Self-adaptive grabbing method based on tactile friction perception and adjustable friction coefficient

Abstract

The invention provides a self-adaptive grabbing method based on tactile friction perception and an adjustable friction coefficient. The self-adaptive grabbing method is based on a grabbing mechanical arm with the adjustable friction coefficient. The surface, used for making contact with a grabbed object, of each finger is a friction surface with an adjustable friction coefficient, and a strain sensing module is embedded in the lower portion of each friction surface to detect the strain amount. The motor output is gradually increased, and meanwhile dynamic optimization of the minimum safe clamping force is conducted; the clamping static friction force Ff1 is obtained according to the motor output at the current moment in combination with the adjustable friction coefficient; a dependent variable epsilon is obtained and converted into a strain friction force Ff2, and a real-time friction force Ff is calculated; whether the minimum safe clamping force is obtained or not is judged according to the real-time friction force; and subsequently, the minimum safe clamping force is adopted for grabbing. According to the method, the minimum safe clamping force capable of overcoming the gravity of the workpiece can be dynamically adjusted, so that the vulnerable workpiece is protected to the greatest extent while reliable grabbing is ensured.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of mechanical arm light clamping, more particularly, to a self-adaptive grasping method based on tactile friction perception and adjustable friction coefficient. BACKGROUND

[0002] Mechanical arms are widely used in industrial automation, logistics sorting, precision manufacturing, and medical surgery, etc. One of the core actions for them to perform tasks is to clamp objects. Standard clamping operation usually consists of two steps: first, the gripper is closed to grasp the object (clamping stage); then, the mechanical arm performs a lifting or moving action (lifting stage). However, this process implies a key problem: if the object is not effectively grasped in the clamping stage, the lifting action will inevitably result in "lifting empty" failure, causing operation interruption, efficiency reduction, and even possible equipment malfunction or workpiece damage.

[0003] To address this problem, existing technologies mainly adopt two strategies, but each has significant limitations:

[0004] (1) Pre-set fixed clamping force: This is the most common method, which sets a fixed clamping force value based on experience or rough estimation. Although this method is simple, it has obvious drawbacks. On the one hand, if the set force value is insufficient, it cannot generate enough friction to overcome the weight of the workpiece, resulting in a high risk of lifting empty. On the other hand, to avoid lifting empty, operators tend to set a higher clamping force, which can cause irreversible damage such as crushing, deformation, or fragmentation to flexible workpieces (such as rubber, sponge, biological tissues) or brittle workpieces (such as glass, ceramics, precision electronic components). This is contrary to the demand for precision operation and diversified material processing in modern industry;

[0005] (2) Indirect judgment based on vision or position sensors: Some systems attempt to use cameras or displacement sensors to detect whether the workpiece is successfully grasped after clamping (such as whether the position changes). However, this method has low reliability due to environmental factors such as changes in lighting conditions and obstruction interference, which can lead to misjudgment. Additionally, image processing or position detection often requires additional calculation time, increasing the operation cycle, and cannot directly perceive and control the clamping force, which cannot solve the damage risk caused by excessive clamping force.

[0006] Furthermore, existing solutions generally lack an effective mechanism to determine the "minimum necessary clamping force" in real-time and dynamically. Ideally, the system should be able to automatically apply a minimum force that is just enough to ensure reliable grasping and is the safest for the workpiece based on the specific properties of the workpiece (material, surface condition, shape) and the real-time clamping state (contact force distribution, friction). Pursuing the minimum clamping force is crucial for protecting fragile workpieces, improving system efficiency, and achieving intelligent adaptive operation.

[0007] Although advanced force sensors (such as wrist six-axis force sensors or fingertip pressure sensors) can provide partial force information, their high cost limits their wide application. More importantly, these sensors mainly sense normal pressure or overall torque, and it is difficult to directly, accurately and quickly obtain the real-time dynamic friction characteristics (especially the effective friction coefficient) of the contact interface between the gripper and the workpiece, which is exactly the core physical basis for judging the gripping stability (whether the maximum friction force is greater than the gravity).

[0008] Therefore, in the face of the risk of lifting, the difficulty of protecting fragile workpieces, and the demand for minimum adaptive gripping force, the current mechanical arm gripping technology urgently needs an innovative solution that can directly and real-time sense the friction force of the gripping interface, accurately predict the gripping success based on this, and dynamically adjust to the minimum safe gripping force before lifting. SUMMARY

[0009] To overcome the shortcomings and deficiencies in the prior art, the purpose of the present application is to provide an adaptive grasping method based on tactile friction perception and adjustable friction coefficient; the method can dynamically adjust to the minimum safe gripping force that can overcome the gravity of the workpiece, thereby ensuring reliable grasping while maximizing the protection of fragile workpieces.

