Multi-dimensional mechanical perception and analysis system based on optical elastomer

Abstract

The invention relates to the field of sensor technology and photoelectric detection, and discloses a multi-dimensional mechanical perception and analysis system based on an optical elastomer, the system comprises a collimation light source, a beam splitter and three spatial dimension detection light paths, a collimation light beam is subjected to beam splitting reflection and then sequentially passes through a polarizer, the optical elastomer and an analyzer, and the three spatial dimension detection light paths are connected with the collimation light source; receiving a signal by a photoelectric detector; the core component optical elastomer contains layered nanometer materials which are oriented in a single direction and arranged in a coplanar mode, and the layered nanometer materials are arranged at the intersection of light paths. Each optical path adopts differential cross-polarization configuration, and the detector is connected to a computing terminal operating a neural network model. According to the invention, the birefringence sensitivity is enhanced through the anisotropic nano composite material, and the multi-channel polarization detection and deep learning algorithm are combined, so that the nonlinear coupling problem of the mechanical response of the soft material is effectively solved, and the high-precision decoupling and anti-interference detection of the multi-dimensional mechanical component are realized.

Description

Technical Field

[0001] This invention relates to the fields of sensor technology and photoelectric detection, specifically to a multidimensional mechanical sensing and analysis system based on optical elastomers. Background Technology

[0002] Multidimensional mechanical sensing technology is crucial in fields such as tactile perception and feedback in intelligent robots and biomedical engineering. Traditional electrical sensors typically use rigid materials, which are difficult to conformally fit with flexible surfaces. Furthermore, they pose interference and safety hazards in strong electromagnetic fields or flammable and explosive environments. Complex circuit wiring also limits their miniaturization and integration.

[0003] Optical sensing is suitable for soft robot perception due to its advantages of interference resistance, intrinsic safety, and passive detection. However, existing solutions are mostly limited to two-dimensional planar analysis. When faced with complex multidimensional force fields, the birefringent light intensity signal exhibits a highly nonlinear response to external forces, and the overlapping optical effects of force components in different directions make it difficult for a single detection optical path to achieve effective separation of force components.

[0004] Furthermore, conventional isotropic soft materials have limited sensitivity and lack microstructures to aid in orientation identification. In addition, the large deformation, hysteresis, and viscoelasticity of polymer materials make accurate physical modeling difficult. Traditional linear regression or stripe counting methods have low accuracy when processing such nonlinear coupled data, making it difficult to meet the requirements for real-time, high-precision multidimensional force decoupling. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a multidimensional mechanical sensing and analysis system based on optical elastomers. This system solves the problems of existing multidimensional force sensors being complex in structure, susceptible to electromagnetic interference, and suffering from severe coupling and low decoupling accuracy of multidimensional force components due to the nonlinear characteristics of force state and optical response in soft material mechanical detection.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solution: a multidimensional mechanical sensing and analysis system based on optical elastomers. This system utilizes the birefringence response of anisotropic nanocomposite materials under external force combined with neural network algorithms to achieve decoupled detection of multidimensional mechanical information.

[0007] The system includes a collimating light source, a beam splitter, a multi-dimensional detection optical path, an optical elastomer, and a computing terminal. The monochromatic beam output from the collimating light source is modulated by the beam splitter and then directed into the detection optical paths in three spatial dimensions. Each detection optical path is equipped with a polarizer, an optical elastomer, an analyzer, and a photodetector in sequence along the light propagation direction, and the optical paths converge at the center of the optical elastomer.

[0008] The optical elastomer, serving as the core sensing element, is positioned at the intersection of the three detection optical paths, with its upper and lower surfaces fixedly connected to a transparent substrate. This optical elastomer is composed of a polymer matrix and dispersed layered nanomaterials, with the internal layered nanomaterials exhibiting a unidirectional orientation and coplanar arrangement. This microstructure endows the material with specific optical anisotropy, allowing the orientation of the nanosheets to change according to the stress state when subjected to external mechanical forces, resulting in characteristic birefringence changes.

[0009] To extract rich and non-redundant mechanical feature information from a single optical elastomer, this invention differentiates the polarization states of the three optical paths. Polarizer 1 and analyzer 1 form an orthogonal polarization structure, with the transmission axis of polarizer 1 parallel to the alignment direction of the nanomaterials and the transmission axis of analyzer 1 perpendicular to this alignment direction. Polarizer 2 and analyzer 2 form an orthogonal polarization structure, with their transmission axes both at a 45-degree angle to the vertical direction. Polarizer 3 and analyzer 3 form an orthogonal polarization structure, with their transmission axes also at a 45-degree angle to the alignment direction of the nanomaterials. This orthogonal design in both spatial and polarization dimensions allows the three photodetectors to capture the light intensity transmission changes of the optical elastomer under the influence of different dimensional force components.

