Multi-gradient rigid-flexible coupling fluid self-adjusting bionic adhesion microstructure processing method

Abstract

The invention relates to a processing method of a multi-gradient rigid-flexible coupling fluid self-adjusting bionic adhesion microstructure, which aims at the problem of high-adaptability adhesion of special-shaped and dynamic surfaces, and comprises the following steps of: acquiring three-dimensional surface topography and stress distribution data, fusing standardized feature vectors, modeling a bionic three-layer structure and embedding a fluid cavity channel. The multi-stage rigid-flexible gradient layout of the microstructure and the integrated optimization of cavity parameters are realized, and the functional gradient microstructure is prepared by using a micro-nano 3D printing and transfer printing process. Real-time monitoring feedback and machine learning continuous optimization are achieved, and the surface adaptability, mechanical uniformity and durability of the bionic adhesion microstructure under variable working conditions are improved.

Description

Technical Field

[0001] This invention relates to the fields of "bionic microstructure design and manufacturing, intelligent adaptive surface engineering and precision fluid control technology", and in particular to a method for fabricating multi-gradient rigid-flexible coupled fluid self-regulating bionic adhesive microstructures. Background Technology

[0002] Currently, biomimetic adhesive materials and microstructures suitable for complex scenarios are showing broad application prospects in fields such as robotic gripping, medical devices, wearable devices, and flexible sensing. Inspired by the multi-scale structures of plants and animals in nature (such as gecko feet and octopus suckers), biomimetic adhesive microstructures typically endow artificial materials with reversible adhesion capabilities and a certain degree of adaptability through methods such as arrayed microstructures, rigid-flexible layouts, and surface microtexturing modifications. In terms of fabrication, advanced technologies such as micro / nano 3D printing, soft / hard multi-material composites, and photolithography / transfer printing all have different applications. With the increasing demand for flexible electronics, smart terminals, and precision medicine, higher requirements are being placed on the high adaptability and stability of adhesive structures in irregular, dynamic, and rough surface environments.

[0003] Currently, the mainstream biomimetic adhesive microstructures can be broadly classified into the following technical solutions:

[0004] A single-layer rigid-flexible composite biomimetic structure. Using an elastic polymer as a substrate, surface micro- and nano-textures mimic gecko toes and insect hairs to enhance the contact area and van der Waals forces with the target surface. A few designs introduce a rigid skeleton to strengthen the overall mechanical support, but the rigid-flexible layering is limited, resulting in insufficient deformation response when adapting to irregular and variable surfaces.

[0005] Arrayed deformable microstructures. Precisely arranging microstructural units in an array enhances overall fit and local fine-tuning capabilities. Some products utilize flexible materials or variable-thickness structures to improve non-planar adaptability. However, most arrayed structures still rely on upper and lower layers or the same material system, resulting in limited adaptability and stability to surfaces with high curvature, high roughness, or rapid dynamic deformation.

[0006] Passively responsive rigid-flexible gradient materials. Some advanced fabrication techniques achieve finite gradient distributions, enhancing adaptability to complex surfaces through gradual changes in the material's elasticity, thickness, or filler distribution. However, the gradient distribution is statically limited and cannot adaptively adjust in real time to rapidly changing or highly variable irregular dynamic surfaces.

[0007] Fluid-assisted or controllable structures. In recent years, some studies have attempted to combine microcavities with fluid / pneumatic actuation to achieve shape and pressure regulation with the help of external control. However, most of these studies are limited to simple geometric control or overall flexible deformation. Fine spatial partitioning response and microscopic-level structural coordination are difficult to achieve.

[0008] Existing biomimetic adhesive microstructures mainly remain at the level of single-layer rigid-flexible, simple material gradient, or passive array, which cannot meet the requirements for efficient adaptation to irregular (large curvature, high roughness, complex topology) and dynamic (real-time surface changes, periodic / non-periodic motion) target surfaces.

[0009] Typical applications of this technology include: flexible grippers for surgical robots, grasping tips for minimally invasive medical catheters, wearable flexible electronic devices, and end effectors for high-end intelligent robotic arms. In these scenarios, the target surface often exhibits significant irregularities and dynamic changes, requiring the adhesive structure to not only possess reversible / high-strength adhesion but also achieve rapid, localized, and intelligent adaptive deformation during contact to ensure maximum contact area, minimum contact stress concentration, and optimal adhesion performance. However, existing structures generally suffer from deficiencies in compliance, dynamic response speed, and local controllability, leading to problems such as fluctuating adhesive force, easy detachment, rapid wear, reduced durability, and high power consumption during use.

[0010] The main problems are analyzed as follows:

[0011] Limited adaptability to curved, rough, and dynamically deformable surfaces. Most biomimetic microstructures are optimized for smooth or low-curvature surfaces. When faced with high-curvature, high-roughness, or irregularly shaped surfaces that change constantly, the overall structure cannot deform sufficiently to match, resulting in a significant decrease in effective adhesion. Especially in dynamic environments (such as the human body in breathing motion or continuously deforming mechanical parts), the real-time adhesion ability of microstructures is very poor, leading to significant fluctuations in adhesion reliability.

[0012] The material's rigidity-flexibility matching and layering are limited, resulting in weak intelligent shape adjustment capabilities. Existing materials are typically just simple stacks of elastic / flexible materials, with discontinuous gradient transitions, leading to stress concentration and limited deformation capacity. The coarse layering makes it difficult to balance mechanical support and flexible compliance, and to meet the complex adaptation requirements under various working conditions.

[0013] The system lacks multi-dimensional real-time sensing and feedback control. Under complex surfaces and dynamic conditions, it lacks real-time detection and adaptive control methods for multi-point surface morphology and pressure. Structural motion is mostly a passive response or global adjustment, making it difficult to support local directional optimization and high-precision dynamic compensation.

[0014] There is a lack of intelligent fluid-assisted mechanisms. Most existing pneumatic or fluid-assisted solutions involve overall pressurization / depressurization or coarse partitioning, lacking fine-grained, spatially controllable fluid cavity adjustment design based on real-time feedback, making it difficult to achieve independent and autonomous pressure and deformation control of each unit in the microstructure array.

[0015] The integration and high degree of integration in the fabrication process are challenging. Most technical approaches employ step-by-step layering, manual transfer, and single-material injection molding, which cannot achieve micro-nano-scale integrated fabrication of multi-material, high-precision gradient structures with embedded fluid cavities. This results in large structures, poor fabrication accuracy, and higher costs.

[0016] Adaptive adhesion performance is unstable and prone to failure and short lifespan due to changes in the operating environment. In long-term applications, the adhesive structure will experience performance degradation due to repeated deformation, wear and tear and environmental interference. Existing structures are difficult to monitor online and intelligently correct and regulate, making them unsuitable for long-term, high-reliability applications in variable environments. Summary of the Invention

[0017] This application provides a method for fabricating multi-gradient rigid-flexible coupled fluid self-regulating biomimetic adhesive microstructures, aiming to solve one of the problems or issues of the prior art mentioned in the background.

[0018] This application provides a method for fabricating multi-gradient rigid-flexible coupled fluid self-regulating biomimetic adhesive microstructures, specifically including:

[0019] S1: Collect the morphological parameters and stress distribution information of the target irregular and dynamic surface, and record multi-point surface data containing different curvatures, roughnesses and dynamic motion states.

[0020] S2: Normalize the collected surface morphology parameters and stress distribution information to eliminate sampling errors, and construct a set of surface feature vectors under multiple working conditions as the input basis for subsequent biomimetic microstructure modeling.

[0021] S3: Based on a set of surface feature vectors under multiple working conditions, perform biomimetic microstructure geometry modeling, generate multi-layer gradient rigid-flexible layout parameters that adaptively match different surface curvatures and stress states, and form three-layer structural design data, including a rigid support zone, an elastic buffer zone, and a flexible coronal end layer.

[0022] S4: For the three-layer structure design data, embed the micro fluid cavity structure model, and map the fluid cavity layout parameters according to different surface features to achieve adaptive distribution mapping between the intermediate elastic buffer fluid cavity and the target surface morphology.

[0023] S5: Using auxiliary micro-nano 3D printing parameters and layer-by-layer transfer, filling and curing process data, according to the three-layer structure design data and micro fluid cavity design information, a multi-layer gradient rigid-flexible coupled crown-shaped biomimetic adhesive microstructure entity is prepared to achieve spatial positive gradient of mechanical properties of different layers of materials.

[0024] S6: The prepared coronal biomimetic adhesive microstructures are distributed and integrated to construct a controllable array unit. At the same time, a microfluidic system channel is configured for external fluid supply and pressure regulation to establish an initial parameter set for fluid cavity control.

[0025] S7: Based on the microfluidic system channel input, the contact state between the array unit and the target complex surface is monitored in real time. The contact pressure, morphology sensing and other multi-point feedback are input into the adaptive fluid regulation algorithm to realize the directional pressure control of the fluid cavity.

[0026] S8: Based on the output of the adaptive fluid regulation algorithm, the state of the fluid cavity inside the microstructure is dynamically adjusted, causing the multi-layer gradient rigid-flexible microstructure at different positions to generate real-time deformation for the current surface, thereby accurately matching the target surface contour and achieving maximum effective contact and adhesion.

[0027] S9: Establish a mapping relationship between the adaptive adhesion results and surface morphology changes and array unit response data. Use pressure feedback and machine learning optimization algorithms to continuously analyze the deviation between microstructure response and environmental adaptation, and update the flow control algorithm and hierarchical material layout to optimize adhesion stability and durability.

[0028] S10: Periodically test and analyze the mechanical properties, adaptability and reversible adhesion-detachment capability of the entire multi-layer gradient rigid-flexible coupled crown-shaped biomimetic adhesive microstructure array, and dynamically optimize the biomimetic structure parameters and flow control strategy based on the test feedback to meet the application requirements of different irregular or dynamic surfaces.

[0029] This application provides a multi-gradient rigid-flexible coupled fluid self-regulating biomimetic adhesive microstructure fabrication method, which has the following beneficial effects:

[0030] (1) This invention employs a multi-layered gradient rigid-flexible coupling structural design, utilizing a spatial distribution of rigid support zones, elastic buffer zones, and flexible coronal end layers. It achieves seamless transition of interlayer properties through a positive gradient of mechanical performance parameters, endowing the microstructure with excellent mechanical compliance and shape control capabilities. Furthermore, the embedded microfluidic channels, working in conjunction with an external microfluidic system, enable real-time dynamic driving of local deformation. This allows the microstructure to adaptively adjust its morphology based on the curvature, roughness, and instantaneous dynamic changes of the target surface, maximizing contact and steady-state adhesion. Simulation and experimental results show that, adapting to different radii of curvature (e.g., 1.5~100mm), surface roughness variations (Ra 0.1~15 μm), and dynamic disturbances, the microstructure contour fitting rate is significantly improved, the dynamic deformation response error is reduced to within 10 μm, and the effective adhesion area is increased by more than 20%, far superior to traditional single rigid / flexible microstructure arrays.

[0031] (2) This patent utilizes a fluid-assisted deformable mechanism, combined with real-time multi-point pressure and morphology feedback, and an intelligent adaptive algorithm, to achieve active control over the dynamic coupling between the microstructure and the surface. Even if the target surface undergoes abrupt changes, vibrations, or deformations, the system can quickly correct the cavity pressure and microstructure deformation, ensuring that the adhesion performance remains within the high-efficiency range. Compared with conventional non-feedback mechanical biomimetic arrays, the fluctuation of the adhesion force is significantly reduced, greatly enhancing the adaptability and controllability of the biomimetic microstructure under different working conditions.

