Three-jaw variable-rigidity gripping clamp based on SMA-MRF cooperative driving and cooperative driving method

Abstract

The invention discloses a three-jaw variable-rigidity grabbing clamp based on SMA-MRF cooperative driving and a cooperative driving method, and belongs to the technical field of robot end executors. The grasping clamp adopts a 120-degree central symmetry layout, each clamping jaw finger is made of superelastic silica gel, a closed pipeline is arranged in each clamping jaw finger and is filled with magnetorheological fluid, a nickel-titanium alloy temperature control wire is embedded in the pipeline, and the two ends of each nickel-titanium alloy temperature control wire are anchored; electromagnetic coils are wound on the outer walls of the clamping jaw fingers and generate a magnetic field vertically penetrating through the fingers after being electrified. Clamping jaw bending driving is achieved by controlling Joule thermally induced phase change of an SMA wire, a magnetic field is used for regulating and controlling MRF to complete liquid-solid reversible conversion in a millisecond level, and thousand-fold dynamic adjustment of the clamping rigidity is achieved within the range of 15 N / m to 15000 N / m. The contradiction that rigidity and flexibility of a traditional grabbing clamp cannot be achieved at the same time is solved, the soft-state self-adaptive enveloping and hard-state precise force control dual-mode grabbing capacity is achieved, + / -0.1 N-level closed-loop force control can be achieved in cooperation with feedback of the strain gauges, and the soft-state self-adaptive enveloping and hard-state precise force control dual-mode grabbing clamp is suitable for precise part assembling and fragile object grabbing.

Description

Technical Field

[0001] This invention relates to a three-jaw variable stiffness gripping technology based on the synergistic drive of shape memory alloy (SMA) and magnetorheological fluid (MRF), belonging to the field of robot dexterity hand and smart material application technology. Background Technology

[0002] Existing robotic grippers are mainly classified into two categories based on their stiffness characteristics: rigid grippers and flexible grippers. Rigid grippers, using metal or hard polymer structures, provide stable high-rigidity support and precise force control, but lack the ability to adapt to complex shapes. When gripping fragile items, they are prone to stress concentration, which can damage the object, and human-robot collaboration safety is poor. Flexible grippers (such as pneumatic soft claws and silicone flexible claws) rely on the flexibility of their material to achieve adaptive envelope to irregular objects, providing good contact safety. However, their stiffness is not adjustable, making them unsuitable for precise force control or delicate operations such as insertion, removal, and rotation. They can only roughly estimate the gripping force through air pressure or displacement, with force control accuracy typically greater than ±1N and response speed in the order of seconds.

[0003] To balance rigidity and flexibility, variable stiffness grippers have emerged in recent years. These grippers primarily achieve stiffness variations through layer interference, particle blockage, and aerodynamic adjustment. However, they suffer from limitations such as limited adjustment range (typically 2-10 times), slow response speed (>100ms), complex structure, and difficulty in achieving independent multi-finger control. Magnetorheological fluids (MRFs), as smart materials, exhibit millisecond-level reversible viscosity changes under magnetic field influence. However, existing MRF grippers mostly employ a single magnetic field drive, lacking organic integration with flexible actuators and failing to form a precise force closed-loop control system with multi-finger collaboration. This makes it difficult to meet the intelligent gripping requirements of precision manufacturing, which involve "soft envelopment followed by hard manipulation." Summary of the Invention

[0004] This invention aims to solve the technical problems of existing gripping technologies, such as the inability to achieve both rigidity and compliance, small stiffness adjustment range, slow response speed, and low precision of multi-finger force control. It provides a three-jaw gripper that can achieve a wide range of dynamic stiffness adjustment, has dual modes of soft adaptive envelope and hard precise force control, and achieves ±0.1N level closed-loop control of gripping force and millisecond-level mode switching.

