Flexible three-dimensional capacitive force sensor and preparation and calculation method thereof

Abstract

The invention discloses a flexible three-dimensional capacitive force sensor and a preparation and calculation method thereof. The flexible three-dimensional capacitive force sensor comprises a flexible base, a force bearing column integrally formed with the flexible base, five capacitive sensing units arranged around the force bearing column and a flexible boss top cover arranged above the flexible base. The top capacitance sensing unit is used for detecting normal force, and the four side wall capacitance sensing units are used for detecting shearing force in two orthogonal directions in pairs; through the spatial distribution of the top and side wall sensing units and the design of a load transmission path between the boss top cover and the bearing column body, the primary decoupling of normal force and shear force on the structural level is realized. In addition, a capacitance signal acquisition module is arranged in a matched mode, feature extraction and neural network mapping are carried out on five paths of capacitance signals, real-time calculation of normal force and shearing force in two orthogonal directions is achieved, and therefore compensation is carried out on non-linear and miniaturized structure residual coupling of the flexible material.

Description

Technical Field

[0001] This invention belongs to the technical field of flexible electronics and multidimensional force sensing. It mainly relates to a flexible three-dimensional capacitive force sensor and its preparation and calculation method. In particular, it relates to a miniaturized flexible sensing structure that can simultaneously detect normal force and shear force in two orthogonal directions, as well as a matching signal acquisition and three-dimensional force calculation scheme. Background Technology

[0002] With the development of technologies such as wearable devices, flexible electronics, rehabilitation medicine, foot monitoring, and robotic tactile sensing, the demand for multidimensional force sensors in human motion monitoring, human-computer interaction, and intelligent sensing is constantly increasing. In applications such as foot force monitoring, diabetic foot risk assessment, gait analysis, and tactile feedback at the end of a robot, sensors not only need to sense normal pressure but also simultaneously sense shear forces in two orthogonal directions to more comprehensively reflect the mechanical state of the contact interface.

[0003] Existing multi-dimensional force sensors typically employ combinations of multiple uniaxial sensing elements, rigid structure layering, or complex microstructure designs to achieve multi-directional force detection. These solutions often suffer from problems such as large device size, complex structure, unclear force path, insufficient flexibility, strong coupling between channels, and difficulty in integration within small areas, making them unsuitable for space-constrained scenarios such as insoles, adhesive flexible devices, or curved tactile surfaces of robots.

[0004] In recent years, flexible capacitive sensors have attracted widespread attention due to their simple structure, low power consumption, and suitability for flexible integration. However, most existing flexible capacitive sensors focus on normal pressure detection, and their ability to simultaneously detect shear forces, especially biaxial shear forces, is limited. On the other hand, flexible materials exhibit nonlinear deformation, local stress concentration, and hysteresis during stress, which can easily lead to residual coupling between loads in different directions. When using traditional linear calibration methods, it is difficult to balance the accuracy of multidimensional force decoupling and measurement stability under miniaturization conditions.

[0005] Therefore, it is necessary to propose a flexible three-dimensional force sensor and its implementation scheme that is compact, compliant, suitable for miniaturization and integration, capable of synchronous detection of normal force and shear force in two orthogonal directions, and capable of high-precision decoupling by combining data-driven methods. Summary of the Invention

[0006] This invention addresses the problems of existing multi-dimensional force sensors, such as large structural volume, insufficient flexibility, complex force paths, significant channel coupling, and difficulty in achieving high-precision simultaneous detection of normal and shear forces under miniaturization conditions. It provides a flexible three-dimensional capacitive force sensor and its fabrication and calculation method, comprising a flexible base, a load-bearing column integrally formed with the flexible base, five capacitive sensing units arranged around the load-bearing column, and a flexible boss top cover disposed above the flexible base. The top capacitive sensing unit detects the normal force, while the four side wall capacitive sensing units are paired to detect shear forces in two orthogonal directions. Through the spatial distribution of the top and side wall sensing units and the load transfer path design between the boss top cover and the load-bearing column, preliminary decoupling of normal and shear forces at the structural level is achieved. This invention also includes a capacitive signal acquisition module, and through feature extraction and neural network mapping of the five capacitive signals, real-time calculation of the normal force and the two orthogonal shear forces is achieved, thereby compensating for the nonlinearity of flexible materials and residual coupling in miniaturized structures. This invention features a compact structure, high flexibility, easy integration, and strong multi-dimensional force detection capabilities, making it suitable for applications such as plantar monitoring, wearable devices, and robotic tactile sensing.