[0032] (3) The invention employs micro-nano 3D printing and layer-by-layer transfer / curing processes to achieve in-situ spatial gradient integration of materials with different levels and multiple properties. This eliminates the cumbersome multi-stage batch manufacturing and manual assembly processes, enabling the manufacturing precision and uniformity of microstructures to reach the micrometer level, and achieving mass production in arrays. At the same time, by integrating fluid channels and structures into a single model, the utilization efficiency of material spatial distribution and structural reliability are significantly improved, reducing the risk of failure caused by discontinuities at heterogeneous interfaces, and greatly increasing the first-pass yield.

[0033] (4) This invention deeply couples high-resolution 3D vision, contact / morphology / stress sensing with adaptive flow control algorithms, and combines machine learning and optimization models to achieve continuous online monitoring, learning and self-optimization of complex and variable surfaces, greatly expanding the applicable scenarios of the adhesion array. It can be widely used in the adhesion and reversible adhesion and detachment of highly irregular dynamic surfaces in many industries such as medical surgical robots, spider robots, non-destructive surface handling, and rescue grasping, significantly improving the overall intelligence level and operational safety of the device. Attached Figure Description

[0034] Appendix Figure 1 This is the main flowchart of a multi-gradient rigid-flexible coupling fluid self-regulating biomimetic adhesive microstructure fabrication method according to this application.

[0035] Appendix Figure 2 This is a sub-flowchart of a multi-gradient rigid-flexible coupling fluid self-regulating biomimetic adhesive microstructure fabrication method according to this application.

[0036] Appendix Figure 3 This is another sub-flowchart of the multi-gradient rigid-flexible coupling fluid self-regulating biomimetic adhesive microstructure fabrication method of this application.

[0037] Appendix Figure 4 This is a magnified micrograph of a biomimetic adhesive microstructure fabricated using a multi-gradient rigid-flexible coupling fluid self-regulating biomimetic adhesive microstructure fabrication method described in this application.

[0038] Appendix Figure 5 This is a magnified micrograph of another biomimetic adhesive microstructure produced by the multi-gradient rigid-flexible coupling fluid self-regulating biomimetic adhesive microstructure fabrication method of this application. Detailed Implementation

[0039] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0040] The following disclosure provides many different embodiments or examples for implementing different structures of the invention. To simplify the disclosure, specific examples of components and arrangements are described below. Of course, these are merely examples and are not intended to limit the invention. Furthermore, reference numerals and / or letters may be repeated in different examples; such repetition is for simplification and clarity and does not in itself indicate a relationship between the various embodiments and / or arrangements discussed.

[0041] As attached Figure 1 As shown, this application provides a method for fabricating multi-gradient rigid-flexible coupled fluid self-regulating biomimetic adhesive microstructures, specifically including:

[0042] S1: Collect the morphological parameters and stress distribution information of the target irregular and dynamic surface, and record multi-point surface data containing different curvatures, roughnesses and dynamic motion states.

[0043] S2: Normalize the collected surface morphology parameters and stress distribution information to eliminate sampling errors, and construct a set of surface feature vectors under multiple working conditions as the input basis for subsequent biomimetic microstructure modeling.

[0044] S3: Based on a set of surface feature vectors under multiple working conditions, perform biomimetic microstructure geometry modeling, generate multi-layer gradient rigid-flexible layout parameters that adaptively match different surface curvatures and stress states, and form three-layer structural design data, including a rigid support zone, an elastic buffer zone, and a flexible coronal end layer.

[0045] S4: For the three-layer structure design data, embed the micro fluid cavity structure model, and map the fluid cavity layout parameters according to different surface features to achieve adaptive distribution mapping between the intermediate elastic buffer fluid cavity and the target surface morphology.

[0046] S5: Using auxiliary micro-nano 3D printing parameters and layer-by-layer transfer, filling and curing process data, according to the three-layer structure design data and micro fluid cavity design information, a multi-layer gradient rigid-flexible coupled crown-shaped biomimetic adhesive microstructure entity is prepared to achieve spatial positive gradient of mechanical properties of different layers of materials.

[0047] S6: The prepared coronal biomimetic adhesive microstructures are distributed and integrated to construct a controllable array unit. At the same time, a microfluidic system channel is configured for external fluid supply and pressure regulation to establish an initial parameter set for fluid cavity control.

[0048] S7: Based on the microfluidic system channel input, the contact state between the array unit and the target complex surface is monitored in real time. The contact pressure, morphology sensing and other multi-point feedback are input into the adaptive fluid regulation algorithm to realize the directional pressure control of the fluid cavity.

[0049] S8: Based on the output of the adaptive fluid regulation algorithm, the state of the fluid cavity inside the microstructure is dynamically adjusted, causing the multi-layer gradient rigid-flexible microstructure at different positions to generate real-time deformation for the current surface, thereby accurately matching the target surface contour and achieving maximum effective contact and adhesion.

[0050] S9: Establish a mapping relationship between the adaptive adhesion results and surface morphology changes and array unit response data. Use pressure feedback and machine learning optimization algorithms to continuously analyze the deviation between microstructure response and environmental adaptation, and update the flow control algorithm and hierarchical material layout to optimize adhesion stability and durability.

[0051] S10: Periodically test and analyze the mechanical properties, adaptability and reversible adhesion-detachment capability of the entire multi-layer gradient rigid-flexible coupled crown-shaped biomimetic adhesive microstructure array, and dynamically optimize the biomimetic structure parameters and flow control strategy based on the test feedback to meet the application requirements of different irregular or dynamic surfaces.

[0052] Step S1: Collect morphological parameters and stress distribution information of the target irregular shape and dynamic surface, and record multi-point surface data including different curvatures, roughnesses, and dynamic motion states. Specifically, this includes:

[0053] S1.1: Conduct preliminary analysis of the target irregular or dynamic surface to clarify the spatial distribution range of the area to be collected, and acquire the original data of the shape perception point cloud based on the industrial vision system to ensure the regional integrity of subsequent multi-point surface parameter measurements.

[0054] The initial input is the coronal biomimetic adhesion object to be detected. The irregular or dynamic surface area to be collected is specified according to the actual application requirements. The input parameters include the preset range of spatial distribution, the physical boundary conditions of the target surface and the corresponding application constraints.

[0055] A spatial distribution range division algorithm (parameters: preset acquisition area boundary, three-dimensional coordinate system origin, coordinate axis orientation) is used to divide the target irregular shape or dynamic surface into regions and form a digital boundary description of the spatial region to be acquired.

[0056] Furthermore, by using a target area localization and spatial mask generation method (parameter: spatial boundary data), the target area on the irregular surface is masked, providing physical saliency constraints for the visual acquisition system, avoiding the influence of external parameter noise, and achieving complete spatial capture of the multi-point measurement area.

[0057] Furthermore, an industrial-grade 3D vision system (such as a high-resolution depth camera or a laser contour camera as the core, with parameters: frame rate ≥ 60Hz and spatial resolution ≤ 10μm) is used to perform area scanning on the irregular or dynamic target surface after spatial masking to acquire the original 3D point cloud data in real time.

[0058] Furthermore, by using the data flow synchronization algorithm embedded in the vision system (parameters: timestamp label, spatial coordinate synchronization compensation), spatial stitching and real-time synchronization of multi-view point cloud data are achieved, which is compatible with the periodic or non-periodic deformation features of dynamic surfaces, and obtains complete, time-series-free multi-point raw point cloud data packages.

[0059] By using the above chain derivation method, the digitization of the spatial boundary of the target area, the physical constraints of visual acquisition, and the acquisition of high-resolution three-dimensional point clouds are fully combined to ensure the spatial integrity, high precision, and dynamic compatibility of the measurement area, and to achieve the preliminary partitioning of complex irregular or dynamic surfaces and the efficient capture of original three-dimensional topographic data.

[0060] For example, in a surgical robot application scenario, data is collected and analyzed on the surface of complex, dynamically deforming human tissue (spatial distribution range approximately 30mm × 30mm, maximum curvature 10mm⁻¹, dynamic deformation speed 0.5mm / s~50mm / s). The actual tissue boundary is input using a region distribution partitioning algorithm, and spatial masking is performed using the coordinate system origin (O(0,0,0)) and principal axis directions (X-axis along the tissue cutting direction, Y / Z axes orthogonal). A structured light 3D vision system with a resolution of 8μm is selected, and the acquisition frame rate is set to 80Hz. After system acquisition and synchronization, a high-density 3D point cloud data of approximately 1.2 million valid points is obtained in a single scan, with a point cloud coverage rate exceeding 99% and a missed detection rate of less than 0.1% within the region boundary. The output point cloud raw data package is used for subsequent high-precision 3D surface reconstruction and feature sampling, achieving the technical goal of spatial integrity of irregularly shaped dynamic surfaces and effectively supporting the upstream data standard for adaptive modeling of biomimetic adhesive microstructures.

[0061] S1.2: Use multi-channel surface profile measurement sensors (such as laser scanning rangefinders and white light interferometers) to perform high-precision three-dimensional reconstruction processing on the raw point cloud data of the target surface perception, forming a three-dimensional surface morphology dataset containing information on curvature changes, surface micro-roughness, and local abrupt changes.

[0062] The raw point cloud data of the target irregular or dynamic surface, acquired by an industrial vision system, is used as input data. A multi-channel surface contour measurement sensor system, including a laser scanning rangefinder and a white light interferometer, is employed. Through high-precision multi-channel synchronous scanning, data completion and resolution enhancement are performed on each surface sensing point in the target area to form a high-density point cloud dataset. Furthermore, through image registration and point cloud spatial alignment algorithms (parameter settings: reference plane coordinates, point cloud density threshold, spatial overlap tolerance), the spatial position unification of the raw point cloud data acquired from different channels and at multiple time points is achieved, and redundant points are removed, generating highly complete 3D point cloud fusion data.

[0063] A multi-scale 3D reconstruction processing algorithm is employed to perform surface interpolation on spatially aligned high-density point cloud data according to a set reconstruction resolution and sampling window, obtaining a high-resolution 3D surface mesh model that includes overall curvature changes, local bends, and abrupt change regions. A surface micro-roughness extraction algorithm is then used to estimate the window sliding distribution of the reconstructed surface mesh model in each local region, based on parameters. (Arithmetic mean roughness) and (Root mean square roughness) is calculated as follows:

[0064]

[0065]

[0066] in, For the first Elevation values ​​of each sampling point This is the regional average. This represents the number of sampling points within the window. Based on this, the micro-roughness index of each grid cell is output.

[0067] The principal curvature is calculated using curvature analysis based on the normal vectors of the three-dimensional surface mesh nodes. and Gaussian curvature With mean curvature The following formulas are used respectively:

[0068]

[0069]

[0070] This allows us to obtain the local curvature feature distribution on each grid node, which is then used to distinguish and identify complex and irregular regions.

[0071] A mutation detection operator is introduced to mark the surface mutation feature region for the node height gradient mutation region in the mesh model by first-order and second-order neighborhood height change threshold analysis, and regard it as the priority area of ​​interest for subsequent adaptive microstructure design.

[0072] By using 3D surface morphology reconstruction and multi-dimensional feature extraction algorithms, the original point cloud data is transformed into a 3D surface morphology dataset containing information on curvature changes, surface micro-roughness, and local abrupt changes. This enables high-precision structured modeling of irregular and dynamic surfaces, providing standardized input for subsequent stress distribution acquisition at key sampling points and adaptive biomimetic modeling of microstructures.

