 Open-circuit voltage (Foe) output in a Ti / PEDOT: PSS coated fabric dynamic DC generator at different relative humidities.Attorney Docket No.: 011520.02056
[0043] Fig. 8 In static metal / PEDOT: PSS coated fabric contact, baseline Foe signal of the (a) Ti and (b) Al under dry condition, and (c) Ti and (d) Al under wet condition (water droplet added to the contact interface).
[0044] Fig. 9. (a) / sc and (b) Foe output of PEDOT: PSS coated fabric / titanium sliding contact and (c) / sc and (d) Foe output of PEDOT: PSS coated fabric / aluminum sliding contact with different sliding speeds from 50.01 mm / s to 116.43 mm / s.
[0045] Fig. 10. (a) Open-circuit voltage (Foe) output in a Ti / PEDOT: PSS coated fabric dynamic DC generator system with different fabric substrates, (b) Two-point resistance of the PEDOT: PSS coated fabric with different fabric substrates, (c) Optical image of different fabrics before coating, after coating (without bending), and after coating (with bending).
[0046] Fig. 11. (a)One million-cycle testing: open-circuit voltage (Foe) output in a Ti / PEDOT: PSS coated fabric dynamic DC generator system. The experiment was carried out continuously, with data collected in 10 consecutive sessions. Time-dependent data features of each session were shown in the figure. SEM images showing comparative images of the PEDOT: PSS coated fabrics sample used for 1,000,000 cyclic sliding test in 3 different areas as indicated in j): (b,e) Worn, (c,f) Edge, and (d,g) Unworn at different magnification. The corresponding EDS mapping is shown in (h), (i), (j).
[0047] Fig. 12. Open-circuit current (Foe) of the Ti / PEDOT: PSS coated fabrics dynamic DC generator system a) before and b) after 1000 cyclic pressing of the PEDOT: PSS coated fabrics. The maximum contact force is IN at 3.33 mm / s speed. SEM images showing PEDOT: PSS coated fabric (c,d) before and (f,g) after 1000 cycles compression testing at different magnification, along with the corresponding EDS mapping showing in (e) and (h).
[0048] Fig. 13. (a) Electrical circuit diagram of the sensor signal readout system and (b) the corresponding block diagram indicating the functionality of each part.
[0049] Fig. 14. (a) Table showing the conduction velocities for different types of mechanoreceptors and their converted response time range for comparison with BTS. (b) table showing the average response time for each velocity.Attorney Docket No.: 011520.02056
[0050] Fig. 15. 3D plots showing (a) Foe and (b) / sc output of PEDOTPSS coated fabric-titanium dynamic Schottky DC contact as a function of applied force (F) and sliding speed (v).
[0051] Fig. 16. Short-circuit current ( sc) generation for the titanium-PEDOTPSS system with varying velocities at applied force (F) of (a) 0.5 N, (b) 1.0 N, (c) 1.5 N and (d) 2.0 N.
[0052] Fig. 17. Short-circuit current ( / sc) generation for the titanium-PEDOTPSS system with varying velocities at applied force (F) of (a) 2.5 N, (b) 3.0 N, (c) 3.5 N and (d) 4.0 N.
[0053] Fig. 18. Short-circuit current ( / sc) generation for the titanium-PEDOTPSS system with varying velocities for applied force (F) of (a) 4.5 N, (b) 5.0 N, (c) force (F) vs short-circuit current ( / sc), the error bars represent the standard deviation (SD) of the sample data collected for each velocity at their respective forces and (d) velocity (v) vs short-circuit current ( / sc), the error bars represent the standard deviation (SD) of the sample data collected for each force at their respective velocities.
[0054] Fig. 19: Flowchart of a robotic control system integrating multi-modal sensing and reinforcement learning. The system includes path planning, trajectory planning with slip detection as an additional degree of freedom, error compensation, and state estimation for enhanced object manipulation. The tactile signal is used in the reinforcement learning (RL) policy to improve grip stability and dexteritv.
[0055] Fig. 20. a) Two-point resistance (R) measurement of the DMSO and the EG treated PEDOTPSS fabric with 3 coatings, b) Short-circuit current ( / sc) and c) open-circuit current (Foe) of the DMSO and the EG treated PEDOTPSS coated fabrics / Ti tribovoltaic DC generator system.
[0056] Fig. 21. Example of Gelsight Mini image data collected to deduce the Gelsight score (a) No contact with object container, (b) Low grasp force before slip, (c) Low grasp force after 1stslip, (d) Low grasp force after 2ndslip, (e) No contact with object container, (f) High grasp force before slip, (g) High grasp force after slip.Attorney Docket No.: 011520.02056
[0057] Figure 22. An exploded perspective view of a tactile sensor assembly according to an embodiment of the present disclosure.
[0058] Figure 23. A cross-sectional view of a tactile sensor assembly according to another embodiment of the present disclosure.
[0059] Figure 24. An enlarged detail of a textile substrate including textile yams / fibers and a conducting polymer coating.
[0060] Figure 25. A chart of a method according to another embodiment of the present disclosure.Detailed Description of the Disclosure
[0061] The present disclosure relates to tactile sensing assemblies, tactile sensors, tribovoltaic textile layers, robotic manipulation systems including tactile sensors, and methods of making and using the same. Various embodiments provide a textile-based tactile sensing structure configured to generate a direct-current electrical signal in response to relative sliding contact between two layers. The direct-current electrical signal may be used to detect slip, to monitor relative motion, to estimate one or more contact-related parameters, and / or to control a robotic gripper or other robotic manipulation device.
[0062] With reference to Figure 22, in some embodiments, a tactile sensor assembly 200 includes a first layer 202 and a second layer 204 disposed adjacent to the first layer. The first layer includes a textile substrate and a tribovoltaic material disposed on fibers of the textile substrate. The second layer includes a second material. The tribovoltaic material of the first layer and the second material of the second layer are arranged for relative sliding contact with one another such that frictional contact between the tribovoltaic material and the second material generates a direct-current electrical signal.
[0063] In some embodiments, a tribovoltaic material may include a material that contributes to generation of a direct-current electrical signal associated with frictional or sliding contact with another material. The tribovoltaic material may include, for example, a conducting polymer, a semiconducting polymer, an organic semiconductor, an inorganic semiconductor, a composite material, a coated material, a doped material, or a combination thereof. In someAttorney Docket No.: 011520.02056embodiments, the tribovoltaic material is disposed on, coated on, impregnated into, wrapped around, adhered to, or otherwise supported by fibers of a textile substrate.
[0064] In some embodiments, the second material comprises a material of the second layer positioned for frictional or sliding contact w ith the tribovoltaic material of the first layer. The second material may be selected so that relative sliding contact between the tribovoltaic material and the second material generates a direct-current electrical signal. The second material may include, for example, an electrically conductive material, a metal, a metal alloy, a conducting composite, a semiconducting material, or another material suitable for forming a tribovoltaic interface with the tribovoltaic material during frictional or sliding contact. In some embodiments, the tribovoltaic material and the second material form a tribovoltaic interface during relative sliding contact.
[0065] The tribovoltaic interface may include a dynamic Schottky interface, a P-N junction, or another interface configured to convert mechanical energy associated with frictional or sliding contact into electrical energy. In some embodiments, mechanical energy from the relative sliding contact provides energy for generating the direct-current electrical signal. In some embodiments, the direct-current electrical signal is generated by conversion of mechanical energy from the relative sliding contact into electrical energy at the tribovoltaic interface.
[0066] In some embodiments, the tribovoltaic interface includes a dynamic Schottky interface. For example, the tribovoltaic material may include a conducting polymer and the second material may include a metal or metal alloy. During relative sliding contact, an interfacial electrical field at the dynamic Schottky interface may drive separation and migration of excited charge carriers, thereby generating a direct-current output. In some embodiments, the dynamic Schottky interface is formed between a conducting polymer-coated textile and a metal foil. In some embodiments, the conducting polymer-coated textile includes poly(3,4-ethylenedioxythiophene) polystyrene sulfonate disposed on fibers of a textile substrate, and the metal foil includes titanium or aluminum.
[0067] In some embodiments, the tribovoltaic interface includes a P-N junction. For example, the first layer and the second layer may include semiconducting materials of complementary conductivity types arranged for relative sliding contact. In some embodiments, a p-type semiconducting material may be disposed on or within the first layer and an n-type semiconducting material may be disposed on or within the second layer. In other embodiments,Attorney Docket No.: 011520.02056an n-type semiconducting material may be disposed on or within the first layer and a p-type semiconducting material may be disposed on or within the second layer. The semiconducting materials may include organic semiconducting materials, inorganic semiconducting materials, doped conducting polymers, doped semiconductor particles, semiconductor coatings, semiconductor composites, or combinations thereof. Relative sliding contact at the P-N junction may generate a direct-current electrical signal by tribovoltaic charge generation at the dynamic interface.
[0068] In some embodiments, where a P-N junction is used as the tribovoltaic interface, the first layer and second layer may include materials selected from semiconducting materials known to generate tribovoltaic direct-current outputs at dynamic junctions. Such materials may be selected based on conductivity type, work function, band alignment, surface condition, mechanical compliance, and compatibility with sliding contact. The P-N junction embodiments may be implemented using any suitable semiconducting materials that, when arranged for relative sliding contact, generate a direct-current electrical signal.
[0069] In embodiments in which the tribovoltaic interface includes a P-N junction, the first layer and the second layer may include complementary semiconductor materials selected to provide a built-in electric field at the sliding interface. For example, one of the first layer or the second layer may include a p-type semiconducting polymer, doped conducting polymer, p-type organic semiconductor, or p-type inorganic semiconductor coating, and the other of the first layer or the second layer may include an n-type semiconducting polymer, n-type organic semiconductor, n-type inorganic semiconductor coating, or doped semiconductor composite. The semiconductor materials may be disposed as coatings on textile fibers, films, foils, sheets, particles in a binder, or composite layers configured to maintain sliding contact during shear loading.
[0070] The direct-current electrical signal may be measured or represented in different ways. In some embodiments, the direct-current electrical signal includes an open-circuit voltage signal. In some embodiments, the direct-current electrical signal includes a short-circuit current signal. In some embodiments, the same underlying tribovoltaic interaction may be characterized by open-circuit voltage, short-circuit current, current density, charge, voltage amplitude, signal rise time, pulse amplitude, signal energy, or another electrical parameter. The direct-current electrical signal may have a substantially non-altemating output component during slidingAttorney Docket No.: 011520.02056contact, and may be distinguished from transient alternating-current outputs produced by some other types of tactile sensors.
[0071] The first layer may include any suitable textile substrate. In some embodiments, the textile substrate is stretchable. In some embodiments, the textile substrate includes spandex, polyester, nylon, cotton, polyethylene terephthalate, elastane, or combinations thereof. In some embodiments, the textile substrate includes spandex and polyester. In some embodiments, the polyester includes polyethylene terephthalate. In some embodiments, the textile substrate includes about 20% spandex and about 80% polyethylene terephthalate. In further embodiments, the textile substrate includes about 5% to about 40% spandex and about 60% to about 95% polyester, about 10% to about 30% spandex and about 70% to about 90% polyester, or about 15% to about 25% spandex and about 75% to about 85% polyester. In some embodiments, the textile substrate includes about 20% spandex and about 80% polyester.
[0072] The textile substrate may have any suitable construction. In some embodiments, the textile substrate includes a woven fabric, a non-woven fabric, a knit fabric, a braided fabric, a tufted fabric, or combinations thereof. In some embodiments, the textile substrate includes hydrophilic cotton. In some embodiments, the textile substrate includes a superhydrophobic fiber. In some embodiments, a textile substrate is selected based on wettability, flexibility, stretchability, surface roughness, fiber density, fiber diameter, porosity, thickness, mechanical compliance, or ability to receive and retain a tribovoltaic material. In some embodiments, the textile substrate is selected to permit intimate surface contact with the second layer while maintaining sufficient flexibility or compliance for mounting to a robotic finger or other contact surface.