[0007] To achieve the above objectives, the technical solution adopted by the present invention is: a flexible three-dimensional capacitive force sensor, comprising a flexible base, a load-bearing column integrally formed with the flexible base, five capacitive sensing units arranged around the load-bearing column, and a flexible boss top cover disposed above the flexible base.

[0008] The five capacitive sensing units include a top capacitive sensing unit located at the top of the load-bearing column and four side capacitive sensing units located on the four side walls of the load-bearing column. The top capacitive sensing unit is mainly used to detect normal force, and the four side capacitive sensing units are paired to detect shear force in two orthogonal directions. Through the three-dimensional arrangement of the top and side capacitive sensing units around the same load-bearing column, the integration of the normal force detection unit and the shear force detection unit is realized in a single small-volume flexible structure.

[0009] Each capacitive sensing unit consists of two layers of flexible electrodes and a flexible dielectric layer disposed between them. The flexible electrodes are fabricated using a flexible printed circuit board, and the flexible dielectric layer is a purchased flexible dielectric film. Under normal load, the top capacitive sensing unit mainly generates a capacitive response through changes in electrode spacing, while the sidewall capacitive sensing unit mainly generates a capacitive response through lateral displacement and local deformation under shear load, thereby distinguishing the different action paths of normal force and shear force at the structural level.

[0010] A flexible boss top cover is positioned above a flexible base. It features a boss structure for bearing external loads and a cavity structure that cooperates with the load-bearing column and the sidewall capacitive sensing unit. When an external normal load acts on the flexible boss top cover, the load is mainly transferred to the top area of ​​the load-bearing column, causing the top capacitive sensing unit to produce a significant compression response. When an external shear load acts on the flexible boss top cover, a relative lateral displacement occurs between the flexible boss top cover and the load-bearing column, causing the sidewall capacitive sensing unit in the corresponding sidewall direction to produce a differentiated response. Through the above load transfer path design, the normal force and shear force are initially separated at the structural level.

[0011] As an improvement of the present invention, the top capacitive sensing unit is denoted as C0, and the four sidewall capacitive sensing units are denoted as C1, C2, C3, and C4. C1 and C3 are located on a pair of opposite sidewalls and are used to characterize the shear force response in the X direction, while C2 and C4 are located on another pair of opposite sidewalls and are used to characterize the shear force response in the Y direction. By jointly analyzing the top capacitive response and the two sets of opposite sidewall capacitive responses, the three-dimensional force components can be distinguished and estimated.

[0012] The present invention also proposes a method for fabricating the above-mentioned flexible three-dimensional capacitive force sensor, including the steps of fabricating an integrally formed flexible base-supporting column structure and a flexible boss top cover, fabricating five capacitive sensing units, fixing the five capacitive sensing units to the top of the support column and the four side walls respectively, and aligning and encapsulating the flexible boss top cover with the flexible base, thereby obtaining a compact integrated flexible three-dimensional force sensor.

[0013] To achieve stable reading of the output signals of the five capacitive sensing units, the present invention also includes a three-dimensional force detection system. This system includes the aforementioned flexible three-dimensional capacitive force sensor, a capacitive signal acquisition module, and a processing unit. Each of the five capacitive sensing units consists of two electrodes, thus a total of ten electrode signal lines are led out. These ten electrode signal lines serve as the two inputs of the five capacitive sensing units and are connected to the five sets of measurement terminals corresponding to the capacitive signal acquisition module, so as to achieve independent reading of the five capacitance values. The processing unit calculates the normal force and the shear force components in two orthogonal directions based on the measurement results of the five capacitive sensing units.

[0014] In a further embodiment of the present invention, a three-dimensional force calculation method for a flexible three-dimensional capacitive force sensor is also included. This method uses the measurement results of five capacitive sensing units as the raw input. First, zero-point calibration is performed to obtain the capacitance change. Then, the capacitance change is characterized by combining the spatial distribution relationship of the top capacitive sensing unit and the four sidewall capacitive sensing units to form a feature vector characterizing the normal response, shear differential response, sidewall coupling compensation, and the proportional relationship between the center and the periphery response. The feature vector is then input into a pre-trained neural network model to obtain the calculation results of the normal force and the shear force components in two orthogonal directions.

