Pneumatic soft finger with variable stiffness phalanges and force feedback and control method

Abstract

The application discloses a kind of pneumatic soft body fingers with variable stiffness finger plate and force feedback and control method, including flexible joint main body, variable stiffness finger plate and fiber bragg grating sensor.Flexible joint main body is made of silicone rubber, inside is provided with wave cavity, is connected external air pressure source by pneumatic interface.Variable stiffness finger plate is arranged at one side of gripping surface, inside is filled with flowable solid particles, is connected external vacuum source by vacuum interface, realizes the fast adjustment of finger plate stiffness in 8 to 15 times by vacuumizing.Light fiber bragg grating sensor is embedded in the back of finger, and bending strain and fingertip contact force are detected in real time.Control method is: the measured value of acquisition fiber wavelength signal and the measured value of bending angle are calculated, the working mode of variable stiffness finger plate is decided according to the comparison result of target gripping force and preset threshold, and the opening of pneumatic proportional valve is adjusted by PID algorithm closed loop, so that gripping force is stabilized in target error range.

Description

Technical Field

[0001] This invention relates to the field of soft robot technology, and in particular to a pneumatic soft finger with variable stiffness finger and force feedback, and a control method thereof. Background Technology

[0002] Soft robotic fingers have garnered significant attention in fields such as grasping irregularly shaped and fragile objects, bionic manipulation, and medical rehabilitation due to their excellent passive compliance and safe human-computer interaction capabilities. Currently, mainstream pneumatic soft fingers are typically integrally cast from highly elastic materials such as silicone rubber, containing multiple wave-shaped cavities or segmented chambers arranged along their length. Inflating these cavities causes the finger to bend and deform, enabling it to wrap around and grasp objects. However, this single-material structure presents inherent contradictions: to ensure sufficiently low contact force (typically less than 5N) when grasping small and fragile objects like eggs and strawberries, the silicone wall thickness must be designed to be thin and the cavity spacing large, resulting in low overall finger stiffness; when grasping objects weighing over 500g (such as tools or metal parts), the finger may overbend or even become laterally unstable under high air pressure, and due to the limited elastic modulus of silicone (typically 0.1~1MPa), it cannot provide sufficient normal contact force, making objects prone to slippage.

[0003] To address the issue of the unadjustable stiffness of soft fingers, researchers have proposed several variable stiffness schemes.

[0004] One approach uses low-melting-point alloys or paraffin as phase change materials, achieving stiffness switching by heating the material to melt it and then cooling and solidifying it. However, this method has a long response time (usually several seconds to tens of seconds) and requires complex heating-cooling cycles and insulation structures, making it difficult to meet the needs of dynamic grasping.

[0005] Another approach uses magnetorheological fluid or electrorheological fluid to fill the cavity, and adjusts its apparent viscosity by applying an external electromagnetic field to change its stiffness. However, magnetorheological fluid is expensive, requires continuous power supply to maintain stiffness, and the magnetic field generating device is large, which is not conducive to the miniaturization and weight reduction of fingers.

[0006] In recent years, variable stiffness structures based on the principle of granular jamming have been introduced into the field of soft robotics due to their advantages such as simple structure, fast response, and large range of stiffness variation. A typical implementation method is to encapsulate micron or millimeter-sized particles (such as coffee grounds, glass microspheres, and polystyrene particles) in a flexible cavity. Under normal conditions, the particles can slide relative to each other, and the cavity is flexible. After the cavity is evacuated, the friction between the particles locks them, and the stiffness of the cavity is significantly improved.

[0007] However, existing particle blocking mechanisms are mostly independent modules. When integrated with pneumatic bending joints on the same soft finger, the following problems generally exist:

[0008] First, the mechanical coupling between the variable stiffness finger plate and the pneumatic drive cavity was not fully considered. During the stiffness switching process, the volume and surface curvature of the finger plate will change slightly, resulting in a sudden change in the fingertip contact force, which may damage the object being grasped or cause the grasping failure.

[0009] Secondly, there is a lack of miniaturized sensing methods that can directly measure fingertip contact force in real time. Existing solutions mostly use air pressure sensors to indirectly estimate contact force. However, due to factors such as the nonlinearity, viscoelasticity, and friction of silicone materials, there is not a one-to-one correspondence between air pressure and contact force, resulting in large estimation errors and making it difficult to achieve coordinated closed-loop control of stiffness switching and gripping force adjustment.

[0010] In terms of grasping force perception, soft fingers currently rely mainly on two methods.

[0011] The first method involves monitoring the air pressure inside the drive cavity and indirectly calculating the contact force using a calibration curve. However, this method is affected by silicone hysteresis and uneven deformation, resulting in limited accuracy and an inability to distinguish the force differences between different parts of the finger.

