Capacitive small six-dimensional force sensor, optimization design and six-dimensional force decoupling method

Abstract

The application discloses a kind of capacitive small six-dimensional force sensor, optimization design method and six-dimensional force decoupling method, the sensor includes three T type beam type elastic body, lower electrode plate, upper electrode plate and base, three T type beam type elastic body is converted into displacement variation to six-dimensional force from outside, upper electrode plate is connected with elastic body, and upper electrode plate and lower electrode plate form six groups of facing electrode, including three groups of horizontal and three groups of vertically arranged electrode, constitute six groups of parallel plate capacitor structure.When sensor is subjected to external force, the deformation of elastic body drives upper electrode plate to generate displacement, changes capacitance value, and the information of force is converted into six capacitance signal outputs.The sensor is compact in design, small in size, and easy to integrate into a robot finger tip and other narrow spaces.Based on the principle of capacitance measurement, it has excellent dynamic response performance and can quickly respond to external force changes.The sensor has high sensitivity, consistent stiffness in each direction, effectively reduces the coupling effect of force, and realizes high-precision six-dimensional force decoupling combined with decoupling algorithm.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of force sensors, in particular to a capacitive small six-dimensional force sensor, an optimized design and a six-dimensional force decoupling method. BACKGROUND

[0002] In recent years, with the rapid development of robot technology, robots are increasingly widely applied in industrial manufacturing, medical care, service robots and other fields. In these applications, force measurement is often required in order to perform higher-level and more sophisticated tasks. In particular, for humanoid robots, force sensors are an important part of the humanoid performance. Among the various components of the robot, the humanoid finger joint exhibits a variety of force information when manipulating tools or objects, and reasonable collection and processing of this force information can greatly improve the performance of the robot, so that force and torque sensing of the robot finger tip becomes a key technology, which directly affects the operational flexibility, safety and interaction quality with the environment of the robot in a complex environment.

[0003] A six-dimensional force sensor can measure all force and torque information at a point in space, i.e. three orthogonal forces and three orthogonal torques. It can measure various types of force information, so it is a suitable force sensor for robot finger joints. Traditional force sensors can meet the needs of general robot operation in most cases, but in such a narrow space as the robot finger tip, the sensor needs to be smaller and more sensitive to achieve high-precision measurement of small forces and torques. In addition, robots often need to contact the environment or people in actual work, and for some special applications such as medical surgical robots, high-precision sensing of the finger tip force is essential. However, designing such a high-performance small sensor still has difficulties in structural design, signal measurement, decoupling design, etc., which hinders the practical application of six-dimensional force sensors in robots. In recent years, small six-dimensional force sensors for robot finger joints have become one of the hotspots of robot technology research.

[0004] It should be noted that the information disclosed in the above background section is only for understanding the background of the present application, and therefore can include information that does not constitute prior art known to those of ordinary skill in the art. SUMMARY

[0005] The main purpose of the present application is to overcome the defects in the background art, and to provide a capacitive small six-dimensional force sensor, an optimized design and a six-dimensional force decoupling method.

[0006] To achieve the above purpose, the application adopts the following technical solutions:

[0007] The application discloses a capacitive small six-dimensional force sensor, which comprises a three-T-shaped beam elastic body, a lower electrode plate, an upper electrode plate and a base; the three-T-shaped beam elastic body is used for converting external six-dimensional force into displacement change; the lower electrode plate is arranged on the base; the upper electrode plate is connected with the three-T-shaped beam elastic body; six groups of opposite electrodes are formed by the upper electrode plate and the lower electrode plate; the six groups of opposite electrodes comprise three groups of horizontally arranged electrodes and three groups of vertically arranged electrodes, forming six groups of parallel plate capacitor structures; the three-T-shaped beam elastic body comprises elastic beams, flexible leaf springs and a loaded circular platform; the three groups of elastic beams and the flexible leaf springs form a T-shaped structure and are used for bearing force; the loaded circular platform is connected with the elastic beams through the flexible leaf springs and is used for transmitting torque; the upper electrode plate is connected with the three-T-shaped beam elastic body; when the sensor is subjected to external force, the deformation of the three-T-shaped beam elastic body drives the upper electrode plate to generate displacement, thereby causing the distance between the electrode plates of the six groups of parallel plate capacitors to change, and then the information of the six-dimensional force is converted into six capacitor signals.

[0008] Further, the lower electrode plate comprises three horizontal electrodes arranged on the upper surface of the lower electrode plate and three vertical electrodes arranged on the side walls of three grooves of the lower electrode plate; the three horizontal electrodes and the three vertical electrodes are uniformly distributed along the circumferential direction.

[0009] Further, the three horizontal electrodes and the three vertical electrodes are alternately and spacedly distributed along the circumferential direction.

[0010] Further, the upper electrode plate comprises three horizontal electrodes arranged on the lower surface of the upper electrode plate and three vertical electrodes vertically extended downwards from the lower surface of the upper electrode plate; the horizontal electrodes and the vertical electrodes of the upper electrode plate are arranged in correspondence with the horizontal electrodes and the vertical electrodes of the lower electrode plate.

[0011] Further, the lower electrode plate is based on a circuit board, and a digital capacitive chip is arranged on the circuit board; the six groups of opposite electrodes are connected with the digital capacitive chip.

[0012] Further, the three-T-shaped beam elastic body further comprises a top cover arranged on the three-T-shaped beam elastic body.

[0013] An optimization design method of a three-T-shaped beam elastic body structure of the capacitive small six-dimensional force sensor is provided.

[0014] A mechanical model of the three-T-shaped beam elastic body is established;

[0015] The mechanical model is improved through a BP neural network; the nonlinear fitting capacity of the BP neural network is utilized to compensate the mechanical model, so as to establish a nonlinear relationship between the size parameters of the elastic body and the error of the mechanical model.