[0015] During the model establishment phase, normal loads, uniaxial shear loads, and combined normal and shear loads are applied to the sensors through a three-dimensional force calibration experiment. Measurement results from five capacitive sensing units and corresponding real three-dimensional force data are simultaneously acquired to form a training sample set. The neural network model learns from this training sample set to establish a nonlinear mapping relationship between capacitive characteristics and three-dimensional force components. After training, the model parameters are deployed in a host computer system or embedded processing unit to achieve real-time calculation of three-dimensional forces. This method can compensate for response deviations under nonlinear conditions of flexible materials, residual coupling in miniaturized structures, and combined stress conditions, thereby improving the accuracy and stability of three-dimensional force measurement.

[0016] Compared with the prior art, the present invention has the following advantages:

[0017] (1) This invention discloses a flexible three-dimensional capacitive force sensor, which achieves synchronous detection of normal force and shear force in two orthogonal directions within a single small-volume flexible structure by using top and side wall capacitive sensing units arranged in a three-dimensional compact layout around the same load-bearing column.

[0018] (2) The flexible three-dimensional capacitive force sensor disclosed in this invention achieves the initial separation of normal force and shear force at the structural level through the load transfer path design between the flexible boss top cover, the load-bearing column and the top / side wall capacitive sensing unit, thereby reducing mechanical crosstalk between channels.

[0019] (3) This invention constructs a three-dimensional force calculation method based on capacitance features and neural network mapping to compensate for response deviations under nonlinear, miniaturized structural residual coupling and composite stress conditions of flexible materials, thereby achieving synergistic optimization that combines structural decoupling and data-driven decoupling.

[0020] (4) The flexible three-dimensional capacitive force sensor disclosed in this invention is composed of flexible materials and flexible electrodes, and has good flexibility, fit and miniaturization integration capabilities. It is suitable for scenarios such as foot monitoring, wearable devices and robot tactile perception. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the overall structure of the flexible three-dimensional capacitive force sensor of the present invention;

[0022] Figure 2 This is an exploded structural diagram of the flexible three-dimensional capacitive force sensor of the present invention;

[0023] Figure 3 This is a schematic diagram showing the distribution of the five capacitive sensing units on the load-bearing column in the flexible three-dimensional capacitive force sensor of the present invention.

[0024] Figure 4 This is a schematic diagram of the deformation of the flexible three-dimensional capacitive force sensor of the present invention under normal load.

[0025] Figure 5 This is a schematic diagram of the force deformation of the flexible three-dimensional capacitive force sensor of the present invention under shear load in the X or Y direction.

[0026] Figure 6 This is a schematic diagram of the mold structure of the flexible base structure and the flexible boss top cover in the flexible three-dimensional capacitive force sensor of the present invention.

[0027] Figure 7 This is a schematic diagram of the flexible electrode structure and lead layout of the capacitive sensing unit in the flexible three-dimensional capacitive force sensor of the present invention.

[0028] Figure 8 This is a schematic diagram of the structure of the five-channel capacitance signal acquisition module in the three-dimensional force detection system of the present invention;

[0029] Figure 9 This is an overall flowchart of the three-dimensional force detection and calculation method of the present invention;

[0030] Figure 10 This is a schematic diagram of the neural network model structure used for three-dimensional force decoupling in this invention;

[0031] Figure 11 This is a physical image of the three-dimensional force calibration experimental platform of the present invention;

[0032] Figure 12 This is a schematic diagram of a physical photograph of the flexible three-dimensional capacitive force sensor of the present invention. Detailed Implementation

[0033] The present invention will be further illustrated below with reference to the accompanying drawings and specific embodiments. It should be understood that the following specific embodiments are for illustrative purposes only and are not intended to limit the scope of the invention.

[0034] Example 1

[0035] A flexible three-dimensional capacitive force sensor, such as Figures 1 to 3As shown, it includes a flexible base, a load-bearing column integrally formed with the flexible base, a capacitive sensing unit, and a flexible boss top cover. The flexible base, load-bearing column, and flexible boss top cover are all made of polydimethylsiloxane material, wherein the load-bearing column is directly integrally formed with the flexible base and is located in the middle area of ​​the flexible base.