[0012] The second type is embedded resistive or capacitive flexible strain sensors. These sensors can be attached to the surface of the finger or embedded in silicone, but they have problems such as creep, poor repeatability and temperature drift, and usually require multi-channel signal conditioning circuits, resulting in insufficient long-term stability.

[0013] Therefore, how to integrate a high-precision, low-drift, and interference-resistant force sensing method within a limited finger structure and form a coordinated control with a variable stiffness mechanism is an unsolved technical problem in this field.

[0014] In summary, existing soft fingers cannot simultaneously achieve the dual functions of gently grasping fragile items and firmly grasping heavy objects, and they lack a closed-loop control strategy that integrates variable stiffness adjustment and real-time force feedback.

[0015] Therefore, developing a pneumatic soft finger with a compact structure, fast stiffness response, and high force control accuracy, and its control method, is a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0016] The main objective of this invention is to propose a pneumatic soft finger with variable stiffness finger plate and force feedback, and a control method thereof, in order to solve the technical problems that existing pneumatic soft fingers cannot simultaneously achieve gentle gripping of fragile items and stable gripping of heavy objects, and lack the technical problems of variable stiffness adjustment and coordinated closed-loop control of gripping force.

[0017] To achieve the above objectives, this invention proposes a pneumatic soft finger with a variable stiffness finger plate and force feedback, comprising:

[0018] The flexible joint body is integrally cast from silicone rubber material and has multiple wave-shaped cavities extending along the length direction inside. One end of each wave-shaped cavity is closed, and the other end is connected to an external air pressure source through a pneumatic interface.

[0019] A variable stiffness finger plate is disposed on one side of the gripping surface of the flexible joint body. A closed particle filling cavity is formed inside the finger plate. The particle filling cavity is filled with flowable solid particles with a particle size of 50μm to 200μm. The particle filling cavity is connected to an external vacuum source through a vacuum interface.

[0020] A fiber Bragg grating sensor is embedded inside the back of the flexible joint body along its length to detect the bending strain of the finger and the contact force of the fingertip in real time.

[0021] The pneumatic interface and vacuum interface are each independently connected to an external controller.

[0022] Preferably, the Shore hardness of the silicone rubber material of the flexible joint body is 10A to 30A, and the wall thickness of the wave cavity is 1.5mm to 2.5mm.

[0023] Preferably, the solid particles in the variable stiffness finger plate are selected from one or more of coffee grounds, glass microspheres, or polystyrene particles; when the vacuum interface is open to the atmosphere, the variable stiffness finger plate is in a low stiffness state, and when the vacuum interface is evacuated to -60kPa to -90kPa, the variable stiffness finger plate is in a high stiffness state, with its stiffness increasing by 8 to 15 times compared to the low stiffness state.

[0024] Preferably, the fiber Bragg grating sensor includes at least two Bragg gratings with different center wavelengths, wherein the first grating is located in the region of maximum bending deformation of the finger, and the second grating is located in the fingertip region; the outer diameter of the optical fiber is 125 μm, and the surface is coated with a polyimide layer.

[0025] Preferably, the fiber Bragg grating sensor is embedded 0.5 mm to 1.0 mm below the inner surface of the back of the flexible joint body, and is arranged in a straight line along the length of the finger.

[0026] Preferably, it also includes an external pneumatic control component, which includes: a pneumatic proportional regulating valve, a miniature vacuum pump, a pneumatic solenoid valve, and a vacuum solenoid valve; the pneumatic proportional regulating valve is connected to the waveform cavity through a pneumatic interface, and the miniature vacuum pump and the vacuum solenoid valve are connected to the particle filling cavity through a vacuum interface.

[0027] Preferably, it also includes an external controller, which is a microcontroller or FPGA connected to a fiber optic demodulation module.

[0028] Preferably, the variable stiffness finger plate and the flexible joint body are fixedly connected by co-vulcanization integral molding or silicone adhesive, and an isolation layer with a thickness of 0.3mm to 0.8mm is provided between them.

[0029] Preferably, the layered structure from the gripping surface to the back surface consists of: a variable stiffness finger plate layer, an isolation layer, a pneumatic drive cavity layer, and a back constraint layer.

[0030] The silicone thickness of the back constraint layer is greater than that of the silicone thickness of the gripping surface.

[0031] The present invention also proposes a control method based on the above-mentioned pneumatic soft finger, comprising the following steps:

[0032] S1: Perception and Solving

[0033] The wavelength offset of the fiber Bragg grating sensor is acquired in real time, and the current measured value of gripping force and measured value of bending angle are calculated according to the pre-stored calibration curve. Different gripping force calibration coefficients are stored for the low stiffness state and high stiffness state of the variable stiffness finger plate.