[0016] wherein, according to the mechanical model and the result of ANSYS finite element analysis, a set of compensation coefficients is generated, wherein a plurality of sets of data are obtained through simulation, the simulation result is compared with the calculation result of the mechanical model, and the compensation coefficient is obtained;

[0017] The data set composed of the size parameters and the corresponding compensation coefficients is divided into a training set and a test set, the BP neural network is trained through the training set, and the compensation effect of the network is verified by using the test set;

[0018] The solving error of the mechanical model is reduced to within a preset threshold through the trained BP neural network, so that an improved mechanical model is obtained;

[0019] The improved mechanical model is used in combination with a particle swarm search algorithm to perform optimal design of the size of the three T-shaped beam type elastic body.

[0020] Four BP neural networks with seven inputs and one output are established, each network corresponds to compensation of the mechanical model under four force conditions; each BP neural network is configured to receive seven size parameters, specifically the length of the elastic beam, the width of the elastic beam, the height of the elastic beam, the length of the flexible leaf spring, the width of the flexible leaf spring, the height of the flexible leaf spring, and the radius of the loaded circular table, as inputs, and output one compensation coefficient; the four force conditions are:

[0021] -F z Case: vertical force applied in the Z-axis direction;

[0022] -M x Case: torque rotating around the X-axis;

[0023] -F y Case: horizontal force applied in the Y-axis direction;

[0024] -M z Case: torque rotating around the Z-axis.

[0025] The constraint condition and the fitness function for the optimal design of the size of the three T-shaped beam type elastic body are:

[0026]

[0027] F = ω1f1 + ω2f2

[0028] wherein, L1 represents the length of the elastic beam, b1 represents the width of the elastic beam, h1 represents the height of the elastic beam, L2 represents the length of the flexible leaf spring, b2 represents the width of the flexible leaf spring, h2 represents the height of the flexible leaf spring, r represents the radius of the loaded circular table, and i = 1 ~ n represents six-dimensional force F x ~Mz D i This indicates the maximum effective displacement caused by the force / torque. ω1 represents the average value of the maximum effective displacement caused by the six-dimensional force, and ω2 represents the weights of the proximity target and the sensitivity target, respectively.

[0029] A method for decoupling six-dimensional forces in a three-T-beam elastic body structure of a capacitive miniature six-dimensional force sensor, comprising:

[0030] S1. Establishing the relationship between capacitance and displacement: When the sensor is subjected to an external force, the deformation of the elastic body causes the moving electrode to shift, while the stationary electrode remains fixed to the base, resulting in capacitance changes corresponding to the displacements in each direction. The three horizontal capacitors correspond to the normal displacement d. n1 ,d n2 ,d n3 The three vertical capacitors correspond to tangential displacements d. s4 ,d s5 ,d s6 ;

[0031] S2. Solving for the vertical force F z The coefficients are determined using a mechanical model based on the change in horizontal capacitance. Calculate the vertical force F z The formula is Simultaneously considering the torque M x and M y To mitigate the impact and ensure that there is no coupling;

[0032] S3. Solving for torque M x The coefficients determined by the mechanical model based on the change in horizontal displacement. Calculate torque M x The formula is Simultaneously considering force F z The coupling effect is mitigated and decoupled.

[0033] S4. Solve for torque M y : Coefficients determined using a mechanical model based on the change in horizontal displacement. Calculate torque M y The formula is Simultaneously considering force F z and torque M x The coupling effect is to be resolved and decoupled.

[0034] S5. Solve for torque M z The coefficients are determined by a mechanical model based on the change in vertical displacement. Calculate torque M z The formula is Simultaneously considering force F xand F y The coupling effect is to be resolved and decoupled.

[0035] S6. Solve for the horizontal force F y : The coefficients determined by the mechanical model using the change in vertical displacement. Calculate the horizontal force F y The formula is Simultaneously considering force F x The coupling effect is to be resolved and decoupled.

[0036] S7. Solve for the horizontal force F x The coefficients are determined by a mechanical model based on the change in vertical displacement. Calculate the horizontal force F x The formula is Simultaneously considering force F y and torque M z The coupling effect is to be resolved and decoupled.

[0037] S8. Decoupling: The parameters in the decoupling matrix are determined by calibration. The displacement change Δd is calculated using the capacitance change. The six-dimensional force is decoupled using the formula F / T=Calib*Δd, where F and T represent force and torque, respectively, and Calib is the decoupling matrix.

[0038] The present invention has the following beneficial effects:

[0039] This invention proposes a capacitive miniature six-dimensional force sensor structure, its optimized design method, and a six-dimensional force decoupling method. When an external six-dimensional force is applied to a three-T-shaped beam elastic body, the elastic body deforms, causing the upper electrode plate to shift. This, in turn, changes the spacing between the plates of the six sets of parallel plate capacitors, resulting in a change in capacitance. By collecting minute capacitance values ​​and using the six-dimensional force decoupling algorithm designed in this invention to decouple the six-dimensional force, accurate measurement of force and torque in three orthogonal directions can be achieved. The main advantages of this invention are: ① The sensor designed in this invention has a compact structure and small size, making it easy to integrate into the narrow space of a robot's dexterous finger joint; ② The sensor designed in this invention is based on the capacitive measurement principle, thus it has good dynamic response performance and can quickly respond to changes in external force; ③ The sensor designed in this invention can measure external force of ±30N and external torque of ±0.3Nm, which can meet the usage requirements of a robot's dexterous fingertip; ④ The three-T beam elastic body structure designed in this invention has high sensitivity and similar stiffness in all directions, which can reduce the coupling effect of six-dimensional forces; ⑤ The capacitor arrangement structure designed in this invention facilitates the decoupling of six-dimensional forces, with a small coupling effect, and combined with the designed decoupling algorithm, it can achieve high-precision six-dimensional force decoupling.

[0040] Other beneficial effects of the embodiments of the present invention will be further described below. Attached Figure Description

[0041] Figure 1 This is an exploded structural diagram of one embodiment of the capacitive miniature six-dimensional force sensor of the present invention;

[0042] Figure 2 This is a three-dimensional schematic diagram of the top cover structure of one embodiment of the capacitive miniature six-dimensional force sensor of the present invention;

[0043] Figure 3 This is a three-dimensional schematic diagram of a three-T beam elastic body structure of one embodiment of the capacitive miniature six-dimensional force sensor of the present invention;

[0044] Figure 4 This is a three-dimensional schematic diagram of the upper electrode plate structure of one embodiment of the capacitive miniature six-dimensional force sensor of the present invention;

[0045] Figure 5 This is a schematic diagram of the lower electrode plate (circuit board) structure of one embodiment of the capacitive miniature six-dimensional force sensor of the present invention;

[0046] Figure 6 This is a schematic diagram of the base structure of one embodiment of the capacitive miniature six-dimensional force sensor of the present invention.