[0073] The tribovoltaic material may be disposed on fibers of the textile substrate in any suitable manner. In some embodiments, the tribovoltaic material coats at least a portion of individual fibers. In some embodiments, the tribovoltaic material wraps around individual fibers. In some embodiments, the tribovoltaic material forms conductive interconnections between adjacent fibers of the textile substrate. In some embodiments, the tribovoltaic material penetrates spaces between fibers. In some embodiments, the tribovoltaic material forms a continuous or partially continuous coating over fibers. In some embodiments, the tribovoltaic material forms one or more conductive pathways along or across the textile substrate.Attorney Docket No.: 011520.02056
[0074] In some embodiments, the tribovoltaic material includes a conducting polymer. Suitable conducting polymers may include poly(3,4-ethylenedioxythiophene) polystyrene sulfonate, polyaniline, polypyrrole, derivatives thereof, blends thereof, doped forms thereof, or combinations thereof. In some embodiments, the conducting polymer includes poly(3,4-ethylenedioxythiophene) polystyrene sulfonate. In some embodiments, the first layer includes a poly(3,4-ethylenedioxythiophene) polystyrene sulfonate-coated fabric. The coating may be applied to the textile substrate by dip coating, spray coating, blade coating, roll coating, drop casting, printing, screen printing, inkjet printing, pad printing, vapor deposition, chemical deposition, physical deposition, impregnation, soaking, or another suitable coating method.
[0075] In some embodiments, the tribovoltaic material is applied to the textile substrate in multiple coating layers. For example, the textile substrate may include one coating layer, two coating layers, three coating layers, four coating layers, five coating layers, or more than five coating layers. In some embodiments, the textile substrate includes three coating layers. In some embodiments, each coating cycle includes applying a coating composition and drying or curing the coated textile substrate. Multiple coating layers may increase conductivity, improve coating continuity, enhance coating penetration into the textile substrate, reduce discontinuities between fibers, or improve signal generation during sliding contact.
[0076] In some embodiments, a conducting polymer coating may be prepared by combining a conducting polymer with one or more conductivity-enhancing additives and / or solvents. In some embodiments, the conducting polymer includes poly(3,4-ethylenedioxythiophene) polystyrene sulfonate. In some embodiments, the coating composition includes poly(3,4-ethylenedioxythiophene) polystyrene sulfonate and ethanol. In some embodiments, the coating composition includes poly(3,4-ethylenedioxythiophene) polystyrene sulfonate, ethanol, isopropanol, and ethylene glycol. In some embodiments, the coating composition includes a conductivity-enhancing additive including ethylene glycol, dimethyl sulfoxide, or a combination thereof. In some embodiments, ethanol promotes dispersion of conducting polymer components, isopropanol adjusts wetting or viscosity, and ethylene glycol or dimethyl sulfoxide enhances electrical conductivity of the conducting polymer coating.
[0077] In one exemplary coating formulation, a first mixture includes poly(3,4-ethylenedioxythiophene) polystyrene sulfonate and ethanol. A second mixture includes ethylene glycol and isopropanol. The first and second mixtures are combined to form a coatingAttorney Docket No.: 011520.02056composition. In some embodiments, a weight ratio of poly(3,4-ethylenedioxythiophene) polystyrene sulfonate:ethanol:isopropanol:ethylene glycol is about 100:50:30:2.05. In other embodiments, the weight ratio may vary. For example, the weight ratio of poly(3,4-ethylenedioxythiophene) polystyrene sulfonate to ethanol may be from about 100:10 to about 100:100, from about 100:25 to about 100:75, or about 100:50. The weight ratio of poly(3,4-ethylenedioxythiophene) polystyrene sulfonate to isopropanol may be from about 100:5 to about 100:75, from about 100: 15 to about 100:50, or about 100:30. The weight ratio of poly(3,4-ethylenedioxythiophene) polystyrene sulfonate to ethylene glycol may be from about 100:0.1 to about 100:10, from about 100:0.5 to about 100:5, or about 100:2.05.
[0078] In some embodiments, the coating composition may be mixed before use. Mixing may be performed by stirring, magnetic stirring, sonication, shaking, vortexing, homogenizing, or combinations thereof. In some embodiments, mixing is performed for about 1 hour to about 72 hours, about 12 hours to about 60 hours, about 24 hours to about 60 hours, or about 48 hours. In some embodiments, the coating composition is prepared to improve penetration of the conducting polymer into the textile substrate and to promote uniform coating of fibers.
[0079] In some embodiments, a textile substrate is dipped in the coating composition. The dipped textile substrate may then be placed under vacuum. For example, the textile substrate may be placed in a vacuum desiccator after applying the coating composition. In some embodiments, vacuum treatment is performed for about 1 minute to about 60 minutes, about 5 minutes to about 30 minutes, or about 15 minutes. Vacuum treatment may assist penetration of the coating composition into the textile substrate and reduce entrapped air between fibers.
[0080] In some embodiments, the coated textile substrate is dried or cured. Drying or curing may include heating the textile substrate. In some embodiments, drying or curing is performed in an oven, a vacuum oven, a convection oven, an infrared oven, a hot plate, or another heating device. In some embodiments, drying or curing is performed at a temperature of about 60° C. to about 180° C., about 80° C. to about 150° C., about 100° C. to about 140° C., or about 120° C. In some embodiments, drying or curing is performed for about 10 minutes to about 5 hours, about 30 minutes to about 2 hours, or about 1 hour. In some embodiments, applying the coating composition and drying or curing the textile substrate are repeated to form multiple coating layers, such as three coating layers.Attorney Docket No.: 011520.02056
[0081] The coated textile substrate may exhibit electrical conductivity that remains stable during mechanical deformation. In some embodiments, the textile substrate coated with the tribovoltaic material exhibits a resistance change of less than about 25%, less than about 20%, less than about 15%, less than about 13%, less than about 10%, or less than about 5% under about 90% tensile strain. In some embodiments, the textile substrate coated with the tribovoltaic material exhibits a resistance change of less than about 13% under about 90% tensile strain. In some embodiments, the coated textile substrate maintains conductive pathways through coating continuity around fibers and interconnections between adjacent fibers.
[0082] The second layer may include any suitable second material selected to generate a direct-current electrical signal during relative sliding contact with the tribovoltaic material. In some embodiments, the second material includes an electrically conductive material. In some embodiments, the electrically conductive material includes a metal, a metal alloy, a conducting composite, a conductive oxide, a conductive coating, a conductive polymer, or a combination thereof. In some embodiments, the electrically conductive material includes titanium, aluminum, copper, stainless steel, nickel, gold, platinum, silver, an alloy thereof, or combinations thereof. In some embodiments, the second layer includes titanium or aluminum. In some embodiments, the second layer includes a titanium foil. In some embodiments, the second layer includes an aluminum foil. Titanium may provide desirable corrosion resistance and stable contact behavior in certain environments.
[0083] The second layer may have any suitable physical form. In some embodiments, the second layer includes a foil, sheet, film, coating, plate, laminate, deposited layer, patterned layer, mesh, textile, or composite. In some embodiments, the second layer is a metal foil. In some embodiments, the second layer is a titanium foil. The second layer may have a thickness selected to provide flexibility, durability, conductivity, contact stability, and compatibility with packaging. In some embodiments, the second layer has a thickness of about 1 pm to about 1 mm, about 5 pm to about 500 pm, about 10 pm to about 250 pm, or about 25 pm to about 150 pm.
[0084] The first layer and the second layer may be arranged in any suitable configuration that permits relative sliding contact between the tribovoltaic material of the first layer and the second material of the second layer. In some embodiments, the first layer and the second layer are disposed face-to-face. In some embodiments, the tribovoltaic material faces the second material. In some embodiments, at least a portion of the tribovoltaic material directly contactsAttorney Docket No.: 011520.02056the second material. In other embodiments, the tribovoltaic material and the second material may contact through surface roughness, asperities, openings, or localized contact regions. In some embodiments, the first and second layers are arranged so that shear motion, slip, vibration, sliding, or relative displacement causes the tribovoltaic material and the second material to undergo frictional contact.
[0085] The first layer may be attached to the second layer. For example, the first layer and the second layer may be adhered at a perimeter (and / or other locations) using an adhesive. The second layer is in slidable contact with the first layer such that friction contact may cause a charge. In some embodiments, the first layer and the second layer are attached to one another by a retainer 250 (e.g., a portion of the first layer and a portion of the second layer, for example, at an edge portion, a central portion, more than one portion, etc.). In some embodiments, the first layer and the second layer are constrained relative to one another at an edge region while permitting relative sliding contact in an interior region. For example, the first layer and the second layer may be coupled at a perimeter or edge region using an adhesive while central or interior portions remain free to undergo relative sliding contact. In some embodiments, the adhesive includes a double-sided adhesive layer. In some embodiments, the adhesive is disposed along one edge, two edges, three edges, four edges, a perimeter, a partial perimeter, or one or more discrete edge regions. In some embodiments, the adhesive is omitted from at least a portion of an interior region so that the tribovoltaic material and the second material may move relative to one another under shear loading.
[0086] The retainer may maintain assembly integrity while permitting motion between the tribovoltaic material and the second material in a sliding region outside the retainer. In some embodiments, the retainer limits separation of the first and second layers while permitting local relative sliding between opposing surfaces in the sliding region. In some embodiments, the retainer controls the amount of relative displacement between a titanium foil and apoly(3,4-ethylenedioxythiophene) polystyrene sulfonate-coated fabric. In some embodiments, the retainer includes adhesive, double-sided adhesive, stitching, heat bonding, welding, mechanical clamping, lamination, encapsulation, a frame structure, a seal, one or more discrete attachment locations, or combinations thereof.
[0087] In some embodiments, a normal -force sensor 230 is disposed adjacent to the second layer. The normal-force sensor may be configured to generate a normal-force signal inAttorney Docket No.: 011520.02056response to normal loading. In some embodiments, the normal-force sensor is a capacitive sensor. In some embodiments, the capacitive sensor may include a dielectric layer and one or more conductive layers. In some embodiments, the capacitive sensor includes an indium-tin-oxide-coated polymer film. In some embodiments, the polymer film includes polyethylene terephthalate. In some embodiments, an indium-tin-oxide-coated polyethylene terephthalate film is disposed beneath a metal layer, such as a titanium foil, to provide a capacitive normal-force sensor.
[0088] In some embodiments, the normal-force sensor generates a signal that varies with applied normal force, pressure, compression, or normal displacement. The tribovoltaic direct-current signal may be more sensitive to dynamic sliding or shear-related motion, while the normal-force sensor signal may be more sensitive to sustained normal loading. In this manner, the tactile sensor assembly may provide multimodal sensing, including a slip or shear-related signal and a normal-force signal. In some embodiments, the tribovoltaic signal is generated in response to dynamic sliding between the first layer and the second layer, while the normal-force signal is generated in response to normal loading of the normal-force sensor.
[0089] In some embodiments, at least a portion of the first layer and the second layer is disposed within or surrounded by an encapsulation layer 220. In some embodiments, the encapsulation layer includes one or more waterproof polymer sheets. In some embodiments, the encapsulation layer includes vinyl sheets, waterproof vinyl adhesive sheets, elastomer sheets, polymer films, silicone layers, polyurethane layers, thermoplastic films, or combinations thereof. The encapsulation layer may protect the first layer, the second layer, the normal-force sensor, electrical contacts, or other components from moisture, dust, abrasion, mechanical damage, or environmental exposure. In some embodiments, the first layer and the second layer are packaged between waterproof polymer sheets.
[0090] In some embodiments, the tactile sensor assembly includes electrical contacts or lead wires 240 electrically coupled to the first layer 202 and second layer 204. In some embodiments, the lead wires are coupled using conductive paste, conductive adhesive, solder, mechanical clamping, crimping, printed conductive traces, conductive tape, or combinations thereof. In some embodiments, Ag / AgCl paste is used to attach lead wires to the sensor. The electrical contacts may be coupled to a processing circuit, data acquisition device, controller, amplifier, filter, analog-to-digital converter, or robotic control system.Attorney Docket No.: 011520.02056
[0091] In some embodiments, the tactile sensor assembly further includes a processing circuit electrically coupled to the first layer and the second layer. The processing circuit may be configured to detect slip based at least in part on the direct-current electrical signal. The processing circuit may include one or more analog circuits, digital circuits, microcontrollers, processors, data acquisition devices, amplifiers, filters, comparators, anal og-to-digi tai converters, memory devices, communication interfaces, or combinations thereof. In some embodiments, the processing circuit receives an open-circuit voltage, short-circuit current, amplified voltage, filtered voltage, digitized voltage, or other signal derived from the direct-current electrical signal.
