A control method of a three-degree-of-freedom mechanical air gripper

Abstract

The application provides a control method of a three-degree-of-freedom mechanical air gripper, and relates to the technical field of industrial automation control. The method comprises the following steps: acquiring current position data and dynamic environment state data of the three-degree-of-freedom mechanical air gripper, identifying a spatial direction deviation of a grasping posture and obtaining a deviation evaluation result; when the deviation exceeds a first preset threshold, a coordinate transformation method is used to calculate a coordination vector and to adaptively adjust the update frequency thereof; based on the corrected coordination vector, when the force is not matched due to the change of the clamping contact area, the air pressure adjustment precision step is dynamically adjusted to generate a distribution scheme; the position and the force parameter are fused and calculated, and an iterative check is performed to obtain a comprehensive matching coefficient; and the final control instruction set is generated by using the coefficient. The application realizes the accurate matching of the position adjustment and the clamping force in the dynamic environment, and improves the adaptability and stability of the grasping.

Description

Technical Field

[0001] This application belongs to the field of industrial automation control technology, specifically a control method for a three-degree-of-freedom mechanical pneumatic gripper. Background Technology

[0002] In the field of modern industrial automation, mechanical grippers, as key components of robot end effectors, are widely used in material handling, assembly, and packaging. For multi-degree-of-freedom mechanical grippers, their control precision directly determines production efficiency and operational quality.

[0003] Existing mechanical gripper control methods typically focus on single-position control or simple force control modes. When faced with complex and dynamic working environments, they often lack real-time response capabilities to environmental disturbances and changes in object characteristics. Especially when handling irregularly shaped or materially varied objects, existing control schemes struggle to establish a precise match between position adjustment and gripping force. For example, when changes in object movement speed alter the gripping contact area, traditional control methods cannot adjust air pressure accuracy in a timely manner, easily leading to object slippage or damage due to improper force distribution. Furthermore, existing technologies often neglect the impact of environmental dynamics on vector update frequency during attitude correction, resulting in significant deviations in the corrected gripping posture.

[0004] Therefore, how to achieve precise matching between the position adjustment and clamping force of a three-degree-of-freedom mechanical pneumatic gripper in a dynamic environment has become a technical problem that urgently needs to be solved. Summary of the Invention

[0005] To address the above problems, this application provides a control method for a three-degree-of-freedom mechanical pneumatic gripper, which solves the problem of difficulty in accurately matching position adjustment and gripping force in dynamic environments in existing technologies.

[0006] To achieve the above objectives, the technical solution adopted in this application is as follows: A control method for a three-degree-of-freedom mechanical pneumatic gripper includes the following steps: acquiring the current position data and dynamic environmental state data of the three-degree-of-freedom mechanical pneumatic gripper, identifying the spatial orientation deviation of the gripping posture based on the current position data, and obtaining an initial gripping posture deviation evaluation result by combining the dynamic environmental state data; when the deviation evaluation result indicates that the orientation deviation exceeds a first preset threshold, calculating a coordination vector for posture correction using a coordinate transformation method, and adaptively adjusting the update frequency of the coordination vector according to the dynamic characteristics of the dynamic environmental state data to obtain a corrected coordination vector; based on the corrected coordination vector... When the gripping contact area changes during the gripping process, causing a mismatch between the actual gripping force and the theoretical requirement, the gripping force is redistributed by dynamically adjusting the precision step size of the air pressure regulation, generating a force distribution scheme adapted to the current gripping contact area. Based on the force distribution scheme, the position adjustment parameters and gripping force parameters of the three-degree-of-freedom mechanical pneumatic gripper are fused and calculated, and iteratively verified according to a preset feedback cycle to obtain a comprehensive matching coefficient. Using the comprehensive matching coefficient, feedback processing is performed according to the update frequency of the coordination vector to generate the final control instruction set for controlling the three-degree-of-freedom mechanical pneumatic gripper to perform gripping actions.

[0007] This application provides a control method for a three-degree-of-freedom (DOF) mechanical pneumatic gripper. By acquiring real-time position and environmental state data, identifying spatial orientation deviations, and evaluating them in conjunction with environmental disturbances, it can accurately perceive deviations in the initial gripping posture. In the posture correction stage, coordinate transformation is used to calculate the coordination vector, and the update frequency is adaptively adjusted according to the dynamic characteristics of the environment, effectively avoiding correction oscillations caused by environmental disturbances and improving the stability of posture adjustment. To address changes in the gripping contact area caused by object movement in dynamic environments, the gripping force is redistributed by dynamically adjusting the air pressure regulation precision step size, solving the problem of mismatch between force and gripping contact area and ensuring precise application of gripping force. Furthermore, by fusing position and force parameters and performing iterative verification, coordinated optimization of position adjustment and force control is achieved. The final generated control command set can precisely drive the three-degree-of-freedom mechanical pneumatic gripper to perform actions. The overall solution achieves high-precision and high-stability control of a three-degree-of-freedom mechanical pneumatic gripper in complex dynamic environments. Attached Figure Description

[0008] Figure 1 A flowchart illustrating a control method for a three-degree-of-freedom mechanical pneumatic gripper provided in this application embodiment; Figure 2 A flowchart illustrating the adaptive adjustment of the coordination vector update frequency based on the dynamic characteristics of dynamic environmental state data, provided in an embodiment of this application. Figure 3A flowchart illustrating the redistribution of clamping force by dynamically adjusting the precision step size of air pressure regulation, as provided in this embodiment of the application; Figure 4 A flowchart illustrating the generation of the final control instruction set for controlling a three-degree-of-freedom mechanical pneumatic gripper to perform gripping actions, provided for embodiments of this application; Figure 5 A flowchart of secondary scanning and instruction update based on clamping force distribution uniformity monitoring provided in this application embodiment; Figure 6 This is a schematic diagram of the backtracking correction control process of a three-degree-of-freedom mechanical pneumatic gripper provided in an embodiment of this application. Detailed Implementation

[0009] To enable those skilled in the art to better understand the technical solution, the present application will be described in detail below with reference to the embodiments. The description in this section is only exemplary and explanatory, and should not be used to limit the scope of protection of the present application in any way.

[0010] See Figure 1 This application provides a flowchart of a control method for a three-degree-of-freedom mechanical pneumatic gripper, as shown in the embodiment. Figure 1 As shown, the method includes the following steps: Step S110: Obtain the current position data and dynamic environmental status data of the three-degree-of-freedom mechanical pneumatic gripper, identify the spatial orientation deviation of the gripping posture based on the current position data, and obtain the deviation evaluation result of the initial gripping posture by combining the dynamic environmental status data.

[0011] In this embodiment, a sensor array can be deployed on a three-degree-of-freedom mechanical pneumatic gripper. This array includes, but is not limited to, photoelectric encoders and vision sensors for acquiring position information, and vibration sensors and force sensors for sensing environmental dynamics. Current position data can refer to the coordinate and attitude angle information of the gripper's end effector in three-dimensional space, collected by the aforementioned sensor array. Dynamic environmental state data can refer to real-time changing physical quantities within the gripper's working area, including but not limited to environmental vibration frequency, external disturbance amplitude, airflow disturbance, and the moving speed of the gripped object.

[0012] In practical applications, the process of acquiring current location data and dynamic environmental status data not only involves reading raw signals from sensors but also implicitly includes data cleaning and preprocessing steps. For example, a moving average filtering algorithm is used to denoise the raw location signal and remove outliers; environmental status data is standardized to ensure it is on the same scale for subsequent calculations.

[0013] Identifying the spatial orientation deviation of the grasping posture involves mapping the current position data to a spatial posture vector, obtaining the current grasping posture vector, and comparing the current grasping posture vector with the preset target grasping posture vector. The degree of deviation is quantified by calculating the angle or Euclidean distance between the vectors. For example, assuming the target posture vector is (1.0, 0.5, 0.3) and the current posture vector is (0.9, 0.6, 0.4), the cosine of the angle between the two vectors is calculated, and the orientation deviation is determined based on this cosine. The deviation angle is found to be 5.2 degrees, indicating a slight shift in the gripper posture. The determined orientation deviation can be used as part of the deviation assessment result. This orientation deviation serves to intuitively quantify the degree of inaccuracy in the grasping posture, directly reflecting the angular magnitude that a three-degree-of-freedom mechanical pneumatic gripper needs to correct. Using this orientation deviation as part of the deviation assessment result means that this angle value will be input into subsequent deviation assessment logic, forming the final deviation assessment result along with other environmental interference factors (such as vibration amplitude, noise weight, etc.).