[0010] To achieve the above purpose, the present application is implemented by the following technical scheme: an adaptive grasping method based on tactile friction perception and adjustable friction coefficient, a grasping robot arm based on adjustable friction coefficient; the fingers of the grasping robot arm are flexible fingers driven by a motor; the surface of the fingers used to contact the grasped object is a friction surface with adjustable friction coefficient μ, and a strain sensing module is embedded below the friction surface to detect the strain ε when the friction surface of the finger locally causes micro-strain due to the interaction between the grasped object and the friction surface of the finger during the gripping of the grasped object by the finger;

[0011] The adaptive grasping method based on tactile friction perception and adjustable friction coefficient is: the fingers of the robot arm move to the side of the grasped object; the motor is started and the motor output is gradually increased to drive the fingers to grip the grasped object;

[0012] During the gradual increase of the motor output, the dynamic optimization of the minimum safe gripping force is carried out: according to the motor output at the current time, the normal load P generated by the gripping is calculated, and then the adjustable friction coefficient μ is combined to obtain the static friction force F f1 ; and the strain ε detected by the strain sensing module is obtained, the compression strain value ε is converted into the strain friction force F f2 ; the static friction force F f1 is the main one, the strain friction force F f2 is the auxiliary one, and the real-time friction force F f is calculated; when the real-time friction force F fWhen F real-time > (1+delta)G, wherein G is the gravity of the object to be grabbed, and delta is a safety margin coefficient, the increase of the motor output is immediately stopped, the friction force at this moment is determined as the minimum safety clamping force, and the corresponding motor output is recorded; after the minimum safety clamping force is obtained, the corresponding motor output is kept to use the minimum safety clamping force to perform grabbing.

[0013] Preferably, the calculation of the real-time friction force F f refers to:

[0014] F f = alpha * F f1 + beta * F f2 .

[0015] Wherein, alpha and beta are positive coefficients, alpha + beta = 1; alpha > beta.

[0016] Preferably, an upper limit threshold of the motor output is provided; in the dynamic optimization process of the minimum safety clamping force, if the motor output is increased to the upper limit threshold of the motor output and still does not satisfy the real-time friction force F f > (1+delta)G, the increase of the motor output is stopped, and an error report is issued.

[0017] Preferably, the visual technology is used to collect the image of the object to be grabbed, and the type of the object to be grabbed is identified, and then the coefficients alpha and beta corresponding to the type of the object to be grabbed and the upper limit threshold of the motor output are obtained in the database.

[0018] Preferably, the friction surface of the finger refers to a friction surface made of an elastic material; the micro-texture profile curve of the friction surface is composed of two cosine waves with different amplitudes and same wavelength, so that the change of the total contact area between the friction surface and the object to be grabbed at different motor outputs is realized, and then the adjustable friction coefficient is realized.

[0019] Preferably, the micro-texture profile curve of the friction surface is represented as z(x):

[0020] z(x) = m(x) * A1 cos(w0x)

[0021]

[0022] Wherein, A1 and A2 are amplitudes of the two cosine waves respectively; w0 represents an angular frequency, m(x) is a modulation function; k1 = A2 / A1, and k1 > 1; N = n1 / n2; n1 and n2 are the number of wave peaks of the two cosine waves in one period respectively; x represents a spatial coordinate position.

[0023] Compared with the prior art, the present application has the following advantages and beneficial effects:

[0024] 1、The present application eliminates the blindness of experience preset fixed clamping force: by comparing the size of the sum of the real-time friction force F f With the gravity G of the grasped object and the safety margin, providing a deterministic success judgment before lifting the grasped object, completely avoiding the risk of lifting empty;

[0025] 2、The present application realizes the minimization and self-adaptation of clamping force: innovatively dynamically finds and maintains the minimum safe clamping force, perfectly solves the core pain point of flexible / brittle grasped objects under traditional large preset force clamping;

[0026] 3、The present application improves the accuracy and efficiency of perception: the friction surface with adjustable friction coefficient combined with strain sensing scheme, low cost, easy integration, can directly, quickly (millisecond level), reliably perceive the key friction physical quantity of the contact interface, not affected by environmental interference such as light and shielding, significantly better than indirect methods such as vision;

[0027] 4、The present application enhances the universality and intelligence of the system: the scheme is based on physical principles, without complex preset programming for each grasped object, can adaptively handle objects of different materials (such as smooth, rough), shapes and surface characteristics, providing key technical support for the wide application of robotic arms in sensitive scenarios such as precision electronic assembly, medical operation and food processing;