[0047] Figure 7 This is a structural parameter diagram of the elastomer according to an embodiment of the present invention.

[0048] Figure 8 This is a schematic diagram of the force applied to an elastic body when F_z is applied alone.

[0049] Figure 9 This is a schematic diagram of the force applied to the elastic body when M_x is applied alone.

[0050] Figure 10 This is a schematic diagram of the force applied to an elastic body when F_y is applied alone.

[0051] Figure 11 This is a schematic diagram of the force applied to the elastic body when M_z is applied alone.

[0052] Figure 12 This is a diagram of the BP neural network structure established to improve the mechanical model.

[0053] Figure 13 This is a diagram showing the compensation effect of the neural network on the mechanical model.

[0054] Figure 14 This is a diagram showing the relationship between the capacitance and the displacement of the upper electrode plate in a capacitive miniature six-dimensional force sensor. Detailed Implementation

[0055] The embodiments of the present invention will be described in detail below. It should be emphasized that the following description is merely exemplary and not intended to limit the scope and application of the present invention.

[0056] It should be noted that when a component is referred to as "fixed to" or "set on" another component, it can be directly on or indirectly on that other component. When a component is referred to as "connected to" another component, it can be directly connected to or indirectly connected to that other component. Furthermore, a connection can be used for fixing, coupling, or communication.

[0057] It should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", and "outer" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing the embodiments of the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the present invention.

[0058] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of embodiments of the present invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0059] See Figures 1 to 7This invention provides a capacitive miniature six-dimensional force sensor, comprising a top cover 1, a three-T-beam elastic body 2, an upper electrode plate 3, a lower electrode plate 4, and a base 5. The three-T-beam elastic body 2 converts external six-dimensional force into displacement changes. The lower electrode plate 4 is placed on the base 5, and the top cover 1 is placed on the three-T-beam elastic body 2. The upper electrode plate 3 is connected to the three-T-beam elastic body 2, and the electrodes of the upper electrode plate 3 and the lower electrode plate 4 correspond to form six sets of opposing electrodes. The six sets of opposing electrodes include three sets of horizontally arranged electrodes and three sets of vertically arranged electrodes, forming six sets of parallel plate capacitors. Structure: The three-T beam elastic body 2 includes elastic beams, flexible leaf springs, and a load-bearing frustum. The three sets of elastic beams and flexible leaf springs form a T-shaped structure to bear the force. The load-bearing frustum located at the center is connected to the elastic beams through the flexible leaf springs to transmit torque. The upper electrode plate 3 is connected to the three-T beam elastic body 2. When the sensor is subjected to an external force, the deformation of the three-T beam elastic body 2 causes the upper electrode plate 3 to displace, thereby causing a change in the spacing between the plates of the six sets of parallel plate capacitors, and thus converting the six-dimensional force information into six capacitive signal outputs.

[0060] like Figure 1 , Figure 4 and Figure 5 As shown, in a preferred embodiment, the lower electrode plate 4 includes three horizontal electrodes arranged horizontally on its upper surface and three vertical electrodes arranged vertically on the sidewalls of three grooves in the lower electrode plate 4. The three horizontal electrodes and the three vertical electrodes are evenly distributed along the circumferential direction. The three horizontal electrodes and the three vertical electrodes are alternately distributed along the circumferential direction. The upper electrode plate 3 includes three horizontal electrodes arranged horizontally on its lower surface and three vertical electrodes extending vertically downward from its lower surface. The horizontal and vertical electrodes of the upper electrode plate 3 are correspondingly arranged with the horizontal and vertical electrodes of the lower electrode plate 4.

[0061] This capacitive miniature six-dimensional force sensor converts the magnitude of an external six-dimensional force into changes in the capacitance of six sets of electrodes within the sensor. By decoupling these changes in capacitance, it enables the sensing of the magnitude of the external six-dimensional force using variations in electrical signals. The sensor boasts advantages such as small size, light weight, high sensitivity, good dynamic response, and ease of fabrication and assembly, making it suitable for use in the dexterous finger joints of robots.

[0062] An optimized design method for the three-T beam elastic body structure of the capacitive miniature six-dimensional force sensor, comprising:

[0063] Establish the mechanical model of the aforementioned three-T beam elastic body;

[0064] The mechanical model is improved by using a BP neural network. The nonlinear fitting capability of the BP neural network is used to compensate for the mechanical model, so as to establish a nonlinear relationship between the elastic body size parameters and the mechanical model error.

[0065] Based on the mechanical model and the results of ANSYS finite element analysis, a set of compensation coefficients is generated. Multiple sets of data are obtained through simulation, and the simulation results are compared with the calculation results of the mechanical model to obtain the compensation coefficients.

[0066] The dataset consisting of the size parameters and the corresponding compensation coefficients is divided into training and test sets. The BP neural network is trained using the training set, and the compensation effect of the network is tested using the test set.

[0067] By using a trained BP neural network, the solution error of the mechanical model is reduced to within a preset threshold, thereby obtaining an improved mechanical model.

[0068] Using the improved mechanical model and the particle swarm optimization algorithm, the dimensions of the three-T beam elastic body are optimized.

[0069] Four 7-input, 1-output backpropagation (BP) neural networks were established, each corresponding to mechanical model compensation under four force conditions. Each BP neural network was configured to receive seven dimensional parameters as inputs: the length, width, and height of the elastic beam; the length, width, and height of the flexible leaf spring; and the radius of the loaded frustum. The output was a compensation coefficient. The four force conditions were as follows:

[0070] -F z Situation: A perpendicular force applied along the Z-axis;

[0071] -M x Condition: Torque rotating about the X-axis;

[0072] -F y Situation: A horizontal force applied along the Y-axis;

[0073] -M z Condition: Torque rotating about the Z-axis.