[0014] In an exemplary embodiment of this application, the deviation evaluation result of the initial grasping posture can refer to a quantitative score generated by a weighted evaluation model, which integrates the spatial orientation deviation and the interference weights in the dynamic environmental state data. Specifically, if the vector deviation weight is set to 0.6 and the environmental interference (such as the amplitude of jitter caused by vibration) weight is set to 0.4, the calculated comprehensive deviation score is 3.5 (out of 10). This result is used to characterize the reliability and urgency of correction of the current grasping posture. By simultaneously acquiring and fusing position and environmental data, this step aims to provide an accurate input benchmark for subsequent posture correction, effectively avoiding misjudgments caused by ignoring the dynamic characteristics of the environment.

[0015] It should be noted that in practical applications, in addition to acquiring current location data and dynamic environmental status data, it is also necessary to obtain the surface hardness information of the target object. Before the gripping process begins, the system first performs a hardness scan on the target object. Specifically, a visual inspection module (such as an industrial camera combined with structured light or stereo vision) acquires the object's three-dimensional contour, surface texture, and material reflectivity to preliminarily determine the object type and hardness range. Simultaneously, a miniature hardness tester or a dot-matrix pressure sensor mounted on the gripper can be optionally used to directly collect hardness values ​​(such as Shore hardness) at multiple sampling points on the object's surface in either slight contact or non-contact measurement modes. The purpose of the first scan is to obtain the baseline hardness distribution of the object's surface, which is used to estimate the initial gripping force threshold required for gripping, set the baseline value for the force distribution, and provide a pre-calibration basis for subsequent force distribution schemes and force requirement matching. Furthermore, the first scan data is also used to assist in the evaluation of deviations in the initial gripping posture; for example, when the hardness is extremely low (easily deformable), the allowable directional deviation threshold can be appropriately relaxed.

[0016] Step S130: When the deviation evaluation result indicates that the directional deviation exceeds the first preset threshold, the coordinate transformation method is used to calculate the coordination vector for attitude correction, and the update frequency of the coordination vector is adaptively adjusted according to the dynamic characteristics of the dynamic environment state data to obtain the corrected coordination vector.

[0017] Specifically, this step is a dynamic correction process for attitude deviations. The first preset threshold is an angle or distance range pre-set based on the mechanical structural accuracy of the three-degree-of-freedom mechanical gripper and the safety redundancy of the gripping task. An orientation deviation exceeding the first preset threshold can mean that the calculated comprehensive deviation score or the included angle exceeds the system's set safety tolerance; for example, the first preset threshold is 10 degrees, while the measured deviation is 15 degrees. When the deviation exceeds this range, the system triggers a correction mechanism. The coordination vector is a multi-dimensional vector used to guide the coordinated movement of the gripper's joints to eliminate attitude deviations. Its calculation relies on coordinate transformation methods, such as using rotation matrices or quaternions to map the target attitude and the current attitude onto the gripper's three-degree-of-freedom coordinate system to solve for the difference. For example, with an initial vector direction (assumed to be x=1.0, y=0.0, z=0.0), the vector values ​​are calculated at 10 degrees, 15 degrees, and 20 degrees with a step size of 5 degrees, resulting in new vector sets such as (x=0.9848, y=0.1736, z=0.0), etc. The proximity of each vector to the target direction is evaluated using the Euclidean distance formula to determine the optimal correction direction.

[0018] One of the core innovations of this application's embodiments lies in the adaptive adjustment of the update frequency of the coordination vector. If the dynamic environmental status data shows severe environmental disturbances (such as high-frequency vibrations), the update frequency is reduced to avoid jitter in the control system; if the environment is relatively stable, the update frequency is increased to pursue a faster response speed. This adaptive mechanism ensures that the corrected coordination vector can both quickly respond to deviations and resist environmental disturbances, achieving a balance between rigidity and flexibility. The corrected coordination vector can refer to the vector data finally determined after frequency adaptation processing, used to adjust the gripper's movements.

[0019] In practical applications, adaptively adjusting the update frequency of the coordination vector based on the dynamic characteristics of dynamic environmental state data can include: dynamically reducing or increasing the calculation and transmission frequency of the coordination vector according to the vibration frequency or disturbance amplitude in the dynamic environment.

[0020] Vibration frequency refers to the number of periodic oscillations caused by external interference sources in a dynamic environment, usually measured in Hertz (Hz). Interference amplitude refers to the maximum deviation of the mechanical pneumatic gripper's position or attitude caused by environmental interference, usually measured in millimeters (mm) or degrees (°). Vibration frequency and interference amplitude are obtained in real-time through spectral analysis using accelerometers, laser displacement sensors, or vision sensors installed on the three-degree-of-freedom mechanical pneumatic gripper or in its surrounding environment. This step dynamically matches the calculation rhythm and transmission rate of the coordination vector according to the severity of environmental disturbances, avoiding excessive consumption of computing resources in calm environments or failure to grasp due to response lag in severe disturbances. Vibration frequency can be set with different frequency adjustment strategies based on preset frequency band thresholds; for example, when the detected vibration frequency is in the low-frequency band, the environment is considered relatively stable. Interference amplitude can be updated to the real-time monitored peak data; when the amplitude exceeds the safety tolerance, a high-frequency response mechanism is triggered. Vibration frequency and disturbance amplitude are used together to characterize the activity level of the dynamic environment: high-frequency and large-amplitude disturbances mean that the system state changes rapidly, requiring an increase in the update frequency of the coordination vector to enhance the tracking sensitivity of the system; while low-frequency and small-amplitude perturbations allow for a reduction in the update frequency, thereby suppressing noise amplification and saving processor computing power.

[0021] For example, when a three-degree-of-freedom mechanical pneumatic gripper is operating on an assembly line, if the accelerometer detects an ambient vibration frequency of 2.0 Hz and an interference amplitude of 0.5 mm, the system determines that there is a moderate level of disturbance in the current environment. In this case, the system dynamically increases the calculation and transmission frequency of the coordination vector from the default 10 Hz to 15 Hz to ensure timely attitude correction. Conversely, if the detected vibration frequency is only 0.5 Hz and the interference amplitude is less than 0.1 mm, the system dynamically reduces the update frequency to 5 Hz to reduce unnecessary command refreshes. This dynamic adjustment method based on physical quantities effectively balances control accuracy and system energy consumption, avoiding the rigidity problem of fixed-frequency strategies when facing complex and changing environments.

[0022] This step aims to address the technical problem that a fixed update frequency cannot adapt to dynamic environmental changes. Based on the aforementioned vibration frequency and interference amplitude data, the processor is configured to execute a frequency adjustment algorithm. Specifically, by comparing the ratio of the real-time monitored value to a first preset threshold, a target update frequency is mapped linearly or non-linearly. This yields a coordination vector calculation and transmission rhythm adapted to the current environmental state. This result provides a timeliness guarantee for subsequent coordinate transformation calculations, thereby significantly improving the attitude stability and response speed of the three-degree-of-freedom mechanical pneumatic gripper under dynamic interference, and effectively avoiding the accumulation of gripping deviations due to untimely updates.

[0023] Further, see Figure 2This is a flowchart illustrating another method for adaptively adjusting the update frequency of the coordination vector based on the dynamic characteristics of dynamic environmental state data, provided in an embodiment of this application. The method further includes: Step S210: Obtain dynamic change data of multi-directional quantities, determine the degree of matching between dynamic change data and coordination vector, and obtain intermediate results of direction correction.

[0024] First, the system acquires dynamic change data of multiple dimensions in real time. This could refer to time-series data such as displacement increments, velocity vectors, or angular velocity rates of change collected in real time along the X, Y, and Z axes during the movement of a three-degree-of-freedom mechanical pneumatic gripper. This data originates from encoders or inertial measurement units installed at each joint of the gripper and characterizes the degree to which the gripper's actual trajectory deviates from the ideal path in space. The system calculates the matching degree between these real-time vectors and the current coordination vector. The matching degree can be quantified by the vector dot product or projected length, thus obtaining a preliminary intermediate result for direction correction. This result reflects the basic correction direction without considering drastic environmental fluctuations. By acquiring dynamic change data of multiple dimensions and performing matching analysis with the coordination vector, the system can utilize the cross-validation mechanism of multi-source sensor data to improve the confidence of the correction basis, avoid misjudgments caused by noise from a single sensor, and provide reliable basic data support for subsequent accurate corrections.