[0073] For example, on a flexible gripper used by a medical surgical robot, the target irregular surface is a rubber sphere with a radius of 60 mm and a surface roughness varying from 0.8 μm to 4.5 μm, with a sudden height step change of 1-2 mm. A laser scanning rangefinder with a wavelength of 650 nm and a point cloud density of 1000 points per square millimeter is used. The acquisition area is 60 mm × 60 mm. After registration and alignment, a 3D point cloud fusion dataset with a total of 3.6 million points is generated. A 2 mm × 2 mm local window is used, and a sliding window algorithm is employed to calculate... In the protruding area at the top =3.5μm and the valley area =1.2μm. Curvature analysis of the surface revealed that the point of maximum principal curvature occurred at the edge of the abrupt step. =0.017 / mm, =0.012 / mm, yielding the Gaussian curvature. mean curvature The first-order height change threshold in the abrupt change region was set to 0.8 mm, and the second-order change threshold was 0.3 mm. Neighborhood analysis identified 18 abrupt change feature regions. The final output 3D surface morphology dataset fully encompasses the aforementioned point cloud, curvature, roughness, and abrupt change features, serving as a guide for microstructure parameter distribution design. Subsequent physical verification showed that the dataset's mean square error compared to the actual rubber sphere surface morphology was less than 9 μm, validating the effectiveness of the high-precision 3D reconstruction technique.

[0074] S1.3: Based on the three-dimensional surface morphology dataset, several representative key sampling points are selected, and an array of stress sensor arrays is used to obtain the original data of dynamic stress distribution at each sampling point, so as to realize the quantitative acquisition of surface load state and its changing trend.

[0075] S1.4: Perform time-series synchronization processing on the raw dynamic stress distribution data collected from the array-type stress sensor, and combine it with the surface point cloud position parameters to establish a stress-morphology correspondence data index for each sampling point in different dynamic operation stages (such as deformation and vibration).

[0076] S1.5: The three-dimensional surface morphology dataset is fused with the synchronous stress-morphology correspondence data index, and a multi-dimensional parameter calibration algorithm is used to generate a set of multi-point surface condition parameters containing different curvatures, roughnesses and dynamic motion states, providing a standardized input benchmark for subsequent surface feature normalization and biomimetic microstructure adaptive modeling.

[0077] Step S2: Normalize the collected surface morphology parameters and stress distribution information to eliminate sampling errors, and construct a set of surface feature vectors under multiple working conditions as the input basis for subsequent biomimetic microstructure modeling. Specifically, this includes:

[0078] S2.1: Perform preliminary calibration processing on the collected raw surface morphology parameter data to eliminate measurement equipment calibration errors and batch-to-batch fluctuations, and output an accurate three-dimensional morphology parameter matrix to provide basic data for subsequent normalization preprocessing.

[0079] For the raw data of the collected three-dimensional surface topography parameters, the measurement equipment error correction algorithm (parameters: equipment calibration parameter set, batch characteristic identifier) ​​is used to correct the offset introduced by equipment system error and time series drift, forming a three-dimensional point cloud dataset of one calibration.

[0080] Furthermore, a multi-source data consistency processing algorithm based on industry batch difference modeling (parameters: historical batch reference template, significant difference threshold) is used to perform batch-to-batch spatial distribution consistency verification on the first calibration data, remove significant outliers and balance the mean and variance between batches, and generate three-dimensional point cloud calibration data after batch difference correction.

[0081] Furthermore, by using a spatial topology reconstruction algorithm (parameters: point cloud density threshold, spatial neighborhood search radius), the local continuity and global topology of the multi-point cloud distribution are reconstructed, and measurement breaks or voids caused by local occlusion or sparse point clouds are corrected, resulting in a three-dimensional surface morphology calibration dataset with enhanced topological consistency.

[0082] Furthermore, a feature statistical archiving method (parameters: spatial resolution setting, data density optimization) is adopted to statistically summarize the topology-enhanced dataset according to the defined sampling region and feature parameter dimensions, forming a three-dimensional topography parameter matrix with complete structure and consistent parameters.

[0083] Through the aforementioned multi-level calibration and processing algorithms, the original topographic point cloud data from multiple devices and batches during the acquisition phase are transformed into a high-precision, highly consistent standardized three-dimensional topographic parameter matrix, thereby providing distortion-free basic data for subsequent normalization processing.

[0084] For example, in the design of a biomimetic array for surgical robot-assisted grasping of the surface of deformed organs, a batch of laser point cloud scanning data was collected. The original data was affected by scanner system offset and software synchronization delay, with the Z-axis mean offset by 23μm and spatial standard deviation of 12μm compared to the standard calibration plate. Using known equipment calibration parameters, an error correction algorithm was used to correct the Z-axis offset, adjusting the mean to within 0μm. A standard sample point cloud template from 10 historical batches was introduced, and the mean and variance of the point distribution in the central region of the current batch were compared. Points with spatial distribution deviations exceeding twice the standard deviation were removed. During spatial topology reconstruction, the lower limit of the point cloud density was set to greater than 200 points per mm², and a spatial search with a radius of 0.5mm was used to reconstruct the local sparse region, completing 32 breaks. The final output three-dimensional morphology parameter matrix contained more than 30,000 effective sampling points within a 1cm² range, with the mean drift controlled within 1μm and the spatial variance all below the target threshold. This matrix provides a standard and error-free input for subsequent Z-score normalization, enabling the comparability and high-reliability modeling of surface parameter data under different batches, equipment, and operating conditions.

[0085] S2.2: Based on the obtained three-dimensional topography parameter matrix, a normalization algorithm (such as Z-score normalization, minimum-maximum normalization, etc.) is used to unify the value range of each parameter dimension, and the three-dimensional topography parameter matrix is ​​converted into a normalized three-dimensional topography parameter matrix to eliminate the influence of dimensions and scale.

[0086] S2.3: Numerical filtering and outlier removal are performed on the collected raw stress distribution information. Sliding window averaging and median filtering schemes are used to obtain a smooth and outlier-free structured stress distribution dataset, laying the foundation for normalization processing.

[0087] S2.4: Perform normalization processing (such as normalized vector mapping) on ​​the smoothed structured stress distribution dataset to obtain the normalized stress distribution matrix, ensuring its numerical consistency and comparability with the normalized three-dimensional topographic parameter matrix.

[0088] S2.5: The normalized three-dimensional topography parameter matrix and the normalized stress distribution matrix are fused together. Through feature selection and dimensionality reduction algorithms (such as principal component analysis PCA), multi-condition surface feature vectors are generated to improve the information density and discrimination ability of the feature carrier.

[0089] S2.6: Perform multi-sample clustering analysis on the generated multi-condition surface feature vectors, using clustering algorithms (such as K-means or DBSCAN) to obtain a set of standard surface feature vectors for multi-conditions, thereby achieving parameter classification and standardization of complex irregular and dynamic surfaces.

[0090] S2.7: Perform format structuring on the multi-condition standard surface feature vector set, and output standardized, machine-readable feature vector set data to provide unified, standardized and highly informative data input for subsequent biomimetic microstructure geometric morphology modeling.

[0091] Step S3: Based on the multi-condition surface feature vector set, perform biomimetic microstructure geometric modeling to generate multi-layer gradient rigid-flexible layout parameters that adaptively match different surface curvatures and stress states, forming three-layer structural design data, including a rigid support zone, an elastic buffer zone, and a flexible coronal end layer. Figure 2 As shown, it specifically includes:

[0092] S3.1: Cluster analysis is performed on the set of surface feature vectors under multiple working conditions. A surface adaptive classification algorithm is used to group different curvatures, roughnesses and stress states into several typical surface types to obtain surface adaptive classification labels, providing a hierarchical basis for subsequent layout parameter generation.

[0093] S3.2: For surface adaptability classification labels, the response simulation of the three-layer material spatial layout is carried out using the biomimetic structural finite element modeling method. The mechanical performance target distribution parameters of the three-layer structure (rigid support area, elastic buffer zone, and flexible crown end layer) under each type of surface adaptability are obtained as parameterized inputs for morphological modeling.

[0094] Using surface adaptability classification labels and a set of standard surface feature vectors for multiple working conditions as input, a biomimetic finite element modeling (FEM) method (parameters: material constitutive parameters, initial template of three-layer spatial layout, and surface action boundary conditions) is adopted to simulate the structural response of a three-layer material spatial layout under target surface working conditions.

[0095] Furthermore, through a parameterized assignment algorithm (parameters: elastic modulus E1 of the rigid support zone, elastic buffer modulus E2, flexible crown end modulus E3, E1>E2>E3; geometric thickness H1, H2, H3), each layer of material is given guiding initial values, and the corresponding curvature, roughness and stress boundary are automatically loaded in combination with classification labels.

[0096] Furthermore, based on multi-field coupled finite element solution, interface contact mechanical boundaries (such as normal pressure, friction coefficient μ, and tangential movement constraint) are set for each type of typical surface working condition. The stress-strain field distribution of each layer in the three-layer material layout is calculated iteratively step by step, and the simulation stability is judged by numerical convergence.

[0097] The following mechanical property target distribution calculation formula is used to obtain the design parameters of each layer under specified surface conditions:

[0098]

[0099] in, For the stress of the i-th layer, Let i be the elastic modulus of the i-th layer. The local strain is obtained from the simulation calculation of the i-th layer.

[0100] Furthermore, the maximum principal strain, effective stress, equivalent stiffness, and deformation amplitude of each layer in the simulation results are evaluated, and a multi-objective optimization algorithm is used to select target distribution parameters that meet the requirements of surface self-adaptation, mechanical safety redundancy, and reversible deformation amplitude.

[0101] By responding to the structured output of the simulation parameter set, the target distribution parameters of mechanical properties of each layer obtained from the above simulation are output as parametric design inputs, providing a quantitative basis for subsequent optimization of the three-layer geometric layout, refinement of interface morphology and gradient positive change scheme, so as to realize the highly adaptable design of biomimetic microstructures in irregular / dynamic surface environments.

[0102] For example, in the structural design scenario of the minimally invasive adhesive claw tip of a surgical robot, for the three representative surfaces that have been clustered and classified (radius of curvature R=2mm / 8mm / 0.2mm, roughness Ra=5μm / 15μm / 0.1μm, typical force state is normal pressure P=30kPa / 50kPa / 10kPa), a rigid support area E1=2GPa, an elastic buffer zone E2=12MPa, and a flexible end layer E3=1MPa are selected, with thickness configurations H1=200μm, H2=300μm, and H3=100μm. A three-layer parametric microstructure model is constructed using the finite element software Abaqus, and each surface condition is used as the load boundary condition and contact interface input.

[0103] Through finite element static simulation, the local stress distribution of each layer under maximum load (rigid zone) is output. =22MPa, elastic zone =1.4MPa, flexible zone =0.08MPa), principal strain (rigid region) =0.011, elastic zone =0.12, Flexible Zone =0.09), and the maximum adaptive deformation amplitude of the coronal terminal layer under different curvatures was calculated to be 34μm, 68μm, and 2μm, respectively. A set of distributed parameters with deformation amplitudes achieving a target contour fit rate >98% was selected as the mechanical target input parameters for the three-layer structure under this type of surface condition. Finally, multiple types of simulation data were output, providing highly reliable parameter support for subsequent multi-layer gradient layout generation and CAD detailed modeling.

[0104] S3.3: Based on the distribution parameters of mechanical performance targets, a multi-layer gradient layout generation algorithm is applied to calculate the interface profile, thickness variation and spatial transition function between the three layers of materials, forming a multi-layer gradient rigid-flexible layout parameter set, and realizing the design data output of positive gradient gradual change between each layer.

[0105] Using the mechanical performance target distribution parameters output from the finite element modeling of the biomimetic structure as input, and for the predetermined requirements of the three-layer structure (rigid support zone, elastic buffer zone, and flexible crown end layer), a multi-layer gradient layout generation algorithm (parameters: target material modulus distribution of each layer, initial estimation of interface transition shape, and structural gradient planning weight) is adopted to realize the parameterized calculation of the interface contour and spatial transition relationship of the three-layer structure.