[0034] S2: Pattern Decision

[0035] Preset the force threshold and bending angle threshold in the controller;

[0036] Obtain the target gripping force, compare the target gripping force with the force threshold, and simultaneously compare the measured bending angle with the bending angle threshold:

[0037] If the target grasping force is less than the force threshold and the measured bending angle is less than the bending angle threshold, it is judged to be a small object grasping mode. The controller controls the variable stiffness finger plate to maintain a low stiffness state and adjusts the pneumatic drive air pressure according to the deviation between the measured grasping force and the target grasping force.

[0038] Otherwise, if the object is judged to be a heavy object or in attitude holding mode, the controller controls the variable stiffness finger plate to switch to high stiffness state, and after a delay until the particle is locked, the pneumatic drive air pressure is adjusted according to the deviation between the measured value of the gripping force and the target gripping force.

[0039] S3: Closed-loop control

[0040] The controller uses a proportional-integral-derivative control algorithm to continuously adjust the opening of the pneumatic proportional control valve and change the driving air pressure based on the deviation between the measured gripping force and the target gripping force.

[0041] S4: Execute in a loop

[0042] Repeat steps S1 to S3 until the measured gripping force stabilizes within the allowable error range of the target gripping force.

[0043] This invention integrates a variable stiffness finger plate based on the particle blocking principle with a pneumatically driven flexible joint body, and embeds a fiber Bragg grating sensor on the back of the finger. For the first time, it achieves dual functions on a single soft finger: compliant wrapping and grasping in a low-stiffness state and stable clamping of heavy objects in a high-stiffness state. Specifically, the variable stiffness finger plate has a compressive elastic modulus of approximately 35 kPa at normal pressure, allowing it to adhere to the surface of fragile objects like a flexible film, controlling the contact force to below 3 N and preventing damage to items such as eggs and tofu. When the vacuum pump evacuates to -70 kPa to -90 kPa, the solid particles in the particle-filled cavity undergo frictional locking, increasing the compressive elastic modulus of the finger plate to over 380 kPa. The stiffness change factor reaches 8 to 15 times, enabling the finger to provide a normal grasping force exceeding 10 N, firmly gripping objects weighing over 500 g without bending instability or slippage. Furthermore, the stiffness switching response time is less than 300 ms, far faster than phase change materials and magnetorheological fluid solutions. This invention utilizes a fiber Bragg grating sensor to measure finger bending strain and fingertip contact force in real time. Combined with force-wavelength conversion coefficients calibrated for both low and high stiffness states, it achieves high-precision closed-loop control of the grasping force. The force measurement resolution reaches 0.1N, and the steady-state error of PID control does not exceed ±0.2N. Compared to the traditional indirect pneumatic estimation method (which typically has an error greater than ±1.5N), this represents an order of magnitude improvement in accuracy, effectively solving the force control problem caused by the nonlinear deformation and hysteresis of silicone. The control method of this invention automatically determines the operating mode of the variable stiffness fingerplate based on the comparison between the target grasping force and a preset threshold. During stiffness switching, a delay is introduced to wait for particle locking, avoiding contact force oscillations caused by sudden stiffness changes. Simultaneously, an incremental PID algorithm continuously adjusts the opening of the pneumatic proportional valve, enabling the grasping force to quickly converge to the target value with a stabilization time of less than 0.8s. This achieves adaptive switching between compliant and stiffness in the grasping process and coordinated force closed-loop control. In addition, the fiber Bragg grating sensor is small in size (125μm outer diameter), free from electromagnetic interference, and has good long-term stability. When embedded in silicone, it does not affect the flexibility of the finger or the smoothness of the gripping surface. The entire system has a compact structure, controllable cost, and good engineering practical value. Attached Figure Description

[0044] Figure 1 This is a cross-sectional view of the pneumatic soft finger structure;

[0045] Figure 2 This is a flowchart of the control method logic. Detailed Implementation

[0046] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0047] It should be noted that if the embodiments of the present invention involve directional indicators (such as up, down, left, right, front, back, top, bottom, inside, outside, vertical, horizontal, longitudinal, counterclockwise, clockwise, circumferential, radial, axial, etc.), the directional indicators are only used to explain the relative positional relationship and movement of the components in a certain specific posture (as shown in the figure). If the specific posture changes, the directional indicators will also change accordingly.

[0048] Furthermore, if the embodiments of this invention involve descriptions such as "first" or "second," these descriptions are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, features defined with "first" or "second" may explicitly or implicitly include at least one of those features. Additionally, the technical solutions of the various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. If the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed by this invention.

[0049] Example 1

[0050] A pneumatic soft finger with variable stiffness finger plate and force feedback has the following specific structure:

[0051] The flexible joint body is made of silicone rubber material (specifically Ecoflex 00-30 or Dragon Skin 20) and is integrally cast using a mold.