[0025] Step S230: Combine the fluctuation data of environmental disturbances, process the intermediate results, and calculate the final coordination vector correction value.

[0026] In practical applications, the intermediate results mentioned above may often contain abrupt changes caused by environmental noise. Therefore, the system further incorporates the fluctuation data of environmental interference, for example, by using a moving average algorithm to smooth the interference data of the past N periods (e.g., N=5), eliminating occasional impulse interference, and calculating the final coordination vector correction value.

[0027] The environmental interference fluctuation data can refer to the statistical characteristics of random variables that cause vibration or positional jitter in a three-degree-of-freedom mechanical pneumatic gripper in a dynamic environment, including but not limited to the temporal variance, peak-to-peak value, spectral energy distribution, or instantaneous amplitude of external impact force of the vibration signal. This data is typically obtained in real-time through accelerometers or force sensors. Processing the intermediate results specifically involves using weighted filtering algorithms, robust estimation methods, or Kalman filtering techniques to incorporate the environmental interference fluctuation data as a noise model or weighting factor into the intermediate results of the direction correction. When the environmental interference fluctuation is large, the response gain of the intermediate results is reduced to suppress oscillations; when the environment is relatively stable, the response gain is increased to accelerate convergence. For example, if a periodic vibration with a frequency of 10Hz and an amplitude of 0.2mm is detected in the current environment, the system can identify this fluctuation characteristic and attenuate the contribution of the corresponding frequency component in the intermediate results when calculating the final correction value, thereby eliminating false motion trends. The final coordination vector correction value is a high-precision control parameter output after noise reduction processing and uncertainty modeling. It not only includes attitude deviation correction information but also incorporates a compensation mechanism for environmental uncertainties. By leveraging environmental disturbance fluctuation data for in-depth processing of intermediate results, the generated correction value possesses strong noise resistance, effectively solving the problem of correction failure caused by sudden disturbances in complex dynamic environments and ensuring the robustness of the correction scheme.

[0028] Step S250: Determine whether the correction scheme generated based on the coordination vector correction value meets the preset stability condition. If it does, determine the update frequency according to the correction scheme.

[0029] In practical applications, a correction scheme can refer to a complete attitude adjustment strategy based on the final coordination vector correction value, including specific execution step size, direction commands, and the expected state trajectory. The preset stability condition is a safety constraint criterion ensuring that the three-degree-of-freedom mechanical pneumatic gripper does not experience divergent oscillations or instability during correction actions. Specific forms may include implicit control theory constraints such as a negative definite Lyapunov function criterion, a phase margin greater than a preset angle (e.g., 45°), a sufficient gain margin, or the corrected system poles being located on the left half of the complex plane. The judgment process involves substituting the correction scheme into the aforementioned stability criteria for verification. If all indicators are within the safe range, the condition is deemed met; otherwise, it is deemed not met. If the preset stability condition is met, the update frequency is determined based on the correction scheme. Specifically, this is manifested in dynamically setting the calculation and transmission rhythm of the coordination vector according to the convergence speed and complexity of the correction scheme. For example, when the correction scheme is determined to be highly stable and converges rapidly, the system can appropriately increase the update frequency to achieve a fast response; while when the correction scheme is at the stability boundary, the update frequency is automatically reduced to prevent overshoot. This mechanism acts as a safety gate for the system, preventing aggressive correction strategies from causing system oscillations. By setting strict stability conditions as a pre-verification step, correction schemes are only adopted and update frequencies are determined after confirming safety, fundamentally avoiding the risk of control failure due to blind adjustments and significantly improving the asymptotic stability of the system under extreme disturbances.

[0030] This application generates intermediate direction correction results with high confidence by acquiring dynamic change data of multi-directional quantities and performing matching degree analysis with the coordination vector. Based on this, it further incorporates environmental disturbance fluctuation data to perform weighted filtering and robust estimation on the intermediate results, outputting a final coordination vector correction value with noise resistance. Furthermore, a preset stability condition is introduced as a safety decision mechanism, determining the update frequency only when the correction scheme meets the stability requirements. Through this three-level collaborative mechanism of multi-dimensional matching, fluctuation compensation, and stability decision, not only is the risk of misjudgment that may arise from relying solely on disturbance indicators overcome, but also the optimal balance between correction accuracy and system stability is achieved. This progressive processing logic enables the three-degree-of-freedom mechanical pneumatic gripper to adaptively adjust its control rhythm when facing complex and ever-changing dynamic environments, ensuring the accuracy of attitude correction while effectively suppressing potential oscillation risks, thereby significantly reducing the correction failure rate and ensuring the continuity and reliability of the grasping task.

[0031] Step S150: Based on the corrected coordination vector, when the actual clamping force does not match the theoretical requirement due to changes in the clamping contact area during the gripping process, the clamping force is redistributed by dynamically adjusting the precision step size of the air pressure regulation, thereby generating a force distribution scheme that adapts to the current clamping contact area.

[0032] In dynamic grasping scenarios, the movement or posture change of an object can alter the effective gripping contact area between the gripper and the object, leading to variations in the actual gripping force under the same air pressure. This step aims to address this mechanical mismatch problem. The change in gripping contact area refers to the variation in the effective contact area between the gripper and the object's surface over time when grasping dynamic objects or objects with uneven surfaces. The mismatch between actual gripping force and theoretical requirements can mean a significant deviation between the ideal gripping force calculated based on the current gripping contact area and the actual gripping force fed back by the sensor. For example, if the initial gripping contact area is 50 square centimeters and the theoretical force is 10 Newtons, when the area is reduced to 48 square centimeters, the theoretical force should be adjusted to 9.6 Newtons, but the actual feedback is only 9.0 Newtons, a deviation of 0.6 Newtons. Redistributing the gripping force by dynamically adjusting the precision step size of air pressure regulation can be achieved by using a PID (Proportional-Integral-Derivative) control algorithm to calculate the incremental step size ΔP of the air pressure output based on the force error value. For example, setting a proportional coefficient K... p =0.5, integral coefficient K i =0.1, differential coefficient K d =0.05, substituting the error e=0.6 Newtons, the air pressure increment ΔP=0.3 Pascals is calculated, and the air pressure is finely adjusted from the initial 5 Pascals to 5.3 Pascals. Generating a force distribution scheme adapted to the current clamping contact area can refer to replanning the force distribution of each joint of the gripper based on the adjusted air pressure parameters, ensuring that the pressure per unit area remains within a safe and effective range, such as adjusting the force per unit area from 0.2 Newtons / cm² to 0.21 Newtons / cm². This step, through dynamic adjustment of air pressure accuracy, solves the problem of clamping force mismatch caused by changes in the clamping contact area, effectively preventing objects from slipping or being damaged due to excessive pressure.

[0033] It should be noted that in the aforementioned mismatch between actual and theoretical clamping forces, the theoretical requirement refers to the ideal clamping force needed to ensure stable gripping. This theoretical requirement is obtained by first acquiring the surface hardness information of the object through visual inspection or hardness scanning, and then combining this information with the current clamping contact area, the estimated mass of the object, and dynamic environmental factors (such as vibration and acceleration). A preset gripping force model is used to calculate the minimum clamping force required per unit area, which is then multiplied by the actual clamping contact area to obtain the theoretically required clamping force value. This theoretical requirement can be dynamically adjusted according to the safety factor of the gripping task and serves as a benchmark reference in subsequent force distribution stages.

[0034] In an exemplary embodiment of this application, the specific implementation process of redistributing the clamping force by dynamically adjusting the precision step size of the air pressure regulation in step S150 is further described in detail. Combined with... Figure 3As shown, this application embodiment addresses the problem of changing clamping contact area caused by object movement in dynamic environments by constructing a precise force allocation mechanism based on speed perception and error feedback. Specifically, it includes: Step S310: Obtain the object's moving speed and calculate the change in the clamping contact area caused by the object's moving speed.

[0035] Specifically, in dynamic grasping scenarios, such as objects on a conveyor belt, the moving speed is a key variable affecting gripping stability. High-speed movement of the object causes a momentary change in the effective gripping contact area between the gripper and the object's surface at the instant the gripper closes. If a constant air pressure output is maintained, the actual gripping force will deviate from the theoretical requirement, potentially causing the object to slip or deform. This embodiment first obtains the relative moving speed of the object through visual detection or a contact sensor installed inside the gripper. Subsequently, the system calls a preset algorithm to calculate the change in gripping contact area caused by this speed.