[0005] The technical solution of this invention is as follows: a three-jaw variable stiffness gripper based on SMA-MRF collaborative drive is provided, comprising: three gripper fingers symmetrically distributed at 120°, each finger being made of superelastic silicone material (Shore A hardness 20-50 degrees) and having a pre-embedded pipe with closed ends, the pipe being filled with magnetorheological fluid; a shape memory alloy temperature control wire is inserted through the pipe and its head and tail are respectively anchored to the two ends of the pipe to form a pre-tightened and straightened state; an electromagnetic coil is independently wound on the outer wall of each gripper finger, and when the electromagnetic coil is energized, it generates a magnetic field that penetrates the gripper finger and the internal pipe vertically, causing the magnetorheological fluid to achieve liquid-solid reversible conversion in milliseconds, thereby driving the gripper to dynamically switch between soft adaptive mode and hard force control mode, enabling the gripper stiffness to be dynamically adjusted by a thousand times in the range of 15 N / m to 15000 N / m.

[0006] In one embodiment of the present invention, the magnetorheological fluid is a carbonyl iron particle suspension with a particle diameter of 1-10 μm; the shape memory alloy temperature control wire is a nickel-titanium alloy wire with a phase transformation induced strain of 3-6%.

[0007] In one embodiment of the present invention, the electromagnetic coil is an independently controllable coil that generates a magnetic induction intensity perpendicular to the pre-embedded pipe at the air gap at the center of the gripper finger. The magnetic induction intensity is continuously adjustable in the range of 0-550mT, and the magnetic field crosstalk between adjacent gripper fingers is less than 8%.

[0008] As one embodiment of the present invention, the geometric constraint anchoring method of the shape memory alloy temperature control wire is as follows: the distance between the two fixed points is 30-60 mm, and the phase change shrinkage stroke is 1.5-5 mm.

[0009] In one embodiment of the present invention, a strain gauge is attached to the inner or outer surface of the root of each gripper finger. The strain gauge detects the bending strain of the gripper in real time through a Wheatstone bridge circuit and establishes a linear calibration relationship between strain, voltage and gripping force to provide feedback on the actual gripping force.

[0010] As one embodiment of the present invention, it also includes a closed-loop control system. The system receives the target grasping force signal and the actual grasping force signal fed back by the strain gauge. The current of the electromagnetic coil of each gripper finger is independently adjusted by the PID controller. The maximum instantaneous output force of a single finger is not less than 20 N, and the maximum instantaneous combined force of the three fingers is not less than 60 N, so as to achieve multi-finger force balance and precise force closed-loop control at the ±0.1 N level.

[0011] This invention also provides a collaborative driving method for a three-jaw variable stiffness gripper based on SMA-MRF collaborative driving, comprising the following stages: Stage 1 - Initial power-off cold state: The shape memory alloy temperature control wire is in the martensitic phase, the magnetorheological fluid is in a zero-field, low-viscosity liquid state, and the three jaws are open and ready; Stage 2 - Soft envelope: The shape memory alloy temperature control wire is energized and heated to above the phase transition temperature As, its contraction drives the bending of the superelastic silicone gripper fingers, and the magnetorheological fluid passively flows to achieve adaptive envelope; Stage 3 - Magnetic field activation: When the strain gauge detects that the contact force reaches the threshold, the electromagnetic coil is activated to generate a vertical magnetic field, causing the magnetorheological fluid to solidify and collaboratively lock the deformation of the gripper, and the stiffness jumps to the hard state; Stage 4 - Force closed-loop maintenance: Based on the actual gripping force fed back by the strain gauge, the closed-loop control system dynamically adjusts the magnetic field strength of each finger to achieve precise maintenance of the target gripping force and anti-disturbance compensation; Stage 5 - Power-off release: The power supply to the shape memory alloy temperature control wire and the electromagnetic coil is cut off, the magnetorheological fluid returns to the liquid state, and the gripper relies on the elastic rebound of the silicone to reset.

[0012] In one embodiment of the present invention, the heating current of the shape memory alloy temperature control wire in stage 2 is 0.5-3.0 A, the heating rate is 30-50℃ / s, and the phase change shrinkage stroke is 2-4 mm; in stage 3, the magnetic field activation time is less than 10 ms, and the peak magnetic induction intensity can reach 200-550 mT.