[0074] The constraints and fitness function for the optimal design of the dimensions of the three-T beam elastic body are as follows:

[0075]

[0076] F=ω1f1+ω2f2

[0077] Where L1 represents the length of the elastic beam, b1 represents the width of the elastic beam, h1 represents the height of the elastic beam, L2 represents the length of the flexible leaf spring, b2 represents the width of the flexible leaf spring, h2 represents the height of the flexible leaf spring, r represents the radius of the loaded frustum, and i = 1 to n represent the six-dimensional force F in sequence. x ~M z D i This indicates the maximum effective displacement caused by the force / torque. ω1 represents the average value of the maximum effective displacement caused by the six-dimensional force, and ω2 represents the weights of the proximity target and the sensitivity target, respectively.

[0078] The following describes specific embodiments of the present invention.

[0079] A small capacitive six-dimensional force sensor includes a three-T beam elastic body 2, a lower electrode plate 4 with a specific capacitor arrangement structure, an upper electrode plate 3 with a specific structure, a top cover 1, and a base 5. The lower electrode plate 4 with the specific capacitor arrangement structure is placed on the base 5. The upper electrode plate 3 is connected to the three-T beam elastic body and forms six sets of opposing electrodes with the lower electrode plate 4. This sensor can convert the magnitude of an external six-dimensional force into changes in the capacitance of the six sets of electrodes in the sensor, decoupling the changes in the six sets of capacitance, thereby realizing the sensing of the magnitude of the external six-dimensional force using changes in electrical signals. The lower electrode plate 4 has six capacitive electrodes evenly arranged on a circular circuit board, with three horizontally arranged on the upper surface of the circuit board and three vertically arranged on the sidewalls of three recesses on the circuit board. All six electrodes are connected to a digital capacitor chip in the center of the circuit board, and the capacitance value is measured through the digital capacitor chip. The structure of the upper electrode plate 3 corresponds to that of the lower electrode plate 4, with six electrodes evenly arranged on the circular electrode plate, three horizontally and three vertically.

[0080] When the sensor is subjected to an external force, the deformation of the three-T beam elastic body 2 causes the upper electrode plate 3 to shift, thereby causing the electrode spacing of the six sets of parallel plate capacitors to change, and then converting the six-dimensional force information into six capacitor signals for output.

[0081] In a specific embodiment, the sensor structure can be assembled as follows: A digital capacitor chip and signal lines are soldered onto the lower electrode plate 4. The lower electrode plate 4 is placed on the base 5, and the signal lines are led out from the bottom of the base 5. The upper electrode plate 3 and the three-T beam elastic body 2 are aligned through grooves and then connected with screws. The three-T beam elastic body 2 and the lower electrode plate 4 on the base 5 are positioned with pins, aligned, and then the base 5 and the three-T beam elastic body 2 are connected with screws, thereby pressing the lower electrode plate 4 tightly and forming six sets of parallel plate capacitors with a plate spacing of 0.2mm between the lower and upper electrode plates. The top cover 1 is aligned with the three-T beam elastic body 2 and then connected with screws. The six-dimensional force sensor is thus assembled.

[0082] Capacitance signal acquisition can be achieved using the AD7147 digital capacitance chip, which can measure capacitance of ±8pF with a resolution of 16 bits and 14 measurable capacitance channels. A single chip can simultaneously measure 6 capacitance signals.

[0083] This sensor has advantages such as small size, light weight, high sensitivity, good dynamic response, and easy processing and assembly, and is suitable for the dexterous finger joints of robots.

[0084] The mechanical model of the three-T beam elastomer was improved using a backpropagation neural network. The dimensions and structure of the three-T beam elastomer were optimally designed using a particle swarm optimization algorithm. The specific dimensions depend on the detection requirements of the six-dimensional force sensor in each direction and the size requirements of the six-dimensional force sensor.

[0085] The dimensions of the three-T beam elastomer structure are optimally designed using an improved mechanical model and a particle swarm optimization algorithm. The BP neural network reduces the relative error of the mechanical model to less than 5%, and the particle swarm optimization algorithm calculates the optimal structural dimensions, enabling the elastomer to achieve high sensitivity and isotropy.

[0086] Method for sizing a three-t beam elastomeric structure

[0087] First, a mechanical model of the three-T beam elastic body structure is established. Then, a BP neural network is used to compensate the model to obtain an improved mechanical model, the solution error of which can be controlled within 5%. Then, combined with the improved mechanical model, a particle swarm search algorithm is used to solve for the optimal size structure of the elastic body with high sensitivity and similar stiffness of each component.

[0088] Establish a mechanical model

[0089] The mechanical model of the three-T-beam elastic body structure has six structural parameters: the length and width of the elastic beam (l1, b1), the length and width of the flexible leaf spring (l2, b2), the heights of the elastic beam and flexible leaf spring (h), and the radius r of the loaded frustum. To optimize the stiffness of the elastic body in each direction and increase its design space, the heights of the elastic beam and flexible leaf spring are distinguished and denoted as h1 and h2, respectively. The structural parameters of the elastic body are as follows: Figure 7 As shown.

[0090] Let the vertical direction be denoted as the Z-direction. When considering a six-dimensional force acting on an elastic body, since the elastic body structure is symmetrical, it can be verified that the force F... y Displacement and force F x The displacement calculation is the same, and the torque M x Displacement calculation and torque M y The displacement calculation is the same. Therefore, the displacement calculation of an elastic body under the action of a six-dimensional force can be simplified to two forces (F). x and F z ) and two torques (M y and M z The displacement calculation was performed. Based on the original model, since the heights of the elastic beam and the flexible leaf spring are distinguished, the mechanical model needs to be adjusted and improved. The results are as follows.

[0091] 1) Apply F alone z

[0092] Apply F alone z The forces acting on Figure 8 As shown.

[0093] remember

[0094] l′1=l1+b2 / 2, l′2=l2-b1, k=5 / 6, S1 = b1h1, S2 = b2h2, (β is the section torsional coefficient), denoted as

[0095] At this time F z and The relationship can be written as

[0096]

[0097] in E is the Young's modulus of the material, and G is the shear modulus of the material.

[0098] 2) Apply M alone x

[0099] Apply M alone x The forces acting on Figure 9 As shown.