[0036] The calculation of the change in clamping contact area caused by the object's moving speed includes: using a pre-established area change model, the change in clamping contact area is calculated based on the proportional relationship between the object's moving speed and the clamping contact area.

[0037] In a specific implementation, the area change model is one of the core computational logics of this application. The area change model can refer to a mathematical mapping relationship pre-constructed through experimental data or theoretical derivation, used to quantify the dynamic correlation between the object's moving speed and the change in the gripping contact area. This model specifically includes input variables, a scaling factor, and an output variable. The input variable is the real-time collected object moving speed; the scaling factor is a constant or function characterizing the coupling relationship between the object's material properties and the gripper structure; and the output variable is the change in the gripping contact area per unit time. This area change model is obtained by offline training using historical grasping data or by fitting physical formulas based on Hertzian contact theory. Its function is to serve as the core algorithm for feedforward control, solidifying the abstract speed-area physical correlation into programmable computational logic, thereby avoiding model drift caused by relying on empirical estimations. The scaling factor can be set according to the material hardness, surface roughness, and elastic modulus of the gripper airbag of different grasped objects. For example, for high-friction objects like rubber, the scaling factor can be set to a larger value to reflect that a small speed change can cause a large contact deformation; for low-friction objects like metals, the scaling factor is set to a smaller value. When the object's speed changes abruptly, the model can instantly update the output area change, thereby triggering a subsequent air pressure regulation mechanism.

[0038] In this embodiment, the model is specifically represented as a linear proportional relationship function. Let the object's moving speed be v, the change in clamping contact area be ΔA, and the proportionality coefficient be k. Then the calculation formula can be expressed as: ΔA = k·v. Here, the proportionality coefficient k is an empirical value obtained through extensive experimental calibration, and its value depends on factors such as the friction coefficient of the gripper fingertips, the surface roughness of the object, and the clamping angle. For example, when the gripper fingertips are made of rubber and the object is a paper packaging box, the value of k can be 0.8; when the gripper fingertips are made of metal and the object is a smooth plastic part, the value of k may be adjusted to 0.5. It should be understood that the area change model is not limited to a linear formula. In other embodiments, a lookup table method or a nonlinear fitting model based on a neural network can also be used, as long as it reflects the mapping relationship between speed and area change. Through this model, the system converts the real-time acquired speed value into a specific area change, providing a quantitative basis for subsequent force adjustment.

[0039] Step S330: Detect the current matching degree between the current vector size and the preset strength requirement.

[0040] After calculating the area change, the system checks the current matching degree between the current vector magnitude and the preset force requirement. The current vector magnitude can be the vector magnitude representing the current actual clamping force state. This vector is synthesized based on the force data from various points fed back by the pressure sensor array and is used to intuitively reflect the current clamping strength. The preset force requirement can be the theoretically optimal clamping force threshold set for the material, weight, and target posture of the object being grasped. It is determined based on the object attribute database or initial calibration parameters. The current matching degree refers to the degree of agreement between the current vector magnitude and the preset force requirement. It is usually expressed by calculating the normalized result of the ratio or difference between the two, and its function is to determine whether the current clamping state meets the requirements for safe grasping. For example, if the preset force requirement is 10 Newtons, and the current vector magnitude synthesized from the sensor feedback is 9.0 Newtons, then the calculated current matching degree is 0.9 (or the deviation is 1.0 Newtons). The system compares this value with the preset matching degree threshold (e.g., 0.95). By detecting this current matching degree in real time, the force mismatch problem caused by changes in the clamping contact area can be accurately identified, providing a criterion for triggering the correction mechanism.

[0041] Step S350: If the current matching degree is lower than the preset matching degree threshold, calculate the incremental step size of air pressure adjustment based on the degree of mismatch, and gradually adjust the air pressure output with the incremental step size to generate a force distribution scheme that adapts to the change of clamping contact area.

[0042] In practical applications, the degree of mismatch can refer to the difference between the current matching degree and the preset matching degree threshold, or the absolute error between the actual clamping force and the theoretically required force. It serves as the direct basis for calculating the air pressure adjustment. The incremental step size of air pressure adjustment refers to the pressure change in a single air pressure adjustment action. This step size is dynamically calculated based on the degree of mismatch, rather than being a fixed constant. Its function is to ensure the smoothness of air pressure adjustment and prevent sudden pressure changes from impacting the object or causing gripper oscillation. The force distribution scheme refers to the set of air pressure settings at the fingertips of each gripper after redistribution. It is generated by gradually adjusting the air pressure output in incremental steps to adapt to changes in the current clamping contact area.

[0043] Specifically, when the current matching degree is lower than a preset matching degree threshold, the system uses a PID control algorithm or a similar linear / nonlinear mapping rule to convert the mismatch degree into an incremental step size. For example, when the force error is 0.6 Newtons, according to the proportional coefficient K... p =0.5 The calculated pressure increment step is 0.3 Pascals. The system control valve gradually increases the pressure output in 0.3 Pa increments until the actual clamping force approaches the theoretical value. This incremental step adjustment mechanism effectively eliminates the risk of slippage caused by a decrease in clamping contact area or overpressure damage caused by an increase in area, ensuring that the clamping force always closely tracks the dynamically changing contact state.

[0044] It should be noted that the innovation of this embodiment lies in the fact that the air pressure regulation is not a simple on / off adjustment, but rather adopts an incremental step dynamic adjustment strategy. The system calculates the incremental step size for air pressure regulation based on the degree of mismatch. The specific logic follows the principle of "large step size for large errors, small step size for small errors": when a large deviation is detected between the actual clamping force and the theoretical requirement (e.g., deviation exceeding 20%), the system calculates a larger incremental step size (e.g., 0.5 Pascals / step) to achieve a rapid response and quickly eliminate the main error; when the deviation is reduced to a smaller range (e.g., deviation less than 5%), the system automatically switches to a tiny incremental step size (e.g., 0.05 Pascals / step) for fine-tuning, avoiding force oscillations caused by excessive adjustment amplitude. This graded adjustment strategy effectively balances response speed and control accuracy.

[0045] Step S170: Based on the force distribution scheme, the position adjustment parameters and clamping force parameters of the three-degree-of-freedom mechanical pneumatic gripper are fused and calculated, and iterative verification is performed according to the preset feedback cycle to obtain the comprehensive matching coefficient.

[0046] In an exemplary embodiment of this application, the fusion calculation may include: weighting and fusing the position adjustment parameter and the clamping force parameter using preset weighting coefficients, and introducing a force calibration factor to correct the fusion result. The position adjustment parameter can refer to the displacement vector or angle increment calculated by the three-degree-of-freedom mechanical pneumatic gripper in the spatial coordinate system to eliminate attitude deviations; it originates from the aforementioned corrected coordination vector. The clamping force parameter can refer to the target pressure value or torque value in the force distribution scheme generated to adapt to changes in the current clamping contact area; it originates from the output result after dynamically adjusting the air pressure regulation accuracy step size. Since position parameters are usually measured in millimeters or degrees, while force parameters are measured in Newtons or Pascals, there are significant differences in their dimensions and dynamic response characteristics; direct superposition would lead to control command distortion. Therefore, this step normalizes and weights and fuses the two heterogeneous parameters using preset weighting coefficients. The weighting coefficients are preset according to the priority strategy of the three-degree-of-freedom mechanical pneumatic gripper under different working conditions. For example, when gripping fragile objects, the force weight is greater than the position weight; while in high-speed assembly line operations, the position weight is greater than the force weight. Specifically, the system constructs a fusion function F = w1·P + w2·F force Where P is the position adjustment parameter, i.e., the displacement or attitude correction amount of the three-degree-of-freedom mechanical pneumatic gripper in space (e.g., the movement distance or angular change in the X, Y, and Z axis directions), used to describe the target position that the gripper needs to be adjusted to; F force The clamping force parameter is the magnitude of the force applied to the object when the grippers are closed (e.g., in Newtons), used to control the tightness of the grip; w1 and w2 are the weighting coefficients for position and force, respectively, and w1+w2=1.