[0092] In some embodiments, the processing circuit identifies slip when an output voltage derived from the direct-current electrical signal satisfies a threshold condition. The threshold condition may include the output voltage being greater than or equal to a threshold voltage, the output voltage crossing a threshold voltage, a filtered voltage crossing a threshold voltage, a moving average crossing a threshold voltage, a derivative or rate of change satisfying a threshold, a peak satisfying a threshold, or a pattern of peaks satisfying a threshold. In some embodiments, the threshold voltage is about 30 mV. In other embodiments, the threshold voltage may be about 1 mV to about 200 mV, about 3 mV to about 100 mV, about 10 mV to about 75 mV, about 20 mV to about 50 mV, or about 30 mV. Thresholds may be selected based on sensor construction, amplification, filtering, noise level, contact materials, intended use, or calibration.
[0093] In some embodiments, slip is detected when a sensor output signal, such as an amplified and / or filtered voltage output, satisfies a threshold condition. In some embodiments, the threshold condition includes the sensor output signal being at least about 30 mV. In other embodiments, the threshold condition may include an error signal based on a reference no-slip signal and the sensor output signal satisfying an error threshold. The sensor-output threshold and the error threshold may be the same or different, and may be selected based on signal amplification, filtering, noise level, sensor configuration, and control objective. In some embodiments, a controller changes a servo angle by a predetermined increment when the threshold condition is satisfied.
[0094] In some embodiments, the processing circuit applies filtering before detecting slip. Filtering may include analog filtering, digital filtering, low-pass filtering, high-pass filtering, band-pass filtering, finite impulse response filtering, infinite impulse response filtering, moving-average filtering, median filtering, smoothing, baseline subtraction, noise thresholding.Attorney Docket No.: 011520.02056peak detection, or combinations thereof. In some embodiments, the processing circuit applies a moving-average fdter to the direct-current electrical signal before detecting slip. The movingaverage filter may be applied over a window of about 2 to about 100 samples, about 3 to about 50 samples, about 5 to about 25 samples, or about 10 samples. In some embodiments, a moving average over previous data points acts as a low-pass filter before threshold-based slip detection.
[0095] In some embodiments, a magnitude of the direct-current electrical signal varies with one or more parameters associated with the relative sliding contact. Such parameters may include sliding velocity, applied force, contact area, current density, frictional power, frictional force, contact pressure, contact state, humidity, surface condition, or combinations thereof. In some embodiments, the direct-current electrical signal increases with increasing sliding velocity. In some embodiments, the direct-current electrical signal increases with increasing applied force. In some embodiments, short-circuit current varies with contact area, and contact area varies with applied force. In some embodiments, current density varies with sliding velocity.
[0096] In some embodiments, a processing circuit, controller, or other computing device determines an estimated value of at least one contact-related parameter using a relationship between the direct-current electrical signal and the parameter. The estimated value may include contact force, sliding velocity, contact area, current density, frictional power, frictional force, contact pressure, or combinations thereof. In some embodiments, the relationship is a calibrated relationship. In some embodiments, the relationship is based on a lookup table, regression model, polynomial fit, physics-based model, empirical model, machine-learning model, or combinations thereof. In some embodiments, the relationship is based at least in part on a Hertzian contact model.
[0097] In some embodiments, a Hertzian contact model relates a contact area to applied force. In some embodiments, the contact radius is modeled as a function of a tip radius, an applied force, and a reduced modulus. The contact area may be determined from the contact radius. Short-circuit current may be modeled as proportional to the contact area multiplied by a current density. In some embodiments, the current density' is a function of sliding velocity.Accordingly, the short-circuit current may be modeled as a function of applied force and sliding velocity. In some embodiments, the short-circuit current is proportional to applied force raised to a power, such as about 2 / 3, under selected modeling assumptions. In some embodiments, theAttorney Docket No.: 011520.02056model is used to correlate the direct-current electrical signal with applied force, sliding velocity, contact area, or current density.
[0098] In some embodiments, the tactile sensor assembly may exhibit rapid response to sliding motion. In some embodiments, in a response-time test in which a voltage from the tactile sensor assembly is compared to a threshold voltage of about 3 mV, the tactile sensor assembly produces a voltage exceeding the threshold voltage within about 0.7 ms to about 40 ms after initiation of sliding motion, depending on sliding velocity. In further embodiments, the response time may be about 0.1 ms to about 100 ms, about 0.5 ms to about 50 ms, about 0.7 ms to about 40 ms, about 1 ms to about 30 ms, or about 1 ms to about 20 ms, depending on sensor construction, sliding velocity, applied force, threshold, filtering, sampling frequency, and contact materials. In some embodiments, a response-time threshold is determined after characterizing noise in the absence of sliding motion.
[0099] In some embodiments, the tactile sensor assembly may exhibit durability under repeated sliding cycles. In some embodiments, the first layer includes a poly(3.4-ethylenedioxythiophene) polystyrene sulfonate-coated fabric and the second layer includes a titanium foil. In some embodiments, when subjected to a sliding test at a load of about 2.5 N, a sliding speed of about 100 mm / s, and a sliding distance of about 5 mm, the tactile sensor assembly is configured to generate an open-circuit voltage of at least about 30 mV after about 400,000 sliding cycles. In some embodiments, the tactile sensor assembly is configured to generate an open-circuit voltage of at least about 20 mV, at least about 30 mV, at least about 35 mV, or at least about 40 mV after about 100,000, 200,000, 400,000, 500,000, 750,000, or 1.000,000 sliding cycles under selected test conditions.
[0100] In some embodiments, a tactile sensor includes the first and second layers, a retainer, a normal-force sensor, and a processing circuit. The first layer may include a textile substrate 310 and a conducting polymer 312 disposed on fibers of the textile substrate (see, e.g., Fig. 24). The second layer may include a metal layer disposed adjacent to the first layer. The retainer may join the first layer and the second layer to one another while permitting at least a portion of the first layer and at least a portion of the second layer outside the retainer to undergo relative sliding contact with one another. The normal-force sensor may be disposed adjacent to the second layer. The processing circuit may be configured to generate a slip signal based at least in part on a direct-current electrical signal generated by relative sliding contact between the firstAttorney Docket No.: 011520.02056layer and the second layer, and to generate a normal-force signal based at least in part on an output of the normal-force sensor.
[0101] In some embodiments, the conducting polymer of the tactile sensor includes poly(3,4-ethylenedioxythiophene) polystyrene sulfonate. In some embodiments, the metal layer includes titanium or aluminum. In some embodiments, the metal layer includes titanium foil. In some embodiments, the normal-force sensor includes a capacitive sensor comprising an indium-tin-oxide-coated polyethylene terephthalate film. In some embodiments, the first layer and the second layer are packaged between waterproof polymer sheets. In some embodiments, the direct-current electrical signal is generated by conversion of mechanical energy7from relative sliding contact into electrical energy.
[0102] In some embodiments, the tactile sensor may be mounted to a robotic finger, robotic gripper, robotic hand, end effector, compliant gripper, soft robot, prosthetic component, haptic device, human-machine interface, wearable device, or other device configured to interact with an object. In some embodiments, the tactile sensor is mounted to a finger of a robotic gripper such that relative sliding between an object and the gripper produces sliding contact between the first layer and the second layer. The direct-current electrical signal may indicate slip of the object relative to the gripper.
[0103] In some embodiments, a robotic manipulation system includes a robotic gripper comprising at least one finger, a tactile sensor mounted to the at least one finger, and a controller configured to adjust operation of the robotic gripper based at least in part on the direct-current electrical signal. The tactile sensor may include a first layer comprising a textile substrate and a tribovoltaic material disposed on fibers of the textile substrate, and a second layer disposed adjacent to the first layer and comprising a second material. The first layer and the second layer may be arranged for relative sliding contact with one another such that frictional contact between the tribovoltaic material and the second material generates the direct-current electrical signal.
[0104] The controller may include one or more processors, microcontrollers, motor controllers, servo controllers, data acquisition devices, analog-to-digital converters, memory7devices, communication interfaces, or combinations thereof. In some embodiments, the controller receives an analog signal derived from the tactile sensor. In some embodiments, the controller receives a filtered or amplified output signal. In some embodiments, the controllerAttorney Docket No.: 011520.02056converts the direct-current electrical signal to a digital signal. In some embodiments, the controller compares the signal to a threshold and detects slip when the threshold is satisfied.
[0105] In some embodiments, the controller is configured to increase a gripping force of the robotic gripper in response to detected slip indicated by the direct-current electrical signal. In some embodiments, the controller is configured to modulate a servo motor to adjust a gripping force of the robotic gripper. In some embodiments, the controller computes a magnitude of an error based on a reference no-slip signal and an output signal derived from the direct-current electrical signal, and changes a servo angle when the magnitude of the error satisfies a threshold condition. In some embodiments, the reference no-slip signal corresponds to zero slip. In some embodiments, the servo angle is changed by a predetermined angular increment in response to detected slip. In some embodiments, the predetermined angular increment is about 0.1° to about 5°, about 0.2° to about 2°, about 0.5° to about 1.5°, or about 0.8°.
[0106] In some embodiments, the robotic manipulation system is configured to reduce linear displacement of an object held by the robotic gripper in response to detected slip. For example, when an object is pulled relative to the gripper, the direct-current electrical signal may indicate sliding between the object and the gripper, and the controller may increase the gripping force to reduce or stop further linear displacement. In some embodiments, the robotic manipulation system is configured to reduce rotational displacement of an object held by the robotic gripper in response to detected slip. For example, when an off-center load or external disturbance causes rotation of an object, the direct-current electrical signal may indicate rotational slip, and the controller may increase gripping force or adjust another gripper parameter to reduce angular displacement.
[0107] In some embodiments, the robotic gripper includes a compliant gripper configured to adjust gripping force independently of grasp width. In some embodiments, the compliant gripper includes a variable-stiffness actuator. In some embodiments, the variablestiffness actuator includes one or more permanent magnets arranged to define an adjustable air gap. The controller may adjust the adjustable air gap in response to detected slip. In some embodiments, decreasing or increasing the air gap changes a stiffness state or gripping force of the compliant gripper. In some embodiments, the controller incrementally adjusts the air gap in response to one or more detected slip events. In some embodiments, an air gap is adjusted byAttorney Docket No.: 011520.02056about 0.1 mm to about 5 mm, about 0.5 mm to about 2 mm, or about 1.3 mm per detected slip event.
[0108] In some embodiments, the controller uses detected slip as an input to trajectory planning for the robotic gripper. For example, the controller may alter a planned motion, grasping trajectory, regrasping strategy, object manipulation path, or force profde based on slip information. In some embodiments, the controller provides detected slip as an input to a control policy for robotic manipulation. The control policy may adjust a grasping parameter based at least in part on detected slip. The grasping parameter may include gripping force, servo angle, air gap, stiffness state, finger position, contact force, grasp width, trajectory, actuation velocity, object retention strategy, or combinations thereof. In some embodiments, a control policy may receive detected slip as a state input and select an action based at least in part on the detected slip.
[0109] In some embodiments, a method of sensing slip includes contacting a first layer with a second layer, permitting relative sliding contact between a tribovoltaic material of the first layer and a second material of the second layer, generating a direct-current electrical signal at an interface between the tribovoltaic material and the second material in response to the relative sliding contact, and detecting slip based at least in part on the direct-current electrical signal. The first layer may include a textile substrate coated with the tribovoltaic material, and the second layer may include the second material. In some embodiments, generating the direct-current electrical signal includes generating the direct-current electrical signal at a dynamic Schottky7interface. In some embodiments, generating the direct-current electrical signal includes generating the direct-current electrical signal at a P-N junction. In some embodiments, generating the direct-current electrical signal includes converting mechanical energy from the relative sliding contact into electrical energy.