[0036] Figure 11 This is a schematic photograph of the flexible three-dimensional capacitive force sensor of the present invention. In this embodiment, the flexible base has a planar dimension of 10 mm × 10 mm and a thickness of 2 mm. The load-bearing column is a square column structure integrally formed with the flexible base, with a side length of 6 mm and a height of 2 mm. The flexible boss top cover matches the flexible base. The planar dimension of the upper boss of the top cover is 6 mm × 6 mm, the outer contour dimension of the lower part of the top cover is 10 mm × 10 mm, and the overall height of the top cover is 1.5 mm. Through the above-mentioned dimensional matching relationship, the external load can be transmitted to the capacitive sensing units on the top and side walls along a predetermined path, thereby realizing the differentiated response of normal force and shear force.

[0037] like Figure 2 and Figure 3 As shown, the capacitive sensing unit comprises five units. One capacitor unit C0 is located at the top for detecting normal pressure. Capacitive units C1, C2, C3, and C4 are located on the four side walls. C1 and C3 are located on opposite side walls and characterize shear force in the X direction, while C2 and C4 are located on another pair of opposite side walls and characterize shear force in the Y direction. The five capacitive sensing units are arranged in a compact, three-dimensional configuration around the same load-bearing column, achieving multi-dimensional force sensing integration within a limited projected area.

[0038] Each capacitive sensing unit consists of two layers of flexible electrodes and a flexible dielectric layer sandwiched between them. The upper and lower electrodes are fabricated using the FPCB process, and the flexible dielectric layer is a polydimethylsiloxane flexible dielectric film. In this embodiment, the electrode size of each capacitive sensing unit is 5 mm × 5 mm, and the thickness of the flexible dielectric film is 0.5 mm. Under normal load, the top capacitive unit mainly increases capacitance by decreasing the electrode spacing; under shear load, the sidewall capacitive units mainly change capacitance through lateral deformation and relative displacement. Because the top and sidewall capacitive units differ significantly in spatial position and response mode, preliminary decoupling of normal and shear forces can be achieved at the structural level.

[0039] like Figure 4As shown in the figure, the displacement contour map of the overall sensor structure under normal load is presented, where the color scale indicates the magnitude of the displacement in mm. Under this condition, the external normal load is applied along the Z direction to the upper surface of the flexible boss top cover, and the load is transmitted downward through the boss structure to the top of the load-bearing column and its surrounding area. The displacement distribution in the figure shows that the overall deformation of the sensor is basically axisymmetric about the center, with a larger displacement in the top loaded area. The contour map color gradually transitions from the lower low displacement area to the upper high displacement area, indicating that the normal load mainly causes compressive deformation in the upper part of the top cover and the top area of ​​the load-bearing column, while the displacement in the base area is relatively small. Since the top capacitor unit C0 is located at the top of the load-bearing column, its location corresponds to the position where the normal compressive deformation is most significant. Therefore, under the normal load, the electrode spacing of C0 changes most significantly, exhibiting higher normal force response sensitivity. In contrast, the areas where the side wall capacitor units C1 to C4 are located are mainly affected by secondary transmitted deformation, and their displacement changes are smaller, which helps to reduce the coupling interference of the normal force on the shear detection channel.

[0040] like Figure 5 As shown in the figure, the displacement contour map of the overall sensor structure under shear load in the X direction is presented, where the color scale indicates the displacement magnitude in mm. Under this condition, an external horizontal load is applied to the upper part of the flexible boss top cover along the X direction. The top cover exhibits a significant lateral displacement relative to the base and the load-bearing column, and the sensor as a whole shows a deformation trend of tilting along the direction of force. As can be seen from the figure, the displacement of the upper structure is significantly greater than that of the lower structure, and the displacement distribution forms a continuous gradient change along the height direction, indicating that the shear load is mainly transmitted through the top cover-load-bearing column interface, and produces more obvious local deformation in the corresponding side wall area. Affected by the shear effect, the relative positions of the electrodes of the side wall capacitor units located on one side of the force direction tend to be closer, and their equivalent capacitance increases; while the relative positions of the electrodes of the side wall capacitor units located on the opposite side tend to be farther apart, and their equivalent capacitance decreases, thus forming a differential response characteristic of one increasing and one decreasing. When the shear load is applied along the Y direction, another pair of oppositely arranged side wall capacitor units will produce a similar differential change pattern as described above. By comparing the response differences of two sets of opposing sidewall capacitor units, it is possible to identify and distinguish shear force components in different directions.