[0106] By using the spatial gradient function modeling method (parameters: initial value of layer thickness, target elastic modulus distribution function, spatial coordinate system), the thickness variation curve of the gradient transition region between each layer is solved, and a mathematical expression for the continuous change of material gradient is established.

[0107] Furthermore, by combining the interface contour fitting algorithm (parameters: surface curvature distribution, stress concentration location, interface smoothness target), the interface morphology between the rigid support area and the elastic buffer zone, and between the elastic buffer zone and the flexible end layer is refined and optimized to form a three-dimensional contour design with smooth interface curves and uniform stress transmission.

[0108] Using a spatial transition function builder, the following multi-layer transition function model is established:

[0109]

[0110] in, For height The elastic modulus of the material at that location, , These are the moduli of the upper and lower layers of material, respectively. , The start and end positions of the interface. This is a piecewise smooth transition function (such as cubic Hermite interpolation or the Sigmoid function). By implementing the above formula in sections, a spatially positive gradual change in the mechanical performance parameters of the three-layer structure as a whole can be achieved.

[0111] Furthermore, a three-dimensional finite element inversion optimization algorithm (parameters: target stress-strain curve, maximum stress tolerance of the interface, thickness variation limit) is used to perform mechanical performance matching correction on the above preliminary gradient layout parameter set, and the material gradient, interface profile and thickness distribution are iteratively adjusted to achieve the optimal rigid-flexible synergistic response of the overall structure under various typical surface adaptation conditions.

[0112] Through the aforementioned multi-layer gradient layout generation algorithm and related numerical optimization strategies, the mechanical performance target distribution parameters are efficiently transformed into interface contours, thickness variations, and spatial transition parameters of a three-layer structure, outputting a multi-layer gradient rigid-flexible layout parameter set, thereby realizing the data-driven design goal of positive gradual change in the mechanical performance gradient of each layer of material.

[0113] For example, for irregularly shaped surfaces with a radius of curvature of 1.5 mm and a surface roughness Ra of 8 μm, the input mechanical performance targets are: elastic modulus of 2.2 GPa and thickness of 12 μm for the rigid support zone; elastic modulus of 280 MPa and thickness of 42 μm for the elastic buffer zone; and elastic modulus of 18 MPa and thickness of 6 μm for the flexible coronal end layer. Assuming a smooth transition of mechanical properties between layers is required, the Sigmoid spatial transition function parameter is set to β=3.4 (adjusting the slope), and the interface start and end position parameters are... Corresponding to the thickness boundary points of the upper and lower layers, a multi-layer gradient layout generation algorithm was used to calculate the surface functions of the rigid / elastic interface and the elastic / flexible interface, with the maximum slope of the overall structural thickness variation curve controlled within 8°. After 3D finite element correction and optimization, the maximum interface stress concentration point was lower than 86% of the design tolerance (maximum value 128 MPa). The final output three-layer structural gradient rigid-flexible layout parameter set was used for 3D modeling and subsequent manufacturing processes. In performance verification tests, under dynamic loading curvature changes within a range of 2 mm and simulated target surface jitter amplitude of ±0.25 mm, the overall deformation capacity and mechanical response of this gradient structure met the steady-state adaptive adhesion index, and the output data possessed high information density and integrated manufacturability characteristics.

[0114] S3.4: Combining the multi-layer gradient rigid-flexible layout parameter set with surface adaptability classification labels, execute the rigid-flexible coupling structure morphology optimization algorithm to refine the boundaries, interface morphology and microstructure geometric details of each layer, so as to maximize the adaptive deformation amplitude of the target surface and generate refined three-layer structure morphology design data.

[0115] Using a multi-layer gradient rigid-flexible layout parameter set and surface adaptability classification labels as input, a region subdivision optimization algorithm (parameters: multi-layer interface contour set, surface adaptability labels, response weight factors) is adopted to achieve accurate positioning and contour fitting of the boundaries of each layer in the three-layer structure, forming refined data for layer boundary division under multi-surface adaptability scenarios.

[0116] Furthermore, by using a multi-scale surface modeling method (parameters: interface fitting function set, thickness variation threshold, target deformation ratio), the continuity and smoothness of each group of interface morphologies at the microscale are optimized, resulting in segmented interface smoothness data that reflects the actual deformation capacity.

[0117] Furthermore, a finite element morphology optimization algorithm (parameters: interlayer gradient modulus distribution, desired deformation vector, elastic limit constraint) is used to perform mechanical response simulation analysis on the geometry of the key interfaces of the three-layer structure. Based on the target surface adaptability label, the thickness, curvature and transition surface morphology of each layer are automatically adjusted to improve the adaptive deformation amplitude and output the optimal interface surface parameter set.

[0118] Furthermore, by combining a local microstructure geometric detail generation algorithm (parameters: representative unit array, coronal terminal boundary features, interface micro-protrusion parameters), microstructure biomimetic detail modeling is performed on the flexible coronal terminal layer. This enables iterative geometric parameter iteration of structures such as biomimetic pawls, micro-pits, or coronal protrusions at the terminal boundary, thereby obtaining microstructure geometric optimization data that maximizes contact compliance.

[0119] By integrating the joint processing flow of layer boundary refinement, interface morphology optimization, and microstructure geometric detail modeling, the multi-layer gradient rigid-flexible layout parameter set is transformed into high-resolution, mechanically positively coupled refined three-layer structural morphology design data, thereby maximizing the adaptive deformation capability of microstructures under complex irregular shapes and dynamic surfaces.

[0120] For example, in the coronal bionic array modeling scenario for surgical micromanipulation robots to grasp flexible heteromorphic internal organs, the target surface adaptive classification label is a combination of high curvature soft tissue and medium stress load; the input multi-layer layout parameter set includes a rigid support area thickness of 0.08 mm, an intermediate elastic buffer layer thickness of 0.18 mm, and a flexible coronal end layer thickness of 0.09 mm, with interlayer gradient moduli of 1950 MPa → 20 MPa → 3 MPa, and the initial interface is a nested quadratic surface.

[0121] By employing a region subdivision optimization algorithm, the three-layer boundary lines are fitted into a parabolic variable cross section according to the input high curvature template, improving the smoothness of the edge transition by 10%.

[0122] Using finite element morphological optimization, the target maximum adaptive curvature radius was set to 3.5 mm, the elastic limit deformation was 15%, and the non-uniform thickness distribution of each layer was output (the thickness at the edge of the rigid support area was reduced to 0.06 mm, and the local elastic buffer was thickened to 0.22 mm), and the maximum interface curvature adjustment was increased to 8%.

[0123] By combining a microstructure geometry generation algorithm, the terminal coronal layer is designed as a micro-claw array with a radius of 0.02 mm and an array spacing of 0.04 mm. Each claw root is provided with a micro-depression of 0.01 mm to conform to the micro-protrusions of soft tissue, thereby maximizing surface matching.

[0124] After the modeling data from this step is output, it is assembled into the subsequent micro-nano 3D modeling and manufacturing process. On-site verification showed that when grasping a high-curvature soft tissue surface, the adaptive deformation amplitude of the fine design reached 17%, the actual maximum contact area increased by 22%, and the uniformity of the structural strain concentration distribution improved by 15%, verifying the superior performance of this step solution in actual complex surface adaptability scenarios.

[0125] S3.5: Input the refined three-layer structure morphology design data into the micro-nano 3D modeling platform, and automatically generate an integrated three-layer structure CAD model using a parameterizable design library, providing an integrated three-layer structure design model output for subsequent microstructure fabrication and fluid cavity embedding.

[0126] Step S4: For the three-layer structure design data, a micro-fluidic cavity structure model is embedded, and fluid cavity layout parameters are mapped according to different surface features to achieve adaptive distribution mapping between the intermediate elastic buffer fluid cavity and the target surface morphology. For example... Figure 3 As shown, it specifically includes:

[0127] S4.1: Using the elastic buffer in the three-layer structure design data as the input object, the spatial distribution and mechanical performance parameters are extracted by the parameter identification algorithm to obtain the spatial mapping template of the intermediate elastic buffer, so as to prepare for the precise positioning of the subsequent fluid cavity embedding.

[0128] The design parameters of the elastic buffer in the three-layer structure design data are used as input objects. Through parameter identification algorithms (parameters: spatial coordinate system, material property distribution function, thickness surface equation), the multidimensional extraction of the spatial distribution parameters of the elastic buffer is realized.

[0129] Furthermore, a spatial feature analysis method based on 3D point cloud reconstruction (parameters: CAD structural surface mesh, local curvature statistics, regional thickness threshold) is adopted to automatically identify the geometric boundary of the elastic buffer zone and extract its key geodesic lines, equal-thickness slices and voxel distribution data.

[0130] Furthermore, by using a mechanical parameter inversion algorithm (parameters: elastic modulus calibration, Poisson's ratio, local stress response curve), the local and overall mechanical performance indicators (such as stiffness distribution and strain tolerance) of the elastic buffer are accurately determined based on the stress-strain output in the biomimetic finite element simulation input.

[0131] Furthermore, a spatial mapping template generation method (parameters: layered mesh subdivision, structural partition weights, geometric boundary functions) is adopted to fuse the obtained spatial distribution data with mechanical performance parameters, generating an intermediate elastic buffer spatial mapping template with spatial region labels, thickness variation, and material property distribution.

[0132] Through the above processing method, the design data of the elastic buffer is transformed into a digital spatial mapping template that can be used for subsequent embedding and distribution optimization of fluid cavities, thus realizing the basis for precise positioning of the elastic buffer fluid cavities in terms of point, shape, and region.

[0133] For example, in the design scenario of microstructures for flexible grasping robots, a 3D CAD design file of a three-layer structure conforming to the surface of a high-curvature organ is input for an intermediate elastic buffer with parameters of 42μm thickness and a spatial projected area of ​​0.36mm². A 3D point cloud reconstruction method is used to mesh the surface of the elastic buffer, outputting 5634 mesh nodes with a 2μm spacing between equal-thickness slices, resulting in a 21-layer dataset of equal thickness. Using a mechanical parameter inversion algorithm, the material elastic modulus E2 = 280MPa and Poisson's ratio 0.47 are loaded. Finite element simulation is used to obtain the principal strain and stress distribution corresponding to each node and map them to structural partition blocks, each with an area of ​​0.02mm². A spatial mapping template generation algorithm is used to fuse the above thickness-mechanical parameters with the spatial distribution, outputting a 42×21 structural partition elastic buffer mapping template, generating a spatial region label set and a local thickness-stress distribution index. In practical applications, this template is used for virtual positioning of multi-point fluid cavity embedding, ensuring precise coupling between the subsequent fluid cavity layout and the deformation capacity of the elastic buffer, thereby improving the adaptive matching performance of the microstructure. The test data shows that, based on the spatial mapping template output in this step, the fluid cavity embedding error is less than 3μm and the elastic buffer deformation amplitude control error is less than 4%, laying a precise parameter foundation for subsequent dynamic fluid control.

[0134] S4.2: Input the set of surface feature vectors under multiple working conditions into the fluid cavity structure modeling unit, calculate the initial distribution parameters of the micro fluid cavity in the elastic buffer based on the surface feature intelligent mapping algorithm, and realize the mapping relationship between cavity position, geometry and surface contour, and pressure distribution.

[0135] The cross-section of the microfluidic cavity is circular, elliptical, or irregular curved, and its spatial distribution corresponds to the mechanical property distribution of the elastic buffer and the pressure distribution on the target surface.

[0136] The elastic buffer space mapping template and the set of surface feature vectors under multiple working conditions output from the previous sub-step are used as inputs to calculate the shape mapping and the initial layout parameters of the fluid cavity.

