Flexible grabbing safety control method based on tactile array and clamping jaw

Abstract

The invention discloses a flexible grabbing safety control method based on a tactile array and a clamping jaw, and relates to the technical field of machine grabbing, and the flexible grabbing safety control method comprises the following specific steps: constructing a flexible tactile sensing layer, a grabbing initial contact stage, a grabbing force establishing stage, a grabbing keeping stage and a releasing stage. The method comprises the steps that firstly, flexible tactile array sensors are attached to the inner sides of the left finger face and the right finger face of a clamping jaw respectively, and a pressure distribution matrix sum is obtained in real time. The method has the advantages that objects are effectively prevented from sliding off and being damaged through full-stage tactile feedback and sliding pre-judgment, and the method is suitable for vulnerable targets such as fresh food, precise components and medical consumables; the shape and rigidity of an object do not need to be known in advance, grabbing posture self-adaption and force-position cooperation can be achieved through tactile guidance, and the debugging threshold is greatly lowered.

Description

Technical Field

[0001] This invention relates to the field of machine gripping technology, specifically to a flexible gripping safety control method and gripper based on a tactile array. Background Technology

[0002] Robot end effectors (grippers) are core components for grasping, manipulating, and placing objects. They are widely used in intelligent manufacturing, logistics sorting, agricultural harvesting, medical rehabilitation, and home services. As application scenarios develop towards refinement, flexibility, and intelligence, higher requirements are placed on the gripper's gripping safety—that is, to stably hold objects while avoiding damage or slippage. Traditional rigid grippers and some existing force control solutions have significant shortcomings when dealing with irregularly shaped, fragile, and soft objects. Therefore, we propose a flexible gripping safety control method and gripper based on a tactile array. Summary of the Invention

[0003] The purpose of this invention is to provide a flexible gripping safety control method and gripper based on a tactile array.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a flexible grasping safety control method based on a tactile array, comprising constructing a flexible tactile sensing layer, an initial grasping contact stage, a grasping force establishment stage, a grasping holding stage, and a release stage. The specific steps of the flexible grasping safety control method are as follows: Step 1: Attach flexible tactile array sensors to the inner sides of the left and right finger surfaces of the gripper to acquire the pressure distribution matrix in real time. and ; Step 2: Based on the contour, centroid, contact area, and pressure center of the contact area, dynamically fine-tune the opening and closing posture and angle of the gripper to make the gripping contact surface fit the object surface, and use tactile-guided impedance control to make the gripper exhibit low stiffness characteristics when approaching the object. Step 3, Grip Force Establishment Stage: Using the pressure distribution matrix as feedback, calculate the total gripping force and pressure distribution uniformity index. A dynamic impedance control model is established, in which the damping parameters are... With stiffness parameters Based on pressure distribution uniformity And the contact area is adaptively adjusted in real time; Step 4, Grab and Hold Phase: Extract time-domain, spatial-domain, and frequency-domain features from the pressure distribution matrix. Input the extracted features into a slip prediction model based on a long short-term memory network, and output the slip probability. and according to Based on the relationship with the preset threshold, a tiered response strategy is executed; Step 5: Adopt a gradual force unloading strategy, and reduce the clamping force at a preset rate gradient according to the real-time pressure distribution.

[0005] As a further aspect of the present invention: In step one, flexible tactile array sensors are attached to the inner sides of the left and right finger surfaces of the gripper, respectively. The sensors are piezoresistive flexible arrays, and each sensing unit independently outputs a pressure value, forming a real-time pressure distribution matrix. and .

[0006] As a further aspect of the present invention: In step two, the controller receives tactile array data, and when the pressure value of any sensing unit exceeds a preset contact threshold... When contact is detected, the system extracts the contour, centroid, and contact area of ​​the contact region. And the pressure center, and based on the symmetry and distribution characteristics of the contact pattern, dynamically fine-tune the opening and closing posture and angle of the gripper to make the gripping contact surface fit the object surface.