[0052] The flexible joint body has three parallel wave-shaped cavities 31 arranged inside along its length. Each wave-shaped cavity has a semi-circular or rectangular cross-section and extends in a periodic wave shape along its length.

[0053] One end of all the waveform cavities is closed at the base of the finger, and the other end converges to a common pneumatic interface 5, which is connected to an external air pressure source via a hose.

[0054] The variable stiffness finger plate is set on one side of the gripping surface of the flexible joint body and is integrally formed with the flexible joint body through a co-vulcanization process.

[0055] The variable stiffness finger plate forms a closed particle-filled cavity 11 inside. The cavity is flat, 2.5 mm thick, and its length is consistent with the gripping surface length of the flexible joint body.

[0056] The particle filling cavity is filled with particle filler 11, such as glass microspheres with a particle size of 80μm to 120μm, and the filling volume occupies 75% to 85% of the cavity volume, leaving some gaps to ensure the initial mobility of the particles.

[0057] The particle-filled cavity is connected to an external vacuum source via a separate vacuum interface 6.

[0058] The fiber Bragg grating sensor is embedded inside the back of the flexible joint body along its length.

[0059] Specifically, before casting the silicone, an optical fiber with an outer diameter of 125 μm and coated with a polyimide layer is placed in the back area of ​​the mold. Two Bragg gratings with different center wavelengths are engraved on the optical fiber (one located at the point of maximum bending deformation in the middle of the finger, and the other at the fingertip).

[0060] After casting, the optical fiber is encased inside the silicone, becoming an integral part of the flexible joint body.

[0061] Pneumatic interface 5 and vacuum interface 6 are independently connected to an external controller, which controls the on / off state and pressure of the pneumatic and vacuum sources, respectively.

[0062] Example 2

[0063] Based on Example 1, the material parameters and cavity wall thickness of the flexible joint body are defined.

[0064] In this embodiment, the silicone rubber material used for the flexible joint body has a Shore hardness of 20A (the measured value is A20±2).

[0065] The minimum wall thickness of the waveform cavity (i.e., the thickness of the silicone layer between the cavity and the outer surface of the finger) is designed to be 2.0 mm.

[0066] This thickness ensures that the fingers have sufficient flexibility while withstanding repeated air pressure loading of 0-200 kPa without irreversible bulging or rupture.

[0067] Tests have shown that when the wall thickness is less than 1.5mm, silicone is prone to excessive local expansion under high pressure.

[0068] When the wall thickness is greater than 2.5 mm, the bending angle is significantly reduced under the same air pressure, affecting the gripping adaptability.

[0069] Example 3

[0070] This embodiment specifically defines the types of solid particles and the stiffness variation performance within the variable stiffness finger plate.

[0071] Specifically, the solid particles in the variable stiffness finger plate are selected from coffee grounds that have been screened and dried, with a particle size range of 50μm to 150μm and an average particle size of about 100μm.

[0072] Before filling, the granules are dried in an 80°C oven for 2 hours to remove moisture and prevent clumping under vacuum.

[0073] When the vacuum interface is open to the atmosphere, the particle filling chamber is at normal pressure, allowing the particles to slide relative to each other. The finger plate exhibits low stiffness, with a compressive modulus of elasticity of approximately 35 kPa. When the particle filling chamber is evacuated to -75 kPa (absolute pressure approximately 26 kPa) using a micro-vacuum pump, the particles are compressed under external atmospheric pressure, locking the friction between them. The finger plate exhibits high stiffness, with a measured compressive modulus of elasticity reaching approximately 380 kPa, representing a stiffness increase of approximately 10.8 times. At -60 kPa, the stiffness increases by approximately 8 times, and at -90 kPa, it increases by approximately 15 times. The vacuum level can be selected according to the gripping requirements.

[0074] Example 4

[0075] This embodiment specifically describes the grating layout and fiber parameters of the fiber Bragg grating sensor.

[0076] Specifically, the fiber Bragg grating sensor uses a single optical fiber with two Bragg gratings of different center wavelengths engraved on it.

[0077] The first grating 41 is located in the area of ​​maximum bending deformation of the finger, that is, about 1 / 3 to 1 / 2 of the total length of the finger from the fingertip to the root. Its center wavelength is 1530nm and the grating area length is 10mm. It is used to measure the tensile strain when the finger is bent, and then calculate the bending angle.

[0078] The second grating 42 is located in the fingertip region, that is, within 5mm to 10mm from the fingertip. Its center wavelength is 1550nm and the grating length is 5mm. It is used to measure the local strain generated when the fingertip contacts an object, and then calculate the gripping force.

[0079] The optical fiber body is coated with polyimide, with an outer diameter of 125μm and a core diameter of 9μm (single-mode fiber). The polyimide coating can withstand casting temperatures up to 300℃ and has good interfacial bonding with silicone rubber, preventing relative slippage during long-term use.