[0110] In some embodiments, the method further includes sensing a normal force using a capacitive sensor disposed adjacent to the second layer. The method may include distinguishing a shear-related slip stimulus from a normal-force stimulus based at least in part on the direct-current electrical signal and an output of the capacitive sensor. In some embodiments, the direct-current electrical signal is used to detect sliding, slip, shear, or dynamic relative motion, while the capacitive sensor output is used to detect normal force, grasp force, compression, or sustained loading.Attorney Docket No.: 011520.02056[OHl] In some embodiments, detecting slip includes comparing an output voltage derived from the direct-current electrical signal to a threshold. In some embodiments, the threshold is about 30 mV. In other embodiments, the threshold is about 1 mV to about 200 mV, about 3 mV to about 100 mV, about 10 mV to about 75 mV, about 20 mV to about 50 mV, or another selected value. In some embodiments, the method further includes filtering the direct-current electrical signal before detecting slip. Filtering may include moving-average filtering, low-pass filtering, smoothing, peak detection, noise suppression, or combinations thereof.
[0112] In some embodiments, the method includes controlling a robotic gripper based at least in part on detected slip. Controlling the robotic gripper may include increasing a gripping force, changing a servo angle, changing a stiffness state, adjusting an air gap between permanent magnets, changing a trajectory, changing an actuation velocity, changing a grasping parameter, or combinations thereof. In some embodiments, changing a servo angle is based on a magnitude of an error between a reference no-slip signal and an output signal derived from the direct-current electrical signal. In some embodiments, changing a stiffness state includes adjusting an air gap between permanent magnets in a variable-stiffness gripper.
[0113] In some embodiments, the method includes determining an estimated value of at least one contact-related parameter using a relationship between the direct-current electrical signal and the contact-related parameter. The contact-related parameter may include contact force, sliding velocity, contact area, current density, frictional force, frictional power, contact pressure, or combinations thereof. In some embodiments, the relationship is based at least in part on a Hertzian contact model. In some embodiments, the relationship is based on calibration data collected by applying known forces, known velocities, known contact geometries, or combinations thereof. In some embodiments, the estimated value may be used to adjust robotic control or to characterize contact conditions.
[0114] In some embodiments, a tribovoltaic textile layer for a tactile sensor includes a textile substrate comprising a plurality of fibers and a conducting polymer coating disposed on the plurality of fibers. The conducting polymer coating is configured to form, during sliding contact with a second material, a tribovoltaic interface that generates a direct-current electrical signal. In some embodiments, the conducting polymer coating includes poly(3,4-ethylenedioxythiophene) polystyrene sulfonate. In some embodiments, the textile substrateAttorney Docket No.: 011520.02056includes a stretchable fabric. In some embodiments, the stretchable fabric includes spandex and polyethylene terephthalate.
[0115] In some embodiments, the conducting polymer coating wraps around individual fibers of the textile substrate. In some embodiments, the conducting polymer coating forms conductive interconnections between adjacent fibers of the textile substrate. In some embodiments, the conducting polymer coating includes three coating layers. The coating may provide conductive pathways across the textile layer while preserving stretchability. flexibility, and conformability of the textile substrate.
[0116] With reference to Figure 25, in some embodiments, a method 400 of making a tribovoltaic textile layer includes preparing 403 a coating composition comprising a conducting polymer, applying 406 the coating composition to a textile substrate, and drying or curing 412 the textile substrate after applying the coating composition. The method thereby forms a conducting-polymer-coated textile substrate configured to generate a direct-current electrical signal during sliding contact with a second material. In some embodiments, the conducting polymer includes poly(3,4-ethylenedioxythiophene) polystyrene sulfonate. In some embodiments, preparing the coating composition includes combining the conducting polymer with a conductivity-enhancing additive including ethylene glycol, dimethyl sulfoxide, or a combination thereof. In some embodiments, preparing the coating composition includes combining poly(3,4-ethylenedioxythiophene) polystyrene sulfonate, ethanol, isopropanol, and ethylene glycol in a weight ratio of about 100:50:30:2.05.
[0117] In some embodiments, applying the coating composition includes dipping the textile substrate in the coating composition. In some embodiments, the method further includes placing 409 the textile substrate under vacuum after applying the coating composition. In some embodiments, drying or curing the textile substrate includes heating the textile substrate. In some embodiments, the method further includes repeating 415 applying the coating composition and drying or curing the textile substrate to form three coating layers. In other embodiments, the coating and drying or curing cycle may be repeated to form one, two, four, five, or more than five coating layers.
[0118] In some embodiments, the method further includes assembling the conducting-polymer-coated textile substrate with a counter-layer 418 (e.g. a metal counter-layer) by coupling the conducting-polymer-coated textile substrate and the metal counter-layer at an edgeAttorney Docket No.: 011520.02056region while permitting relative sliding contact between the conducting-polymer-coated textile substrate and the metal counter-layer. The metal counter-layer may include titanium, aluminum, or another metal or metal alloy. In some embodiments, the metal counter-layer includes titanium foil. In some embodiments, coupling at the edge region includes applying an adhesive along a perimeter or edge of the metal counter-layer or the conducting-polymer-coated textile substrate.
[0119] In some embodiments, the method further includes disposing 421 the conducting-polymer-coated textile substrate and the metal counter-layer adjacent to a capacitive normalforce sensor. In some embodiments, the capacitive normal-force sensor includes an indium-tin-oxide-coated polyethylene terephthalate film. In some embodiments, the method further includes encapsulating 424 the conducting-polymer-coated textile substrate, the metal counter-layer, and the capacitive normal-force sensor in one or more waterproof polymer sheets. In some embodiments, the one or more waterproof polymer sheets include vinyl adhesive sheets.
[0120] In some embodiments, one or more operations described herein as being performed by a controller, processor, processing circuit, computing device, or robotic manipulation system may be implemented using instructions stored on one or more non-transitory computer-readable media. The instructions, when executed by one or more processors, may cause the one or more processors to receive a direct-current electrical signal from a textilebased tribovoltaic tactile sensor, determine that slip has occurred based at least in part on the direct-current electrical signal, compare an output voltage derived from the direct-current electrical signal to a threshold condition, filter the direct-current electrical signal, compute an error based on a reference no-slip signal and the output voltage, and adjust a control parameter of a robotic gripper in response to detected slip. The control parameter may include gripping force, servo angle, gripper stiffness, an air gap of a variable-stiffness gripper, a trajectory-planning input, a control-policy input, or another robotic manipulation parameter.
[0121] The components described herein may be combined in various configurations. For example, a tactile sensor assembly may include the first and second layers without a normalforce sensor in some embodiments, and may include a normal-force sensor, encapsulation layer, processing circuit, robotic gripper, or combinations thereof in other embodiments. A robotic manipulation system may include one tactile sensor or a plurality of tactile sensors. Tactile sensors may be mounted on one finger, multiple fingers, an opposing pair of fingers, a palm region, a gripper contact surface, or another robotic contact surface. Each tactile sensor mayAttorney Docket No.: 011520.02056provide a direct-current electrical signal, and one or more processing circuits may process the signals individually or collectively.
[0122] In some embodiments, dimensions of the tactile sensor assembly may be selected based on an intended application. The textile substrate, second layer, normal-force sensor, and encapsulation layer may each be cut, patterned, shaped, laminated, or otherwise formed to match a robotic finger, gripper pad. curved surface, planar surface, or compliant substrate. In some embodiments, the first layer and second layer have a lateral dimension of about 1 mm to about 200 mm, about 5 mm to about 100 mm, about 10 mm to about 50 mm, or another suitable size. In some embodiments, the tactile sensor assembly may be flexible, bendable, conformable, stretchable, or washable, depending on material selection and encapsulation.
[0123] In some embodiments, the tactile sensor assembly may operate under dry conditions, humid conditions, or wet-contact conditions. Material selection, encapsulation, and use of corrosion-resistant second materials may improve stability in humid or wet environments. In some embodiments, titanium is used as the second material to improve stability relative to aluminum in selected environments. In some embodiments, waterproof polymer sheets reduce moisture exposure at the sliding interface or protect electrical connections.
[0124] In some embodiments, the tactile sensor assembly may be used in robotic manipulation, manufacturing, industrial automation, exploration, medical robotics, healthcare robotics, human-machine interfaces, prosthetics, wearable tactile devices, smart textiles, or other systems that benefit from detection of slip, shear, relative motion, grasp force, or contact state. The tactile sensor assembly may provide a direct-current signal generated by contact mechanics at the sliding interface and may reduce or eliminate the need for externally powered sensing elements at the tribovoltaic interface itself. External power may still be used by associated processing circuits, controllers, amplifiers, data acquisition systems, communication circuits, robotic actuators, or other system components.
[0125] In another aspect, the present disclosure may be embodied as a tactile sensor. Such a tactile sensor may be useful in robotic applications, but not limited to such applications. The sensor includes a first layer comprising a first material. A second layer is adjacent the first layer. The second layer comprises a second material. The first material and the second material are selected configured to generate an electric charge upon frictional contact with a second layer. The second layer is disposed adjacent to the first layer. The materials of the first layer and theAttorney Docket No.: 011520.02056second layer may be selected to form a tribovoltaic interface (e.g, dynamic Schottky interface or P-N junction) upon frictional contact with each other. For example, the second layer may be a conductor (e.g.. a metal, metal alloy, conducting composite, etc.) or other suitable material, and the first layer may include a material which forms a dynamic Schottky generator or P-N junction with the material of the second layer. The first layer may be attached to the second layer. For example, the first layer and the second layer may be adhered at a perimeter (and / or other locations) using an adhesive. The second layer is in slidable contact with the first layer such that friction contact may cause a charge.
[0126] In an example, the first layer may include Poly(3,4-ethylenedioxythiophene) polystyrene sulfonate (PEDOT: PSS) and the second layer may include titanium or aluminum. The first layer may be configured as a fabric such as, for example, a woven fabric, non-woven fabric, knit fabric, braided fabric, or other fabric configuration. The material of the fabric may include a material creating a Schottky interface or P-N junction with the second material (e.g., PEDOT: PSS. etc.) The fabric may further include other materials such as, polyester, elastane, cotton, wool, etc., and combinations of these and / or other materials. In a particular example, a fabric of the first layer is an elastane (Spandex) polyester blend which is coated with PEDOT: PSS. The fabric may be a stretchable fabric (e.g., including elastane, etc.)
[0127] Some embodiments of the tactile sensor include a processor in electronic communication with the first layer and the second layer. The processor may be configured to provide a slip signal based on a voltage between the first layer and the second layer.
[0128] In some embodiments, the tactile sensor includes a dielectric layer adjacent to the second layer and a conductive layer adjacent to the insulative layer. In this way, a capacitive structure is formed by the second layer, the dielectric layer, and the conductive layer (sometimes referred to as the capacitive layer). The dielectric may be any suitable material (or combinations of materials) such as. for example, PET. The processor may be further configured to provide a force signal based on the capacitive layer.
[0129] In some embodiments, the tactile sensor includes an encasement. For example, an inner vinyl layer and an outer vinyl layer may cooperate to envelope at least the first layer and the second layer (see, for example. Figure 2(c)). The encasement may protect the enveloped components. For example, the encasement may be water resistant / waterproof, UV protectant.Attorney Docket No.: 011520.02056dust resistant, washable, humidity controlling, wear protecting, and / or may provide additional benefits or combinations.
[0130] In another aspect, the present disclosure may be embodied as a first layer, for example, for use in a tactile sensor. The first layer may be configured as described herein.
[0131] In another aspect, the present disclosure may be embodied as a method for making a tribovoltaic layer. The method includes coating a fabric with a tribovoltaic material. For example, the method may include mixing PEDOT: PSS and ethanol, mixing ethylene glycol and isopropanol, and combining the two mixtures, and coating a fabric using the combined mixtures. The method may further include drying the coated fabric. For example, the coated fabric may be dried using a vacuum desiccator. The method may include repeating the steps of coating the fabric and drying the coated fabric. For example, the steps may be repeated such that the fabric includes two, three, or more coats of the tribovoltaic material.