[0028] 5、The present application, when the motor outputs different radial forces on the grasped object, the total contact area between the finger friction surface and the grasped object also changes, and the friction coefficient is adjusted in coordination, which can meet the two technical requirements of anti-skid and anti-damage at the same time. BRIEF DESCRIPTION OF DRAWINGS

[0029] Figure 1 is the flowchart of the adaptive grasping method based on tactile friction perception and adjustable friction coefficient of the present application;

[0030] Figure 2 is a structural schematic diagram of the grasping robot arm based on adjustable friction coefficient of the present application;

[0031] Figure 3 is a finger structure schematic diagram of the grasping robot arm based on adjustable friction coefficient of the present application;

[0032] Figure 4 is a cross-sectional view of the finger of the grasping robot arm based on adjustable friction coefficient of the present application. DETAILED DESCRIPTION

[0033] The present application will be described in further detail below in conjunction with the drawings and specific embodiments.

[0034] EMBODIMENT

[0035] The embodiment is a self-adaptive grasping method based on tactile friction perception and adjustable friction coefficient, and the flowchart is as shown in Figure 1 The grasping robot with adjustable friction coefficient is as shown in Figure 2 and Figure 3 The finger 1 is a flexible finger and is driven by a motor.

[0036] The surface of the finger 1 for contacting the grasped object 5 is a friction surface with adjustable friction coefficient μ. The friction surface of the finger refers to a friction surface 2 made of elastic material. The micro-texture profile curve of the friction surface 2 is composed of two cosine waves with different amplitudes and the same wavelength, as shown in Figure 4 The friction coefficient can be adjusted by changing the total contact area between the friction surface and the grasped object at different motor outputs.

[0037] The micro-texture profile curve of the friction surface is represented by z(x):

[0038] z(x) = m(x) · A1 cos(w0x)

[0039]

[0040] wherein A1 and A2 are the amplitudes of the two cosine waves respectively; w0 represents the angular frequency, m(x) is a modulation function; k1 = A2 / A1, and k1>1; N = n1 / n2; n1 and n2 are the number of wave peaks of the two cosine waves in one period respectively; and x represents the spatial coordinate position.

[0041] The strain sensing module 3 is embedded below the friction surface 2, as shown in Figure 4 When the finger clamps the grasped object, the grasped object interacts with the friction surface of the finger, causing local micro-strain of the friction surface of the finger, and the strain ε is detected. The strain sensing module 3 is arranged along the thickness direction of the finger, and there is more than one strain sensing module 3. In this embodiment, there is one strain sensing module in the middle; in actual application, there can be multiple strain sensing modules, and the more the number of strain sensing modules, the higher the precision.

[0042] The self-adaptive grasping method based on tactile friction perception and adjustable friction coefficient is as follows:

[0043] The finger of the robot moves to the side of the grasped object; the motor is started, and the motor output is gradually increased to drive the finger to clamp the grasped object; the motor output refers to the motor output torque or output current;

[0044] In the process of gradually increasing motor output, dynamic optimization of minimum safe clamping force is carried out: according to the motor output at the current time, the normal load P generated by clamping is calculated according to the mechanical arm transmission mechanism parameters; the normal load P can be calculated by using the existing algorithm; then the adjustable friction coefficient μ is combined to obtain the clamping static friction force F f1 ; the relationship curve of the clamping static friction force F f1 , the normal load P and the adjustable friction coefficient μ can be obtained through experimental data,

[0045] At the same time, the strain amount ε detected by the strain sensing module is obtained, and the compression strain value ε is converted into the strain friction force F f2 based on the pre-stored calibration parameters;

[0046] The clamping static friction force F f1 is mainly used, and the strain friction force F f2 is used as a supplement to calculate the real-time friction force F f :

[0047] F f =α×F f1 +β×F f2 ;

[0048] Wherein, α and β are positive coefficients, α+β=1; α>β; different α and β values for different grasping objects;

[0049] When the real-time friction force F f >(1+δ)G, where G is the weight of the grasping object, and δ is a safety margin coefficient, the motor output is immediately stopped, the friction force at this time is determined as the minimum safe clamping force, and the corresponding motor output is recorded; after obtaining the minimum safe clamping force, the corresponding motor output is maintained to use the minimum safe clamping force for grasping.

[0050] In the dynamic optimization process of the minimum safe clamping force, after the motor output is increased, the strain amount ε of the strain sensor can be collected after waiting for 10ms, and then the clamping static friction force F f1 , the strain friction force F f2 , and the real-time friction force F f are calculated; and it is judged whether F f >(1+δ)G.