[0100] Remember λ1=l′1 / r, λ2=l′2 / r,

[0101] At this time M x and The relationship can be written as

[0102]

[0103] in

[0104] 3) Apply F alone y

[0105] Figure 10 To apply F individually y A force diagram.

[0106] remember remember

[0107]

[0108] At this time F y and The relationship can be written as

[0109]

[0110] in

[0111] 4) Apply M separately z

[0112] Figure 11 To apply M individually z A force diagram.

[0113] remember

[0114] At this time M z and The relationship can be written as

[0115]

[0116] in Due to the symmetry of the elastic body structure, it is easy to obtain

[0117]

[0118] in

[0119] Improved mechanical model

[0120] By comparing the solution results of the above mechanical model with those of the Ansys finite element analysis, it can be seen that the solution error of the above mechanical model is approximately 5% to 20%, which is relatively large. For the optimal design of the elastic structure, a highly accurate mechanical model is required. The chosen method for model optimization is to use an artificial neural network for compensation.

[0121] Backpropagation (BP) neural networks are a common type of artificial neural network model. They are widely applicable and possess strong nonlinear fitting capabilities, allowing the neural network to fit complex nonlinear relationships by introducing nonlinear mappings through activation functions. These advantages make BP neural networks highly suitable for solving the current problem of establishing nonlinear relationships between the dimensional parameters of an elastic body and the errors in a mechanical model.

[0122] For the three-T beam elastic body structure shown, there are a total of 7 dimensional parameters. Therefore, four "7-input-1-output" BP neural networks can be established, corresponding to F respectively. z M x F y M z Compensation of the mechanical model under four conditions. When building the BP neural network, it is necessary to select an appropriate number of hidden layers and nodes. Since the designed neural network has relatively few input parameters, one hidden layer is sufficient. Typically, the number of nodes is twice the number of input parameters. To avoid overfitting, the number of nodes is set to 10. The structure of the established BP neural network is as follows... Figure 12 As shown.

[0123] 200 sets of data were generated using ANSYS simulation. The simulation results were compared with the results calculated by the existing mechanical model to obtain 200 sets of compensation coefficients. These 200 sets of dimensional parameters and their corresponding 200 sets of compensation coefficients were combined into a dataset, with the dimensional parameters as input and the compensation coefficients as output. 150 sets were randomly selected as the training set, and the remaining 50 sets as the test set. After training with the training set, the test set was input into the neural network to examine the compensation effect of the neural network on the mechanical model. The results are as follows. Figure 13 As shown in the figure, the blue curve represents the solution error of the mechanical model before adding neural network compensation to the test set parameters, and the red curve represents the solution error of the mechanical model after adding neural network compensation. It can be seen that the solution error has decreased significantly, with F... z For example, the solution error of the mechanical model decreased from about 20% to less than 5%. Similarly, the error in other cases was also controlled within 5%. Therefore, the neural network demonstrated excellent performance on the test set, indicating that the model has the ability to generalize on new data.

[0124] Optimal design of elastomer size and structure

[0125] Based on the optimized elastic body mechanical model described above, optimization algorithms can be used to optimize the dimensions of the elastic body structure. Common optimization algorithms include gradient descent, genetic algorithms, simulated annealing, and particle swarm optimization (PSO). Among these, PSO is suitable for complex multi-dimensional problems, is easy to implement, and is applicable to the optimization design of the elastic body structure in this invention.

[0126] The fitness function is crucial to the optimization process, and its design must reflect the optimization objective. Furthermore, certain constraints must be satisfied regarding the particle's position and velocity. In this study, the constraints and optimization objective for the optimization design of the elastic body are defined as follows:

[0127] • Size constraints: There are strict range requirements for every dimensional parameter in the elastomer structure;

[0128] • Deformation constraint: The morphology of the elastic body cannot be too large or too small;

[0129] • Proximity target: When an elastic body is subjected to the same magnitude of force / torque in different directions, the resulting deformation is similar;

[0130] • Sensitivity target: Maximize the deformation within the constraints.

[0131] Based on the above four points, the constraints and fitness function for the optimization design are given as follows:

[0132]

[0133] F=ω1f1+ω2f2#(4-8)

[0134] Where L1 represents the length of the elastic beam, b1 represents the width of the elastic beam, h1 represents the height of the elastic beam, L2 represents the length of the flexible leaf spring, b2 represents the width of the flexible leaf spring, h2 represents the height of the flexible leaf spring, r represents the radius of the loaded frustum, and i = 1 to n represent the six-dimensional force F in sequence. x ~M z D i This indicates the maximum effective displacement caused by the force / torque. ω1 represents the average value of the maximum effective displacement caused by the six-dimensional force, and ω2 represents the weights of the proximity target and the sensitivity target, respectively.

[0135] Six-dimensional force decoupling method

[0136] The small capacitive six-dimensional force sensor has a corresponding six-dimensional force decoupling algorithm designed based on its specific capacitor arrangement, and the relative error of its decoupling is within 0.5%.

[0137] When the designed six-dimensional force sensor is subjected to an external force, the elastic body deforms, and the moving electrode plate fixed to the elastic body also displaces accordingly, while the stationary electrode plate remains stationary on the base. Therefore, the displacement of the upper electrode plate will cause a change in the electrode spacing, which in turn will cause a change in capacitance. Decoupling is performed in the following order:

[0138] C→d→F / T

[0139] In such Figure 14 In the designed structure, each capacitor corresponds one-to-one with the displacement of each point on the upper electrode plate, where the horizontal capacitors C1, C2, and C3 correspond to the normal displacement d at each point. n1 d n2 d n3 The vertical capacitors C4, C5, and C6 correspond to the tangential displacement d at each location. s4 d s5 d s6 Therefore, the step of solving for the displacement change from the capacitance change does not involve coupling issues; only the coupling between the displacement change and the six-dimensional force needs to be considered.

[0140] 1) Solve for F z

[0141] According to the mechanical model:

[0142]

[0143] Furthermore,

[0144]

[0145] Considering torque M x The resulting coupling problem: when F is applied z When M is applied simultaneously x , by M x Caused Δd n1 , Δd n2 , Δd n3 The following relationship must be satisfied:

[0146] Δd n1 =-2Δd n2 =-2Δd n3

[0147] Therefore, the calculation result of equation (4-7) will not change, that is, there is no coupling.