[0132] In another aspect, the present disclosure may be embodied as a controller for a robotic gripper. The controller may include a processor configured to use slip (for example, in trajectory planning). For example, slip detection may be used as an additional degree of freedom in trajectory planning. In this way, a robotic gripper may be controlled to adapt to various object manipulation tasks with higher precision and reduced errors.
[0133] In some instances, a suitable processor for any of the disclosed embodiments may include one or more modules and / or components. Each module / component executed by the processor can be any combination of hardware-based module / component (e.g, a field-programmable gate array (FPGA), an application specific integrated circuit (ASIC), a digital signal processor (DSP)), software-based module (e.g., a module of computer code stored in the memory and / or in the database, and / or executed at the processor), and / or a combination of hardware- and software-based modules. Each module / component executed by the processor is capable of performing one or more specific functions / operations as described herein. In some instances, the modules / components included and executed in the processor can be, for example, a process, application, virtual machine, and / or some other hardware or software module / component. The processor can be any suitable processor configured to run and / or execute such modules / components. The processor can be any suitable processing device configured to run and / or execute a set of instructions or code. For example, the processor can be a general-purpose processor, a central processing unit (CPU), an accelerated processing unitAttorney Docket No.: 011520.02056(APU), a field-programmable gate array (FPGA), an application specific integrated circuit (ASIC), a digital signal processor (DSP), and / or the like.
[0134] Additional Discussion
[0135] The following additional discussion describes certain illustrative and non-limiting examples, experimental configurations, materials, fabrication conditions, test conditions, results, and operating principles that may be used in connection with one or more embodiments of the present disclosure. The examples and results are provided to aid in understanding the disclosed subject matter and are not intended to limit the scope of the disclosure or the claims to any particular material, layer configuration, fabrication parameter, test condition, performance value, mechanism of operation, robotic platform, control algorithm, or combination of features. Unless expressly required by the claims, features described in connection with one example may be used separately from, omitted from, substituted for, or combined with features described in connection with another example.
[0136] Results
[0137] Tribovoltaic dynamic DC generation as a bionic friction sensing mechanism
[0138] To demonstrate this concept, a modeled contact mimicking an artificial finger was designed with a piece of PEDOTPSS coated stretchable E-textile fabric attached to a polydimethylsiloxane (PDMS) soft hemisphere (diameter d= 15 mm). In the meantime, a capacitive force sensor made of polyester (PET) thin film with Indium-Tin oxide (ITO) electrodes is placed on the linear actuator stage beneath the titanium (Ti) sheet (inset of Fig. If) (See Fig. 6 for characterization of the capacitive sensor). The synthesis and characterization of the E-textile will be discussed in the later section. The hemisphere is then brought into contact with a Ti foil fixed on a linear actuator stage with wired electrode. Fig. 1c and Id respectively show different steps of a representative frictional contact process with human finger and our modeled system, where step #1-7 respectively correspond to: step #1- engaging (no contact), step #2- compressing up to a force (F) of 5.0 N, step #3- standing still (no motion), step #4- linear reciprocal sliding with a force variation of <0.25 N, step#5- standing still (no motion), step#6-lifting (releasing the applied force, but the probe still in contact), step #7- lifting (out of contact). In step #1, the artificial finger is attached to a mechanical tester and moved down with a constant speed of 0.16 mm / s before contact. In step #2, it can be observed that as soon as the artificialAttorney Docket No.: 011520.02056finger contacts the Ti sheet and continuously exerts increasing F in normal direction, the output of capacitive sensor shows similar trend to the force applied. In the meantime, the BTS sensing signal (i.e., open-circuit voltage, Foe) remains negligible, indicating trivial slip between the E-textile fabric and the Ti when an increasing static force is applied in the normal direction. In step #3. the artificial finger stands still without any normal or longitudinal movements under a constant F of 5.0 N. When the linear actuator drives the Ti sheet at a speed of 53.63 mm / s for 10 seconds in step #4, a minor disturbance in the F with a small variation of <0.25 N is observed, corresponding to a <5% fluctuation of F. During this step, the artificial finger slides back and forth on the Ti sheet. It can be seen that the BTS shows a significant signal of ~ 50 mV (Fig. If), whereas the capacitive sensor shows negligible variation in the signal. Furthermore, steps #5, #6, and #7 are analogous to #3, #2, and #1 respectively, except the force is removed.
[0139] In the tribovoltaic effect, mechanical energy is converted into electrical energy through mechano-electronic excitation at dynamic Schottky interfaces. As the metal-conducting polymer interface undergoes continuous frictional contact, the interfacial electric field drives the separation and migration of the excited electrons and holes, resulting in the generation of a DC output. The results (Fig. If) prove that the tribovoltaic DC generation is substantially more sensitive to the dynamic sliding along the shear force direction compared to the normal force variation under static contact. From a bionic perspective, the characteristics of BTS signal resembles the RA MRs, which can operate in conjunction with the capacitive sensors as a SA MRs counterpart to feed input for robotic control algorithm similar to the process in the human sensory system (Fig. le).
[0140] Dynamic DC generator-integrated E-textile array as one-piece BTS unit
[0141] In order to implement the concept for robotic manipulation, it can be advantageous to integrate the dynamic Schottky generators into a one-piece form factor. Noting that E-textile materials have advantages to be woven, sewn, or manufactured into desirable structures, we developed a material system for dynamic DC generation featuring high electrical conductivity, stretchability, and manufacturability. To fabricate this structure, we developed an exemplary embodiment using a highly conductive PEDOT: PSS ink for coating onto a stretchable fabric substrate made of 20% spandex and 80% PET.
[0142] Fig. 2a shows the fabrication process (see Methods for more details). After three repeated coatings and thermal treatment, the PEDOT: PSS-coated fabric was cut into small,Attorney Docket No.: 011520.02056desired pieces for device integration (Fig. 2b). Fig. 2c shows the image of a fully packed BTS sensor. In the BTS sensor prototype, a Ti foil was used with a PEDOT: PSS coated fabric placed on the top. To ensure the degree of movement between the Ti foil and the PEDOT: PSS coated fabric, as well as the constraint on the edge, a thin layer of double sided adhesive was applied along the edges of the foil. A thin ITO-coated PET sheet was placed underneath as a capacitive normal force sensor. To further improve humidity stability and anti-wearing properties, the sensor was packaged in between two waterproof vinyl adhesive sheets. In this work, Ti has been chosen over Al for constructing the dynamic Schottky contact. Compared to Al, Ti is more cathodic and less vulnerable to electrochemical corrosion, which is evident from the DC output characteristics with water added to the metal-conducting polymer interface (See Figs. 7-9). We have also tested other fabric substrates (e.g., hydrophilic cotton, elastane, non-woven fabrics, nylon, PET, superhydrophobic fiber, tufting fabrics) for PEDOT: PSS coating (See Fig. 10). The results indicate that both the wettability and the mechanical properties of the fabrics are beneficial factors for the success of the conducting polymer coating.
[0143] To confirm the effectiveness of coating, Scanning Electron Microscopy (SEM) characterization was carried out. Fig. 2d and 2e show the morphology of fabric surface before and after 1-3 coatings at different magnifications, revealing a two-phase evolution of the PEDOT: PSS coating. Firstly, it is evident from the SEM results that a continuous PEDOT: PSS coating is developed and wrapped around each fiber in the fabric after one coating cycle. After the second and third coatings, additional thin film of PEDOT: PSS is developed on top of the previously coated fibers, which may act as a conducting pathway and further enhance the overall electrical conductivity. Elemental mapping by Energy Dispersive Spectroscopy (EDS) confirms a homogeneous distribution of sulfur across the textile surface, which is a characteristic element of PEDOT: PSS. (Fig. 2f and 2g) The SEM images (Fig. 2h and 2i) at a higher magnification further confirm the resulting coating of the PEDOT: PSS on the fibers, revealing good penetration and adherence of the ink to the fabric. To test the stretchability and electrical conductivity of the as-prepared E-textile, a tensile test was performed with a mechanical tester, while the electrical resistance ( / ?) of the fabric was measured simultaneously. The results demonstrate that the increase in resistance (AT?) was less than 13% with an overall R in the range of 0.019 kQ / cm under a tensile strain ty) of 90% (Fig. 2j). The stability of the electrical conductivity stems from the homogenous and uniform coating of the ink, and penetration of the ink into the fabric.Attorney Docket No.: 011520.02056
[0144] To investigate the interfacial electronic structure of the contact, a current-voltage (Z-F) curve characterization was carried out. The I-V data (Fig. 21) shows that the contact between the Ti and PEDOT: PSS coated fabric exhibits non-linear characteristics at different applied loads. At relatively lower forces, rectified I-V curves are observed. When F is increased, the I-V curves become more symmetrical while still showing non-linear features. The results indicate the existence of an energy barrier between the PEDOT: PSS and the Ti, which is in line with the work function (φ) difference between Ti ( T, = 4.33 eV) and PEDOT: PSS ( PEDOT:PSS = 5.0 eV). It is noted that the Schottky-like I-V characteristics of the dynamic contact is a key feature for the tribovoltaic dynamic DC generation (inset of Fig. 21).
[0145] To test the long-term stability of the BTS, a 1,000,000-cycle sliding test was performed for -98,350 s (Fig. 2m). A 2.5 N load was applied using a force gauge and the linear actuator was set to a speed of 100 mm / sec with a sliding distance of 5 mm. Due to the limitation of the data storage of electrometers, the complete data was collected in 18 consecutive sessions. (See Fig. 11 for details). The Foe first exhibits fluctuation and then becomes stabilized to -40 mV after -400,000 cycles. This may be due to the PEDOT: PSS-Ti system reaching its equilibrium in contact conditions (force, temperature, contact configuration, etc.) at contact asperities. In addition, we carried out cyclic compressing tests (1000 cycles with maximum contact force of 1 N and compression velocity of 3.33 mm / s) to investigate the robustness of our E-textile. The DC output before and after the test remains similar, though a small extent of coating fracture has been observed from SEM images (See Fig. 12).
[0146] Demonstration of robotic manipulation with BTS integrated robotic gripper
[0147] The application of the BTS system for robotics was demonstrated for manipulation tasks. We attached BTS sensor to a robotic gripper for demonstrating the robotic control based on DC generation. The block diagram (Fig. 3a) outlines a feedback control system used to maintain a secure grip on an object, preventing slipping despite external disturbances such as gravity or pulling forces. For robotic control experiments, a customized circuit has been made to generate filtered and amplified output signal (Fout) as sensor output, which is used in par with the Analog to Digital converter (ADC) (See Fig. 13).
[0148] Here, two sets of experiments have been designed: sensing and mitigation of slip under: i) linear pulling of the object with a mechanical tester (Fig. 3b-3g) and ii) gravitational rotation of the object from robotic gripper (Fig.3h-3m). The system is operated by calculating anAttorney Docket No.: 011520.02056error based on the difference between the desired no-slip condition and any detected slip, which is identified by the BTS (Fig. 3b). To distinguish the signal generated by the BTS and the noise, the threshold set for the Tout was 30 mV. The BTS provides real-time feedback, allowing the control algorithm to compute the adjustments in grip force. The motor controller then executes these adjustments by modulating the servo motors to apply the appropriate grip force (Fgrip), thereby counteracting the slip and stabilizing the object. This closed-loop process ensures continuous adaptation to disturbances, maintaining the grip and preventing the object from slipping.