[0007] As a further aspect of the present invention: In step three, after entering the force establishment stage, the controller calculates the total gripping force using the pressure distribution matrix as feedback. ; in, For unit pressure, For unit area.

[0008] As a further aspect of the present invention: in step three, the pressure distribution uniformity index is calculated: ; in Standard deviation, This is the mean.

[0009] As a further aspect of the present invention: in step three, a dynamic impedance control model is established: ; in For positional deviation, For the target inertia, and Uniformity of distribution with pressure The damping and stiffness matrices adaptively adjust to changes in contact area. When At lower levels, increase Reduce clamping speed to avoid localized overload, and when the contact area... When increasing, appropriately decrease To maintain a constant clamping force and prevent rigid compression.

[0010] As a further aspect of the present invention: In step four, during the grasping and holding phase, the system continuously analyzes the dynamic changes of the tactile array and constructs a multi-dimensional stability evaluation index. Time-domain characteristics: time series fluctuations of pressure values ​​of each unit, and drift rate of pressure center.

[0011] Spatial characteristics: changes in pressure gradient at the contact edge and changes in the topology of the contact region.

[0012] Frequency domain features: Fourier transform is performed on the pressure center trajectory to extract the slip precursor frequency; A glide prediction model based on a long short-term memory network is adopted, with the input being past data. The pressure distribution sequence within the time window is output as the slip probability. ,when Exceeding the warning threshold At this time, the controller performs a fine-tuning action, increasing the clamping force to the preset safety upper limit. And adjust the impedance parameters to suppress vibration, when Exceeding the danger threshold When it is determined that slippage is about to occur, the system immediately activates an active anti-slip strategy, including instantly increasing the clamping force, adjusting the gripper posture to redistribute the contact surface, and issuing an audible and visual alarm.

[0013] A flexible gripping safety control gripper based on a tactile array includes a gripper body, flexible grippers, a control unit, and a flexible tactile sensor, with three sets of the flexible grippers mounted on the bottom of the gripper body.

[0014] As a further embodiment of the present invention: the control unit is installed at the bottom of the gripper body, and the flexible tactile sensor is installed on the inner side of the flexible gripper.

[0015] Compared with the prior art, the beneficial effects of the present invention by adopting the above technical solution are as follows: 1. This invention effectively prevents objects from slipping and getting damaged through full-stage tactile feedback and slip prediction. It is suitable for fragile targets such as fresh food, precision components, and medical consumables. It can achieve adaptive grasping posture and force-position coordination through tactile guidance without prior knowledge of the object's shape and rigidity, greatly reducing the debugging threshold. 2. This invention achieves force control accuracy of 0.1N by utilizing pressure distribution information, and can control the uniformity of pressure distribution on the contact surface, which is superior to traditional single force sensors. 3. This invention can complete an evaluation and decision in a short time through a slip prediction model based on edge computing, which meets the needs of dynamic grasping scenarios. Attached Figure Description

[0016] Figure 1This is a flowchart of the flexible gripping security control method in an embodiment of the present invention; Figure 2 This is a perspective view of the flexible gripping safety control gripper in an embodiment of the present invention; Figure 3 This is a side view of the flexible gripping safety control gripper in an embodiment of the present invention. Detailed Implementation

[0017] The specific embodiments of the present invention will be further described below with reference to the accompanying drawings. It should be noted that the description of these embodiments is for the purpose of helping to understand the present invention, but does not constitute a limitation of the present invention.