[0080] Example 5

[0081] This embodiment defines the embedding location of the fiber Bragg grating sensor within the flexible joint body.

[0082] Specifically, in the mold for casting the flexible joint body, two 0.2mm diameter metal wires are placed as positioning elements on the back area, with the metal wires 0.8mm from the inner surface of the mold (i.e., the outer surface of the back of the finger). An optical fiber with an etched grating is placed tightly against the metal wires. After the silicone is cast and cured, the metal wires are removed, and the optical fiber is fixed at a depth of 0.8mm from the outer surface of the back of the finger. The optical fiber is arranged in a straight line along the length of the finger, without bending or twisting.

[0083] Tests have shown that this embedding depth allows the optical fiber to effectively sense the bending strain of the silicone (too shallow and it is easily exposed or damaged, too deep and the strain transmission loss increases), without affecting the smoothness and gripping performance of the finger's outer surface.

[0084] Example 6

[0085] This embodiment specifically describes the composition and connection relationship of the external pneumatic control components.

[0086] Specifically, the external pneumatic control assembly includes the following components:

[0087] The pneumatic proportional control valve, model SMC ITV0030-2BL, has a range of 0 to 200 kPa and an input signal of 0 to 10V analog voltage or PWM. It is used to precisely control the air pressure input to the pneumatic interface.

[0088] Miniature vacuum pump, model VBY series miniature vacuum pump, maximum vacuum degree -85kPa, no-load flow rate 6L / min, working voltage 24V DC;

[0089] Pneumatic solenoid valve, two-position three-way type, used to connect or disconnect the passage between the air source and the pneumatic interface;

[0090] A vacuum solenoid valve, a two-position normally closed type, is installed between a miniature vacuum pump and a vacuum interface to control the on / off state of vacuuming.

[0091] Connections: The inlet of the pneumatic proportional control valve connects to an external compressed air source (pressure 200kPa~400kPa), and the outlet connects to a pneumatic interface via a pneumatic solenoid valve. The suction port of the miniature vacuum pump connects to a vacuum interface via a vacuum solenoid valve, and the exhaust port of the miniature vacuum pump is directly open to the atmosphere. All valve bodies and pump bodies are installed in the same control box and connected to an external controller via cables.

[0092] Example 7

[0093] This embodiment specifically describes the selection of the external controller and the configuration of the fiber Bragg grating demodulation module.

[0094] Specifically, the external controller uses an STM32F407 series microcontroller with a main frequency of 168MHz and a built-in 12-bit ADC and PWM output module. This microcontroller communicates with the fiber Bragg grating demodulation module via an RS232 serial port.

[0095] The fiber Bragg grating demodulation module uses a commercially available module based on the scanning laser principle (such as the MOI SM130), with a wavelength resolution of 1 pm and a scanning frequency of 100 Hz, capable of demodulating four grating channels simultaneously. The wavelength data output by the demodulation module is sent to the STM32 microcontroller in real time via a serial port. The microcontroller runs a real-time operating system, reading the wavelength data at a frequency of 100 Hz and performing subsequent calculations of bending angles and gripping forces.

[0096] Example 8

[0097] This embodiment specifically defines the connection method between the variable stiffness finger plate and the flexible joint body, as well as the setting of the isolation layer.

[0098] Specifically, the variable stiffness finger plate and the flexible joint body are manufactured using a co-vulcanization integral molding process.

[0099] The specific steps are as follows: First, a preform (made of thermoplastic film and filled with particles) with a particle-filled cavity is placed in the mold. Then, the silicone material of the flexible joint body is cast. During the vulcanization process, the silicone and the outer wall of the particle-filled cavity undergo chemical cross-linking to form a strong bond. A 0.5mm thick insulating layer made of thermoplastic film is naturally formed between the two.

[0100] This isolation layer serves two purposes:

[0101] Firstly, it prevents tiny particles (minimum particle size 50μm) in the particle-filled cavity from seeping into the waveform cavity through the micropores or interface gaps of the silicone during long-term use, causing blockage of the pneumatic channel;

[0102] Secondly, during vacuum suction, the isolation layer can evenly transmit pressure, preventing localized depressions.

[0103] The thickness of the isolation layer should be controlled within the range of 0.3mm to 0.8mm: less than 0.3mm is prone to damage, while more than 0.8mm will significantly increase the overall bending stiffness of the finger and reduce its flexibility.

[0104] Example 9

[0105] This embodiment describes the cross-sectional arrangement of the finger from the perspective of layered structure.

[0106] Specifically, from the gripping surface of the fingers (the side in contact with the object) to the back surface, the order is as follows:

[0107] First layer: Variable stiffness finger plate layer 1. This layer is about 3mm thick and consists of a flexible cavity wall and internal particles. Its outer surface is the gripping surface, and anti-slip textures can be added to the surface.