[0047] Furthermore, to eliminate systematic deviations caused by sensor zero-point drift, mechanical transmission backlash, or changes in ambient temperature, this step introduces a force calibration factor to correct the preliminary fusion results. The force calibration factor is a compensation coefficient calculated based on historical calibration data or real-time sensor feedback, and its function is to offset nonlinear mechanical losses. For example, when a slight leak is detected in the pneumatic pipeline, causing the actual output force to be 5% lower than the theoretical value, the force calibration factor is automatically set to 1.05, amplifying the fusion result to compensate for the loss. Through this combination of weighted fusion and factor correction, the characterization scale of multi-source heterogeneous parameters can be unified, significantly reducing control errors caused by dimensional mismatch and ensuring the consistency of position and force at the control command level.

[0048] In an exemplary embodiment of this application, iterative verification includes: recalculating the fusion result after each feedback loop cycle until the deviation of the comprehensive matching coefficient calculated in adjacent cycles is less than a second preset threshold or reaches a preset maximum number of iterations. The feedback loop cycle can refer to the time window during which the control system performs a complete data acquisition, fusion calculation, and instruction generation process, the duration of which is adaptively determined based on the update frequency of the dynamic environmental state data. This step aims to address the potential local optima or convergence instability issues in single fusion through multiple approximations over time. Specifically, at the end of each feedback loop cycle, the system re-executes the aforementioned weighted fusion and correction process using the sensor measurement data of the current cycle to obtain a new comprehensive matching coefficient. Subsequently, the system compares the comprehensive matching coefficient calculated in the current cycle with the calculation result of the previous cycle, calculating the absolute value of the difference as the deviation.

[0049] If the deviation is less than the preset convergence threshold (e.g., 0.001), the fusion result is considered to have reached stable convergence, the iteration stops, and the final comprehensive matching coefficient is output. If the deviation is still greater than the convergence threshold, the system continues to the next feedback loop cycle for recalculation. To prevent getting stuck in an infinite loop under extreme interference, this step sets a maximum iteration limit (e.g., 10 times). Once the iteration count reaches this limit, regardless of whether the deviation meets the convergence threshold requirement, the system forcibly terminates the iteration and uses the result of the last calculation as the output, ensuring real-time control. For example, in a certain capture task, the initial comprehensive matching coefficient obtained by fusion is 0.82. After the first iteration, it becomes 0.88 with a deviation of 0.06; after the second iteration, it becomes 0.895 with the deviation reduced to 0.015; after the third iteration, it becomes 0.898 with a deviation of 0.003. If the second preset threshold is 0.005, then convergence is determined, and 0.898 is output as the final coefficient. This iterative verification mechanism effectively filters out transient noise interference, allowing the comprehensive matching coefficient to gradually approach the true optimal solution, thus significantly improving the accuracy and robustness of control commands.

[0050] This application embodiment achieves high-precision dynamic matching of position adjustment and clamping force through the synergistic effect of the aforementioned technical features. By employing preset weighting coefficients to weight and fuse position adjustment parameters and clamping force parameters, the problem of difficulty in uniformly representing multi-source heterogeneous parameters due to their different dimensions is solved. This allows the control system to flexibly balance position accuracy and clamping stability according to task requirements. Furthermore, a force calibration factor is introduced to correct the fusion result, effectively eliminating nonlinear errors caused by systematic hardware deviations and environmental factors, further improving the accuracy of the fused data. Moreover, by recalculating the fusion result after each feedback loop cycle and using the deviation of the comprehensive matching coefficient between adjacent cycles as a convergence criterion, a closed-loop iterative verification mechanism is constructed. This ensures continuous optimization and stable convergence of the fusion result in dynamically changing environments, thereby significantly improving the grasping success rate and operational stability of the three-degree-of-freedom mechanical pneumatic gripper in complex dynamic scenarios.

[0051] Step S190: Using the comprehensive matching coefficient, feedback processing is performed according to the update frequency of the coordination vector to generate the final control instruction set for controlling the three-degree-of-freedom mechanical pneumatic gripper to perform gripping actions.

[0052] In the exemplary embodiments of this application, utilizing the comprehensive matching coefficient can refer to using the obtained matching coefficient as the core basis to determine the final instruction generation strategy. For example, if the comprehensive matching coefficient is 0.47, the instruction generation frequency is adjusted to 47% of the base frequency to match the current control accuracy requirements. Feedback processing according to the update frequency of the coordination vector can refer to strictly following the determined adaptive update rhythm to convert the calculated control parameters into specific drive signals. For example, if the update frequency is determined to be 1.5Hz, the system outputs a complete instruction packet every approximately 0.67 seconds. Generating the final control instruction set for controlling the three-degree-of-freedom mechanical pneumatic gripper to perform gripping actions can refer to converting parameters such as the movement distance of each degree of freedom (e.g., 10 mm on the X-axis, 5 mm on the Y-axis, and 3 mm on the Z-axis) and gripping force (e.g., 20 Newtons) into binary instruction streams or specific communication protocol messages. This step aims to transform all the aforementioned computational results of perception, evaluation, correction, adaptation, and fusion into executable physical actions, forming the final link of closed-loop control, ensuring that the three-degree-of-freedom mechanical pneumatic gripper can accurately and stably complete the gripping task.

[0053] In one optional embodiment of this application, such as Figure 4 The flowchart shown is a process for generating a final control instruction set for controlling a three-degree-of-freedom mechanical pneumatic gripper to perform a gripping action, according to an embodiment of this application. The process specifically includes the following steps: Step S410: Determine the initial value of the comprehensive matching coefficient based on the attitude error and environmental adaptation data, and dynamically correct the initial value through a feedback loop mechanism to obtain the corrected matching parameters.

[0054] In practical applications, attitude error refers to the deviation between the current actual attitude vector of a three-degree-of-freedom mechanical pneumatic gripper and the target grasping attitude vector. It is typically acquired and calculated in real-time by a three-axis accelerometer or vision sensor mounted on the gripper's end. Environmental adaptation data refers to a set of parameters reflecting the stability characteristics of the current working environment, including but not limited to ambient temperature, vibration frequency, and the amplitude of external disturbances. The initial value of the comprehensive matching coefficient is a quantitative index calculated using a preset weighted evaluation model based on the attitude error and environmental adaptation data. It characterizes the theoretical matching degree of the gripper's grasping task under the current state. For example, when the attitude error is 0.3 degrees and within the preset ±0.5 degree tolerance range, and the ambient temperature is within the suitable range of 20 to 30 degrees Celsius, the calculated initial comprehensive matching coefficient might be 0.85, indicating a relatively ideal current state.

[0055] The feedback loop mechanism is a dynamic adjustment strategy based on an iterative algorithm. Its function is to compare the execution feedback data from the previous cycle with the current initial value. If the attitude error exceeds a preset attitude error tolerance (e.g., the maximum allowable spatial orientation deviation is ±5 degrees) or a sudden change in the environment occurs, the initial value is corrected using a preset compensation model. Specifically, the system can set a fixed correction step size (e.g., 0.05) to increase or decrease in each iteration, repeating this process multiple times until the coefficients converge and stabilize. By co-inputting attitude error and environmental adaptation data, and combining this with continuous correction through the feedback loop mechanism, errors caused by single-measurement noise can be eliminated, resulting in highly confident corrected matching parameters, providing a precise basis for subsequent command generation.

[0056] It should be noted that the attitude error tolerance refers to a pre-set allowable error range (e.g., ±5 degrees) set by the system. When the real-time monitored attitude spatial orientation error (e.g., the angle between the gripper's current pointing direction and the target's pointing direction) exceeds this tolerance, it indicates that the current grasping attitude has deviated to a degree that may affect the grasping success rate. Therefore, it is necessary to activate the compensation model to correct the matching parameters. Conversely, if the error is within the tolerance range, the grasping attitude is considered to be within an acceptable range, and no correction is required. This tolerance value can be dynamically adjusted according to the precision requirements of the grasping task (e.g., precision assembly requires a smaller tolerance, while rough handling can be appropriately relaxed) or the surface characteristics of the object (e.g., fragile objects require a stricter tolerance).

[0057] Step S430: Generate an instruction parsing scheme based on the corrected matching parameters, optimize instruction generation efficiency based on the parsing speed of the instruction parsing scheme, and obtain optimized control instructions.

[0058] The instruction parsing scheme refers to a set of rules that map the modified matching parameters to specific executable logic, including instruction priority sorting rules, parallel parsing channel allocation strategies, and data format conversion standards. The scheme's function is to transform abstract matching parameters into a sequence of actions that a three-degree-of-freedom pneumatic gripper can recognize. Optimizing instruction generation efficiency based on the parsing speed of the instruction parsing scheme can refer to dynamically adjusting the frequency and complexity of instruction generation according to the current system's processing power and real-time requirements. For example, if the modified matching parameters indicate that the system is currently in a high-precision fine-tuning phase, the system can automatically reduce the instruction generation frequency (e.g., from 500 instructions per second to 235 instructions per second) to reserve more time for complex algorithm calculations and avoid control delays caused by parsing bottlenecks; conversely, the generation frequency is increased during rapid movement phases.