[0137] A surface feature intelligent mapping algorithm (parameters: spatial mapping template, surface feature vector, pressure distribution index, fluid cavity layout rules) is adopted to quantitatively map the multidimensional feature vector of the surface to the internal spatial coordinates of the elastic buffer based on the curvature, roughness and local stress state of the target surface.

[0138] Furthermore, through the intelligent layout generation function, representative surface features (such as the maximum curvature point, stress concentration area, and surface abrupt change location) are projected onto the spatial index of the elastic buffer mapping template to realize the fixed-point identification of feature regions. Based on the local thickness tolerance and mechanical deformation requirements, the candidate embedding region of the micro fluid cavity is initialized.

[0139] Furthermore, a constrained geometric parameterization method (parameters: minimum / maximum cavity size, distance between adjacent cavities, safety boundary setting, etc.) is adopted to automatically generate preliminary geometric templates for cavities within the candidate region, and output the spatial coordinates, cross-sectional dimensions, axis orientation, and distribution density parameters of each cavity.

[0140] Furthermore, by combining the original surface pressure distribution data and using a local pressure mapping function, a strong correlation is established between high-pressure areas and cavity embedding schemes. The embedding locations are then optimized and selected using the following formula:

[0141]

[0142] in, For the first The comprehensive selection score of each candidate cavity For surface curvature, The parameters for the high stress region at this location are... For regional spatial redundancy, These are the weighting coefficients.

[0143] Based on the comprehensive score, several cavity layout points are selected from high to low to avoid repeated embedding of cavities in areas that are too dense or weak in stress, and the cavity index data that overlaps in space or conflicts with the elastic area is automatically corrected.

[0144] Through the above mapping and filtering process, the initial distribution parameters of the microfluidic channels in the elastic buffer space are output, including the three-dimensional geometric parameter set of each channel, the mapping set with surface features, and preliminary interval pressure distribution suggestions, which serve as the input basis for subsequent finite element simulation and iterative layout optimization.

[0145] The elastic buffer is made of an elastomeric material, which is selected from silicone rubber, polyurethane or similar elastic polymers, and the elastic modulus is distributed in a gradient distribution that gradually decreases in the thickness direction.

[0146] For example, in the adaptive biomimetic adhesive microstructure design for the end face of a minimally invasive surgical instrument with a curvature radius of 1.5 mm and a roughness of Ra 8 μm, the input consists of a spatial structure mesh of an elastic buffer layer with 21 layers of equal thickness, each 2 μm thick, and 12 sets of representative surface feature vectors (curvature, roughness, and maximum stress) for various working conditions. A surface feature intelligent mapping algorithm is used to map the curvature peak and the maximum stress point to the high deformation capacity region of the 5th-7th layers of the elastic buffer layer, initially selecting 9 candidate cavity projection points. A cavity comprehensive optimization formula is then used to set... , , The comprehensive score for each point was calculated (highest score 0.87, lowest score 0.59), and five cavity distribution points with scores greater than 0.75 were retained. Then, based on a minimum center-to-center distance of 10 μm and a minimum wall thickness of 1.5 μm between adjacent cavities, safety boundary correction was performed, automatically adjusting two overlapping points caused by insufficient spatial redundancy, and outputting the final cavity layout parameters. This step outputs five sets of three-dimensional coordinates on the elastic buffer mapping template, the cross-sectional diameter of each cavity (2.4 μm to 3.1 μm), the orientation determined by the surface normal and the direction of maximum stress, and includes the preliminary design pressure target range for each cavity (e.g., 20–60 kPa), providing strong data support for finite element simulation and subsequent structural forming. In actual trial production, the initial cavity distribution scheme achieved a spatial overlap of over 97% with the target surface contour, and the maximum deformation deviation in the pressure response area was controlled within 6%, fully verifying the technical feasibility and superior accuracy of this step.

[0147] S4.3: Based on the initial distribution parameter information of the fluid cavity obtained in S4.2, the finite element structural simulation technology is applied to simulate the influence of the micro fluid cavity on the mechanical response of the entire three-layer structure under different deformation conditions, and optimize the geometric layout of the micro fluid cavity to ensure the structural integrity and deformation capacity of the microstructure under dynamic response.

[0148] Using the initial distribution parameters of the microfluidic cavity and the space mapping template of the elastic buffer obtained by S4.2 as input, the multiphysics finite element structural simulation method (parameters: three-dimensional geometric data of the elastic buffer, geometric shape of the microfluidic cavity, nonlinear constitutive parameters of the material, layered boundary stress scenario, fluid-solid coupling boundary conditions) is adopted to realize the multi-field simulation calculation of the mechanical response of the entire three-layer structure of the microfluidic cavity under different typical deformation conditions (such as compression, bending, shear and surface bonding dynamic evolution).

[0149] Furthermore, by setting multiple sets of microfluidic cavity geometric variables (including cavity cross-sectional radius, directional curvature, distribution spacing, etc.) and fluid internal pressure loading parameters within the finite element model, a full-factor simulation experiment was designed to obtain a cavity parameter-structure response mapping dataset. Using Von Mises equivalent stress, local principal strain, and maximum allowable deformation amplitude as evaluation indicators, key response data affecting microstructure integrity and adaptive deformation were extracted.

[0150] Furthermore, based on the structural response index output by simulation, a multi-objective optimization algorithm (such as genetic algorithm or multi-objective particle swarm optimization algorithm, with parameters: cavity geometric variables, fluid pressure variables, structural bearing capacity and target deformation coupling weight) is used to iteratively optimize the cavity geometric parameters and their spatial distribution, generating an optimal microfluidic cavity geometric layout scheme that takes into account both the adaptive deformation capability of the drive and the overall structural mechanical safety.

[0151] Furthermore, by re-simulating the optimized cavity layout scheme under different typical surface adaptive conditions, the following joint criterion formula for structural integrity and adaptive performance is used for judgment:

[0152]

[0153] in, For comprehensive performance indicators, To simulate the maximum deformation, Match the deformation amount to the target contour. For the local maximum equivalent stress, For the material's allowable ultimate stress, To obtain the maximum effective contact area from the simulation, For the required contact area, It is an adjustable weighting factor.

[0154] The best output by optimizing the algorithm The corresponding microfluidic cavity geometry is determined to define parameters such as the cavity cross-section, orientation, distribution range, and spatial topology within the elastic buffer zone.

[0155] Through the above finite element simulation and multi-objective optimization, the initial fluid cavity distribution parameters are transformed into optimal micro-cavity geometric layout parameters that take into account both the mechanical performance of the rigid-flexible coupling structure and the adaptive deformation response under fluid drive, thereby realizing the intelligent design of the three-layer gradient structure internal cavity system under irregular and dynamic surface conditions.

[0156] For example, in a scenario involving adaptation to a medical irregular surface with a radius of curvature of 5 mm, a local surface dynamic deformation of 100 μm, and a normal loading pressure of 40 kPa, the input parameters are: an elastic buffer thickness of 120 μm, an initial fluid cavity design of three parallel spirals, an initial cavity radius of 18 μm, a spacing of 38 μm, and a maximum internal fluid pressure of 80 kPa. Finite element simulation parameters are set as follows: elastic modulus of the elastic region 35 MPa, Poisson's ratio 0.42, and a fully sliding boundary at the fluid-solid coupling interface. Simulation is performed using the ABAQUS-CFD co-coupling module with an iteration step size of 5 μs and a total of 500 steps. The maximum Von Mises stress, principal strain, maximum fit deformation, and contact area of ​​the structure under different cavity parameter configurations are extracted. A multi-objective optimization algorithm is used, setting the cavity radius range to 10–24 μm, the spacing to 25–50 μm, and the internal fluid pressure to 50–90 kPa, evaluating each combination. The optimal cavity layout was ultimately obtained: a spiral distribution with a radius of 22 μm and a spacing of 40 μm, and a fluid working pressure of 75 kPa. Simulation output showed a maximum fitting deformation of 117 μm, a principal strain of 0.143, a maximum Von Mises stress of 28 MPa at the interface (less than the material limit of 35 MPa), and an effective contact area increase of 17%. This layout was confirmed as the cavity scheme with excellent structural integrity and adaptive deformation performance under current irregular dynamic surfaces. Detailed geometric parameters were output for integration into the unified design model using S4.4, and the application results achieved the specified performance targets.

[0157] S4.4: Based on the finite element simulation optimization results, the final determined fluid cavity structure model is seamlessly integrated into the three-layer structure design data using the embedded CAD parametric modeling method, and the output is a microstructure parameter model that integrates the global gradual rigid-flexible layout and the local micro fluid cavity.

[0158] S4.5: Perform adaptive distribution mapping verification on the microstructure parameter model with integrated fluid cavities, use surface feature response analysis technology to evaluate the adaptability of micro fluid cavities driven by surface features under multiple working conditions, output adaptive distribution identification results, and provide complete parametric microstructure design data for the next step (physical entity manufacturing).

[0159] Step S5: Using auxiliary micro / nano 3D printing parameters and layer-by-layer transfer, filling, and curing process data, and according to the three-layer structure design data and microfluidic cavity design information, a multi-layered gradient rigid-flexible coupled crown-shaped biomimetic adhesive microstructure is prepared to achieve a spatially positive gradient of the mechanical properties of different material layers. Specifically, this includes:

[0160] S5.1: Using the three-layer structure design data as input, the reverse layer modeling algorithm is used to generate a microscale geometric digital model of the bottom rigid support area, the middle elastic buffer and the flexible coronal terminal layer, and obtain a multi-layer three-dimensional digital structure model containing the structural dimensions of each layer, the gradient boundary and the preset micro-cavity embedding position, which is used for subsequent manufacturing path verification.

[0161] S5.2: Perform slicing process parameter generation based on material property partitioning on the three-layer three-dimensional digital structure model. Use gradient material parameter mapping algorithm to obtain multi-material 3D printing process files corresponding to each structural region, including process parameter files for printing path speed, supporting material distribution and cavity region auxiliary structure, to prepare for subsequent printing with multi-layer mechanical property space positive gradient.

[0162] The multi-layer three-dimensional digital structure model output by S5.1 is used as the input object. The input data includes the spatial geometric parameters of the bottom rigid support area, the middle elastic buffer zone and the flexible coronal terminal layer, the partition labels, the preset micro fluid cavity embedding positions and the material property settings of each region.

[0163] By adopting a partitioning and slicing strategy (parameters: hierarchical partitioning labels, spatial projection coordinates, structural boundary functions), the three-dimensional digital structural model is processed into independent slices according to material properties and spatial functional requirements, thereby realizing the partitioning of material printing process for each structural sub-region and obtaining a partitioned slice set.

[0164] Furthermore, through a gradient material parameter mapping algorithm (parameters: regional elastic modulus, thickness gradient, stiffness-flexibility transition function), corresponding 3D printing material parameters are mapped for each structural partition, including printing material type, target density, mechanical property gradient coefficient, and process control variables, generating a parameterized material property mapping table.

[0165] Furthermore, a multi-material 3D printing slice file generation method is adopted (parameters: printing path, step thickness, cavity support structure shape). Multi-material printing path files (G-code or equivalent process files) are generated using partitioned slice sets and material property mapping tables. Specifically, the process parameter files include partitioned printing path speed, variable layer height, amount of support material in different regions, and local auxiliary structures required for fluid cavity embedding.

[0166] Furthermore, by using printing process path simulation and collision detection technology (parameters: printhead motion trajectory, auxiliary structure generation rules, cavity morphology constraints), the feasibility of multi-material printing files in the actual manufacturing process is verified, the consistency of path overlap, auxiliary support structure and cavity connection in fluid cavity embedded area is corrected, and finally an optimized set of process control files that can be directly prepared is output.