[0041] It should be noted that the above dimensions are merely a set of specific examples in this embodiment, used to illustrate the structural relationships and implementation of the present invention, and do not constitute a limitation on the scope of protection of the present invention.

[0042] Example 2

[0043] The method for fabricating a flexible three-dimensional capacitive force sensor based on Example 1 mainly includes the fabrication of an integrally formed flexible structural component, the fabrication of a capacitive sensing unit, and the overall assembly and packaging steps.

[0044] First, a one-piece flexible base-supporting column structure and a flexible boss top cover are fabricated, such as... Figure 6 As shown, negative molds for molding are prepared using 3D printing or machining, and the draft angle and dimensional stability of the elastomer after curing are considered in the mold design. Polydimethylsiloxane prepolymer and curing agent are mixed evenly at a mass ratio of 10:1, then degassed, and then poured into corresponding molds. After curing and demolding, an integrally formed flexible base-supporting column structure and a flexible boss top cover are obtained.

[0045] Next, five capacitive sensing units are fabricated. The upper and lower electrodes are made using the FPCB process, with corresponding gold finger connection terminals pre-reserved. A flexible polydimethylsiloxane dielectric film is cut to the required size and sandwiched between the upper and lower electrodes as a dielectric layer. The layers are then bonded together using a flexible insulating adhesive to form independent capacitor units. In this embodiment, the electrode size of the capacitor unit is 5mm × 5mm, and the thickness of the flexible dielectric film is 0.5mm.

[0046] Finally, the entire assembly is performed. The top capacitor unit is fixed to the top of the load-bearing column, and the four side-wall capacitor units are fixed to the four side walls of the load-bearing column. The capacitor units are mainly fixed to the load-bearing column and the flexible base, rather than directly connected to the flexible boss top cover, to reduce non-target coupling when subjected to forces in different directions. The flexible boss top cover is fixedly connected to the flexible base, thus forming an encapsulation structure that covers the load-bearing column and the side-wall capacitor units, thereby reducing non-target coupling when subjected to forces in different directions. Subsequently, the flexible boss top cover and the flexible base are aligned and encapsulated, so that its internal cavity covers the load-bearing column and the side-wall capacitor units, thus forming a complete flexible three-dimensional force sensor.

[0047] Example 3

[0048] A three-dimensional force detection system includes the aforementioned flexible three-dimensional capacitive force sensor, a capacitive signal acquisition module, and a processing unit.

[0049] The flexible three-dimensional force sensor in this embodiment includes five capacitor units, such as... Figure 7 and Figure 8 As shown, each capacitor unit consists of two electrodes, thus a total of ten electrode signal lines are led out. These ten lines are respectively connected to five sets of measurement terminals corresponding to the capacitance signal acquisition module, thereby enabling independent measurement of five capacitance values.

[0050] The capacitance signal acquisition module independently reads the capacitance signals from five capacitance sensing units and transmits the measurement results to the processing unit. The processing unit calculates the normal force and the shear force components in two orthogonal directions based on the measurement results from the five capacitance sensing units. The acquisition module can use a small acquisition board based on a capacitance measurement chip and connect to an external microcontroller via a communication interface.

[0051] In the PCB layout, the sensor input interface is placed on one side of the acquisition chip to shorten the capacitance input trace; the communication interface is placed on the other side to reduce the coupling effect of digital signals on the capacitance measurement channel. The two leads corresponding to the five capacitance units are routed in pairs and kept as equal in length as possible. At the same time, large areas of copper plating are avoided in the input area to reduce the impact of parasitic capacitance on the measurement results.

[0052] Example 4

[0053] A three-dimensional force calculation method based on the mapping of structural prior features to neural networks, such as... Figure 9 As shown, this is used to further compensate for the nonlinearity of flexible materials and the residual coupling of miniaturized structures, based on the sensor's existing preliminary decoupling capability at the structural level.