[0149] For the linear pulling experiment, the mechanical tester was set to move upward with a constant speed (v) of 7.25 mm / s. A threshold of pulling force (Fpull)= 4.5 N was set as a force limit, considering the dynamixel servo used to control the grasp having a load constraint of 5 N. It was observed that the safety feature of the mechanical tester activated at about 75% of the force threshold, causing the speed to automatically decrease (step #4 in Fig. 3c). This safety feature is a failsafe mechanism set by the manufacturers. Here, the experiments were performed under two scenarios: in the first case, no feedback was activated when the object was pulled away using the mechanical tester. In this scenario, no increase was observed in grasp force even when a signal was generated from the sensor output (Fig. 3c). The steps for experiments are: step # 1 -initial position of the finger and the object, step #2- the finger is engaged so that the object does not slip / fall under its own weight, step #3- the mechanical tester is initiated and the object moves up with Fpull, step #4- the Fpuii reaches ~1.5 N and a displacement (77) of 15 mm, and the object continues to move up as the signal generated is not used to increase the grasp force, step#5- the final D of -30 mm is seen without any increase of Fgrip(AFgrip= 0 N). When the feedback was turned off, the gripper did not increase the FgriPby adjusting the servo angle (Fig.3b). Corresponding optical images demonstrate the displacement (D) during each phase, highlighting the magnitude of movement of the object in response to gripper's applied forces (Fig. 3d). The total D was observed to be -30 mm since there was no feedback control. In the second scenario, with the feedback from the BTS activated, precise movements were successfully executed by the gripper as shown in the schematic and corresponding photographs (Fig. 3e-3g). The steps for experiments are: step #1- initial position of the finger and the object, step#2-the finger is engaged, step #3- the mechanical tester is initiated, and the object moves up with Fpuii, step #4- the Fpuii reaches -4.0 N and a D of 4.1 mm, and the object stops as the grasp force of AFgrip= 4.0 N is reached with the movement between the object and the finger,Attorney Docket No.: 011520.02056generating DC signal by the BTS system, step #5- the final D of 4.1 mm is seen. Data plots confirmed the gripper's capability' to accurately control D to under 4.1 mm and consistently increase force throughout the task whenever a DC signal was generated from the tactile sensor (Fig. 3f). The optical image displays the successful mitigation of the object's movement due to the increase in grasp force by the gripper (Fig. 3g).
[0150] For the second set of experiments, rotational control experiments were conducted to further demonstrate the sensor's capability' to detect and mitigate rotational slip: step #1- initial position of the finger and the object, step #2- the finger is engaged so that the object does not rotate under its own weight due to the unequal center of mass, step #3- the weight (W) of 0.26 N made of a 3D printed part and a copper bar is added, due to which the object started to rotate. Initially, when the gripper grasped the object, no rotation or motion was detected by the sensor. As external weights were added, the object began to rotate (Fig. 3h and 3i). Without feedback, the gripper exhibited significant angular displacement ( / )= 70°) (Fig. 3j). When the feedback is activated as soon as the Tout passes the 30-mV threshold, the servo is initiated and the \Vgrip increases (Fig. 3k and 31) to ~1.2 N to stop the angular motion of the object. With this implementation of feedback from the BTS sensor, precise control was achieved by the gripper, reducing angular deviation to a minimal value (6>=7°) (Fig. 3m). This result underscores the effectiveness of the sensor in enhancing the accuracy and reliability of robotic manipulation tasks.
[0151] Furthermore, experiments were performed to investigate the relationship between displacement D and Vgripper at a constant pulling velocity (vPuii=18.33 mm / s) (Fig. 4a). In all experiments, the values of D were recorded when Fgrip reaches 3 N. It can be seen thatD decreases as vgriPper increases at constant vpull, which can be rationalized by a faster actuation of the gripper in response to the detected slip. Also, experiments were conducted to examine the relationship between / ) and vpullat a constant vgripper=126.75 mm / s) (Fig. 4b). Fig. 4b shows that D also decreases as vpullincreases within the tested vpullrange. Such a trend may be rationalized by the fact that the input frictional power ( / ’friction) is proportional to sliding velocity:^friction f ' ^pull (1) where f is the frictional force. The increasing / ’friction would result in a faster rising of the BTS sensor signal, leading to an earlier actuation of the gripper. However, we predict that aAttorney Docket No.: 011520.02056substantial further increase of vpullwould eventually cause D to increase again, the gripper may no longer be able to react quickly enough to such rapid motion. We have also investigated the final rotation angle (0) of the object as a function of vgriPPerin the angular displacement experiment (Fig. 4c). As the vgriPPer increases, 6 was found to decrease, which is in line with the expectation as a faster gripper response would leads to a quicker stop of the rotational motion.
[0152] It is noted that the response time of human MRs depend on their types. The conduction speed of neurological signal for Meissner’s corpuscles (RA-I), Pacinian corpuscles (RA-II), Merkel cells (SA-I), and Ruffini endings (SA-II) has been reported to be -26-91 m / s, -30-90 m / s, -16-96 m / s, and -20-100 m / s, respectively. Accordingly, the calculated response times for a l m traveling distance of human MRs fall into the range of 10-200 ms (milliseconds). It can be seen that the estimated response time of BTS approaches the limit of human MRs, showing great promise for enhancing the dexterity in future humanoid robots (Fig. 4d). (See Fig. 14 for more details)
[0153] In some experiments, we demonstrated a closed-loop control system by integrating the BTS with an adjustable stiffness gripper which is capable of modifying grip force independent of grasp width. To assess slip detection performance, we attached BTS and a visionbased GelSight sensor to separate gripper fingers. GelSight, commonly used in robotic manipulation, estimates slip by tracking geometric contact features, making its performance highly dependent on the presence of such features. To ensure fairness in our comparison, we used a 3D-printed container with geometric patterns for the grasping experiments.
[0154] During the experiments, the robotic gripper was set to the minimum grip force needed to prevent slip. To induce slip, we dropped 7 metal beads (— 17g each) from a height of 12 cm into the container (Fig. 4h-i). The impact caused the container to slide within the gripper. We recorded data from GelSight, BTS, and a 6-axis Torque / Force (T / F) sensor embedded in the robotic arm. The Lab Streaming Layer ensured precise synchronization. GelSight captured images at 20 frames per second, analyzed by computing the average pixel difference between consecutive frames, while BTS and force data were recorded at 1000 Hz. The T / F sensor identified the moment of impact, allowing us to estimate the latency of BTS and GelSight in detecting post-impact slip.
[0155] The results showed that impact-induced slip generated spikes in the FT, GelSight, and BTS outputs, detected using a threshold on the local maxima of a moving windowAttorney Docket No.: 011520.02056(Fig. 4j -41). BTS-detected slip triggered the gripper controller to incrementally increase grasp force and stop further slippage (Fig. 4m). Without feedback loop activated, in contrast, the gripper remained at the same position and the object fell after the 5thconsecutive bead was added. The largest displacement occurred after the first impact when the grip force was minimal. With each subsequent impact, the accumulated beads reduced the drop height and impact force, while the pre-impact grasp force increased, leading to less slippage.
[0156] To compare slip detection times, we conducted experiments without the controller, maintaining a constant grip force. We tested 10 initial grasp forces ranging from 0.69 to 5.48 N, with each trial involving a single bead drop and 15 repetitions under identical conditions. The spike in T / F sensor data identified the exact impact time, while slip detection time was defined as the time difference between impact and the corresponding sensor’s detection spike. The comparison results shown in Fig. 4g. ANOVA analysis indicated no significant effect (p-value 0.85) of initial grip force on detection delay (p = 0.801), allowing data from all trials to be combined. As shown in Fig. 4g, BTS achieved a significantly lower mean detection time (34 ms) compared to GelSight (96 ms), with an average difference of 62 ms. BTS demonstrates an improved detection time of slippage that would enable real-time slip detection and object retaining in robotic manipulation in the presence of external impact, while systems relying on vision-based tactile sensors such as GelSight would fail due to slower detection time and permitting larger post-impact displacement.
[0157] Contact mechanics analysis of the parameterized BTS output
[0158] The mathematical relationship between the BTS output and the mechanical input parameters is critical for facilitating advanced robotic manipulation including delicate manipulations and in-hand reorientation, where understanding the force and speed of contact is the key to avoiding damage to objects or changing the grasping strategy through object reorientation. Here, a modeled contact of BTS is used to perform this analysis. The experimental setup included a linear actuator on which a Ti sheet was placed, and a PEDOT: PSS fabric on a PDMS finger of d = 15 mm attached to a force gauge. The experiments were performed for 30 reciprocal cycles at a moving distance of 8 mm. Fig. 5a and 5b show the performance of BTS under varying velocities and applied forces. The sensor generates short-circuit current ( / sc) in response to an increase in both velocity of motion (v) (Fig. 5 a) and the applied force (F) (Fig. 5b). The v-dependent current response spans from 0.75 mm / s to 116 mm / s, with higherAttorney Docket No.: 011520.02056velocities generating significantly larger currents up to 20 nA, whereas the lower velocities generating ~5 nA. Similarly, the F-dependent current response (Fig. 5c) ranges from 0.5 N to 5.0 N, indicating the sensor's capability to distinguish between different loads (see Figs. 15- 18).
[0159] Fig. 5d further elucidates the relationship between / sc and F at different velocities. As anticipated analytically, the / sc increases with increasing velocity for each corresponding F. Here, the Hertz contact mechanics model is used to analyze the relationship between / sc, F, and v (Fig. 5c). The contact radius (aHertz) and contact area (A) for Hertz model is given by:3aHertz=(2)N totwhere, Rtipis the radius of the tip in contact, F is the force applied, and Etotis the reduced young’s modulus. The / ?tipis measured as 7.5 mm. Etotcan be expressed by:\Where vsub. Eguband vtip, Ftjpare the Poisson’s ratio and Young’s modulus for the substrate and the tip respectively. The contact area A can be further approximated as,Considering the fact that the current output is proportional to contact area with a linear coefficient of current density:Isc= / sc ' (5) Assuming the intrinsic parameter Jsc varies as a function of velocity, i.e., Jsc(v), and combining Eq. 2 to 5, one can obtain:fsc —77■ JscAttorney Docket No.: 011520.02056
[0160] From Eq. 6, it can be predicted that the / sc is proportional to F273. The solid lines in Fig. 5d show the fitting results based on Eq. 6, where the Poisson’s ratio for Ti and PDMS of 0.34 (vsub) and 0.495 (vtip) and Young’s modulus of 116.0 GPa (Esub) and 2.1 MPa (Etip) are used. It can be seen that the theoretical prediction is in good accordance with the experimental data of / sc vs F at all v. From the data fitting in Fig. 5d, the values of Jsc(v) can be extracted. In Fig. 5e. Jsc vs v is plotted, which is fitted by a second order polynomial approach for approximation. It can be seen that as the v increases from 0.75 mm / s to 116 mm / s, the Jsc increases from 0.27 mA / m2to 1.72 mA / m2. Such trend may be explained by the increasing frictional energy input expressed by (Fig. 5e):where / is the frictional force, the product of fv is the frictional power, and t is the time.According to the tribovoltaic effect theory, the more energy is input, the more intensive electronic excitation of electron-hole pairs is expected at the sliding interface, resulting in higher DC output. In order to explicitly explain the Jsc vs v relationship, a deeper understanding of the tribovoltaic physical mechanism is needed in future work.
[0161] Discussion
[0162] Inspired by the human mechanoreceptors (MR), a self-powered bionic textile sensing (BTS) system was designed and developed based on the tribovoltaic effect. A layered structure was tailored and integrated with a robotic gripper, acting as a rapid- adapting (RA) MR-like friction sensing element. In conjunction with slow-adapting (SA) MR-like normal force sensing, the BTS system has been proved to be effective in providing DC output as input data for robotic control. In our demonstration, we have successfully shown the sensing and mitigation of object slip from the robotic gripper under external linear pulling force or intrinsic gravitational force. A Hertzian contact mechanics model has also been used to reveal the dependency of DC output on the force and velocity, paving the way for future reinforcement learning-based control systems. The BTS’s ability to generate distinguishable signals corresponding to different applied forces and velocities underscores its potential for real-world robotic applications. The flowchart (Fig. 19) outlines an example of a comprehensive robotic control system. The system could utilize the tactile signals in the reinforcement learning policy to enhance grip stability and dexterity. By incorporating slip detection as an additional degree of freedom in trajectoryAttorney Docket No.: 011520.02056planning, the system can also adapt to various object manipulation tasks with higher precision and reduced errors.