[0167] Through the above-mentioned chain-like partitioning mapping and parameterized process file generation, the material distribution, micro-cavity layout and control logic of the three-layer digital model are integrated and mapped with the 3D printing equipment, ensuring the accurate execution of the subsequent multi-layer spatial forward gradient solid printing stage.

[0168] For example, for a deformable crown-shaped adhesive array with a bending radius of 2 mm, the bottom rigid support area has a thickness of 40 μm and a designed elastic modulus of 1100 MPa; the middle elastic buffer layer has a thickness of 90 μm and an elastic modulus of 45 MPa; three microfluidic channels are preset; the minor axis of the cross-section ellipse is 12 μm; and the flexible crown-shaped end layer has a thickness of 5 μm and an elastic modulus of 1.3 MPa. Using a partitioning and slicing strategy, the input 3D model is divided into three regions: rigid, elastic, and ultra-flexible. The thickness of the sliced ​​layers in each region is set to 1 μm, 2 μm, and 0.5 μm, respectively. An in-plane mapping algorithm is used to output a partition label matrix. Gradient material parameter mapping is employed: high-modulus polymer A is assigned to the rigid region; semi-rigid polymer B is assigned to the elastic region; low-modulus silicone rubber C is used for the flexible end layer; and soluble material D is assigned to the cavity auxiliary structure. A material property table is then output. When generating the G-code file, the printing speed was set to 8 mm / s for the rigid zone, 7 mm / s for the elastic zone, and 6 mm / s for the flexible end layer. Auxiliary support strips were automatically added to the micro-cavity region, with a cavity filling density of 40% to ensure smooth subsequent filling processes. Through printing simulation analysis, the auxiliary structure in the cavity inlet overlap area was adjusted to avoid path overlap, ultimately outputting a multi-material 3D printing process file for solid manufacturing. The multi-layer gradient rigid-flexible structure printed using the above process file, after microstructure testing and verification, showed that the interface thickness error of each layer was less than 1 μm, the cavity penetration rate was 100%, and the material performance conformity of each zone was better than the design value by 5%, achieving the technical effect of this step in preparing a high-precision process for multi-layer gradient rigid-flexible coupled solids.

[0169] The surface of the flexible coronal terminal layer is provided with arrayed biomimetic microstructure details, which are coronal protrusions, micropits or pseudopodia, with a single unit diameter of 1 to 200 micrometers, preferably 10 to 50 micrometers, and an array spacing of 5 to 300 micrometers, preferably 20 to 100 micrometers.

[0170] S5.3: Based on the printing process document of the underlying rigid support area, a high modulus rigid polymer material is selected, and the underlying structure is manufactured through micro-nano 3D printing technology to obtain a rigid support area microstructure containing cavity interface reservations, providing a basic entity and positioning reference for the embedded integration of intermediate elastic buffer and micro fluid cavity.

[0171] The rigid support region is made of a polymer with an elastic modulus between 1 GPa and 5 GPa, preferably polyetheretherketone (PEEK) or a similar high-modulus polymer.

[0172] Using a three-layer 3D digital structural model that has been partitioned and sliced ​​according to material properties as input, the 3D modeling parameters and process documents of the bottom rigid support area are used as the basic data for solid manufacturing.

[0173] A high-precision micro-nano 3D printing method (parameters: high modulus rigid polymer material, micron-level printing resolution, specified printing path speed, auxiliary support structure configuration) is adopted to achieve the layer-by-layer stacking of the bottom rigid support area and ensure that a solid interface matching the intermediate elastic buffer and fluid cavity interface is formed at the preset structural position.

[0174] Furthermore, by setting process document parameters (parameters: structural grid size, printing layer thickness, cavity interface reserved cavity geometry data), the print head is controlled to perform void retention molding at a predetermined position, thereby achieving accurate pre-setting of cavity ports and positioning grooves within the rigid support area, providing auxiliary positioning references for subsequent transfer of elastic buffer material and embedding of fluid cavity templates.

[0175] Furthermore, a process monitoring feedback algorithm (parameters: real-time displacement monitoring sensor data, printing layer self-correction algorithm, and process deviation allowable threshold) is adopted to dynamically adjust the printing path and positioning accuracy, ensuring the spatial consistency between the actual manufactured entity and the three-dimensional digital model, and automatically compensating for material shrinkage or local stacking errors.

[0176] Furthermore, by verifying the interface dimensions, positioning groove geometry, and cavity port accuracy of the printed structure (parameters: laser confocal microstructure parameter detection, three-dimensional contour scanning), quality assessment data of the manufacturing process is generated, providing acceptance criteria for subsequent multi-layer material integration.

[0177] Through the aforementioned micro-nano 3D printing manufacturing process, the preceding process files and digital structural models are transformed into three-dimensional entities with high-precision positioning grooves, cavity interfaces, and standard-sized rigid support areas. This enables reliable spatial docking and embedding of the downstream elastic buffer zone with the micro-fluidic cavity, thereby ensuring the spatial accuracy and structural integration effect of the multi-layer gradient structure fabrication.

[0178] For example, in the manufacturing scenario of the basic unit of the coronal biomimetic adhesive array for a medical abbreviated grasping robot, the thickness of the rigid support area is set to 120 μm, the lateral dimension to be 2 mm × 2 mm, the target material to be high-modulus polyetheretherketone (PEEK), the printing resolution to be 1 μm, and the printing path to be a honeycomb woven structure. Using a micro-nano 3D printing device, the cavity interface area, marked as 0.18 mm × 0.09 mm × 0.12 mm in the structural model, is left empty and formed, with the spatial layout error of the positioning groove kept within 2 μm. During the monitoring process, a laser confocal microscopy imaging system is used to reconstruct the three-dimensional contour of the structural interface after each layer is processed, and the local contour error is detected in real time to be less than 3 μm. The final output is a rigid entity that meets the integration requirements of the elastic buffer and fluid cavity. After subsequent interlayer integration, the cumulative misalignment error of the overall structure is less than 5 μm, and the cavity interface pathways are 100% connected, meeting the stringent requirements of spatial precision manufacturing and effectively improving the controllability of the biomimetic array's overall adaptation to abbreviated surfaces and the structural integration strength.

[0179] S5.4: Based on the solid printed in the rigid support area, the elastomeric material of the intermediate elastic buffer is transferred sequentially using the layer-by-layer transfer technology, and a prefabricated template of fluid cavity is embedded inside it. Through microstructure embedding to assist positioning, the micro-cavities are accurately distributed in the elastic buffer space, creating conditions for the preparation of the fluid response region.

[0180] S5.5: After the intermediate elastic buffer is transferred and formed, the fluid cavity filling material is injected and cured. The embedded template is removed by template dissolution or selective ablation process, thereby obtaining a micro fluid cavity with a dense structure, unobstructed channels and controllable layout, realizing the overall forming of the coupling between the elastic buffer and the fluid cavity.

[0181] S5.6: Based on the integrated structure of the bottom rigid support area and elastic buffer zone-fluid cavity, the micro-nano 3D printing of the top flexible coronal end layer is performed. Low modulus flexible polymer material is used. Through high resolution and layer-by-layer stacking, a flexible coronal end microstructure that is precisely connected with the bottom structure is formed, and a complete structural closed loop with gradient rigid-flexible space positive gradient is completed.

[0182] S5.7: After the multi-layer gradient rigid-flexible coupled coronal biomimetic adhesive microstructure is completed, the uniformity of gradient material distribution, fluid cavity unobstructedness and coronal microstructure forming accuracy are evaluated based on mechanical performance testing and microstructure characterization methods. Based on the test feedback, the 3D printing process parameters and layer-by-layer transfer template design are closed-loop corrected to output a standardized multi-layer gradient rigid-flexible coupled coronal biomimetic adhesive microstructure that meets the requirements of subsequent array integration.

[0183] Step S6: The prepared coronal biomimetic adhesive microstructure is distributed and integrated to construct a controllable array unit. Simultaneously, a microfluidic system channel is configured for external fluid supply and pressure regulation, establishing an initial parameter set for fluid cavity control. Specifically, this includes:

[0184] S6.1: Modularize the physical characteristic parameters of the prepared multi-layer gradient rigid-flexible coupled crown-shaped biomimetic adhesive microstructure module, and generate a spatial arrangement matrix and unit interconnection configuration table through an array topology design scheme for use as target input for subsequent distributed integration operations.

[0185] S6.2: Based on the spatial arrangement matrix, the array layout assembly of microstructure modules is performed. High-precision positioning and alignment technology is used to achieve seamless coupling connection between units, obtain the physical model of the distributed array, and provide a structural foundation for the integration and docking of flow control channels.

[0186] The array topology spatial arrangement matrix and the unit interconnection configuration table are used as input objects. The input data includes the physical characteristic parameters of the standardized multi-layer gradient rigid-flexible coupled crown-shaped biomimetic adhesive microstructure module and the target array unit layout position information.

[0187] A high-precision picking and automatic positioning assembly method (parameters: microstructure module identification code, array substrate positioning coordinates, and six-degree-of-freedom pose alignment algorithm) is adopted to realize the automatic identification and assembly request of the local spatial coordinates of each microstructure module and the preset unit position in the target array arrangement matrix.

[0188] Furthermore, through submicron-level three-dimensional alignment technology (parameters: high-resolution machine vision-assisted system, laser displacement sensor, dual-axis servo micro-manipulation platform), precise measurement of spatial position error between modules is achieved, and the module attitude is automatically corrected in six degrees of freedom to ensure that the static positioning error between units is less than the set thresholds Δx, Δy, and Δz (e.g., Δx = 1 μm).

[0189] Furthermore, a face-to-face docking and lateral limiting auxiliary structure design is adopted (parameters: gradient steps at the module connection interface, self-centering limiting groove at the boundary). Under the action of linear propulsion and lateral clamping force field, the interface structure between adjacent units is self-positioned and coupled. Through mechanical compliance limiting and shape locking, a seamless docking effect is achieved, and the contact signal of the coupling interface is sensed in real time.

[0190] Furthermore, relying on the ultrasonic-assisted micro-docking pressurization platform (parameters: ultrasonic frequency, pressurization force F, pressurization time T), vibration-assisted micro-pressurization is applied to the coupling interface to achieve leveling and curing of the interface bonding layer (such as micro-nano gap reactive adhesive or molecular self-assembly interface agent) in a short time, so that a continuous mechanical connection body is formed between the modules, and its interface connection strength and microstructure integrity meet the technical requirements of the subsequent process integration of the flow control channel.

[0191] Through the above-mentioned multi-level high-precision positioning, six-degree-of-freedom attitude automatic correction, docking interface mechanical coupling and ultrasonic-assisted curing chain process, the multi-layer gradient rigid-flexible coupled crown-shaped biomimetic adhesive microstructure module is assembled into a distributed array physical model, realizing seamless coupling between modules and regular spatial arrangement, laying a reliable structural foundation for subsequent precise integration of flow control channels and array-level fluid dynamic response.

[0192] For example, when constructing a 40×40-unit crown-shaped biomimetic adhesive microstructure array, the spacing between units is 60 μm. The arrangement matrix automatically constrains the boundaries according to the spatial requirements of the task area, and the units in the central area must be compatible with dynamic lateral expansion and contraction functions. Each unit module is numbered using a two-dimensional spatial index, and the target arrangement coordinates are determined by high-resolution machine vision during assembly, achieving a pose control accuracy of ±0.8 μm. The lateral limiting groove width is 20 μm, the gap is controlled to be less than 0.5 μm, the ultrasonic-assisted pressurization frequency is set to 32 kHz, the pressure is 0.25 N, and the pressurization time is 2 s, using a reaction-curing micro-interface agent. After assembly, the maximum gap at the array interface is less than 0.3 μm. Interface scanning and structural acoustic analysis verify that the mechanical strength of the seamless coupling area interface reaches more than 80% of the parent structure, the center-to-center distance error of each unit in the array physical model is less than 1 μm, and the spatial arrangement matrix perfectly matches the theoretical model. In subsequent integration of the flow control channel, the array achieved 100% connectivity between the fluid port and the array, and the response time delay of the multi-point synchronous fluid empowerment link was less than 500 μs. The assembly performance met the system integration requirements for dynamic capture and controllable adhesion of highly adaptable irregular surfaces.