[0054] First, the measurements from the five capacitance sensing units were zero-point calibrated to obtain the capacitance changes ΔC0, ΔC1, ΔC2, ΔC3, and ΔC4 relative to the initial state. Among them, ΔC0 showed the most significant response to the top normal load; ΔC1 and ΔC3 mainly corresponded to shear information in the X direction; and ΔC2 and ΔC4 mainly corresponded to shear information in the Y direction.

[0055] Then, feature vectors are constructed based on the sensor spatial layout. The constructed features include:

[0056] ;

[0057] ;

[0058] ;

[0059] ;

[0060] ;

[0061] .

[0062] Among them, x1 is used to characterize the normal principal response, x2 and x3 are used to characterize the shear differential responses in two orthogonal directions, x4 and x5 are used to characterize the coupling compensation information in the sidewall channels, and x6 is used to reflect the proportional relationship between central compression and peripheral response, so as to enhance the model's ability to identify composite stress states and nonlinear deformation modes of the structure. The above features constitute the input feature vector.

[0063] .

[0064] In this embodiment, the neural network model adopts a multi-layer feedforward neural network structure. For example... Figure 10 As shown, the input layer receives a six-dimensional feature vector X, and the output layer outputs the normal force Fz, the shear force Fx in the X direction, and the shear force Fy in the Y direction. The network has three hidden layers with 64, 128, and 64 neurons in the fully connected layers, respectively, and each hidden layer uses the ReLU nonlinear activation function. The output layer uses a linear output to perform continuous value regression prediction of the three-dimensional force components.

[0065] In the data processing stage, the collected feature vectors and corresponding real three-dimensional force data are used to construct a training sample set. Before training, the input features are normalized to reduce the impact of differences in the dimensions and magnitudes of different features on the stability of model training; the output three-dimensional force components are also normalized accordingly to improve the network convergence speed and numerical stability.

[0066] During the model training phase, a three-dimensional force calibration platform is used to apply normal force, uniaxial shear force, and combined normal and shear loads to the sensor. Simultaneously, five capacitance changes and corresponding real three-dimensional force values ​​are collected to construct a training sample set. The training samples include both single-direction load conditions and various composite force conditions, allowing the model to fully learn the sensor's response characteristics under different force states. In this embodiment, the training set, validation set, and test set are divided according to a certain ratio for model training, parameter tuning, and performance evaluation, respectively.

[0067] During training, the mean squared error loss function was used, the Adam optimization algorithm was employed, the learning rate was set to 0.001, the batch size to 64, and the number of training epochs to 200. The model training status was monitored based on the changes in validation set error to avoid significant overfitting. After training, the model parameters were deployed to a host computer or embedded processing platform, and the five-channel capacitor input was calculated in real time during actual measurements.

[0068] In the actual calculation process, the processing unit first receives the real-time measurement values ​​from the five capacitive sensing units and calculates ΔC0, ΔC1, ΔC2, ΔC3 and ΔC4; then it constructs the feature vector X according to the above formula; finally, it inputs the feature vector into the trained neural network model to obtain the output results of the normal force Fz, the X-direction shear force Fx and the Y-direction shear force Fy, thereby realizing the real-time calculation of the external three-dimensional contact force.

[0069] In terms of error evaluation, the normal force and shear force components output by the neural network are compared with the measurement results of the standard three-dimensional force sensor, and the model is evaluated using indicators such as root mean square error, mean absolute error, relative error, and coupling error.

[0070] This embodiment uses, as follows: Figure 11 The calibration experimental platform shown is used to calibrate and test the flexible three-dimensional capacitive force sensor. This platform includes a vertical normal loading mechanism, a horizontal shear loading mechanism, a sensor clamping mechanism, and a multi-axis displacement adjustment mechanism.

[0071] The vertical normal loading mechanism includes a digital push-pull force gauge mounted on a vertical Z-axis moving frame, used to apply a normal load to the sensor along the Z-direction. The horizontal shear loading mechanism includes a digital push-pull force gauge mounted on an X / Y-axis translation platform, used to apply a horizontal shear load to the sensor along the X or Y direction. The sensor is fixed to the Z-axis lifting platform by a clamp. By adjusting the position of the Z-axis lifting platform, the force-bearing surface of the sensor is made to be at the same height and substantially parallel to the horizontal shear loading end, thereby reducing the influence of the additional normal component on the shear force test results during loading.