[0013] As one embodiment of the present invention, the closed-loop control in stage 4 is characterized by using a multi-finger independent PID controller for adjustment, which achieves a force error of less than 5% for grasping irregularly shaped objects, and a force fluctuation recovery time of less than 20ms under external disturbances.

[0014] As one embodiment of the present invention, the PID controller in stage 4 adopts an incremental algorithm and includes feedforward lookup table, anti-integral saturation and output limiting circuits, so that the steady-state force error is ≤ ±0.1 N and the disturbance recovery time is ≤ 20 ms.

[0015] Compared with the prior art, the present invention achieves the following beneficial effects:

[0016] 1. Dynamically adjustable stiffness over a wide range: It achieves continuous stiffness adjustment from 15 N / m to 15000 N / m, combining the self-adaptability of soft grippers with the high stiffness support capability of rigid grippers.

[0017] 2. Millisecond-level fast response: MRF liquid-solid transition time <10ms, SMA drive to phase change completion only takes 2-3 seconds, mode switching speed far exceeds existing pneumatic solutions;

[0018] 3. High-precision force control: Through strain gauge feedback and PID closed-loop control, the gripping force accuracy reaches ±0.1N, and the force fluctuation recovery time is <20ms, making it suitable for precision assembly;

[0019] 4. Multi-finger independent coordination: The design of three independent coils and strain gauges reduces magnetic field crosstalk between adjacent fingers to <8%, enabling balanced gripping of irregular objects with force error of <5% for each finger.

[0020] 5. Compact and integrated structure: The SMA wire is pre-embedded in the silicone tube and the MRF is sealed in the body. The overall structure is compact with no exposed drive components and the protection level reaches IP65. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the overall structure of the present invention.

[0022] Figure 2 This is a schematic diagram of the internal structure of the present invention.

[0023] Figure 3 This is a diagram showing the gripper before it grasps an object.

[0024] Figure 4 This is a schematic diagram illustrating the bending and changing of the fingers during the gripper's grasping of an object.

[0025] Figure 5 This is a flowchart of the PID force closed-loop control algorithm of the present invention.

[0026] Explanation of the attached diagram labels: No. 1 nickel-titanium alloy temperature control wire tube 1, No. 2 nickel-titanium alloy temperature control wire tube 2, No. 3 nickel-titanium alloy temperature control wire tube 3, No. 1 clamping finger 4, No. 2 clamping finger 5, No. 3 clamping finger 6. Detailed Implementation

[0027] Example 1: Standard Three-Jaw Variable Stiffness Grip

[0028] This embodiment fully implements all the technical features described in claims 1-10 and is a standard configuration scheme.

[0029] 1. Structural composition as follows Figure 1The diagram shows the overall structure of the invention. The gripper includes a three-claw body, a drive unit, a stiffness control unit, a sensing unit, and a control system. The three claw fingers (4 is the first gripper finger, 5 is the second gripper finger, and 6 is the third gripper finger) are made of super-elastic silicone material with a Shore A hardness of 30, and are cast using a 3D printing mold. Medical-grade silicone tubes (1 is the first nickel-titanium alloy temperature control wire tube, 2 is the second nickel-titanium alloy temperature control wire tube, and 3 is the third nickel-titanium alloy temperature control wire tube) with a diameter of 3mm are pre-embedded inside, with a wall thickness of 0.5mm to ensure a sealed connection with the body. Each finger has a base thickness of 8mm, a tip thickness of 5mm, and a length of 55mm. The three fingers are symmetrically distributed at a 120° center on a 60mm diameter triangular aluminum alloy base. Figure 2 This is a schematic diagram of the internal structure of the present invention. Figure 3 This is a diagram showing the gripper before it grasps an object.

[0030] 2. Material selection and parameters

[0031] MRF magnetorheological fluid: MRF magnetorheological fluid is used, with carbonyl iron particles accounting for 32% of the volume, particle diameter of 3-5 μm, zero-field viscosity of 0.3 Pa·s, saturation magnetization of 1.5 T, completely filling the pre-embedded pipe without any residual air bubbles.