[0193] S6.3: Based on the distributed array physical model, an array-level flow control interface mapping table is established. Through precision micromachining technology and microchannel integration process, the micro fluid cavity port of each adhered microstructure module is connected to the corresponding channel of the external micro flow control system to obtain the software and hardware collaborative framework of the flow control system.

[0194] S6.4: For the flow control interface mapping table, perform functional testing of the fluid supply path, correct the initial configuration of the micro flow control system channels through standard flow calibration and pressure response verification, realize the physical matching parameter set of multi-channel synchronous liquid supply and pressure distribution, and provide basic physical boundary conditions for the preliminary cavity control algorithm.

[0195] S6.5: Combining the array physical model with the software and hardware collaborative framework of the flow control system, the array unit multi-channel control parameter initialization is performed. The initial volume, pressure threshold and response sensitivity of each micro fluid cavity are calibrated using a distributed control algorithm. Finally, the array-level fluid cavity control initial parameter set is output as the input for the subsequent dynamic adaptive fine control algorithm.

[0196] Step S7: Based on the channel input of the microfluidic system, the contact state between the array unit and the target complex surface is monitored in real time. Multi-point feedback, including contact pressure and morphology sensing data, is input into the adaptive fluid control algorithm to achieve directional pressure and intensity control of the fluid cavity. Specifically, this includes:

[0197] S7.1: Sample and acquire the fluid pressure signal output from the microfluidic system channel to achieve real-time acquisition of the basic operating parameters of the fluid cavity inside the coronal biomimetic adhesion array unit.

[0198] S7.2: Based on a multi-point arrangement of contact pressure sensors, it acquires contact pressure distribution data between the array unit and the complex surface of the target in the contact area, serving as the basic signal acquisition for the surface contact state.

[0199] S7.3: Using topography sensing technology, data is collected on the micro-contour changes in the contact area between the array unit and the target surface, thereby obtaining real-time surface topography parameters and processing them together with pressure sensing data.

[0200] An integrated high-resolution surface topography sensor array (such as a white light interferometer, a laser confocal sensor, and a micro cantilever beam micro-profile detection device) is used to collect the micro-three-dimensional surface profile of the contact area between the coronal biomimetic adhesion array unit and the complex surface of the target in real time.

[0201] Using a synchronously triggered data acquisition mechanism, a complete surface height matrix is ​​obtained for each contact area according to the dynamic temporal sequence of the contact state between the array unit and the target surface. And the rate of change of local micro-undulations, forming high-density point cloud data with spatial resolution better than the micro-structural unit scale (e.g., <5μm).

[0202] The multi-channel topography sensor signal digitization processing module employs an adaptive sliding window filtering algorithm and a spatial outlier correction mechanism to remove external measurement noise and occasional small particulate interference, outputting a noise-suppressed purified topography dataset. This will improve the practical usability and accuracy of topographic data.

[0203] Based on a structured topography dataset, spatial gradient calculation and curvature extraction algorithms are performed to extract the local normal curvature of the contact area. Distribution of concave and convex extreme points and roughness index The microscopic geometric parameters are quantized to construct a topography parameter matrix that corresponds one-to-one with the physical location of the array elements:

[0204]

[0205] A multi-dimensional time synchronization mechanism is employed to synchronize the real-time surface topography parameter matrix with the multi-point contact pressure distribution data from the previous step. Perform temporal alignment and establish point-to-point spatial-temporal mapping relationships based on the data index table to achieve multi-dimensional joint characterization of the actual contact state between the array unit and the target surface.

[0206] By using multi-channel data fusion and feature matching algorithms, the purified morphology parameters and contact pressure data are combined and transformed into a standardized time-series feedback dataset, enabling a quantitative description of the local contact microenvironment of microstructure array units on complex irregular or dynamic surfaces.

[0207] For example, in a scenario where a micromanipulation robot grasps the surface of an irregularly shaped biological tissue, a three-dimensional white light interferometric topography sensing module is selected, with a data acquisition frame rate as high as 200Hz and a spatial sampling interval of 2μm. The 1024×1024 point height matrix obtained from a single frame in the contact area of ​​the array unit is processed using a sliding window (5×5) mean filter to remove burst noise. A gradient algorithm is used to calculate the spatial curvature of each sampling point to obtain the maximum normal curvature. 0.25μm -1 roughness The thickness is 1.2 μm. It is then aligned with the contact pressure distribution matrix (output in real-time by a 32×32 pressure sensing unit) via an internal bus synchronization timing marker, for single-point... A joint characterization vector is constructed to fully describe the actual contact microenvironment at a given point in time, serving as the standardized input for the subsequent core algorithm of adaptive fluid regulation. When abrupt changes in irregular regions are captured, the morphology parameter set can dynamically reflect surface changes. The time-series feedback data output by the system is reduced from the original spatial information to highly correlated input features, significantly improving the sensitivity and accuracy of adaptive regulation. In actual testing, under the condition that the maximum dynamic surface change frequency in the gripping area is 30Hz, the input delay of the joint characterization of morphology and pressure data is less than 5ms, ensuring rapid and precise response of the microstructure and effective adhesion control.

[0208] S7.4: Synchronously align and preprocess the collected contact pressure distribution data and surface morphology parameter data to generate multi-channel time-series feedback feature vectors, which will be uniformly standardized for the input of the subsequent adaptive fluid regulation algorithm.

[0209] S7.5: Input the standardized multi-channel time-series feedback feature vector into the adaptive fluid control algorithm, and analyze the difference between the current state of each microfluidic cavity and the target contact pressure through the optimization algorithm to generate directional pressure control command parameters.

[0210] S7.6: Based on the pressure control command parameters output by the adaptive fluid regulation algorithm, the microfluidic control system is closed-loop regulated to realize the directional dynamic distribution of local pressure inside the microstructure cavity, so as to regulate the local rigid / flexible response of the array unit and realize feedback control for dynamic matching of microstructure in complex surface environment.

[0211] Step S8: Based on the output of the adaptive fluid regulation algorithm, the state of the fluid cavity inside the microstructure is dynamically adjusted, causing the multi-layered gradient rigid-flexible microstructures at different locations to undergo real-time deformation in response to the current surface, thereby accurately matching the target surface contour and achieving maximum effective contact and adhesion. Specifically, this includes:

[0212] S8.1: Analyze the output parameter set of the adaptive fluid regulation algorithm to extract the target fluid pressure distribution characteristics under the current surface contour mapping of different array units, generate the target parameter set for dynamic regulation of the fluid cavity, and provide parameter input for subsequent fluid cavity state adjustment.

[0213] S8.2: Based on the target parameter set for dynamic control of fluid cavities, the multi-point parallel pressure / flow regulation operation of the main control unit of the microfluidic system is executed, driving the microfluidic cavity embedded in the intermediate elastic buffer to perform spatially distributed pressure dynamic empowerment, thereby realizing the real-time reconstruction of the fluid distribution inside the cavity.

[0214] S8.3: By using an embedded deformation sensor module, the real-time local deformation state of the multi-layer gradient rigid-flexible microstructure under the filling and discharging regulation of the fluid cavity is monitored, and the local deformation variables and the deviation vector from the desired matching contour are extracted to realize continuous quantitative evaluation of the adaptive deformation behavior of the structure.

[0215] Based on the target parameter set of dynamic control of fluid cavity, the numbering mapping and spatial distribution information of target multi-layer gradient rigid-flexible microstructure unit are obtained, and the real-time monitoring initialization parameters of each embedded deformation sensor module are set to realize local deformation sampling for the current fluid cavity energized area.

[0216] High-sensitivity strain gauge arrays, flexible piezoresistive membranes, or MEMS miniature triaxial stress sensors are embedded in key regions of the three-layer structure in a spatially distributed manner. Combined with the three-dimensional geometric model of the biomimetic microstructure, the local strain, normal displacement, and shear deformation signals of each sampling point are output in real time, forming high-resolution real-time raw observation data of the deformation field inside the microstructure.

[0217] By using a multi-channel synchronous acquisition algorithm (parameters: sampling period Δt, number of channels N, sampling resolution δ), the deformation data of the microstructure during the filling and discharging adjustment process of the fluid cavity is acquired in parallel. The sensor output of each channel at each moment is mapped one-to-one with the spatial distribution of the array unit to form a multi-level, regionalized original matrix of deformation.

[0218] Furthermore, based on the microstructure CAD design model, the acquired multi-channel local deformation variables are compared with the preset target surface contour reference parameters in real time. A spatial rigid-flexible coupling deformation reconstruction algorithm is used to calculate the actual deformation variable ΔL at each monitoring point. i(i=1,2,...,N) The expected deformation of the contour matching the target The difference between them yields the deviation vector: in, Let i be the real-time deformation of the i-th monitoring point. Let the expected profile shape variable be at this point. To address local fitting bias, a moving average and Gaussian filtering algorithm within a continuous time window is used to perform temporal filtering on the original deformation variables and bias vectors, removing high-frequency noise and instantaneous abnormal fluctuations. Based on a set fitting degree threshold, a quantitative evaluation index of the continuous deformation of the current microstructure to the target contour is output, and the bias vector matrix is... This serves as the closed-loop control input for downstream adaptive pressure correction and contour dynamic iterative optimization. Through the aforementioned collaborative monitoring algorithm based on an embedded sensor array, a full-process quantitative evaluation of the local and overall adaptive deformation behavior of multi-layered gradient rigid-flexible microstructures under real-time fluid cavity adjustment is achieved, providing high-precision data support for continuous dynamic optimization of microstructure pressure distribution and contour fitting. For example, for a set of multi-layered gradient rigid-flexible crown-shaped biomimetic microstructures composed of a 12×12 unit array, a high-sensitivity strain gauge array (250Ω resistance coefficient, 10ms sampling period, ±1000με measurement range) and a MEMS piezoresistive sensor (0.5mV / V / kPa sensitivity, 0.1kPa resolution) are integrated in the interface region between the rigid and elastic layers of each unit structure. Local deformation is continuously collected within the fluid cavity filling and discharging range (0 ~ 200μL, maximum output pressure 20kPa). Assuming the target surface contour elevation variation range is Δh = 0 ~ 350μm, the expected deformation of each unit is obtained based on the morphology reconstruction model. ΔL acquired by real-time sensor array i Through the above formula and To perform point-by-point comparison, the deviation vector is output.

[0219] For a single deformation response of a 12×12 array, the deviation matrix The mean square error is less than 12 μm, and the proportion of cells exceeding the predetermined threshold is less than 2%. Real-time noise suppression is achieved using a sliding window width of 50 ms and a Gaussian filter with σ=7, and a continuous contour fitting rate index is output. Ultimately, the system achieves high-resolution quantization of the local adaptive deformation of multi-layered gradient rigid-flexible microstructures under rapid fluid regulation, accurately supporting subsequent full-array dynamic pressure correction and contour iterative optimization tasks.

[0220] S8.4: Based on the fitting results of local deformation and deviation vector with the target surface profile, the target parameter set for dynamic control of the fluid cavity is corrected in real time. The multi-point pressure output is optimized through iterative PID control algorithm, and the gradient rigid-flexible deformation curve of the microstructure unit is further refined, so that the microstructure profile and the target surface can achieve high-precision dynamic fitting.