[0163] Methods
[0164] Materials
[0165] 1.1% in H2O, surfactant-free, high conductivity grade Poly(3,4-ethylenedioxythiophene) polystyrene sulfonate (PEDOTPSS), Ethylene Glycol (EG), isopropanol (IP A), ethanol, and 99.6 % pure titanium (Ti) foil were procured from Sigma-Aldrich. The spandex fabric substrate was procured from Echoine Store through Amazon. Sylgard 184 (PDMS) was procured from Sigma- Aldrich. PET-ITO film was procured from Sigma-Aldrich. 3D printing filament was procured from Prusa. Ecoflex-0030 was procured from Smooth-On Store through Amazon.
[0166] Fabrication of the BTS sensor
[0167] For conducting polymer ink preparation, mixture A comprised 25 g of the PEDOTPSS mixed with 12.5 g of ethanol. Mixture B was composed of 0.512 g of ethylene glycol (EG) and 7.5 g of isopropanol (IP A). The weight ratio of PEDOTPSS: ethanol: IPA: EG = 100: 50:30:2.05. The ethylene glycol is added, as it has shown to increase the electrical conductivity7due to the restructuring of the PEDOT chains and removal of PSS shells increasing the charge carrier mobility and density. The electrical conductivity of PEDOTPSS can also be enhanced by the addition of other chemicals such as dimethyl sulfoxide (DMSO) (See Fig. 20). Heating the mixture to above 90 °C removes the water replacing it with EG molecules.Alongside, adding solvents such as isopropyl alcohol (IPA) to PEDOTPSS has show n to decrease the size of the insulating PSS agglomerates and to develop a more uniform film morphologically along with changing the viscosity of the prepared ink. The IPA has also shown to increase the Foe and short-circuit current density (Jsc) in PEDOTPSS based photovoltaic cells. Ethanol helps promoting a homogeneous dispersion of PEDOT and PSS by reducing the phase separation between PEDOT and PSS molecules. The two mixtures A and B were combined and subjected to magnetic stirring for 48 hours to achieve a smooth, homogeneous solution for better ink penetration into the fabric and uniform coating. For coating treatment, a 6x6 inch cut of pristine spandex fabric was dipped in the prepared ink solution. The fabric wasAttorney Docket No.: 011520.02056then placed under vacuum for 15 minutes in a vacuum desiccator and cured in a vacuum oven for 1 hour at 120 °C. This process of dip and dry coating was repeated 3 times in total.
[0168] For the integration of the electrodes, the PEDOT: PSS coated fabric (Fig. 2b) was cut into strips. The metal electrode of the fabric and the Ti foil were cut to required dimensions. Here, the junction betw een the PEDOT: PSS fabric and the Ti foil acts as a DC generator node. Both the cut PEDOT PSS fabric and the Ti foil were then bonded together with a thin layer of adhesive. This was then placed on top of the capacitive sensor and encased with waterproof vinyl layer. The prepared BTS sensor was then placed on the fabricated composite finger. Lead wires for the sensor w ere prepared by applying Ag / AgCl paste and attaching wires for the reception of signals generated from the prepared sensor. The composite finger fabricated was made of 3D printed polylactic acid (PLA).
[0169] Material Characterization of PEDOT: PSS-Coated fabric
[0170] To verify the integrity of the ink coating on the spandex substrate. Scanning Electron Microscopy (SEM) and Energy Dispersive X-ray Spectroscopy (EDS) were performed using Focused Ion Beam Scanning Electron Microscope (FIB-SEM) - Carl Zeiss AURIGA CrossBeam.
[0171] Mechanical Characterization
[0172] The mechanical characterization was performed using an F- 105 model mechanical tester with IntelliMESUR® software to record the data with the force sensors FS05-05 manufactured by MARK-10. The tensile test jig was manufactured in-house using a Prusa MK43D printer and the fabric w as attached to the mechanical tester. The lead wires w ere attached using Ag / AgCl paste, and the resistance was measured using a Keithley DMM65006I / 2-digit digital multimeter. Data for load and strain was collected from the IntelliMESUR® software.
[0173] DC power characterization
[0174] DC powder characterization w as performed using a linear actuator from Fourien. Inc and force gauge (Torbal). The PEDOT: PSS coated spandex fabric was fixed on a 15 mm diameter sphere made of PDMS sphere and the Ti foil w as placed on the platform of the actuator. For the cyclic testing, the load was fixed at 4 N and sliding speed of 53.63 mm / s, andAttorney Docket No.: 011520.02056the data was collected using a Keithley DMM6500 61 / 2-digit digital multimeter. The I-V curve characterization was performed at different loads of 0.5 N, 1.0 N, 1.5 N, 2.0 N and 3.0 N in static conditions without sliding and to collect the data with a Keithley 2450 SourceMeter.
[0175] Robotic demonstration
[0176] Mitigation of slip was achieved by utilizing a control law, which works on the principle of increasing the grip force by modulating the servo motor (DYNAMIXEL XC330-M288-T) angle to grasp the object. The closed-loop feedback, used for the demonstration uses reference signal of no slip (Fref=0). The BTS sensor provides a DC signal (Tout) of the relative motion between the object and the gripper. The disturbance to the system such as pulling force (Mark-10 F105 with FS05-10 force sensor) and gravity (0.26 N offset weight), induces slip between gripper and the object. The continuous analog signal generated by the BTS is converted to digital signal by NI-DAQ (NI-DAQ USB-6003). The error computed between the sensor signal and the reference signal is used by the control law. The control law processes the error and modifies the servo angle to increase the grip force.(8) ( 0 error < 30 mF)Servo Angle Change = 1 (servo angl,e +,nO.Oo0error >o3n0 mTVZ) 1
[0177] Robotic manipulation with compliant gripper system
[0178] The compliant gripper system utilizes variable stiffness to adjust the grip force independent from the finger position. In other words, we can modulate the force without moving the fingers. This is achieved by adjusting airgaps between permanent magnets that act as nonlinear springs. The gripper is equipped with a linear potentiometer to measure the airgap. The grip force in Fig. 4m is determined using the airgap and the data-driven model for the variable stiffness system as described and validated in our previous work.
[0179] The slip detection in the BTS data is based on applying a threshold on the moving average of the previous 10 data points with a rectangular weight. This is equivalent to applying a low pass FIR filter with a cut-off frequency of ~44 Hz. For the GelSight the average of absolute difference between the previous two frames of the image is used as a slip score (See Fig. 21 for more details). This calculation utilizes OpenCV's absdiff function, optimized with SIMDAttorney Docket No.: 011520.02056acceleration and GPU execution using CUD A streams for maximum efficiency. The computational time is approximately 50 ps, eliminating the need for additional filtering. The system configuration used for data processing included a 12th Generation Intel i7 CPU (14 cores), an Nvidia GeForce RTX 4070 GPU, and 32 GB DDR4 RAM, ensuring high-speed execution. Slip is detected through simple thresholding of the computed score. Slip is detected through simple thresholding of the computed score.
[0180] Upon slip detection, the controller incrementally reduces the air gap by 1.3 mm per detected slip, modulating the grip force according to the following equation:Where d is the airgap that is equal to the initial airgap d0needed for minimum grasp and the incremented displacement after the kthslip:d = dQ+ 0.0013 k (10)?i, P2, P3are determined and validated based on a data-driven model described in our previous work. Their values are equal to 5.5xl0'6N m?, 6.85><10'5m2and 2.22xl0'7m3.
[0181] Additional Detail - Response Time Estimation
[0182] For the response time estimation, a total of 150 experiments were performed with various speeds to compute the response time. The experimental setup comprised a servo motor which can run at different speeds, BTS was mounted on a force gauge (Torbal FB-010) and connected with a data acquisition device (DAQ) (NI-DAQ USB-6003). The data was collected at 50,000 Hz which ensures no aliasing and gives us the abi li ty to compute the response time in the order of microseconds (ps) for better accuracy. The experiment started with the sensor data collection, while there was no servo movement, the noise data from the sensor was collected by the DAQ. The time was recorded while providing the command to the servo to move (isovo). Further the time was recorded when the sensor crosses the threshold value of 3 mV ( / sensor). BTS response time was computed as the time duration sensor takes to generate a voltage greater than the threshold value:sensor - / servoAttorney Docket No.: 011520.02056
[0183] After analyzing 150 experiments, the response times for BTS were in the range of 0.768 to 38.366 ms depending on the velocity of sliding. The 3.0 mV threshold voltage was determined after analyzing the noise signal which had a mean value of 0.4 mV and a max value of 1.7 mV. This led us to set the signal to noise threshold to 1.765. It was observed that in the experiments conducted, the maximum voltage generated in each setup was well above the max noise of 1.7 mV ( Vmax / Vnoise in range of 5 to 20). Thus, setting the signal to noise ratio of 1.765 was acceptable to compute the response time, as any signal greater than 3 mV in the setup is only generated when motion is detected.
[0184]
[0185] Although the present disclosure has been described with respect to one or more particular embodiments, it will be understood that other embodiments of the present disclosure may be made without departing from the spirit and scope of the present disclosure.
Claims
Attorney Docket No.: 011520.02056What is claimed is:
1. A tactile sensor assembly comprising:a first layer comprising a textile substrate and a tribovoltaic material disposed on fibers of the textile substrate; anda second layer disposed adjacent to the first layer and comprising a second material; wherein the tribovoltaic material of the first layer and the second material of the second layer are arranged for relative sliding contact with one another such that frictional contact between the tribovoltaic material and the second material generates a direct-current electrical signal.
2. The tactile sensor assembly of claim 1, wherein the tribovoltaic material and the second material form a tribovoltaic interface during the relative sliding contact.
3. The tactile sensor assembly of claim 2, wherein the tribovoltaic interface comprises a dynamic Schottky interface.
4. The tactile sensor assembly of claim 2, wherein the tribovoltaic interface comprises a P-N junction.
5. The tactile sensor assembly of claim 1, wherein the tribovoltaic material comprises a conducting polymer.
6. The tactile sensor assembly of claim 5, wherein the conducting polymer comprises poly(3,4-ethylenedioxythiophene) polystyrene sulfonate.
7. The tactile sensor assembly of claim 1, wherein the textile substrate comprises a stretchable textile.
8. The tactile sensor assembly of claim 7, wherein the stretchable textile comprises spandex, polyester, nylon, cotton, or a combination thereof.
9. The tactile sensor assembly of claim 7, wherein the textile substrate comprises a woven fabric, a non-woven fabric, a knit fabric, a braided fabric, a tufted fabric, or a combination thereof.
10. The tactile sensor assembly of claim 7, wherein the textile substrate comprises hydrophilic cotton or a superhydrophobic fiber.Attorney Docket No.: 011520.0205611. The tactile sensor assembly of claim 1, wherein the textile substrate comprises spandex and polyester.
12. The tactile sensor assembly of claim 11, wherein the polyester comprises polyethylene terephthalate.
13. The tactile sensor assembly of claim 12, wherein the textile substrate comprises about 20% spandex and about 80% polyethylene terephthalate.
14. The tactile sensor assembly of claim 1, wherein the second material comprises an electrically conductive material.
15. The tactile sensor assembly of claim 14, wherein the electrically conductive material comprises a metal or metal alloy.
16. The tactile sensor assembly of claim 15, wherein the metal or metal alloy comprises titanium, aluminum, or a combination thereof.
17. The tactile sensor assembly of claim 16, wherein the second layer comprises a titanium foil.
18. The tactile sensor assembly of claim 1, wherein the first layer and the second layer are constrained relative to one another at an edge region while permitting relative sliding contact in an interior region.
19. The tactile sensor assembly of claim 18, wherein the first layer and the second layer are coupled at the edge region by an adhesive.
20. The tactile sensor assembly of claim 1, further comprising a normal-force sensor disposed adjacent to the second layer.
21. The tactile sensor assembly of claim 20, wherein the normal-force sensor comprises a capacitive sensor.
22. The tactile sensor assembly of claim 21, wherein the capacitive sensor comprises an indium-tin-oxide-coated polymer film.