[0018] Furthermore, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0019] Please see the appendix Figure 1 -Appendix Figure 3 This invention discloses a flexible grasping safety control method based on a tactile array, comprising constructing a flexible tactile sensing layer, an initial grasping contact stage, a grasping force establishment stage, a grasping holding stage, and a release stage. The specific steps of the flexible grasping safety control method are as follows: Step 1: Attach flexible tactile array sensors to the inner sides of the left and right finger surfaces of the gripper to acquire the pressure distribution matrix in real time. and ; Step 2: Based on the contour, centroid, contact area, and pressure center of the contact area, dynamically fine-tune the opening and closing posture and angle of the gripper to make the gripping contact surface fit the object surface, and use tactile-guided impedance control to make the gripper exhibit low stiffness characteristics when approaching the object. Step 3, Grip Force Establishment Stage: Using the pressure distribution matrix as feedback, calculate the total gripping force and pressure distribution uniformity index. A dynamic impedance control model is established, in which the damping parameters are... With stiffness parameters Based on pressure distribution uniformity And the contact area is adaptively adjusted in real time; Step 4, Grab and Hold Phase: Extract time-domain, spatial-domain, and frequency-domain features from the pressure distribution matrix. Input the extracted features into a slip prediction model based on a long short-term memory network, and output the slip probability. and according to Based on the relationship with the preset threshold, a tiered response strategy is executed; Step 5: Adopt a gradual force unloading strategy, and reduce the clamping force at a preset rate gradient according to the real-time pressure distribution.

[0020] In one embodiment of the present invention: in step one, flexible tactile array sensors are respectively attached to the inner sides of the left and right finger surfaces of the gripper. The sensors are piezoresistive flexible arrays, and each sensing unit independently outputs a pressure value, forming a real-time pressure distribution matrix. and .

[0021] In one embodiment of the present invention: in step two, the controller receives tactile array data, and when the pressure value of any sensing unit exceeds a preset contact threshold... When contact is detected, the system extracts the contour, centroid, and contact area of ​​the contact region. And the pressure center, and based on the symmetry and distribution characteristics of the contact pattern, dynamically fine-tune the opening and closing posture and angle of the gripper to make the gripping contact surface fit the object surface.

[0022] In one embodiment of the present invention: In step three, after entering the force establishment stage, the controller calculates the total gripping force using the pressure distribution matrix as feedback. ; in, For unit pressure, For unit area.

[0023] In one embodiment of the present invention: in step three, the pressure distribution uniformity index is calculated: ; in Standard deviation, This is the mean.

[0024] In one embodiment of the present invention: In step three, a dynamic impedance control model is established: ; in For positional deviation, For the target inertia, and Uniformity of distribution with pressure The damping and stiffness matrices adaptively adjust to changes in contact area. When At lower levels, increase Reduce clamping speed to avoid localized overload, and when the contact area... When increasing, appropriately decrease To maintain a constant clamping force and prevent rigid compression.