[0051] The preferred scheme is: there is a motor output upper threshold; in the dynamic optimization process of the minimum safe clamping force, if the motor output is increased to the motor output upper threshold and still does not satisfy the real-time friction force F f >(1+δ)G, then stop increasing the motor output and issue an error report.

[0052] The visual technology is used to collect the image of the grasped object, and the type of the grasped object is recognized, and then the coefficients α and β corresponding to the type of the grasped object and the upper limit threshold of the motor output are obtained in the database. Different types of grasped objects can be intelligently matched. A calibration interface can also be provided to manually adjust the coefficients α and β and the upper limit threshold of the motor output to match different types of grasped objects.

[0053] The application can be deployed in two-finger / three-finger grippers and electric / pneumatic driving architectures.

[0054] The above embodiments are the preferred embodiments of the application, but the embodiments of the application are not limited to the above embodiments, and any changes, modifications, substitutions, combinations, simplifications made without departing from the spirit and principles of the application should be equivalent replacement methods and are included in the protection scope of the application.

Claims

1. A method for adaptive grasping based on tactile friction perception and adjustable friction coefficient, characterized in that: The grasping mechanical arm based on adjustable friction coefficient; the fingers of the grasping mechanical arm are flexible fingers driven by a motor; the surface of the fingers for contacting the grasped object is a friction surface with adjustable friction coefficient μ, and a strain sensing module is embedded below the friction surface, so as to detect the strain ε when the local micro-strain of the friction surface of the finger is caused by the interaction between the grasped object and the friction surface of the finger when the finger clamps the grasped object; The adaptive grasping method based on tactile friction perception and adjustable friction coefficient is as follows: the fingers of the mechanical arm are moved to the side of the grasped object; the motor is started, and the motor output is gradually increased to drive the fingers to clamp the grasped object. In the process of gradually increasing motor output, dynamic optimization of minimum safe clamping force is carried out: according to the motor output at the current time, the normal load P generated by clamping is calculated, and then the adjustable friction coefficient μ is combined to obtain the clamping static friction force F f1 ; and the strain amount ε detected by the strain sensing module is obtained, and the compression strain value ε is converted into the strain friction force F f2 ; the clamping static friction force F f1 is mainly used, and the strain friction force F f2 is used as a supplement, and the real-time friction force F f is calculated; when the real-time friction force F f >(1+δ)G, wherein G is the gravity of the grasped object, and δ is a safety margin coefficient, the motor output is immediately stopped from increasing, it is determined that the friction force at this time is the minimum safe clamping force, and the corresponding motor output is recorded; after obtaining the minimum safe clamping force, the corresponding motor output is maintained, so that the minimum safe clamping force is used for grasping.

2. The adaptive grasping method based on tactile friction perception and adjustable friction coefficient according to claim 1, characterized in that: The calculation of the real-time friction force F f refers to: F f = a x F f1 + b x F f2 ; Wherein, α and β are positive coefficients, and α+β=1; α>β.

3. The adaptive grasping method based on tactile friction perception and adjustable friction coefficient according to claim 2, characterized in that: The motor output upper limit threshold is provided; in the dynamic optimization process of the minimum safety clamping force, if the motor output is increased to the motor output upper limit threshold and still does not satisfy the real-time friction force F f >(1+δ)G, the motor output is stopped from being increased, and an error report is issued.

4. The adaptive grasping method based on tactile friction perception and adjustable friction coefficient according to claim 3, characterized in that: The image of the grasped object is collected by using visual technology, and the type of the grasped object is identified, and then the coefficients α and β corresponding to the type of the grasped object are obtained from the database, and the upper limit threshold of the motor output is obtained.

5. The adaptive grasping method based on tactile friction perception and adjustable friction coefficient according to claim 1, characterized in that: The friction surface of the finger refers to a friction surface made of an elastic material; the micro-texture profile curve of the friction surface is composed of two cosine waves with different amplitudes and same wavelength, so as to realize the adjustment of the friction coefficient by changing the total contact area between the friction surface and the grasped object at different motor outputs.

6. The adaptive grasping method based on tactile friction perception and adjustable friction coefficient according to claim 5, characterized in that: The micro-texture profile curve of the friction surface is represented as z(x): z(x)=m(x)·A1cos(w0x) wherein A1 and A2 are the amplitudes of the two cosine waves, respectively; w0 represents the angular frequency, m(x) is a modulation function; k1 = A2 / A1, and k1 > 1; N = n1 / n2; n1 and n2 are the number of peaks of the two cosine waves in one period, respectively; and x represents the spatial coordinate position.

Images