[0059] This optimization process ensures that the rate of control command generation matches the actual response capability of the robotic arm. Through flexible configuration of the command parsing scheme and adaptive optimization of the parsing speed, the computational load and control real-time performance can be effectively balanced, resulting in optimized control commands that meet motion planning requirements and possess efficient execution characteristics.

[0060] Step S450: After verification, the optimized control command is converted into the action signal to drive the three-degree-of-freedom mechanical pneumatic gripper.

[0061] In practical applications, verification refers to the process of verifying the integrity and correctness of optimized control commands. This includes Cyclic Redundancy Check (CRC), syntax checking, and logical boundary determination, aiming to eliminate command data with transmission errors or illegal formats. Action signals can refer to electrical or pneumatic signals that, after digital-to-analog conversion or pulse modulation, directly drive the joint motors or pneumatic valves of a three-degree-of-freedom mechanical pneumatic gripper. The process of converting optimized control commands into action signals involves decoding the binary or hexadecimal command stream into specific voltage values, current values, or PWM (Pulse Width Modulation) duty cycles through an underlying action execution mapping table. For example, an optimized command containing a 10mm X-axis movement and a 20N gripping force, after verification, is converted into a specific pulse sequence driving the X-axis servo motor and an analog voltage signal controlling the valve opening. Through rigorous verification and precise signal conversion, it is ensured that the control strategy generated by the host computer is transmitted to the actuator with zero errors, guaranteeing the accuracy and safety of the three-degree-of-freedom mechanical pneumatic gripper's movements.

[0062] This application constructs a complete closed-loop control architecture encompassing perception, evaluation, correction, adaptation, fusion, feedback, and execution through the synergistic effect of the aforementioned technical features. First, by synchronously acquiring and fusing position and environmental data, an accurate input benchmark is provided for attitude correction. Second, by utilizing the synergy of coordinate transformation and frequency adaptation, the rhythm of attitude correction can dynamically adapt to environmental interference characteristics, significantly improving the system's anti-interference capability. Furthermore, through dynamic adjustment of the air pressure accuracy step size, a fine response of the clamping force to instantaneous changes in the clamping contact area is achieved, avoiding the risk of slippage or pressure damage. Further, by employing a weighted fusion and iterative verification mechanism, the temporal and dimensional differences between position control and force control are effectively bridged, improving the convergence accuracy of the comprehensive matching coefficient. Finally, a closed-loop instruction set is generated according to the update frequency of the coordination vector, balancing attitude accuracy and force control stability. This series of interconnected steps jointly solves the problems of gripping failure and control lag caused by posture deviation, changes in gripping contact area and environmental interference in unstructured and highly dynamic scenarios for three-degree-of-freedom mechanical pneumatic grippers, achieving high-precision and highly robust adaptive integrated gripping control.

[0063] Based on the above embodiments, referring to Figure 5 As shown in the figure, this application embodiment also provides a flowchart of secondary scanning and instruction update triggered by clamping force distribution uniformity monitoring, which specifically includes the following steps: Step S510: Execute the gripping action according to the final control instruction set, and monitor the uniformity of the gripping force distribution in real time during the execution process; if the uniformity of distribution is detected to be lower than the third preset threshold, a second scan of the surface hardness of the object is triggered to obtain hardness distribution information.

[0064] In practical applications, the uniformity of gripping force distribution is a key indicator for measuring gripping quality, especially for irregularly shaped objects or objects with uneven material distribution (such as parts with a shifted center of gravity or partially filled packaging boxes). A single force control method often fails to guarantee balanced force on the contact surface. In this embodiment, executing the gripping action according to the final control instruction set can refer to the control system converting the binary instruction stream generated in the preceding steps into drive signals, controlling the coordinated movement of the joint motors and pneumatic valve groups of the three-degree-of-freedom mechanical pneumatic gripper to achieve physical contact and envelopment of the target object. Real-time monitoring of the uniformity of gripping force distribution can refer to continuously collecting real-time force data at each contact point at a high-frequency sampling rate (e.g., 1kHz) during the dynamic process of gripping action execution using a high-precision pressure sensor array integrated inside the gripper fingers. Distribution uniformity is quantified by calculating the standard deviation or coefficient of variation of the force values ​​at all effective contact points at the same moment. Its function is to identify the risk of localized stress concentration or contact failure caused by abrupt changes in the curvature of the object's surface, uneven material hardness, or a shifted center of gravity. The purpose of setting up this monitoring mechanism is to introduce immediate feedback after the open-loop command is executed, so as to prevent the grasping slippage caused by the initial model error.

[0065] If the detected uniformity of the surface distribution is below a third preset threshold (e.g., the maximum pressure difference exceeds 30%), a secondary scan of the object's surface hardness is triggered. Unlike the initial rapid scan, the secondary scan uses a high-resolution tactile sensor array to densely sample the contact area with smaller step sizes (e.g., 0.1 mm), accurately acquiring the surface hardness distribution information. This information specifically includes estimated Young's modulus values ​​or Shore hardness grades for different regions of the object's surface, revealing the local deformation characteristics of the object. This process can identify minute hardness differences on the object's surface, providing a precise basis for subsequent force compensation.

[0066] Step S530: Update the final control instruction set based on the hardness distribution information and environmental feedback data to obtain the optimized gripping action parameters.

[0067] After acquiring hardness distribution information, the system combines environmental feedback data (such as current vibration and noise levels, temperature drift, and the gripper's own pose error compensation) to correct the original control commands. Specifically, it performs multi-source fusion of the newly acquired hardness map and environmental feedback data, and uses gradient descent or model predictive control (MPC) algorithms to recalculate the target positions for each degree of freedom and the output pressure values ​​for each air path. This process upgrades the gripping strategy from uniform application to material-specific force application. Specifically, it appropriately reduces the stroke in hard areas to prevent damage and increases the gripping contact area in soft areas to ensure friction, thereby obtaining an optimized command dataset and generating optimized gripping motion parameters. This, in turn, adjusts the robotic arm's execution logic, updates the gripping force distribution, and achieves more uniform gripping. In this way, stable gripping can be guaranteed even if the object's surface hardness is uneven.

[0068] For example, based on the detected uneven force distribution, the system initiates a secondary scan, finding that the hardness of the area corresponding to gripper finger B is 60 HRC (high hardness), while the area corresponding to gripper finger A is only 45 HRC (low hardness). Combining this with the minor vibration interference detected in the environmental feedback, the system updates the instruction set: for the high-hardness area, the target gripping force is reduced from 6.8 N to 6.0 N and micro-motion compensation is added to eliminate gaps; for the low-hardness area, the target gripping force is increased from 3.5 N to 5.8 N to increase the effective gripping contact area. After 10 iterations of optimization, the force values ​​at each contact point tend to be balanced, and the standard deviation drops to within 0.8 N, thereby generating optimized gripping motion parameters that include new position coordinates and new air pressure values. This step, by introducing a new dimension of material property perception, significantly improves the gripper's adaptability to complex and heterogeneous objects and its gripping stability.

[0069] This application embodiment constructs an enhanced closed-loop control mechanism of monitoring, diagnosis, re-sensing, and re-optimization through the synergistic effect of the above steps. Specifically, the real-time monitoring mentioned above acts as the system's sensory nerve, quickly identifying deviations in the execution level of the initial control commands, especially for uneven force distribution problems that are difficult to predict during the planning stage. Once the third preset threshold condition is triggered, the above-mentioned secondary scanning mechanism is immediately activated, which is like adding tactile discrimination to the system, obtaining key information on the surface hardness distribution of the object. On this basis, the system deeply integrates the hardness information with real-time environmental feedback data, dynamically correcting the final control command set, so that the generated optimized grasping action parameters can accurately match the local physical characteristics of the object. This mechanism not only solves the problem that single force control cannot take into account the grasping stability of mixed soft and hard objects, but also effectively suppresses grasping posture instability caused by environmental interference. Through this adaptive adjustment, the system can achieve a leap from coarse grasping to precise locking in dynamic environments, significantly reducing the risk of object slippage and ensuring the reliability and safety of operation under complex three-degree-of-freedom motion trajectories.