[0148] Considering torque M y The resulting coupling problem: when F is applied z When M is applied simultaneously y , by M y Caused Δd n1 , Δdn2 , Δd n3 The following relationship must be satisfied:

[0149] Δd n1 =0

[0150] Δd n2 =-Δd n3

[0151] Therefore, the calculation result of equation (4-7) will not change, that is, there is no coupling.

[0152] In addition, F x F y M z Neither will cause Δd n1 , Δd n2 , Δd n3 Since the variables change, there is no coupling. Equation (4-7) is F. z The decoupling formula.

[0153] 2) Solve for M x

[0154] According to the mechanical model:

[0155]

[0156] Furthermore,

[0157]

[0158] Considering torque M y Coupling problems caused by M: when applying M x When M is applied simultaneously y , by M y Caused Δd n1 , Δd n2 , Δd n3 Satisfy the following relationship

[0159] Δd n1 =0

[0160] Δd n2 =-Δd n3

[0161] Therefore, the calculation result of equation (4-8) will not change, that is, there is no coupling.

[0162] Considering force F z Coupling problems caused by M: when applying M x When, if F is applied simultaneously z , by F z Caused Δd n1 , Δd n2, Δd n3 Satisfy the following relationship

[0163] Δd n1 =Δd n2 =Δd n3

[0164] This will change the calculation result of equation (4-8), requiring decoupling. The method is to first calculate F from equation (4-7). z Then F z The resulting displacement change is calculated and substituted into equation (4-8) to eliminate it in advance. The specific process is as follows:

[0165]

[0166] In addition, F x F y M z Neither will cause Δd n1 , Δd n2 , Δd n3 Since the variables change, there is no coupling. Equation (4-9) is M. x The decoupling formula.

[0167] 3) Solve for M y

[0168] According to the mechanical model:

[0169]

[0170] Furthermore,

[0171]

[0172] Considering force F z Coupling problems caused by M: when applying M y When, if F is applied simultaneously z , by F z Caused Δd n1 , Δd n2 , Δd n3 Satisfy the following relationship

[0173] Δd n1 =Δd n2 =Δd n3

[0174] Therefore, the calculation result of equation (4-10) will not change, that is, there is no coupling.

[0175] Considering torque M x Coupling problems caused by M: when applying M y When M is applied simultaneously x , by Mx Caused Δd n1 , Δd n2 , Δd n3 Satisfy the following relationship

[0176] Δd n1 =-2Δd n2 =-2Δd n3

[0177] Therefore, the calculation result of equation (4-10) will not change, that is, there is no coupling.

[0178] In addition, F x F y M z Neither will cause Δd n1 , Δd n2 , Δd n3 Since the variables change, there is no coupling problem. Equation (4-10) is M. y The decoupling formula.

[0179] 4)M z Solve

[0180] According to the mechanical model:

[0181]

[0182] Furthermore,

[0183]

[0184] Considering force F x Coupling problems caused by M: when applying M z When, if F is applied simultaneously x , by F x Caused Δd s4 , Δd s5 , Δd s6 Satisfy the following relationship

[0185] Δd s6 =-2Δd s4 =-2Δd s5

[0186] Therefore, the calculation result of equation (4-11) will not change, that is, there is no coupling.

[0187] Considering force F y Coupling problems caused by M: when applying M z When, if F is applied simultaneously y , by F y Caused Δd s4 , Δd s5 , Δds6 Satisfy the following relationship

[0188] Δd s4 =-Δd s5

[0189] Δd s6 =0

[0190] Therefore, the calculation result of equation (4-11) will not change, that is, there is no coupling.

[0191] Considering torque M x Coupling problems caused by M: when applying M z When M is applied simultaneously x , by M x Caused Δd s4 , Δd s5 , Δd s6 Satisfy the following relationship

[0192] Δd s4 =-Δd s5

[0193] Δd s6 =0

[0194] Therefore, the calculation result of equation (4-11) will not change, that is, there is no coupling.

[0195] Considering torque M y Coupling problems caused by M: when applying M z When M is applied simultaneously y , by M y Caused Δd s4 , Δd s5 , Δd s6 Satisfy the following relationship

[0196] Δd s6 =-2Δd s4 =-2Δd s5

[0197] Therefore, the calculation result of equation (4-11) will not change, that is, there is no coupling.

[0198] In addition, F z It will not cause Δd s4 , Δd s5 , Δd s6 The change in M ​​is such that there is no coupling. Equation (4-11) is M. z The decoupling formula.

[0199] 5)F y Solve

[0200] According to the mechanical model:

[0201]

[0202] Furthermore,

[0203]

[0204] Considering force F x The resulting coupling problem: when F is applied y When, if F is applied simultaneously x , by F x Caused Δd s4 , Δd s5 , Δd s6 Satisfy the following relationship

[0205] Δd s6 =-2Δd s4 =-2Δd s5

[0206] Therefore, the calculation result of equation (4-12) will not change, that is, there is no coupling.

[0207] Considering torque M z The resulting coupling problem: when F is applied y When M is applied simultaneously z , by M z Caused Δd s4 , Δd s5 , Δd s6 Satisfy the following relationship

[0208] Δd s6 =Δd s4 =Δd s5

[0209] Therefore, the calculation result of equation (4-12) will not change, that is, there is no coupling.

[0210] Considering torque M y The resulting coupling problem: when F is applied y When M is applied simultaneously y , by M y Caused Δd s4 , Δd s5 , Δd s6 Satisfy the following relationship

[0211] Δd s6 =-2Δd s4 =-2Δd s5

[0212] Therefore, the calculation result of equation (4-12) will not change, that is, there is no coupling.