[0072] During the experiment, two digital push-pull force gauges were used to provide reference values ​​for the normal load and horizontal shear load, respectively. The capacitance signal acquisition module simultaneously acquired the measurement values ​​of five capacitance sensing units for subsequent training and verification of the three-dimensional force calculation model.

[0073] The flexible three-dimensional force sensor of this invention was calibrated and tested using the aforementioned three-dimensional force calibration experimental platform under room temperature experimental conditions. During the test, normal load, X-direction shear load, Y-direction shear load, and a combination of normal and shear loads were applied, and the measured values ​​of the five capacitive sensing units and the corresponding reference three-dimensional force data were collected simultaneously. Based on the same set of experimental data, the linear matrix calibration method and the neural network solution method in this embodiment were compared. Table 1 below shows the comparison results between the linear matrix calibration method and the neural network solution method.

[0074] Table 1

[0075]

[0076] As shown in Table 1 above, the neural network method outperforms the linear matrix calibration method in terms of root mean square error (RMSE), mean absolute error (MAE), and goodness of fit for both normal and shear force components. Specifically, in the X-direction shear force calculation, the RMS error decreased from 1.006 to 0.721, and the MAE decreased from 0.677 to 0.375; in the Y-direction shear force calculation, the RMS error decreased from 1.080 to 0.648, and the MAE decreased from 0.796 to 0.408; and in the normal force calculation, the RMS error decreased from 1.565 to 0.364, and the MAE decreased from 1.084 to 0.217. Meanwhile, the goodness of fit R² for the X-direction shear force, Y-direction shear force, and normal force calculations reached 0.564, 0.791, and 0.985, respectively, all of which are superior to the linear matrix calibration method. The above results demonstrate that the neural network solution method in this embodiment can more effectively compensate for the nonlinear deformation, local hysteresis, and residual coupling under miniaturized structural conditions of flexible materials. In particular, it shows a more significant error reduction effect in the solution of normal force, thereby improving the decoupling accuracy of normal force and shear force and the overall measurement stability.

[0077] It should be noted that the above content merely illustrates the technical concept of the present invention and should not be construed as limiting the scope of protection of the present invention. For those skilled in the art, various improvements and modifications can be made without departing from the principle of the present invention, and all such improvements and modifications fall within the scope of protection of the claims of the present invention.

Claims

1. A flexible three-dimensional capacitive force sensor, characterized in that: It includes a flexible base, a load-bearing column integrally formed with the flexible base, five capacitive sensing units, and a flexible boss top cover disposed above the flexible base. The five capacitive sensing units are located on the top of the load-bearing column and on the four side walls of the load-bearing column, respectively. The top capacitive sensing unit is used to characterize the normal force response. Among the four side wall capacitive sensing units, a pair of side wall capacitive sensing units in opposite positions are used to characterize the shear force response in the X direction, and another pair of side wall capacitive sensing units in opposite positions are used to characterize the shear force response in the Y direction. The flexible boss top cover is provided with a boss structure for bearing external loads. When the flexible boss top cover is closed with the load-bearing column and the four side walls of the load-bearing column, it has an internal cavity structure. When an external normal load is applied to the flexible boss top cover, the load is transferred to the top capacitive sensing unit at the top of the load-bearing column; when an external shear load is applied to the flexible boss top cover, a relative lateral displacement occurs between the flexible boss top cover and the load-bearing column, thereby causing the side capacitive sensing unit in the corresponding side wall direction to produce a differentiated response, so as to realize the detection of normal force and shear force in the X and Y orthogonal directions.

2. The flexible three-dimensional capacitive force sensor as described in claim 1, characterized in that: The capacitive sensing unit consists of a flexible electrode, a flexible dielectric layer, and another flexible electrode from top to bottom. The flexible electrode is fabricated using a flexible printed circuit board, and the flexible dielectric layer is composed of a polydimethylsiloxane flexible dielectric film.