[0032] SMA temperature control wire: NiTi-01 nickel-titanium alloy wire, 0.3mm in diameter, phase change temperatures As=68℃, Af=78℃, Ms=55℃, Mf=45℃, latent heat of phase change 24J / g, maximum recoverable strain 4.5%, the two ends of the wire are anchored to the beginning and end of the pipe through crimp terminals, the anchoring spacing is 48mm, and the preload is set to 5N.

[0033] Electromagnetic coil: 800 turns are wound independently for each finger. The enameled copper wire has a diameter of 0.2mm, a coil length of 15mm, an inner diameter that fits tightly against the outer wall of the finger, an outer diameter of 12mm, and a DC resistance of 8.5Ω / finger.

[0034] 3. Magnetic Circuit Design: The coil frame is made of pure iron in a U-shaped magnetic yoke. A 2mm wide isolation slot is cut into the yoke to ensure that the measured crosstalk between adjacent magnetic fields is 6.5% (<8%). When the coil is energized at 1.5A, the central air gap magnetic induction intensity B_max = 350mT, penetrating vertically through the MRF channel. The magnetic leakage flux ratio of the magnetic circuit is optimized to 4.2% (<5%) through finite element simulation.

[0035] 4. The sensing and control system uses strain gauges attached to the inner side of the base of each finger, with a range of ±5000με and a sensitivity coefficient of 2.1. These gauges output a 0-10V analog signal via a Wheatstone bridge module. The control system employs an STM32F407 microcontroller with an operating frequency of 168MHz and a PID control cycle of 1ms. With a force control target value F_target = 5N, the measured steady-state error is ±0.08N (<±0.1N), and the force fluctuation recovery time is 18ms (<20ms).

[0036] 5. Driving Process and Performance: The process for gripping a ceramic part with a diameter of 30mm is as follows:

[0037] Phase 1: Initial state, current I_SMA=0A, I_coil=0A, three-jaw opening angle 120°;

[0038] Stage 2: SMA is energized with I_SMA=1.8A, heating rate is 42℃ / s, temperature reaches 70℃ after 2.1 seconds, phase change shrinkage is 3.2mm, driving the finger to bend to contact the surface of the part; Figure 4 This is a schematic diagram illustrating the bending changes of the fingers during the gripper's grasping of an object;

[0039] Stage 3: The strain gauge detected F_actual=0.6N>th threshold, the starting coil current I_coil=1.2A, the magnetic field strength B=260mT, the MRF curing time is 8ms, and the stiffness jumps from 20N / m to 7850N / m;

[0040] Phase 4: PID control maintains F_target=5N for 30 seconds, with force fluctuation range of 4.92-5.08N during the grasping process;

[0041] Phase 5: Power off, MRF viscosity recovery time 4ms, SMA cooling to below Mf temperature takes 4.5 seconds, silicone elastic rebound causes the three claws to reset.

[0042] Example 2: Miniature Precision Assembly Type Gripper

[0043] This embodiment focuses on size reduction and parameter optimization for high-precision force control scenarios.

[0044] 1. Structural differences: The gripper finger length is shortened to 35mm, the root thickness is 5mm, the fingertip thickness is 3mm, and the three-finger base diameter is 40mm. The pre-embedded pipe diameter is 2mm, the SMA temperature control wire diameter is reduced to 0.15mm, and the phase change temperature As is optimized to 55℃ (low stress condition). The electromagnetic coil turns are increased to 1000 turns, and the wire diameter is 0.15mm to reduce volume.

[0045] 2. Parameter Adjustment

[0046] MRF: Select MRF-140CG for better flowability, with a zero-field viscosity of 0.18 Pa·s;

[0047] SMA: Maximum contractile force decreased to 8N, but strain remained at 4.5%;

[0048] Magnetic field: When the coil current is 0.8A, B_max = 220mT, which meets the miniaturization requirements.

[0049] 3. Performance indicators: When grasping a 2mm×2mm micro gear, the target force F_target=0.5N, the measured force control accuracy is ±0.03N, the response time is 6ms, the stiffness adjustment range is 15N / m to 3200N / m (213 times), and the three-finger force imbalance is <4%.