[0213] Considering torque M x The resulting coupling problem: when F is applied y When M is applied simultaneously x , by M x Caused Δd s4 , Δd s5 , Δd s6 Satisfy the following relationship

[0214] Δd s4 =-Δd s5

[0215] Δd s6 =0

[0216] This will change the calculation result of equation (4-12), requiring decoupling. Let the distance between the centroids of the stationary and moving plates be D, then applying M alone... x The following relationships are valid at times

[0217]

[0218] remember Then there is

[0219]

[0220] During decoupling, M is first calculated using equation (4-9). x Then M x The resulting displacement change can be calculated and substituted into equation (4-12) to eliminate it in advance. The specific process is as follows:

[0221]

[0222] In addition, F z It will not cause Δd s4 , Δd s5 , Δd s6 The change in [the variable] means there is no coupling. Equation (4-15) is F. y The decoupling formula.

[0223] 6)F x Solve

[0224] According to the mechanical model:

[0225]

[0226] Furthermore,

[0227]

[0228] Considering force F y The resulting coupling problem: when F is appliedx When, if F is applied simultaneously y , by F y Caused Δd s4 , Δd s5 , Δd s6 Satisfy the following relationship

[0229] Δd s4 =-Δd s5

[0230] Δd s6 =0

[0231] Therefore, the calculation result of equation (4-16) will not change, that is, there is no coupling.

[0232] Considering torque M z The resulting coupling problem: when F is applied x When M is applied simultaneously z , by M z Caused Δd s4 , Δd s5 , Δd s6 Satisfy the following relationship

[0233] Δd s6 =Δd s4 =Δd s5

[0234] This will change the calculation result of equation (4-16), requiring decoupling. The method is to first calculate M from equation (4-11). z Then M z The resulting displacement change is calculated and substituted into equation (4-16) to eliminate it in advance. The specific process is as follows:

[0235]

[0236] Considering torque M x The resulting coupling problem: when F is applied x When M is applied simultaneously x , by M x Caused Δd s4 , Δd s5 , Δd s6 Satisfy the following relationship

[0237] Δd s4 =-Δd s5

[0238] Δd s6 =0

[0239] Therefore, the calculation result of equation (4-17) will not change, that is, there is no coupling.

[0240] Considering torque M y The resulting coupling problem: when F is applied y When M is applied simultaneously y , by M y Caused Δd s4 , Δd s5 , Δd s6 Satisfy the following relationship

[0241] Δd s6 =-2Δd s4 =-2Δd s5

[0242] This will also change the calculation results of equation (4-17), requiring decoupling. Applying M separately y The following relationships are valid at times

[0243]

[0244] remember Then there is

[0245]

[0246] During decoupling, M is first calculated using equation (4-10). x and M y Then M y The resulting displacement change can be calculated and substituted into equation (4-17) to eliminate it in advance. The specific process is as follows:

[0247]

[0248] In addition, F z It will not cause Δd s4 , Δd s5 , Δd s6 The change in [the variable] means there is no coupling. Equation (4-21) is F. y The decoupling formula.

[0249] In summary, the decoupling algorithm for six-dimensional forces is summarized as follows:

[0250]

[0251] It can also be expressed in the following form

[0252] F / T=Calib*Δd#(4-22)

[0253] In practical applications, the parameters in the decoupling matrix can be obtained through calibration. Δd can be calculated from the capacitance change, and six-dimensional force decoupling can be achieved using equation (4-22).

[0254] Example

[0255] like Figures 1 to 6 As shown, the sensor includes a top cover 1, a three-T beam elastic body 2, an upper electrode plate 3, a lower electrode plate 4, and a base 5. The top cover 1 is fixed to the three-T beam elastic body 2 with screws; the three-T beam elastic body 2 is positioned to the upper electrode plate 3 by a positioning block and fixed with screws; the three-T beam elastic body 2 is positioned to the lower electrode plate 4 by a pin; the lower electrode plate 4 is positioned to the base 5 by a groove; and the three-T beam elastic body 2 is fixed to the base 5 with screws.

[0256] like Figures 1 to 6 As shown, SUS304 stainless steel is selected as the material for the top cover 1, upper electrode plate 3, and base 5, and is processed using machining methods. A212 aluminum alloy is selected as the material for the three-T beam elastomer, and is processed using machining methods. FR-4 sheet metal is selected as the material for the lower electrode plate 4 to fabricate the PCB.

[0257] like Figure 1 As shown, the overall diameter of the top cover 1 is 15mm and the thickness is 1mm;

[0258] like Figure 2 As shown, the overall diameter of the three-T beam type elastomer 2 is 15mm and the thickness is 4mm;

[0259] like Figure 4 As shown, the overall diameter of the lower electrode plate 4 is 13.8 mm and the thickness is 2 mm. The horizontally arranged capacitor electrodes in the lower electrode plate are processed by a windowing process, and the vertically arranged capacitor electrodes are processed by pasting 50 nm copper foil onto the side wall.

[0260] like Figure 5 As shown, the base 5 has an overall diameter of 15mm and a thickness of 5mm.

[0261] The above description provides a further detailed explanation of the present invention in conjunction with specific / preferred embodiments, and it should not be construed that the specific implementation of the present invention is limited to these descriptions. For those skilled in the art, various substitutions or modifications can be made to these described embodiments without departing from the concept of the present invention, and all such substitutions or modifications should be considered within the scope of protection of the present invention. In the description of this specification, the reference to terms such as "an embodiment," "some embodiments," "preferred embodiment," "example," "specific example," or "some examples," etc., indicates that the specific features, structures, materials, or characteristics described in connection with that embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in any suitable manner in one or more embodiments or examples. Without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification and the features of different embodiments or examples. Although the embodiments of the present invention and their advantages have been described in detail, it should be understood that various changes, substitutions, and modifications can be made herein without departing from the scope of protection of the patent application.