[0025] Example 1, please refer to the appendix. Figure 1 -Appendix Figure 3 Application in the end effector of tomato harvesting robots: 1. Equipment preparation: The system employs a flexible gripper for safety control. Three sets of flexible grippers 2 are mounted on the bottom of the gripper body 1, symmetrically distributed at 120°. Each flexible gripper 2 has a 16×16 dot matrix piezoresistive flexible tactile sensor 4 attached to its inner side. The sensor has a sampling frequency of 200Hz and a pressure measurement range of 0N-20N. The control unit 3 uses an STM32H743 microcontroller with a built-in LSTM slip prediction model based on TensorFlow LiteMicro. 2. Pre-fetch settings: The target object is a ripe tomato (approximately 70mm in diameter, 150g in weight, and with a surface friction coefficient of approximately 0.35). The system has a preset contact threshold. =0.1N, target grasping force =3.0N, warning threshold =0.6, danger threshold =0.85; 3. Fetching process: Step 1: Constructing a flexible tactile sensing layer The gripper slowly approaches the tomato under the guidance of the robotic arm. When the pressure value of any sensing unit in the flexible tactile sensor 4 exceeds 0.1N, the control unit 3 determines that contact has occurred and records the pressure distribution matrix of the left and right finger surfaces at this time. and The initial contact points are two points located at the tomato equator, and the pressure distribution shows two isolated peaks; Step Two: Grasping the Initial Contact Phase Control unit 3 extracts the contact area contour and calculates the contact area. =85mm² pressure center is located slightly below the left claw. Based on the asymmetry of the contact pattern, the system determines that the tomato is in an eccentric posture in the gripper. Control unit 3 sends a command to drive gripper body 1 to fine-tune the opening and closing angle: the left claw opens outward by 0.5mm, and the right claw closes inward by 0.3mm. At the same time, the end effector of the robotic arm adjusts the pitch angle by 2°. After three iterations of adjustment, the ratio of the left and right contact areas becomes 1:0.95, and the pressure center tends to be in the center position. During this process, the initial damping of the impedance control model... =0.5 N·s / m, stiffness =200N / m, exhibiting low stiffness characteristics, avoiding impact damage; Step 3: Grasping Force Establishment Phase The system enters the force build-up phase and performs the following calculations with a period of 20ms: Total grasping power ,in =1mm 2 Pressure distribution uniformity ,initial =0.48, indicating that the pressure distribution is extremely uneven; Dynamic impedance control model: ,Pick =0.1kg, positional deviation Feedback from the encoder, the controller according to Adjust parameters in real time: when When <0.6, increase Reduce the clamping speed to 1.2 N·s / m (from 5 mm / s to 2 mm / s) while maintaining... =200N / m; As the tomato surface gradually adheres... The value rose to 0.82, and the system gradually decreased. Up to 0.6 N·s / m, and appropriately increase The force was increased to 250 N / m, allowing the total gripping force to smoothly reach 3.0 N. The entire force build-up process took 1.2 seconds with no overshoot. Step 4: Grab and Hold Phase The gripper maintains its grasping state, and the system continuously analyzes the dynamic changes of the tactile array: Time-domain characteristics: Extract the standard deviation volatility (<0.05N / s) and pressure center drift rate (0.2mm / s) of each unit pressure value. Spatial characteristics: The pressure gradient at the contact edge was calculated, and the rate of decrease from the center to the outer edge was 0.3 N / mm. The topology of the contact area remained stable. Frequency domain characteristics: A fast Fourier transform of the pressure center trajectory shows a small fluctuation at 0.8 Hz, but the amplitude is lower than the warning threshold; The above features are input into the LSTM slip prediction model. Model structure: Input window =0.5s, i.e., 100 time steps, 64 hidden layers, sigmoid activation in the output layer, and the model outputs the slip probability. =0.32, which is lower than the warning threshold of 0.6, indicating that the system is stable; Five seconds after grasping, due to the accelerated movement of the robotic arm, the tactile sensor detected a sudden increase in the pressure center drift rate to 0.8 mm / s, and the pressure gradient at the contact edge rose to 0.6 N / mm. The LSTM model output... =0.72, exceeding the warning threshold of 0.6, the controller performs a fine-tuning action: increasing the clamping force from 3.0N to the safe upper limit. =4.5N, and experimental verification showed that the crushing threshold of tomato skin is 5.0N. The