[0108] Second layer: Isolation layer 2. Thickness 0.5mm, made of PET film, used to separate the particle chamber from the pneumatic chamber.

[0109] The third layer: pneumatic drive cavity layer 3. This layer includes a wave-shaped cavity and a surrounding silicone wall. The cavity height is approximately 2 mm, and the silicone thickness above the cavity (i.e., the distance from the top of the cavity to the isolation layer) is 1 mm.

[0110] Fourth layer: Back restraint layer 4. This layer is made of solid silicone, with a thickness of 3mm from the bottom of the cavity to the outer surface of the back of the finger, which is significantly greater than the thickness of the silicone on the gripping side (1mm).

[0111] Because the back constraint layer is thicker and has a relatively higher tensile modulus, when air is inflated into the wave-shaped cavity, the elongation deformation of the back constraint layer is less than that of the gripping surface side, thus generating a bending moment towards the gripping surface side, causing the fingers to bend towards the palm. This asymmetric structural design is a common technique for achieving directional bending in pneumatic soft fingers.

[0112] Example 10 This example describes in detail the control method of the pneumatic soft finger based on any of the above examples, including specific parameters, algorithm flow and examples.

[0113] Step S1: Perception and Solving

[0114] Two calibration curves are pre-stored in the controller's internal memory and obtained through experimental calibration.

[0115] Calibration method: The pneumatic soft finger is fixed on the test bench, and the wavelength offset Δλ_b of the first grating is measured at different bending angles. The bending angle θ = k1·Δλ_b + c1 is obtained by linear fitting. Typical values ​​are k1=1.2° / pm and c1=0.5°.

[0116] For the gripping force, calibration was performed under both low-stiffness (normal pressure) and high-stiffness (vacuum -70 kPa) conditions of the variable stiffness fingerplate. Simultaneous acquisition of the fingertip contact force by the fingertip sensor and the second grating of the FBG yielded two distinct linear relationships:

[0117] Low stiffness state: F = k2_soft·Δλ_f + c2_soft, typical values ​​k2_soft=0.08 N / pm, c2_soft=0.1N.

[0118] High stiffness state: F = k2_stiff·Δλ_f + c2_stiff, typical values ​​k2_stiff=0.12 N / pm, c2_stiff=0.2N.

[0119] During operation, the controller first reads the current state of the variable stiffness index plate (determined by the on / off state of the vacuum solenoid valve and feedback from the pressure sensor), selects the corresponding calibration coefficient, and converts the real-time acquired Δλ_f into the measured value of the gripping force F_measured. Simultaneously, it continuously uses the Δλ_b of the first grating to calculate the measured value of the bending angle θ_measured.

[0120] Step S2: Pattern Decision

[0121] The controller has a preset force threshold F_th1 = 5N and a bending angle threshold θ_th = 30°. The user sets the target gripping force F_target through the host computer, for example, F_target = 2.5N when gripping an egg, and F_target = 12N when gripping a 500g iron block.

[0122] The controller reads F_target and the current θ_measured, and performs the following judgment:

[0123] Case A (Small Object Grabbing Mode): If F_target < 5N and θ_measured < 30°, the controller outputs the following command: Keep the vacuum solenoid valve closed, allowing the particle filling chamber to communicate with the atmosphere, and maintain the low stiffness state of the variable stiffness finger plate. Simultaneously, based on the deviation between F_target and F_measured, it directly enters the PID closed-loop adjustment in step S3.

[0124] Case B (Heavy Object or Attitude Holding Mode): If F_target ≥ 5N or θ_measured ≥ 30°, the controller outputs the command: immediately start the micro vacuum pump and simultaneously open the vacuum solenoid valve to evacuate the particle filling chamber to -70kPa (controlled by a pressure switch or timer). Maintain this vacuum level for a delay of 150ms (this time is experimentally determined to be the typical time required for the particles to go from a movable state to complete locking). After the delay, the PID closed-loop adjustment in step S3 is initiated based on the deviation between F_target and F_measured.

[0125] Step S3: Closed-loop control

[0126] The controller uses an incremental PID control algorithm. Let the error be e(t) = F_target - F_measured(t), and the sampling period be T = 0.02s (corresponding to 50Hz).

[0127] The PID control output is:

[0128] u(t) = Kp·e(t) + Ki·∑[e(j)·T] + Kd·[e(t)-e(t-1)] / T

[0129] After engineering tuning, Kp = 0.5, Ki = 0.05, and Kd = 0.1 were selected. These coefficients are dimensionless proportional factors, which are ultimately converted into PWM duty cycles (0–100%). The PWM signal output by the microcontroller's timer controls the opening of the pneumatic proportional control valve. The PWM duty cycle has an approximately linear relationship with the output air pressure: 0% duty cycle corresponds to 0 kPa, and 100% corresponds to 200 kPa.