Claims

1. A capacitive miniature six-dimensional force sensor, characterized in that, The system includes a three-T-beam elastomer, a lower electrode plate, an upper electrode plate, and a base. The three-T-beam elastomer converts external six-dimensional force into displacement changes. The lower electrode plate is placed on the base, and the upper electrode plate is connected to the three-T-beam elastomer. The electrodes of the upper and lower electrode plates correspond to form six sets of opposing electrodes. The six sets of opposing electrodes include three sets of horizontally arranged electrodes and three sets of vertically arranged electrodes, forming six sets of parallel plate capacitor structures. The three-T-beam elastomer includes elastic beams, flexible leaf springs, and a load-bearing frustum. The three sets of elastic beams and flexible leaf springs form a T-shaped structure to bear the force. The load-bearing frustum located at the center is connected to the elastic beams through the flexible leaf springs to transmit torque. The upper electrode plate is connected to the three-T-beam elastomer. The sensing... When the device is subjected to external force, the deformation of the three-T beam elastic body causes the upper electrode plate to displace, thereby causing a change in the spacing between the plates of the six sets of parallel plate capacitors, and thus converting the information of the six-dimensional force into six capacitor signals for output; the lower electrode plate includes three horizontal electrodes arranged horizontally on the upper surface of the lower electrode plate, and three vertical electrodes arranged vertically on the sidewalls of the three grooves of the lower electrode plate. The three horizontal electrodes and the three vertical electrodes are evenly distributed along the circumferential direction, and are alternately distributed along the circumferential direction. The upper electrode plate includes three horizontal electrodes arranged horizontally on the lower surface of the upper electrode plate, and three vertical electrodes extending vertically downward from the lower surface of the upper electrode plate. The horizontal and vertical electrodes of the upper electrode plate are correspondingly arranged with the horizontal and vertical electrodes of the lower electrode plate.

2. The capacitive miniature six-dimensional force sensor as described in claim 1, characterized in that, The lower electrode plate is based on a circuit board structure, on which a digital capacitor chip is provided, and the six sets of opposing electrodes are all connected to the digital capacitor chip.

3. The capacitive miniature six-dimensional force sensor as described in any one of claims 1 to 2, characterized in that, It also includes a top cover placed on the three-T beam elastomer.

4. An optimized design method for a three-T-beam elastic body structure of a capacitive miniature six-dimensional force sensor according to any one of claims 1 to 3, characterized in that, include: Establish the mechanical model of the aforementioned three-T beam elastic body; The mechanical model is improved by using a BP neural network. The nonlinear fitting capability of the BP neural network is used to compensate for the mechanical model, so as to establish a nonlinear relationship between the elastic body size parameters and the mechanical model error. Based on the mechanical model and the results of ANSYS finite element analysis, a set of compensation coefficients is generated. Multiple sets of data are obtained through simulation, and the simulation results are compared with the calculation results of the mechanical model to obtain the compensation coefficients. The dataset consisting of the size parameters and the corresponding compensation coefficients is divided into training and test sets. The BP neural network is trained using the training set, and the network's compensation effect is tested using the test set. By using a trained BP neural network, the solution error of the mechanical model is reduced to within a preset threshold, thereby obtaining an improved mechanical model. Using the improved mechanical model and the particle swarm optimization algorithm, the dimensions of the three-T beam elastic body are optimized.

5. The optimized design method for the three-T-shaped beam elastic body structure of the capacitive miniature six-dimensional force sensor as described in claim 4, characterized in that, Four 7-input, 1-output backpropagation (BP) neural networks were established, each corresponding to mechanical model compensation under four force conditions. Each BP neural network was configured to receive seven dimensional parameters as inputs: the length, width, and height of the elastic beam; the length, width, and height of the flexible leaf spring; and the radius of the loaded frustum. The output was a compensation coefficient. The four force conditions were as follows: - Situation: A perpendicular force applied along the Z-axis; - Condition: Torque rotating about the X-axis; - Situation: A horizontal force applied along the Y-axis; - Condition: Torque rotating about the Z-axis.

6. The optimized design method for the three-T-shaped beam elastic body structure of the capacitive miniature six-dimensional force sensor as described in claim 4, characterized in that, The constraints and fitness function for the optimal design of the dimensions of the three-T beam elastic body are as follows: ； ； ； in, Indicates the length of the elastic beam, Indicates the width of the elastic beam, Indicates the height of the elastic beam, Indicates the length of the flexible leaf spring, Indicates the width of the flexible leaf spring, Indicates the height of the flexible leaf spring. Indicates the radius of the loaded frustum. The six-dimensional forces are represented sequentially. , This indicates the maximum effective displacement caused by the force / torque. This represents the average value of the maximum effective displacement caused by the six-dimensional force. and These represent the weights of proximity targets and sensitivity targets, respectively.

7. A method for decoupling six-dimensional forces in a three-T-beam elastic body structure of a capacitive miniature six-dimensional force sensor according to any one of claims 1 to 3, characterized in that, include: S1. Establishing the relationship between capacitance and displacement: When the sensor is subjected to an external force, the deformation of the elastic body causes the moving electrode to shift, while the stationary electrode remains fixed to the base, resulting in capacitance changes corresponding to the displacements in each direction. The three horizontal capacitors correspond to the normal displacements, respectively. The three vertical capacitors correspond to tangential displacements respectively. ; S2. Solving for the perpendicular force The coefficients are determined using a mechanical model based on the change in horizontal capacitance. Calculate vertical force The formula is Considering torque at the same time and To mitigate the impact and ensure that there is no coupling; S3. Solving for torque The coefficients determined by the mechanical model based on the change in horizontal displacement. Calculate torque The formula is At the same time, force The coupling effect is decoupled, and the coupling is decoupled, where Indicates the radius of the loaded frustum; S4. Solving for torque : Coefficients determined using a mechanical model based on the change in horizontal displacement. Calculate torque The formula is At the same time, force and torque The coupling effect is mitigated and decoupled. S5. Solving for torque The coefficients are determined by a mechanical model based on the change in vertical displacement. Calculate torque The formula is At the same time, force and The coupling effect is mitigated and decoupled. S6. Solve for horizontal forces : The coefficients determined by the mechanical model using the change in vertical displacement. Calculate horizontal force The formula is ， Simultaneously considering forces The coupling effect is to be resolved and decoupled. S7. Solving for horizontal forces The coefficients are determined by a mechanical model based on the change in vertical displacement. Calculate horizontal force The formula is ， Simultaneously considering forces and torque The coupling effect is to be resolved and decoupled. S8. Achieve decoupling: Determine the parameters in the decoupling matrix through calibration, and calculate the displacement change using the capacitance change. Through formula To achieve decoupling of six-dimensional forces, where and T They represent force and torque, respectively. This is the decoupling matrix.

Images