23. The tactile sensor assembly of claim 22, wherein the polymer film comprises polyethylene terephthalate.Attorney Docket No.: 011520.0205624. The tactile sensor assembly of claim 1, further comprising an encapsulation layer disposed around at least a portion of the first layer and the second layer.
25. The tactile sensor assembly of claim 24, wherein the encapsulation layer comprises one or more waterproof polymer sheets.
26. The tactile sensor assembly of claim 2, wherein the direct-current electrical signal is generated by conversion of mechanical energy from the relative sliding contact into electrical energy at the tribovoltaic interface.
27. The tactile sensor assembly of claim 17, wherein the first layer comprises a poly(3.4-ethylenedioxythiophene) polystyrene sulfonate-coated fabric and the second layer comprises the titanium foil, and wherein, when subjected to a sliding test at a load of about 2.5 N, a sliding speed of about 100 mm / s, and a sliding distance of about 5 mm, the tactile sensor assembly is configured to generate an open-circuit voltage of at least about 30 mV after about 400,000 sliding cycles.
28. The tactile sensor assembly of claim 6, wherein the textile substrate coated with the tribovoltaic material exhibits a resistance change of less than about 13% under about 90% tensile strain.
29. The tactile sensor assembly of claim 1, wherein the direct-current electrical signal comprises an open-circuit voltage signal or a short-circuit current signal.
30. The tactile sensor assembly of claim 1, further comprising a processing circuit electrically coupled to the first layer and the second layer and configured to detect slip based at least in part on the direct-current electrical signal.
31. The tactile sensor assembly of claim 30, wherein the processing circuit is configured to identify slip when an output voltage derived from the direct-current electrical signal satisfies a threshold condition.
32. The tactile sensor assembly of claim 31, wherein the threshold condition comprises the output voltage being at least about 30 mV.
33. The tactile sensor assembly of claim 30, wherein the processing circuit is configured to apply a moving-average filter to the direct-current electrical signal before detecting slip.Attorney Docket No.: 011520.0205634. The tactile sensor assembly of claim 30, wherein a magnitude of the direct-current electrical signal varies with at least one of a sliding velocity, an applied force, a contact area, a current density, or a frictional power associated with the relative sliding contact.
35. The tactile sensor assembly of claim 30, wherein the processing circuit is configured to generate a slip signal based at least in part on the direct-current electrical signal and a normalforce signal based at least in part on an output of a normal-force sensor, wherein the slip signal is generated in response to dynamic sliding between the first layer and the second layer, and wherein the normal-force signal is generated in response to normal loading of the normal-force sensor.
36. The tactile sensor assembly of claim 1, wherein, in a response-time test in which a voltage from the tactile sensor assembly is compared to a threshold voltage of about 3 mV, the tactile sensor assembly produces a voltage exceeding the threshold voltage within about 0.7 ms to about 40 ms after initiation of sliding motion, depending on sliding velocity.
37. A tactile sensor comprising:a first layer comprising a textile substrate and a conducting polymer disposed on fibers of the textile substrate;a second layer disposed adjacent to the first layer and comprising a metal layer;a retainer joining the first layer and the second layer to one another, wherein at least a portion of the first layer and at least a portion of the second layer outside the retainer are configured to undergo relative sliding contact with one another;a normal-force sensor disposed adjacent to the second layer; anda processing circuit configured to generate a slip signal based at least in part on a direct- current electrical signal generated by relative sliding contact between the first layer and the second layer and to generate a normal-force signal based at least in part on an output of the normal-force sensor.
38. The tactile sensor of claim 37, wherein the conducting polymer comprises poly(3,4-ethylenedioxythiophene) polystyrene sulfonate.
39. The tactile sensor of claim 37, wherein the metal layer comprises titanium or aluminum.
40. The tactile sensor of claim 37, wherein the metal layer comprises titanium foil.Attorney Docket No.: 011520.0205641. The tactile sensor of claim 37, wherein the normal-force sensor comprises a capacitive sensor comprising an indium-tin-oxide-coated polyethylene terephthalate film.
42. The tactile sensor of claim 37, wherein the first layer and the second layer are packaged between waterproof polymer sheets.
43. The tactile sensor of claim 37, wherein the direct-current electrical signal is generated by conversion of mechanical energy from the relative sliding contact into electrical energy.
44. A robotic manipulation system comprising:a robotic gripper comprising at least one finger;a tactile sensor mounted to the at least one finger, the tactile sensor comprising:a first layer comprising a textile substrate and a tribovoltaic material disposed on fibers of the textile substrate; anda second layer disposed adjacent to the first layer and comprising a second material, wherein the first layer and the second layer are arranged for relative sliding contact with one another such that frictional contact between the tribovoltaic material and the second material generates a direct-current electrical signal; anda controller configured to adjust operation of the robotic gripper based at least in part on the direct-current electrical signal.
45. The robotic manipulation system of claim 44, wherein the controller is configured to increase a gripping force of the robotic gripper in response to detected slip indicated by the direct-current electrical signal.
46. The robotic manipulation system of claim 44, wherein the controller is configured to modulate a servo motor to adjust a gripping force of the robotic gripper.
47. The robotic manipulation system of claim 44, wherein the controller is configured to compute a magnitude of an error based on a reference no-slip signal and an output signal derived from the direct-current electrical signal and to change a servo angle when the magnitude of the error satisfies a threshold condition.
48. The robotic manipulation system of claim 47, wherein the controller is configured to change the servo angle by a predetermined angular increment in response to detected slip.Attorney Docket No.: 011520.0205649. The robotic manipulation system of claim 44, wherein the controller is configured to reduce linear displacement of an object held by the robotic gripper in response to detected slip.
50. The robotic manipulation system of claim 44, wherein the controller is configured to reduce rotational displacement of an object held by the robotic gripper in response to detected slip.
51. The robotic manipulation system of claim 44, wherein the robotic gripper comprises a compliant gripper configured to adjust gripping force independently of grasp width.
52. The robotic manipulation system of claim 51, wherein the compliant gripper comprises a variable-stiffness actuator.
53. The robotic manipulation system of claim 52, wherein the variable-stiffness actuator comprises one or more permanent magnets arranged to define an adjustable air gap.
54. The robotic manipulation system of claim 53, wherein the controller is configured to adjust the adjustable air gap in response to detected slip.
55. The robotic manipulation system of claim 44, wherein the controller is configured to use detected slip as an input to trajectory planning for the robotic gripper.
56. The robotic manipulation system of claim 44, wherein the controller is configured to provide detected slip as an input to a control policy for robotic manipulation.
57. The robotic manipulation system of claim 56, wherein the control policy is configured to adjust a grasping parameter based at least in part on the detected slip.
58. A method of sensing slip, comprising:contacting a first layer with a second layer, the first layer comprising a textile substrate coated with a tribovoltaic material and the second layer comprising a second material; permitting relative sliding contact between the tribovoltaic material of the first layer and the second material of the second layer;generating a direct-current electrical signal at an interface between the tribovoltaic material and the second material in response to the relative sliding contact; anddetecting slip based at least in part on the direct-current electrical signal.Attorney Docket No.: 011520.0205659. The method of claim 58, wherein generating the direct-current electrical signal comprises generating the direct-current electrical signal at a dynamic Schottky interface.
60. The method of claim 58, wherein generating the direct-current electrical signal comprises generating the direct-current electrical signal at a P-N junction.
61. The method of claim 58, wherein generating the direct-current electrical signal comprises converting mechanical energy from the relative sliding contact into electrical energy.
62. The method of claim 58, further comprising sensing a normal force using a capacitive sensor disposed adjacent to the second layer.
63. The method of claim 62, further comprising distinguishing a shear-related slip stimulus from a normal-force stimulus based at least in part on the direct-current electrical signal and an output of the capacitive sensor.
64. The method of claim 58, wherein detecting slip comprises comparing an output voltage derived from the direct-current electrical signal to a threshold.
65. The method of claim 64, wherein the threshold is about 30 mV.
66. The method of claim 58, further comprising filtering the direct-current electrical signal before detecting slip.
67. The method of claim 58, further comprising controlling a robotic gripper based at least in part on the detected slip.
68. The method of claim 67, wherein controlling the robotic gripper comprises increasing a gripping force of the robotic gripper.
69. The method of claim 67, wherein controlling the robotic gripper comprises changing a servo angle based on a magnitude of an error between a reference no-slip signal and an output signal derived from the direct-current electrical signal.
70. The method of claim 67, wherein controlling the robotic gripper comprises changing a stiffness state of a compliant gripper by adjusting an air gap between permanent magnets.Attorney Docket No.: 011520.0205671. The method of claim 58, further comprising determining an estimated value of at least one of a contact force, sliding velocity, contact area, or current density using a relationship between the direct-current electrical signal and the at least one of the contact force, sliding velocity, contact area, or current density.
72. The method of claim 71, wherein the relationship is based at least in part on a Hertzian contact model.
73. A tribovoltaic textile layer for a tactile sensor, comprising:a textile substrate comprising a plurality of fibers; anda conducting polymer coating disposed on the plurality of fibers;wherein the conducting polymer coating is configured to form, during sliding contact with a second material, a tribovoltaic interface that generates a direct-current electrical signal.
74. The tribovoltaic textile layer of claim 73, wherein the conducting polymer coating comprises poly(3,4-ethylenedioxythiophene) polystyrene sulfonate.
75. The tribovoltaic textile layer of claim 73, wherein the textile substrate comprises a stretchable fabric.
76. The tribovoltaic textile layer of claim 75, wherein the stretchable fabric comprises spandex and polyethylene terephthalate.
77. The tribovoltaic textile layer of claim 73, wherein the conducting polymer coating wraps around individual fibers of the textile substrate.
78. The tribovoltaic textile layer of claim 73, wherein the conducting polymer coating forms conductive interconnections between adjacent fibers of the textile substrate.
79. The tribovoltaic textile layer of claim 73, wherein the conducting polymer coating comprises three coating layers.Attorney Docket No.: 011520.0205680. A method of making a tribovoltaic textile layer, comprising:preparing a coating composition comprising a conducting polymer;applying the coating composition to a textile substrate; anddrying or curing the textile substrate after applying the coating composition, thereby forming a conducting-polymer-coated textile substrate configured to generate a direct-current electrical signal during sliding contact with a second material.
81. The method of claim 80, wherein the conducting polymer comprises poly(3,4-ethylenedioxythiophene) polystyrene sulfonate.
82. The method of claim 80, wherein preparing the coating composition comprises combining the conducting polymer with a conductivity-enhancing additive comprising ethylene glycol, dimethyl sulfoxide, or a combination thereof.
83. The method of claim 80, wherein preparing the coating composition comprises combining poly(3,4-ethylenedioxythiophene) polystyrene sulfonate, ethanol, isopropanol, and ethylene glycol in a weight ratio of about 100:50:30:2.05.
84. The method of claim 80, wherein applying the coating composition comprises dipping the textile substrate in the coating composition.
85. The method of claim 80, further comprising placing the textile substrate under vacuum after applying the coating composition.
86. The method of claim 80, wherein drying or curing the textile substrate comprises heating the textile substrate.
87. The method of claim 80, further comprising repeating applying the coating composition and drying or curing the textile substrate to form three coating layers.
88. The method of claim 80, further comprising assembling the conducting-polymer-coated textile substrate with a metal counter-layer by coupling the conducting-polymer-coated textile substrate and the metal counter-layer at an edge region while permitting relative sliding contact between the conducting-polymer-coated textile substrate and the metal counter-layer.
89. The method of claim 88, further comprising disposing a capacitive normal-force sensor adjacent to one of the conducting-polymer-coated textile substrate and the metal counter-layer.Attorney Docket No.: 011520.0205690. The method of claim 89, further comprising encapsulating the conducting-polymer-coated textile substrate, the metal counter-layer, and the capacitive normal-force sensor in one or more waterproof polymer sheets.
91. The method of claim 80, wherein the coating composition comprises poly(3,4-ethylenedioxythiophene) polystyrene sulfonate, ethanol, isopropanol, and ethylene glycol, and wherein a weight ratio of poly(3.4-ethylenedioxythiophene) polystyrene sulfonate to ethylene glycol is from about 100:0.1 to about 100:10.