[0130] To prevent integral saturation, an integral limit is set: when u(t) exceeds 95% or falls below 5%, the accumulation of the integral term stops.

[0131] Step S4: Execute repeatedly

[0132] The controller repeats steps S1 to S3 at a frequency of 50Hz. The allowable error is set to δ = 0.2N. If F_measured is within the range of [F_target - 0.2N, F_target + 0.2N] for 5 consecutive sampling periods (i.e., 0.1s), and this state is maintained for at least 0.5s, the gripping force is considered stable, and the controller can enter the holding phase or end the adjustment.

[0133] Specific application examples

[0134] Example 1: Egg handling. Set F_target = 2.5N. Since 2.5N < 5N, the controller does not activate the vacuum pump, and the finger remains flexible. The PID controller adjusts the pneumatic proportional valve, and the PWM gradually increases from 0 to approximately 25%, allowing the finger to slowly bend and contact the egg. F_measured increases to 2.5N ± 0.2N within 1.5 seconds and stabilizes without the egg breaking.

[0135] Example 2: Grabbing a 500g iron block. Set F_target = 12N. Since 12N ≥ 5N, the controller immediately starts the vacuum pump and evacuates to -70kPa, with a delay of 150ms. Then, the PID controller adjusts the pneumatic proportional valve, and the PWM controller rapidly increases from 0 to approximately 65%. F_measured reaches 12N within 0.8s and stabilizes. The iron block is firmly gripped without slippage.

[0136] This invention integrates a variable stiffness finger plate based on the particle blocking principle with a pneumatically driven flexible joint body, and embeds a fiber Bragg grating sensor on the back of the finger. For the first time, it achieves dual functions on a single soft finger: compliant wrapping and grasping in a low-stiffness state and stable clamping of heavy objects in a high-stiffness state. Specifically, the variable stiffness finger plate has a compressive elastic modulus of approximately 35 kPa at normal pressure, allowing it to adhere to the surface of fragile objects like a flexible film, controlling the contact force to below 3 N and preventing damage to items such as eggs and tofu. When the vacuum pump evacuates to -70 kPa to -90 kPa, the solid particles in the particle-filled cavity undergo frictional locking, increasing the compressive elastic modulus of the finger plate to over 380 kPa. The stiffness change factor reaches 8 to 15 times, enabling the finger to provide a normal grasping force exceeding 10 N, firmly gripping objects weighing over 500 g without bending instability or slippage. Furthermore, the stiffness switching response time is less than 300 ms, far faster than phase change materials and magnetorheological fluid solutions. This invention utilizes a fiber Bragg grating sensor to measure finger bending strain and fingertip contact force in real time. Combined with force-wavelength conversion coefficients calibrated for both low and high stiffness states, it achieves high-precision closed-loop control of the grasping force. The force measurement resolution reaches 0.1N, and the steady-state error of PID control does not exceed ±0.2N. Compared to the traditional indirect pneumatic estimation method (which typically has an error greater than ±1.5N), this represents an order of magnitude improvement in accuracy, effectively solving the force control problem caused by the nonlinear deformation and hysteresis of silicone. The control method of this invention automatically determines the operating mode of the variable stiffness fingerplate based on the comparison between the target grasping force and a preset threshold. During stiffness switching, a delay is introduced to wait for particle locking, avoiding contact force oscillations caused by sudden stiffness changes. Simultaneously, an incremental PID algorithm continuously adjusts the opening of the pneumatic proportional valve, enabling the grasping force to quickly converge to the target value with a stabilization time of less than 0.8s. This achieves adaptive switching between compliant and stiffness in the grasping process and coordinated force closed-loop control. In addition, the fiber Bragg grating sensor is small in size (125μm outer diameter), free from electromagnetic interference, and has good long-term stability. When embedded in silicone, it does not affect the flexibility of the finger or the smoothness of the gripping surface. The entire system has a compact structure, controllable cost, and good engineering practical value.

[0137] The above description is merely a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural transformations made using the contents of the present invention's specification and drawings under the inventive concept of the present invention, or direct / indirect applications in other related technical fields, are included within the patent protection scope of the present invention.

Claims

1. A pneumatic soft finger with variable stiffness finger plate and force feedback, characterized in that, include: The flexible joint body is integrally cast from silicone rubber material and has multiple wave-shaped cavities (31) extending along the length direction inside. One end of the wave-shaped cavity is closed and the other end is connected to an external air pressure source through a pneumatic interface (5). A variable stiffness finger plate is set on one side of the gripping surface of the flexible joint body. A closed particle filling cavity (11) is formed inside the plate. The particle filling cavity is filled with flowable solid particles (12) with a particle size of 50μm to 200μm. The particle filling cavity is connected to an external vacuum source through a vacuum interface (6). A fiber Bragg grating sensor is embedded inside the back of the flexible joint body along its length to detect the bending strain of the finger and the contact force of the fingertip in real time. The pneumatic interface (5) and vacuum interface (6) are each independently connected to an external controller.