impedance parameters were adjusted accordingly. Increased to 1.0 N·s / m to suppress vibration; after 0.3 s, The temperature dropped to 0.45, and the gripping stabilized, with no slippage or damage to the peel occurring throughout the process; Step 5: Release Phase Upon receiving the release command, the control unit 3 adopts a gradual force unloading strategy: the clamping force is gradually reduced at a rate of 0.5 N / s while monitoring the pressure distribution change. When the total force drops to 1.0 N, the system pauses unloading for 0.2 s to confirm that the tomato's posture is stable and the pressure center drift is <0.1 mm / s. Then, it continues to unload to 0 N, the grippers are fully opened, and the tomato is placed stably in the collection basket. 4. Comparison of effects: Using the method described in this embodiment, a grasping test was conducted on 100 ripe tomatoes, and the results are as follows: Breakage rate: 0% (the control group used a traditional force / position hybrid control gripper, with a breakage rate of 14%). Slippage rate: 1% (slippage rate of control group was 6%); Average grasping cycle: 2.3s (2.8s for the control group, which required multiple attempts); This embodiment fully demonstrates the superiority of the present invention in the non-destructive grasping of fruits and vegetables; Example 2, please refer to the appendix. Figure 1 -Appendix Figure 3 Applications in surface mount technology for microchips: 1. Device and parameter adjustment The same three-finger gripper structure as in Example 1 is adopted, but the flexible tactile sensor 4 is upgraded to a 32×32 dot matrix capacitive sensor (sampling rate 500Hz), and the control unit 3 operates in high-precision mode, increasing the target gripping force. =0.8N, contact threshold =0.02N, warning threshold =0.5, danger threshold =0.7. Due to the smooth surface of the chip, the coefficient of friction is about 0.15, and it is easily affected by electrostatic adsorption, so the uniformity of pressure distribution needs to be strictly controlled. 2. Key points of the crawling process Initial contact phase: The grippers approach the chip at a low speed of 0.5 mm / s. When the tactile array detects contact, the initial contact area is only 0.5 mm². 2 (Equivalent to 4 sensing units), the pressure distribution is extremely uneven. =0.21), the system expands the contact area to 2.5mm² and improves the pressure distribution uniformity to 0.65 by finely adjusting the gripper posture (three fingers move independently, each finger adjusts by 0.02mm). Force building phase: The dynamic impedance control model employs a higher gain: =0.02kg, initial =0.1 N·s / m, =50N / m, due to the fragility of the chip, the system sets the pressure distribution uniformity target to... =0.9, the control algorithm calculates the pressure value of each sensing unit in real time. If the pressure of any unit exceeds 0.05N, the stiffness is immediately reduced. The clamping speed was slowed down, and after a fine adjustment of 0.8 seconds, the total gripping force reached 0.8N, while the pressure of all units was between 0.02N and 0.04N. =0.92, which meets the requirements; Maintenance phase: Due to the extremely light weight and low inertia of the chip, slippage mainly originates from vibration or the acceleration and deceleration of the robotic arm. The input features of the LSTM slippage prediction model include high-frequency fluctuations (>20Hz) at the pressure center. During a sudden stop of the robotic arm, the model detected that the drift rate of the pressure center jumped from 0.05mm / s to 0.4mm / s within 10ms, and the output... =0.68 (exceeding the warning threshold of 0.5), the controller immediately performs fine-tuning: increasing the clamping force to 1.0N (the safety upper limit is tested to be 1.2N), and adjusting the impedance parameters to increase the system damping to 0.3N·s / m. After 0.1 seconds, the slippage probability drops to 0.3, and the chip remains stable; Release phase: During release, a gentler unloading gradient (0.2N / s) is used, with a 0.1-second pause in the last 0.2N phase. A tactile sensor is used to detect whether the chip has failed to detach due to electrostatic adhesion. If residual contact force (>0.02N) is detected, the system performs a micro-blowing-assisted release. Test results: Statistics were compiled from 1000 crawling cycles: Chip microcrack rate: 0.1% (control group with rigid gripper + single-point force feedback, microcrack rate 3.2%). Success rate of pickup: 99.7% (compared to 96.5% in the control group); Position repeatability accuracy: ±0.03mm (control group ±0.1mm) This embodiment demonstrates that the present invention can meet the stringent requirements of precision electronics manufacturing for the distribution of gripping force and contact pressure.