[0070] Furthermore, referring to Figure 6 As shown in the figure, this application embodiment also provides a schematic diagram of the backtracking correction control process of a three-degree-of-freedom mechanical pneumatic gripper, which specifically includes the following steps: Step S610: Execute the grasping action based on the optimized grasping action parameters, and monitor the spatial orientation error of the grasping posture in real time.

[0071] This step refers to the process where, during the actual grasping operation of a three-degree-of-freedom mechanical pneumatic gripper based on the optimized grasping motion parameters generated in the previous steps, a built-in attitude sensor array continuously collects real-time attitude data of the gripper's end effector in three-dimensional space. This real-time attitude data is then compared with the target grasping attitude data to calculate the current spatial orientation error. Spatial orientation error refers to the angular deviation or vector difference between the gripper's current actual orientation vector and the ideal target orientation vector. Its sources include instantaneous attitude drift caused by the object's center of gravity shift in a dynamic environment, external vibration interference, or mechanical transmission backlash. This error serves as a key indicator for judging grasping stability, quantifying the degree to which the current grasping motion deviates from the expected trajectory.

[0072] Specifically, spatial orientation error can be determined by calculating the cosine of the angle between the current attitude vector and the target attitude vector. For example, when the target attitude vector is (1.0,0,0) and the actual attitude vector monitored in real time is (0.96,0.28,0), the calculated angle is approximately 16.3 degrees, which is the current spatial orientation error. Through this real-time monitoring mechanism, the system can instantly perceive the risk of slippage or attitude instability that objects may experience in complex scenarios, providing accurate data input for subsequent risk response.

[0073] Step S630: If the spatial direction error exceeds the preset environmental adaptation threshold, then backtrack to the step of adaptively adjusting the update frequency of the coordination vector according to the dynamic characteristics of the dynamic environmental state data, re-adaptively adjust the update frequency of the coordination vector according to the current dynamic environmental state data, and recalculate the corrected coordination vector.

[0074] Specifically, when executing optimized actions, if a sudden strong disturbance (such as an external collision or airflow impact) occurs, causing the spatial orientation error of the grasping posture to suddenly increase and exceed the environmental adaptation threshold (e.g., the error exceeds 5 degrees), traditional control logic often chooses to stop and alarm or start calculating from scratch, resulting in low efficiency. The backtracking mechanism adopted in this embodiment has the core logic of precise jump, that is, the system does not backtrack to the initial step S110 to re-collect all data, but directly jumps to the intermediate link in step S120, namely the "adaptive adjustment of the update frequency of the coordination vector" step. This is because the initial position data and attitude deviation did not change abruptly in a short period of time; the only change was the dynamic characteristics of the environment. The system directly uses the latest dynamic environmental state data to re-evaluate the vibration frequency and disturbance amplitude, adaptively adjust the update frequency (for example, if a high-frequency impact is detected, the update frequency is immediately reduced to smooth the jitter), and quickly recalculate the corrected coordination vector based on this. This breakpoint-resume calculation logic greatly shortens the anomaly recovery time and avoids the waste of resources from full-process recalculation.

[0075] Step S650: Based on the corrected coordination vector, generate the final accurate grasping scheme.

[0076] This step involves using a recalculated and corrected coordination vector, combined with the current force distribution scheme and position adjustment parameters, to generate a final, precise grasping scheme that adapts to the current extreme dynamic environment. The corrected coordination vector contains multi-directional motion commands optimized for the latest environmental disturbances. Its function is to guide the three-degree-of-freedom mechanical pneumatic gripper to make fine adjustments at the correct posture and frequency to compensate for the identified spatial orientation errors. The process of generating a precise grasping scheme involves mapping the corrected coordination vector into specific joint drive signals and pneumatic control commands, ensuring that the gripper maintains a uniform distribution of gripping force while adjusting its posture. For example, based on the recalculated coordination vector, the system generates a comprehensive instruction set including adjusting the first joint angle from 30 degrees to 32.5 degrees, synchronously compensating the second joint by 1.2 degrees, and dynamically increasing the gripping pressure by 0.5 Newtons, forming the final precise grasping scheme. This scheme not only corrects the current posture deviation but also improves the system's response speed to subsequent disturbances through optimized update frequency, thereby significantly reducing the risk of object slippage and ensuring a high success rate for grasping tasks in complex scenarios.

[0077] Specifically, through the synergistic effect of the above-mentioned secondary scanning and backtracking mechanisms, this embodiment constructs a closed-loop control system with self-sensing, self-diagnosis and self-repair capabilities, which significantly reduces the grasping failure rate and the risk of object slippage in complex dynamic environments while ensuring grasping efficiency.

[0078] This embodiment applies the above technical solution to a specific industrial automated production line scenario to verify its technical effectiveness. For example, the scenario is set as a conveyor belt with adjustable operating speed, on which irregularly shaped metal parts are randomly placed. This embodiment aims to demonstrate how a three-degree-of-freedom mechanical pneumatic gripper can achieve precise and stable gripping actions in a dynamic environment where parts move at high speed with the conveyor belt.

[0079] In this scenario, the system first executes step S110 to acquire the current position data and dynamic environmental status data of the three-degree-of-freedom mechanical pneumatic gripper. Specifically, the conveyor belt speed sensor provides real-time feedback that the current conveyor belt speed is 2.5 m / s, the vision sensor captures the initial image of the target part, and the image processing algorithm identifies the contour coordinates and surface hardness distribution of the part. Assuming the target part is made of cast iron with an average surface hardness of 3.2 N / mm², the system first executes step S110 to acquire the current position data and dynamic environmental status data of the three-degree-of-freedom mechanical pneumatic gripper. 2 The current position coordinates are (x=12.5, y=8.3, z=5.7) cm. Simultaneously, a vibration sensor mounted on the robotic arm base detects an environmental vibration frequency of 2.1 Hz, generated by the mechanical vibration of the conveyor belt motor. The system integrates the above position data, hardness information, and vibration data to calculate the deviation assessment result of the initial grasping posture. For example, by calculating the cosine of the vector angle, it is found that there is a 5.2-degree directional deviation between the current grasping posture and the target posture, and that environmental vibration may cause a 0.3 mm positional jitter. This step demonstrates the comprehensiveness of this application's dynamic environment perception, not only focusing on the target object itself but also quantifying environmental interference factors.

[0080] Subsequently, the system executes step S130 to correct the deviation. Since the deviation assessment result shows a directional deviation of 5.2 degrees, exceeding the preset first threshold of 3 degrees, the system immediately initiates attitude correction logic. The system uses a coordinate transformation method to calculate the coordination vector for attitude correction. Considering the environmental vibration frequency of 2.1Hz, which is considered low-to-medium frequency interference, the system adaptively adjusts the update frequency of the coordination vector based on the dynamic characteristics of the dynamic environmental state data. Specifically, the system determines that interference at this frequency may cause control jitter, therefore reducing the update frequency of the coordination vector from the default 50Hz to 30Hz to smooth the control curve, resulting in the corrected coordination vector. This process demonstrates how this application maintains the stability of the control system under dynamic interference through a frequency adjustment strategy, avoiding resonance or jitter problems that may be caused by traditional fixed-frequency control.

[0081] Next, the system executes step S150 to handle the clamping force distribution issue. During the gripping process, because the conveyor belt is still running, the part has a moving speed relative to the grippers, causing the clamping contact area to change. The system obtains the object's moving speed as 2.5 m / s, and combined with the pre-established area change model (proportionality coefficient k=0.8), calculates the change in clamping contact area to be 2.0 cm. 2 At this point, the system detects a 0.6N deviation between the clamping force corresponding to the current vector size and the theoretical requirement, indicating that the current matching degree is below the preset matching degree threshold. The system then triggers the air pressure adjustment mechanism, calculating the incremental step size for air pressure adjustment based on the degree of mismatch. Due to the small deviation, the system adopts a fine adjustment mode, setting the incremental step size to 0.05 Pascals / step, gradually adjusting the air pressure output to generate a force distribution scheme adapted to the current clamping contact area. This process demonstrates the precision and adaptability of this application in force control, effectively addressing the risk of slippage caused by sudden changes in the clamping contact area during dynamic grasping.