[0010] To adapt to the mechanical detection needs of different scenarios, the system of this invention supports simultaneous sensing and analysis of multiple mechanical types. The system can simultaneously sense, analyze, and calculate at least two different types of mechanical information acting on the optical elastomer; the mechanical information includes basic mechanical quantities and biomechanical property parameters; wherein the basic mechanical quantities are selected from at least one of shear force, pressure, tension, torque, and bending moment; and the biomechanical property parameters are selected from at least one of tissue elastic modulus, biological tissue viscoelasticity, cell traction force, cell adhesion force, surface friction coefficient, and dynamic physiological load.

[0011] In terms of material preparation, the polymer matrix can be selected from polyacrylamide hydrogel, polyvinyl alcohol hydrogel, polydimethylsiloxane, silicone rubber, or polyurethane; the layered nanomaterials can be selected from titanium phosphate nanosheets, graphene oxide, layered silicate clay nanosheets, or transition metal carbonitrides. The material hardness and elastic modulus of the optical elastomer can be controlled by adjusting the ratio of monomers to crosslinking agents in its polymer matrix, so that the force sensing value of the system can cover the range of 0.001N to 1000N, to meet the needs of cross-scale applications from minimally invasive medical tactile sensing (millinews) to industrial robot grasping (hundreds of Newtons).

[0012] The orientation structure of the internal nanomaterials is formed through processes such as magnetic field induction, electric field induction, rheological shear induction, or mechanical stretching induction. The optical elastomer is designed as a polyhedral structure, ensuring that the surface area of ​​each of its light-transmitting surfaces is sufficient to cover the incident light spot.

[0013] The system's signal analysis relies on data processing software running on a computing terminal. This software incorporates a neural network model to process the light intensity data acquired by the photodetector. This neural network model comprises an input layer, a feature extraction layer, a feature fusion layer, and an output layer. The feature extraction layer is designed with multiple parallel, specialized subnetworks, each independently processing the light intensity signal from a single detection direction to extract specific light response features for that direction. The feature fusion layer uses a joint parameter allocation mechanism to weightedly fuse the features output from each subnetwork, thereby achieving decoupled output of the mechanical components.

[0014] In addition, the data processing software includes a user interface for real-time visualization of at least one of the following information: the transmitted light intensity signal acquired by the photodetector in each detection optical path; multidimensional mechanical parameters calculated by a calculation model based on the transmitted light intensity signal, including spatial resultant force, component forces in each direction, torque and bending moment; and stress distribution diagram (or force state distribution diagram) of the optical elastic body reconstructed based on the mechanical parameters.

[0015] Furthermore, the system features automated data acquisition and calibration. Data processing software controls the loading device to apply horizontal and vertical displacement components to the optical elastomer at preset precision steps, while simultaneously recording the transmitted light intensity values ​​of each photodetector. This process establishes a mapping database between the mechanical state and optical response, used to train a neural network model, enabling it to accurately predict unknown multidimensional mechanical signals. The collimated light source's output wavelength covers the ultraviolet to near-infrared band, and the spot size is adjustable to accommodate optical elastomers of different sizes and detection requirements.

[0016] This invention provides a multidimensional mechanical sensing and analysis system based on optical elastomers. It has the following beneficial effects:

[0017] 1. This invention uses an optical elastomer containing a single-direction oriented and coplanarly arranged layered nanomaterial as the sensing core. This anisotropic microstructure endows the material with a birefringence effect when subjected to force, enabling a single solid element to produce a highly sensitive optical response to external mechanical force. This avoids the dependence of traditional multidimensional force sensors on complex mechanical decoupling structures and realizes the miniaturization and integration of the detection system.

[0018] 2. This invention constructs three spatially intersecting detection optical paths and configures differentiated polarization states with specific angular relationships to the nanomaterial arrangement direction. By combining parallel, perpendicular, and 45-degree angled polarization measurement modes, the system can comprehensively capture optical phase delay information induced by different force components. This multi-dimensional and multi-view detection method eliminates signal ambiguity in single-path measurement and provides complete optical feature data for the accurate decoupling of multi-dimensional force information.