2. The pneumatic soft finger with variable stiffness finger plate and force feedback as described in claim 1, characterized in that: The silicone rubber material of the flexible joint body has a Shore hardness of 10A to 30A, and the wall thickness of the wave-shaped cavity is 1.5mm to 2.5mm.

3. The pneumatic soft finger with variable stiffness finger plate and force feedback as described in claim 1, characterized in that: The solid particles in the variable stiffness finger are selected from one or more of coffee grounds, glass microspheres, or polystyrene particles. When the vacuum interface is open to the atmosphere, the variable stiffness finger plate is in a low stiffness state. When the vacuum interface is evacuated to -60kPa to -90kPa, the variable stiffness finger plate is in a high stiffness state, and its stiffness is increased by 8 to 15 times compared with the low stiffness state.

4. The pneumatic soft finger with variable stiffness finger plate and force feedback as described in claim 1, characterized in that: The fiber Bragg grating sensor includes at least two Bragg gratings with different center wavelengths, wherein the first grating (41) is located in the region of maximum bending deformation of the finger and the second grating (42) is located in the fingertip region; the outer diameter of the optical fiber is 125 μm and the surface is coated with a polyimide layer.

5. The pneumatic soft finger with variable stiffness finger plate and force feedback as described in claim 4, characterized in that: The fiber Bragg grating sensor is embedded 0.5 mm to 1.0 mm below the inner surface of the back of the flexible joint body, and is arranged in a straight line along the length of the finger.

6. The pneumatic soft finger with variable stiffness finger plate and force feedback as described in claim 1, characterized in that: It also includes an external pneumatic control assembly, which includes: a pneumatic proportional regulating valve, a miniature vacuum pump, a pneumatic solenoid valve, and a vacuum solenoid valve; the pneumatic proportional regulating valve is connected to the waveform cavity through a pneumatic interface, and the miniature vacuum pump and the vacuum solenoid valve are connected to the particle filling cavity through a vacuum interface.

7. The pneumatic soft finger with variable stiffness finger plate and force feedback as described in claim 1, characterized in that: It also includes an external controller, which is a microcontroller or FPGA connected to a fiber optic demodulation module.

8. The pneumatic soft finger as described in claim 1, characterized in that: The variable stiffness finger plate and the flexible joint body are fixedly connected by co-vulcanization integral molding or silicone adhesive, and an isolation layer with a thickness of 0.3mm to 0.8mm is provided between them.

9. The pneumatic soft finger with variable stiffness finger plate and force feedback as described in claim 1, characterized in that: The layered structure from the gripping surface to the back side is as follows: variable stiffness finger plate layer (1), isolation layer (2), pneumatic drive cavity layer (3), and back constraint layer (4). The silicone thickness of the back constraint layer is greater than that of the silicone thickness of the gripping surface.

10. A control method for a pneumatic soft finger with variable stiffness finger plate and force feedback based on any one of claims 1 to 9, characterized in that, Includes the following steps: S1: Perception and Solving The wavelength offset of the fiber Bragg grating sensor is acquired in real time, and the current measured value of gripping force and measured value of bending angle are calculated according to the pre-stored calibration curve. Different gripping force calibration coefficients are stored for the low stiffness state and high stiffness state of the variable stiffness finger plate. S2: Pattern Decision Preset the force threshold and bending angle threshold in the controller; Obtain the target gripping force, compare the target gripping force with the force threshold, and simultaneously compare the measured bending angle with the bending angle threshold: If the target grasping force is less than the force threshold and the measured bending angle is less than the bending angle threshold, it is judged to be a small object grasping mode. The controller controls the variable stiffness finger plate to maintain a low stiffness state and adjusts the pneumatic drive air pressure according to the deviation between the measured grasping force and the target grasping force. Otherwise, if the object is judged to be a heavy object or in attitude holding mode, the controller controls the variable stiffness finger plate to switch to high stiffness state, and after a delay until the particle is locked, the pneumatic drive air pressure is adjusted according to the deviation between the measured value of the gripping force and the target gripping force. S3: Closed-loop control The controller uses a proportional-integral-derivative control algorithm to continuously adjust the opening of the pneumatic proportional control valve and change the driving air pressure based on the deviation between the measured gripping force and the target gripping force. S4: Execute in a loop Repeat steps S1 to S3 until the measured gripping force stabilizes within the allowable error range of the target gripping force.

Images