[0082] After completing attitude correction and force distribution, the system executes step S170 to perform parameter fusion and iterative verification. The system fuses the position adjustment parameters (corrected coordination vector) and the clamping force parameters (force distribution scheme) for calculation. Assuming the weight of the position adjustment parameters is set to 0.6 and the weight of the force parameters is set to 0.4, a force calibration factor of 1.05 is introduced for correction. The system performs iterative verification according to a preset 10ms feedback loop cycle. After the first iteration, the calculated comprehensive matching coefficient is 0.85; after the second iteration, the coefficient is updated to 0.87, and the deviation between adjacent cycles is 0.02, which is less than the second preset threshold of 0.05, and the system determines that the verification passes. This step demonstrates how this application ensures optimal coordination of control commands in both position and force dimensions through multi-parameter fusion and a fast-converging iterative algorithm.

[0083] Subsequently, the system executes step S190 to generate the final control instruction set. Utilizing a comprehensive matching coefficient of 0.87 and an update frequency of 30Hz, the system performs feedback processing to generate a control instruction set containing target position values, target force values, and timestamps. This instruction set is parsed and converted into PWM signals, driving the air valve and motor to work in tandem. The mechanical pneumatic gripper responds rapidly and accurately grasps irregular parts on the conveyor belt. During execution, the system monitors the uniformity of the gripping force distribution in real time. If a sudden change in local hardness is detected on the part's surface, causing the force distribution uniformity to fall below a third preset threshold, the system immediately triggers a secondary scan of the object's surface hardness to obtain more refined hardness distribution information and updates the control instruction set accordingly, fine-tuning the gripping force of each fingertip. This closed-loop feedback mechanism ensures the reliability of the gripping action and effectively prevents gripping failures caused by uneven surface characteristics of the object.

[0084] Finally, if a sudden external impact occurs during the grasping process, causing the attitude error to exceed the environmental adaptation threshold, the system will activate a backtracking mechanism. Instead of recalculating from scratch, the system will directly backtrack to the frequency adjustment step in S130, readjusting the update frequency based on the latest interference data and quickly generating a new, precise grasping scheme. For example, when a sudden impact is detected causing the attitude error to instantly increase to 8 degrees, the system immediately increases the update frequency to 50Hz for a rapid response, recalculates the coordination vector, and performs corrections, successfully mitigating the risk of slippage. Through the complete simulation of the above industrial scenario, this embodiment verifies the high adaptability and stability of the method provided in this application in dynamic environments. Compared to traditional fixed-parameter control methods, the grasping success rate is increased by approximately 15%, and the object slippage rate is reduced by approximately 20%, significantly improving the operating efficiency of industrial automated production lines.

[0085] It should be noted that, in this document, the terms "comprising," "including," and any other variations are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Specific examples have been used in this document to illustrate the principles and implementation methods of the technical solutions of this application. The above examples are only for the purpose of helping to understand the methods and core ideas of this application. The above descriptions are merely preferred embodiments of this application. It should be pointed out that, due to the limitations of written expression and the objective existence of infinite specific structures, those skilled in the art can make several improvements, modifications, or changes without departing from the principles of this application, and can also combine the above technical features in an appropriate manner; these improvements, modifications, changes, or combinations, or the direct application of the concept and technical solutions of this application to other situations without modification, should all be considered within the scope of protection of this application.

Claims

1. A control method for a three-degree-of-freedom mechanical pneumatic gripper, characterized in that, Includes the following steps: The current position data and dynamic environmental state data of the three-degree-of-freedom mechanical pneumatic gripper are obtained, and the spatial orientation deviation of the gripping posture is identified based on the current position data. Combined with the dynamic environmental state data, the deviation evaluation result of the initial gripping posture is obtained. When the deviation evaluation result indicates that the directional deviation exceeds the first preset threshold, a coordinate transformation method is used to calculate the coordination vector for attitude correction, and the update frequency of the coordination vector is adaptively adjusted according to the dynamic characteristics of the dynamic environment state data to obtain the corrected coordination vector. Based on the modified coordination vector, when the gripping contact area changes during the gripping process, causing the actual gripping force to mismatch with the theoretical requirement, the gripping force is redistributed by dynamically adjusting the precision step size of the air pressure regulation, thereby generating a force distribution scheme that adapts to the current gripping contact area. Based on the force distribution scheme, the position adjustment parameters and clamping force parameters of the three-degree-of-freedom mechanical pneumatic gripper are fused and calculated, and iteratively verified according to the preset feedback cycle to obtain the comprehensive matching coefficient. Using the comprehensive matching coefficient, feedback processing is performed according to the update frequency of the coordination vector to generate the final control instruction set for controlling the three-degree-of-freedom mechanical pneumatic gripper to perform gripping actions.

2. The method according to claim 1, characterized in that, The step of adaptively adjusting the update frequency of the coordination vector based on the dynamic characteristics of the dynamic environment state data includes: dynamically reducing or increasing the calculation and transmission frequency of the coordination vector according to the vibration frequency or interference amplitude in the dynamic environment.

3. The method according to claim 1, characterized in that, The adaptive adjustment of the update frequency of the coordination vector based on the dynamic characteristics of the dynamic environment state data further includes: Acquire dynamic change data of multi-directional quantities, determine the degree of matching between the dynamic change data and the coordination vector, and obtain intermediate results of direction correction; By combining the fluctuation data of environmental disturbances, the intermediate results are processed to calculate the final coordination vector correction value. Determine whether the correction scheme generated based on the coordination vector correction value meets the preset stability condition. If it does, determine the update frequency according to the correction scheme.

4. The method according to claim 1, characterized in that, The method of redistributing clamping force by dynamically adjusting the precision step size of air pressure regulation includes: Obtain the object's moving speed and calculate the change in the clamping contact area caused by the object's moving speed; Detect the current matching degree between the current vector size and the preset strength requirement; If the current matching degree is lower than the preset matching degree threshold, the incremental step size of air pressure adjustment is calculated based on the degree of mismatch, and the air pressure output is gradually adjusted with the incremental step size to generate the force distribution scheme that adapts to the change of clamping contact area.

5. The method according to claim 4, characterized in that, The calculation of the change in clamping contact area caused by the moving speed of the object includes: Using a pre-established area change model, the change in the clamping contact area is calculated based on the proportional relationship between the object's moving speed and the clamping contact area.

6. The method according to claim 1, characterized in that, The fusion calculation includes: using preset weighting coefficients to perform weighted fusion of position adjustment parameters and clamping force parameters, and introducing a force calibration factor to correct the fusion result; The iterative verification includes: recalculating the fusion result after each feedback cycle until the deviation of the comprehensive matching coefficient calculated in adjacent cycles is less than the second preset threshold or the preset maximum number of iterations is reached.

7. The method according to claim 1, characterized in that, The final control instruction set generated for controlling the three-degree-of-freedom mechanical pneumatic gripper to perform gripping actions includes: The initial value of the integrated matching coefficient is determined based on the attitude error and environmental adaptation data, and the initial value is dynamically corrected through a feedback loop mechanism to obtain the corrected matching parameters. Based on the corrected matching parameters, an instruction parsing scheme is generated, and the instruction generation efficiency is optimized based on the parsing speed of the instruction parsing scheme to obtain optimized control instructions. After verification, the optimized control commands are converted into action signals to drive the three-degree-of-freedom mechanical pneumatic gripper.

8. The method according to claim 1, characterized in that, The method further includes: The grasping action is executed according to the final control instruction set, and the uniformity of the gripping force distribution is monitored in real time during the execution process; if the uniformity of the distribution is detected to be lower than the third preset threshold, a second scan of the surface hardness of the object is triggered to obtain hardness distribution information. The final control instruction set is updated based on the hardness distribution information and environmental feedback data to obtain optimized gripping action parameters.

9. The method according to claim 8, characterized in that, The method further includes: The grasping action is executed based on the optimized grasping action parameters, and the spatial orientation error of the grasping posture is monitored in real time. If the spatial orientation error exceeds the preset environmental adaptation threshold, the process goes back to the step of adaptively adjusting the update frequency of the coordination vector based on the dynamic characteristics of the dynamic environmental state data, and then adaptively adjusts the update frequency of the coordination vector again based on the current dynamic environmental state data, and recalculates the corrected coordination vector. Based on the revised coordination vector, a final accurate crawling scheme is generated.

10. The method according to any one of claims 1-9, characterized in that, The spatial orientation deviation of the identified grasping posture includes: Compare the current grasping posture vector with the target grasping posture vector, and calculate the cosine value of the angle between the two vectors; The directional deviation is determined based on the cosine value of the included angle, and the directional deviation is used as part of the deviation evaluation result.

Images