[0019] 3. This invention utilizes a neural network model containing parallel specialized subnetworks and feature fusion layers to process light intensity signals. It can selectively extract feature information from each optical path and perform weighted fusion, solving the complex nonlinear mapping problem between the stress state and transmitted light intensity during the birefringence change process under external force. Furthermore, the system supports synchronous decoupling of multiple mechanical modes such as shearing, tension, and torque, has a wide and adjustable range, and possesses intuitive visualization analysis functions, greatly enhancing its practical value. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the structure of a multidimensional mechanical sensing and analysis system based on an optical elastomer according to an embodiment of the present invention;

[0021] Figure 2 This is a schematic diagram of the deformation of the optical elastomer under different stress conditions according to an embodiment of the present invention;

[0022] Figure 3 The present invention provides a curve showing the change of normalized birefringent light intensity as a function of force when detecting force in the horizontal direction, as detected by a photodetector in direction one.

[0023] Figure 4 This is a curve showing the change of normalized birefringent light intensity with force detected by photodetector two in direction two when detecting force in the horizontal direction, provided by an embodiment of the present invention.

[0024] Figure 5 This is a curve showing the change of normalized birefringent light intensity with force detected by photodetector three in direction three when detecting force in the horizontal direction, provided by an embodiment of the present invention.

[0025] Figure 6 This is a graph showing the change in training loss of the neural network model as a function of the number of training rounds in an embodiment of the present invention.

[0026] Figure 7 This is a graph showing the change in model test accuracy as a function of the number of training epochs in an embodiment of the present invention.

[0027] Figure 8This is a curve showing the change of normalized birefringent light intensity over time detected by a photodetector in direction one when real-time detection of a force of unknown magnitude and direction. This is an embodiment of the present invention.

[0028] Figure 9 This is a curve showing the change of normalized birefringent light intensity over time detected by photodetector two in direction two when real-time detection of a force of unknown magnitude and direction provided by an embodiment of the present invention.

[0029] Figure 10 This is a curve showing the change of normalized birefringent light intensity over time when a force of unknown magnitude and direction is detected in real time by a photodetector in direction three. This is an embodiment of the present invention.

[0030] Among them, 1. Collimating light source; 2. Beam splitter; 3. Mirror 1; 4. Polarizer 1; 5. Analyzer 1; 6. Mirror 2; 7. Polarizer 2; 8. Analyzer 2; 9. Mirror 3; 10. Mirror 4; 11. Polarizer 3; 12. Analyzer 3; 13. Optical elastomer; 14. Substrate; 15. Photodetector 1; 16. Photodetector 2; 17. Photodetector 3. Detailed Implementation

[0031] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0032] See attached document Figure 1 The system of the present invention includes a collimating light source 1, a beam splitter 2, a set of reflectors (reflector 1 3, reflector 2 6, reflector 3 9, reflector 4 10), three sets of polarization / analyzing components, an optical elastomer 13, a substrate 14, and photodetectors (photodetector 1 15, photodetector 2 16, photodetector 3 17).

[0033] Collimated light source 1 outputs collimated monochromatic light with a wavelength covering the ultraviolet to near-infrared band (specifically selectable from 200nm to 1600nm) and an adjustable spot size. The light source wavelength can be flexibly selected based on the actual application scenario and the light transmission characteristics of the selected optical elastomer matrix material. For example, in biomechanical detection scenarios, more penetrating near-infrared light (such as 850nm or 1310nm) can be selected; visible light can be selected in conventional laboratory scenarios.

[0034] The core of the system lies in the differentiated polarization configuration of each optical path:

[0035] In the optical path of direction one, polarizer-4 and analyzer-5 are orthogonal, and the transmission axis forms an angle of 0° and 90° with the arrangement direction of the nanomaterials inside the optical elastomer 13, respectively.

[0036] In the second optical path, polarizer 7 and analyzer 8 are orthogonal, and their transmission axes form angles of +45° and -45° with the vertical direction, respectively.

[0037] In the three-directional optical path, polarizer 311 and analyzer 312 are orthogonal, and their transmission axes form angles of +45° and -45° with the nanomaterial arrangement direction, respectively.

[0038] An optical elastomer 13 is placed at the intersection of the three optical paths, with its upper and lower surfaces connected to a substrate 14 having a transmittance greater than 50%. Photodetectors (photodetector one, 15; photodetector two, 16; photodetector three, 17) are located at the ends of each optical path, their response wavelengths covering the light source range, and their outputs are connected to a computing terminal. This system is based on the principle of polarized light interference, synchronously acquiring the intensity signals of the three transmitted light paths through data processing software, and calculating multidimensional mechanical information based on a mapping model.

[0039] See attached document Figure 1 and attached Figure 2 The optical elastomer 13 is configured as a polyhedral structure (including but not limited to right-angled square prisms, hexagonal prisms, and other structures with light-transmitting surfaces), and the surface area of ​​each light-transmitting surface is preferably set to be four times or more than the area of ​​the incident light spot to eliminate edge effects. The optical elastomer 13 is composed of a polymer matrix (selected from polyacrylamide / polyvinyl alcohol hydrogel, PDMS, silicone rubber, or polyurethane) and layered nanomaterials (selected from titanium phosphate, graphene oxide, layered silicate clay, or MXene) dispersed therein.