[0026] The above two sets of embodiments demonstrate the implementation details and beneficial effects of the method of the present invention from two typical fields: fruit and vegetable harvesting and precision manufacturing. Those skilled in the art should understand that, without departing from the spirit and scope of the present invention, appropriate adjustments can be made to the parameters, structure, and application scenarios, and these adjusted solutions still fall within the protection scope of the present invention.

[0027] While the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the invention. Any variations and modifications can be made by those skilled in the art without departing from the spirit and scope of the invention. Therefore, any modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention, without departing from the scope of the invention, fall within the protection scope defined by the claims of the present invention.

Claims

1. A flexible grasping safety control method based on a tactile array, comprising constructing a flexible tactile sensing layer, an initial grasping contact stage, a grasping force establishment stage, a grasping holding stage, and a release stage, characterized in that: The specific steps of the flexible gripping security control method are as follows: Step 1: Attach flexible tactile array sensors to the inner sides of the left and right finger surfaces of the gripper to acquire the pressure distribution matrix in real time. and ; Step 2: Based on the contour, centroid, contact area, and pressure center of the contact area, dynamically fine-tune the opening and closing posture and angle of the gripper to make the gripping contact surface fit the object surface, and use tactile-guided impedance control to make the gripper exhibit low stiffness characteristics when approaching the object. Step 3, Grip Force Establishment Stage: Using the pressure distribution matrix as feedback, calculate the total gripping force and pressure distribution uniformity index. A dynamic impedance control model is established, in which the damping parameters are... With stiffness parameters Based on pressure distribution uniformity And the contact area is adaptively adjusted in real time; Step 4, Grab and Hold Phase: Extract time-domain, spatial-domain, and frequency-domain features from the pressure distribution matrix. Input the extracted features into a slip prediction model based on a long short-term memory network, and output the slip probability. and according to Based on the relationship with the preset threshold, a tiered response strategy is executed; Step 5: Adopt a gradual force unloading strategy, and reduce the clamping force at a preset rate gradient according to the real-time pressure distribution.

2. The flexible grasping safety control method based on a tactile array according to claim 1, characterized in that: In step one, flexible tactile array sensors are attached to the inner sides of the left and right finger surfaces of the gripper. The sensors are piezoresistive flexible arrays, and each sensing unit independently outputs a pressure value, forming a real-time pressure distribution matrix. and .

3. The flexible grasping safety control method based on a tactile array according to claim 2, characterized in that: In step two, the controller receives tactile array data, and when the pressure value of any sensing unit exceeds a preset contact threshold... When contact is detected, the system extracts the contour, centroid, and contact area of ​​the contact region. And the pressure center, and based on the symmetry and distribution characteristics of the contact pattern, dynamically fine-tune the opening and closing posture and angle of the gripper to make the gripping contact surface fit the object surface.

4. The flexible grasping safety control method based on a tactile array according to claim 3, characterized in that: In step three, after entering the force establishment phase, the controller calculates the total gripping force using the pressure distribution matrix as feedback: ； in, For unit pressure, For unit area.

5. The flexible grasping safety control method based on a tactile array according to claim 4, characterized in that: In step three, the pressure distribution uniformity index is calculated: ； in Standard deviation This is the mean.

6. The flexible grasping safety control method based on a tactile array according to claim 5, characterized in that: In step three, a dynamic impedance control model is established: ； in For positional deviation, For the target inertia, and Uniformity of distribution with pressure The damping and stiffness matrices are adaptively adjusted according to changes in contact area.

7. The flexible grasping safety control method based on a tactile array according to claim 6, characterized in that: In step four, during the grasping and holding phase, the system continuously analyzes the dynamic changes of the tactile array and constructs a multi-dimensional stability evaluation index: Time-domain characteristics: time series fluctuations of pressure values ​​of each unit, and drift rate of pressure center. Spatial characteristics: changes in pressure gradient at the contact edge and changes in the topology of the contact region. Frequency domain features: Fourier transform is performed on the pressure center trajectory to extract the slip precursor frequency; A glide prediction model based on a long short-term memory network is adopted, with the input being past data. The pressure distribution sequence within the time window is output as the slip probability. ,when Exceeding the warning threshold At this time, the controller performs a fine-tuning action, increasing the clamping force to the preset safety upper limit. And adjust the impedance parameters to suppress vibration, when Exceeding the danger threshold When it is determined that slippage is about to occur, the system immediately activates an active anti-slip strategy, including instantly increasing the clamping force, adjusting the gripper posture to redistribute the contact surface, and issuing an audible and visual alarm.

8. A flexible gripping safety control gripper applicable to the flexible gripping safety control method based on a tactile array as described in any one of claims 1-7, comprising a gripper body (1), a flexible gripper (2), a control unit (3), and a flexible tactile sensor (4), characterized in that: The three sets of flexible grippers (2) are installed at the bottom of the gripper body (1).

9. A flexible gripping safety control gripper based on a tactile array according to claim 1, characterized in that: The control unit (3) is installed at the bottom of the gripper body (1), and the flexible tactile sensor (4) is installed inside the flexible gripper (2).

Images