[0221] S8.5: For the spatial morphology of the multi-layer gradient rigid-flexible microstructure after fine-tuning, perform real-time detection of maximum contact area and adhesion force, including measuring the interface normal contact force, adhesion area distribution and unit contact pressure response, and output the digital characterization results of the maximum effective contact area and adhesion strength as input for the system's automatic adhesion-detachment decision and subsequent learning feedback.

[0222] Step S9: Establish a mapping relationship between the adaptive adhesion results and surface morphology changes and array unit response data; continuously analyze the deviation between microstructure response and environmental adaptation using pressure feedback and machine learning optimization algorithms; and update the flow control algorithm and hierarchical material layout to optimize adhesion stability and durability. Specifically, this includes:

[0223] S9.1: Real-time collection of contact pressure distribution and dynamic response data of adaptive adhesive microstructure array units under different surface morphology changes. Multi-point synchronous pressure sensors and morphology monitoring system are used to obtain the response characteristic parameters of array units to form a pressure feedback matrix that includes adaptive deformation state.

[0224] S9.2: Based on the collected pressure feedback matrix and surface morphology change parameters, the dynamic response signal is decomposed into time-series features using a feature extraction algorithm to generate a set of surface-response mapping feature vectors that reflect the coupling characteristics between adhesion state and environmental changes.

[0225] S9.3: Input the set of surface-response mapping feature vectors into a machine learning optimization model (such as a reinforcement learning neural network architecture), train it to predict the adaptive adhesion performance of microstructure array units under different surface morphologies and pressure feedback conditions, and output the response bias estimation results of the biomimetic adhesion system.

[0226] S9.4: Based on the response deviation estimation results of the biomimetic adhesion system, dynamic backtracking analysis is performed on the existing fluid control regulation algorithm and gradient material layout parameters. The parameter adaptive adjustment algorithm is applied to continuously optimize the fluid cavity pressure regulation strategy and microstructure hierarchical layout scheme, thereby improving the adhesion stability for complex surfaces and dynamic environments.

[0227] S9.5: The control parameters of the microstructure array unit are updated by a new flow control algorithm and gradient material layout scheme. Online closed-loop feedback optimization is implemented, and the optimized fluid pressure control results are continuously mapped to the adaptive response characteristics of the array unit. Finally, a highly stable and durable adhesion performance optimization model is output.

[0228] Step S10: Periodically perform performance tests and durability analyses on the mechanical properties, adaptability, and reversible adhesion-detachment capability of the entire multilayer gradient rigid-flexible coupled crown-shaped biomimetic adhesive microstructure array. Based on the test feedback, dynamically optimize the biomimetic structural parameters and flow control strategies to meet the application requirements of different irregular or dynamic surfaces. Specifically, this includes:

[0229] S10.1: Periodic mechanical performance tests were conducted on a multi-layer gradient rigid-flexible coupled coronal biomimetic adhesive microstructure array. A standardized pressure loading device and multi-point micro force sensors were used to achieve high-precision measurement of the local area bearing capacity, the overall maximum adhesion force, and the reversible desorption critical force of the array, and to obtain a set of mechanical performance parameters for subsequent performance adaptability analysis.

[0230] S10.2: Based on the obtained set of mechanical performance parameters, the adaptive response capability of the adhesive array on surfaces with curvature, roughness and dynamic motion under multiple working conditions is tested in multiple dimensions. Combined with the intelligent surface contour tracking system, the geometric matching degree of array units, stress distribution uniformity and maximum effective contact area under different surface conditions are identified, and adaptive performance indicators are output.

[0231] S10.3: Using adaptive performance indicators as input, dynamic testing of multi-point reversible adhesion-detachment capability is carried out. Through an external micro-fluid control unit, the cavity pressure and fluid composition are progressively adjusted to evaluate the number of reversible adhesion-detachment cycles, residual adhesion and instantaneous recovery capability of the array under different biomimetic structural parameters, and to extract reversible durability evaluation data.

[0232] S10.4: Perform inductive statistical analysis on the reversibility and durability evaluation data and the historical mechanical performance parameter set. Use a multi-objective optimization algorithm to associate array structure parameters, hierarchical stiffness-flexibility ratio, flow control strategy and performance to form structural parameter optimization suggestions and flow control adjustment mapping relationships for different application scenarios, and realize a feedback loop between testing and optimization.

[0233] S10.5: Based on the structural parameter optimization suggestions and flow control adjustment mapping relationship generated in the feedback closed loop, the multi-layer gradient rigid-flexible layout parameters, embedded cavity spatial distribution model and micro-flow control strategy parameters are dynamically updated to guide subsequent array iterative fabrication, system integration and application adaptation, so as to achieve continuous improvement of the performance of the adhesive array under irregular or dynamic surfaces and meet the subdivided requirements.

[0234] For those skilled in the art, various other corresponding changes and modifications can be made based on the technical solutions and concepts described above, and all such changes and modifications should fall within the protection scope of the claims of this invention.

[0235] Unless otherwise defined, the technical or scientific terms used herein shall have the ordinary meaning as understood by one of ordinary skill in the art to which this application pertains. The terms “first,” “second,” “third,” and similar terms used in this patent application specification and claims do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Similarly, the terms “an” or “a” and similar terms do not indicate a quantity limitation, but rather indicate the presence of at least one. The terms “comprising” or “including” and similar terms mean that the elements or objects preceding “comprising” or “including” encompass the elements or objects listed following “comprising” or “including” and their equivalents, and do not exclude other elements or objects. The “multiple” mentioned in the embodiments of this application refers to two or more. A and / or B indicate three possibilities: A; B; and A and B.

[0236] The above description is merely an exemplary embodiment of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and such modifications or substitutions should all be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for fabricating multi-gradient rigid-flexible coupled fluid self-regulating biomimetic adhesive microstructures, specifically comprising: S1: Collect the morphological parameters and stress distribution information of the target surface, and record multi-point surface data including different curvatures, roughness and dynamic motion states; S2: Normalize the collected surface morphology parameters and stress distribution information to construct a set of surface feature vectors for multiple working conditions; S3: Based on the multi-condition surface feature vector set, perform biomimetic microstructure geometry modeling, generate multi-layer gradient rigid-flexible layout parameters that adaptively match different curvatures and stress states, and form three-layer structural design data of rigid support area, elastic buffer and flexible crown end layer. S4: For the three-layer structure design data, embed a micro fluid cavity structure model and map the fluid cavity layout parameters according to different surface features; S5: Based on the three-layer structure design data and the microfluidic cavity design information, prepare a multi-layer gradient rigid-flexible coupled crown-shaped biomimetic adhesive microstructure entity; S6: Distribute and integrate the prepared coronal biomimetic adhesive microstructures to construct a controllable array unit, configure the microfluidic system channel, and establish the initial parameter set for fluid cavity control; S7: Based on the microfluidic system channel input, the contact state between the array unit and the target complex surface is monitored in real time, and the contact pressure, morphology sensing and other multi-point feedback are input into the adaptive fluid regulation algorithm. S8: Based on the output of the adaptive fluid regulation algorithm, dynamically adjust the state of the fluid cavity inside the microstructure, causing the multi-layer gradient rigid-flexible microstructure to deform in real time according to the current surface.

2. The fabrication method for multi-gradient rigid-flexible coupled fluid self-regulating biomimetic adhesive microstructures according to claim 1, further comprising the following after step S8: S9: Establish a mapping relationship between the adaptive adhesion results and surface morphology changes and array unit response data, and continuously correct the flow control algorithm and material layout by using pressure feedback and machine learning optimization algorithms.

3. The fabrication method for multi-gradient rigid-flexible coupled fluid self-regulating biomimetic adhesive microstructures according to claim 1, characterized in that, The elastic buffer is made of an elastomeric material, which is selected from silicone rubber, polyurethane or similar elastic polymers, and the elastic modulus is distributed in a gradient distribution that gradually decreases in the thickness direction.

4. The fabrication method for multi-gradient rigid-flexible coupled fluid self-regulating biomimetic adhesive microstructures according to claim 1, characterized in that: In step S1, an industrial vision system is used to acquire the original point cloud data for shape perception. A multi-channel surface contour measurement sensor is used to perform high-precision three-dimensional reconstruction on the original point cloud data to form a three-dimensional surface shape dataset containing curvature, roughness and local mutations.

5. The fabrication method for multi-gradient rigid-flexible coupled fluid self-regulating biomimetic adhesive microstructures according to claim 1, characterized in that, Step S3 specifically includes: Cluster analysis was performed on the set of surface feature vectors under multiple working conditions to classify different curvatures, roughnesses and stress states into several typical surface types to obtain surface adaptability classification labels. Based on the surface adaptability classification labels, response simulation is performed on the spatial layout of the three-layer structure material to obtain the target distribution parameters of the mechanical properties of the three-layer structure under each surface adaptability case. Based on the distribution parameters of mechanical performance targets, the interface profile, thickness variation and spatial transition function between the three-layer structural materials are calculated to form a multi-layer gradient rigid-flexible layout parameter set. By combining the multi-layer gradient rigid-flexible layout parameter set with surface adaptability classification labels, a rigid-flexible coupling structure morphology optimization algorithm is executed to refine the boundaries, interface morphology and microstructure geometric details of each layer, generating refined three-layer structure morphology design data. The refined three-layer structure morphology design data is input into the micro-nano 3D modeling platform, and an integrated three-layer structure CAD model is automatically generated using a parametric design library.

6. The fabrication method for multi-gradient rigid-flexible coupled fluid self-regulating biomimetic adhesive microstructures according to claim 1, characterized in that, Step S4 specifically includes: Using the elastic buffer in the three-layer structure design data as the input object, a parameter identification algorithm is used to extract its spatial distribution and mechanical performance parameters in order to obtain the spatial mapping template of the intermediate elastic buffer. The set of surface feature vectors under multiple working conditions is input into the fluid cavity structure modeling unit to calculate the initial distribution parameters of the micro fluid cavity in the elastic buffer. Based on the obtained initial distribution parameter information of the fluid cavity, the finite element structural simulation technology is applied to simulate the influence of the micro fluid cavity on the mechanical response of the entire three-layer structure under different deformation conditions, and optimize the geometric layout of the micro fluid cavity. Based on the finite element simulation optimization results, the final determined fluid cavity structure model is seamlessly integrated into the three-layer structure design data, and the output is a microstructure parameter model that integrates the global gradual rigid-flexible layout and the local micro fluid cavity. For the microstructure parameter model integrating fluid cavities, an adaptive distribution mapping verification is performed to evaluate the adaptive performance of the micro fluid cavities driven by surface features under multiple operating conditions, and the adaptive distribution identification results are output.

7. The fabrication method for multi-gradient rigid-flexible coupled fluid self-regulating biomimetic adhesive microstructures according to claim 1, characterized in that: The microfluidic system channel in step S6 is used for external fluid supply and pressure regulation.

8. The fabrication method for multi-gradient rigid-flexible coupled fluid self-regulating biomimetic adhesive microstructures as described in claim 1, characterized in that: The rigid support region is made of a polymer material with an elastic modulus between 1 GPa and 5 GPa, and the polymer material is selected as polyether ether ketone.

9. The fabrication method for multi-gradient rigid-flexible coupled fluid self-regulating biomimetic adhesive microstructures as described in claim 1, characterized in that: The surface of the flexible coronal terminal layer is provided with arrayed biomimetic microstructure details, which are coronal protrusions, micropits, or pseudopodia.

10. The fabrication method for multi-gradient rigid-flexible coupled fluid self-regulating biomimetic adhesive microstructures as described in claim 1, characterized in that: The cross-section of the microfluidic cavity is circular, elliptical, or an irregular curved surface.

Images