[0050] Example 3: Heavy-duty industrial gripper

[0051] This embodiment achieves both high stiffness and high output force.

[0052] 1. The thickness of the gripper fingers at the base has been increased to 12mm, the fingertip thickness to 8mm, and the length to 70mm. The base is made of stainless steel. The pre-embedded pipe diameter is 5mm, the SMA temperature control wire diameter is 0.5mm, and the phase change driving force has been increased to 40N.

[0053] 2. Magnetic circuit upgrade

[0054] Coil parameters: 600 turns, wire diameter 0.3mm, maximum current 2.5A;

[0055] Magnetic yoke material: High saturation magnetic induction intensity silicon steel sheet (B_s=2.0T) is used, with a peak magnetic induction intensity of 500mT.

[0056] 3. Performance When gripping a 5kg metal part (weight 49N), the hard stiffness reaches 15000N / m, the holding force F_target=60N±0.5N, the three-finger force imbalance is <5%, and the magnetorheological fluid curing time is <8ms, which can meet the requirements of stable clamping and precise assembly of heavy workpieces.

[0057] Example 4: PID Control Algorithm Example

[0058] This embodiment provides a multi-finger independent PID force closed-loop algorithm running on an STM32F407 microcontroller with a period of 1 ms, which enables the three-jaw gripper to have a steady-state force error ≤ ±0.1 N, a disturbance recovery time ≤ 20ms, and an adjacent finger force imbalance ≤ 5% in "hard force control mode".

[0059] Its control objectives and performance indicators include:

[0060] ① Steady-state error: ±0.1 N (3σ, 25 ℃) ② Rise time: ≤ 25 ms (10%-90%) ③ Recovery time of external 2N step disturbance: ≤ 20 ms ④ Overshoot: σ ≤ 5% ⑤ Difference between adjacent finger strengths: ≤ 5%

[0061] The discrete model of the controlled object, measured by the step response method at a sampling rate of 1 kHz, has a single-finite "magnetic field-force" discrete transfer function (ZOH) as: G(z) = (0.087 z⁻¹ + 0.076 z⁻²) / (1 - 1.65 z⁻¹ + 0.74 z⁻²). This model is only used for offline pole placement and does not depend on the model during online operation.

[0062] Its controller structure adopts a three-loop cascade architecture of "feedforward + incremental PID + anti-integral saturation", such as Figure 5 As shown.

[0063] Its offline tuning procedure is the ZN-λ method, and the measured open-loop phase margin is 48°, which meets the requirements.

[0064] The above four embodiments verify the feasibility of the technical solution of the present invention from different dimensions, covering the full range of needs from micro-precision gripping to heavy-duty industrial applications, and fully realizing the various technical features and beneficial effects described in the claims.

Claims

1. A three-jaw variable stiffness gripper based on SMA-MRF cooperative drive, characterized in that, include: Three gripper fingers are centrally symmetrically distributed at 120°. Each finger is made of super-elastic silicone material and has a pre-embedded channel with closed ends, which is filled with magnetorheological fluid. A shape memory alloy temperature control wire is inserted through the channel and its head and tail are anchored to the two ends of the channel, forming a pre-tightened and straightened state. An electromagnetic coil is independently wound on the outer wall of each gripper finger. When the electromagnetic coil is energized, it generates a magnetic field that penetrates the gripper finger and the internal channel vertically, so that the magnetorheological fluid achieves liquid-solid reversible conversion in milliseconds. This drives the gripper to dynamically switch between a soft adaptive mode and a hard force control mode. The gripper stiffness can be continuously adjusted in the range of 15 N / m to 15000 N / m.

2. The three-jaw variable stiffness gripper based on SMA-MRF cooperative drive according to claim 1, characterized in that, The magnetorheological fluid is a carbonyl iron particle suspension with a particle diameter of 1-10 μm; the shape memory alloy temperature control wire is a nickel-titanium alloy wire with a phase transformation induced strain of 3-6%.