3. The flexible three-dimensional capacitive force sensor as described in claim 1, characterized in that: The five capacitive sensing units are connected to the five surfaces of the load-bearing column, and the lower surface of the flexible boss top cover is connected to the upper surface of the base, completely covering the cross-section of the top of the load-bearing column.

4. A flexible three-dimensional capacitive force sensor as described in claim 1, characterized in that: The flexible base has a planar dimension of 10 mm × 10 mm, the load-bearing column has a square with a side length of 6 mm, the flexible boss top cover has a planar dimension of 6 mm × 6 mm, and the lower outer contour dimension of the flexible boss top cover is 10 mm × 10 mm.

5. A method for fabricating a flexible three-dimensional capacitive force sensor as described in any one of claims 1 to 4, characterized in that, Includes the following steps: S1, Prepare an integrally molded flexible base-supporting column structure and a flexible boss top cover. The flexible base-supporting column structure and the flexible boss top cover are obtained by casting polydimethylsiloxane material into a mold and then curing and demolding. S2, five capacitive sensing units are prepared. Each capacitive sensing unit consists of two layers of flexible electrodes and a flexible dielectric layer disposed therebetween. S3, fix one capacitive sensing unit to the top of the load-bearing column, and fix the other four capacitive sensing units to the four side walls of the load-bearing column respectively. S4, align and encapsulate the flexible boss top cover with the flexible base, so that the cavity structure of the flexible boss top cover covers the load-bearing column and the side wall capacitive sensing unit, thereby obtaining a flexible three-dimensional capacitive force sensor.

6. A three-dimensional force detection system, characterized in that, Includes the flexible three-dimensional capacitive force sensor, capacitive signal acquisition module, and processing unit as described in any one of claims 1 to 4; The capacitance signal acquisition module is connected to five capacitance sensing units and is used to independently read the capacitance signals of the five capacitance sensing units and transmit the measurement results to the processing unit. The processing unit is used to calculate the normal force and the shear force components in the X and Y orthogonal directions based on the measurement results of the five capacitive sensing units.

7. A three-dimensional force detection system as described in claim 6, characterized in that: The five capacitance sensing units are each composed of two electrodes and have a total of ten electrode signal lines. These ten electrode signal lines serve as the two inputs of the five capacitance sensing units and are connected to the five sets of measurement terminals corresponding to the capacitance signal acquisition module to complete the reading of capacitance signals.

8. A three-dimensional force calculation method for a flexible three-dimensional capacitive force sensor, using the flexible three-dimensional capacitive force sensor as described in claim 1, characterized in that, Includes the following steps: S1, collects real-time measurement values ​​from five capacitance sensing units and performs zero-point calibration to obtain the corresponding capacitance changes ΔC0, ΔC1, ΔC2, ΔC3, and ΔC4; where ΔC 0对应 Top normal load; ΔC1 and ΔC3 correspond to shear information in the X direction; ΔC2 and ΔC4 correspond to shear information in the Y direction; S2, based on the spatial distribution relationship between the top capacitive sensing unit and the four sidewall capacitive sensing units, the capacitance change is characterized to obtain a feature vector for characterizing the normal response, shear differential response, sidewall coupling compensation, and the ratio of the center to the periphery response. S3, input the feature vector into the pre-trained neural network model to obtain the solution results of the normal force and the shear force components in two orthogonal directions.

9. The three-dimensional force calculation method for a flexible three-dimensional capacitive force sensor as described in claim 8, characterized in that: The feature vector used in step S2 to characterize the normal response, shear difference response, sidewall coupling compensation, and the ratio of the center to the periphery response is specifically as follows: ； ; ; ； ； ； Among them, x1 is used to characterize the normal response, x2 and x3 are used to characterize the shear differential response, x4 and x5 are used to characterize the sidewall coupling compensation information, and x6 is used to reflect the proportional relationship between the center and the periphery response.

10. The three-dimensional force calculation method for a flexible three-dimensional capacitive force sensor as described in claim 9, characterized in that: The neural network model in step S3 is a multi-layer feedforward neural network structure, wherein the hidden layer is set to three layers, the number of neurons in the fully connected layer is 64, 128 and 64 respectively, each hidden layer adopts the ReLU nonlinear activation function, and the output layer adopts a linear output mode, which is used to complete the continuous value regression prediction of the three-dimensional force components.

Images