3. The three-jaw variable stiffness gripper based on SMA-MRF cooperative drive according to claim 1, characterized in that, The electromagnetic coil is an independent controllable coil that generates a magnetic induction intensity perpendicular to the pre-embedded pipe at the air gap in the center of the gripper fingers. This magnetic induction intensity is continuously adjustable in the range of 0-550mT, and the magnetic field crosstalk between adjacent gripper fingers is less than 8%.

4. The three-jaw variable stiffness gripper based on SMA-MRF cooperative drive according to claim 1, characterized in that, The geometric constraint anchoring method of the shape memory alloy temperature control wire is as follows: the distance between the two fixed points is 30-60 mm, and the phase change shrinkage stroke is 1.5-5 mm.

5. The three-jaw variable stiffness gripper based on SMA-MRF cooperative drive according to claim 1, characterized in that, Each gripper finger has a strain gauge attached to the inner or outer surface of its root. The strain gauge detects the gripper bending strain in real time through a Wheatstone bridge circuit and establishes a linear calibration relationship between strain, voltage, and gripping force to provide feedback on the actual gripping force.

6. The three-jaw variable stiffness gripper based on SMA-MRF cooperative drive according to claim 5, characterized in that, It also includes a closed-loop control system, which receives the target grasping force signal and the actual grasping force signal fed back by the strain gauge, and independently adjusts the current of the electromagnetic coil of each gripper finger through a PID controller. The maximum instantaneous output force of a single finger is not less than 20 N, and the maximum instantaneous combined force of the three fingers is not less than 60 N, realizing multi-finger force balance and precise force closed-loop control at the ±0.1 N level.

7. A collaborative driving method for a three-jaw variable stiffness gripper based on SMA-MRF collaborative driving as described in any one of claims 1-6, characterized in that, The system comprises the following stages: Stage 1 - Initial Power-Off Cold State: The shape memory alloy temperature control wire is in the martensitic phase, the magnetorheological fluid is in a zero-field, low-viscosity liquid state, and the three claws are open and ready to engage; Stage 2 - Soft Envelope: The shape memory alloy temperature control wire is heated to above the phase transition temperature As, and its contraction drives the bending of the superelastic silicone claw fingers, allowing the magnetorheological fluid to passively flow and achieve adaptive envelope; Stage 3 - Magnetic Field Activation: When the strain gauge detects that the contact force reaches the threshold, the electromagnetic coil is activated to generate a vertical magnetic field, causing the magnetorheological fluid to solidify and collaboratively lock the deformation of the claws, resulting in a jump in stiffness to the hard state; Stage 4 - Force Closed-Loop Maintenance: Based on the actual gripping force fed back by the strain gauge, the closed-loop control system dynamically adjusts the magnetic field strength of each finger to achieve precise maintenance of the target gripping force and anti-disturbance compensation; Stage 5 - Power-Off Release: The power supply to the shape memory alloy temperature control wire and the electromagnetic coil is cut off, the magnetorheological fluid returns to a liquid state, and the claws rely on the elastic rebound of the silicone to reset.

8. The collaborative driving method for a three-jaw variable stiffness gripper based on SMA-MRF collaborative driving according to claim 7, characterized in that, In stage 2, the heating current of the shape memory alloy temperature control wire is 0.5-3.0 A, the heating rate is 30-50℃ / s, and the phase change shrinkage stroke is 2-4mm; in stage 3, the magnetic field activation time is less than 10ms, and the peak magnetic induction intensity can reach 200-550 mT.

9. The collaborative driving method for a three-jaw variable stiffness gripper based on SMA-MRF collaborative driving according to claim 7, characterized in that, In stage 4, the closed-loop control uses a multi-finger independent PID controller for adjustment, which achieves a force error of less than 5% for grasping irregularly shaped objects, and the force fluctuation recovery time under external disturbances is less than 20ms.

10. The collaborative driving method for a three-jaw variable stiffness gripper based on SMA-MRF collaborative driving according to claim 9, characterized in that, The PID controller described in Stage 4 adopts an incremental algorithm and includes feedforward lookup table, anti-integral saturation and output limiting components, so that the steady-state force error is ≤ ±0.1 N and the disturbance recovery time is ≤ 20 ms.

Images