[0040] To meet the detection requirements of different ranges, the mechanical properties of the matrix can be precisely controlled by adjusting the molar ratio of monomer to crosslinking agent in the precursor solution during the preparation of optical elastomer 13. For example, in a polyacrylamide hydrogel matrix, increasing the proportion of crosslinking agent (such as N,N'-methylenebisacrylamide) can improve the elastic modulus (hardness) of the material.

[0041] If applied to micro-biomechanical detection, the proportion of crosslinking agent can be reduced to prepare low-modulus elastomers, enabling the system to sense a range from 0.001N to 1N.

[0042] If applied to robotic tactile analysis, the proportion of crosslinking agent can be increased to cover a sensing range of 10N to 1000N.

[0043] The detection principle of this system is based on the birefringence response under external force. The data processing software runs on a computing terminal and performs the following core functions:

[0044] Multidimensional mechanical decoupling: The system is not limited to detecting simple pressure or shear force, but can also simultaneously sense and analyze complex mechanical combinations acting on the optical elastomer, including shear force, pressure, tension, torque, and bending moment. During the calibration phase, a combined load containing the above mechanical types is applied through a multi-degree-of-freedom loading platform, and the neural network is trained to have the ability to decouple these five mechanical components.

[0045] Interactive Data Visualization: The data processing software includes a user interface that provides real-time data monitoring and analysis capabilities. The interface features three display modules:

[0046] Raw signal area: Real-time waveform display of transmitted light intensity signals acquired by the three photodetectors helps users determine signal stability;

[0047] Mechanical parameters area: Displays the calculated multi-dimensional mechanical parameters (magnitude of spatial resultant force, direction angle, values ​​of each component force, torque value, etc.) in the form of a digital dashboard or vector arrows;

[0048] State Reconstruction Zone: Based on the calculated boundary mechanical conditions, combined with finite element analysis or pre-trained field mapping models, the stress distribution cloud map inside the optical elastic body is drawn and displayed in real time, intuitively showing the area of ​​concentrated stress.

[0049] Its internal microstructure has a unidirectional orientation and coplanar arrangement, which endows the material with birefringence properties. );in, Indicates the birefringence of a material. The intrinsic birefringence coefficient (depends on the inherent properties of the material). The volume concentration of nanomaterials. This is the orientation orderliness parameter.

[0050] This oriented structure is primarily prepared using a strong external field induction process, preferably magnetic field induction or electric field induction. Electric field induction involves applying a high-voltage DC or alternating electric field (e.g., 1 kV / mm to 5 kV / mm) before the polymer precursor cures. Utilizing the dielectric anisotropy or induced dipole moment of the nanomaterials, the nanosheets are driven to rotate and align along the electric field direction, and then cured and shaped under these conditions. Magnetic field induction involves placing the mixture in a strong magnetic field (>0.5T) to achieve orientation using magnetic torque. Additionally, depending on the specific preparation conditions, mechanical methods such as rheological shear induction or mechanical stretching induction can be used as supplementary methods, employing shear flow fields or macroscopic deformation of the matrix to drive the directional alignment of the nanomaterials.

[0051] The optical elastomer 13 is attached to a rigid transparent substrate 14 (such as glass, PMMA, or PC) with a light transmittance greater than 50%. The optical elastomer has reversible elastic deformation capability under an external force of 0N to 100N, and the optical signal error remains within 0.01% after 1000 cycles of loading test.

[0052] The detection principle of this system is based on the birefringence response and polarized light interference under external force. The optical elastomer 13 possesses initial anisotropy due to its internal oriented nanomaterials. When subjected to external force, material deformation and microstructural rearrangement lead to a change in the birefringence state. The intensity of the transmitted light received by the photodetector... Follow the formula: ;in, The angle between the optical principal axis and the polarization axis. The phase delay is related to the difference in mechanical state and the material thickness.

[0053] like Figure 2 As shown, when subjected to horizontal shear force At this time, the material undergoes shear deformation, causing the internal nanomaterials to rotate (change). ) and changes in orientation (changes) When subjected to vertical pressure At this time, the material is compressed, changing the optical path. and volume concentration Since a single optical path cannot distinguish force components in different directions (due to multi-valuedness), this system constructs three paths with differentiated polarization angles. The detection optical paths (with different preset values) allow the three detectors to capture the optical response characteristics of the force vector in different dimensions, forming a complementary signal combination. Given the highly nonlinear coupling between light intensity and stress state under large deformation of soft materials, which cannot be directly inverted using analytical formulas, this system employs a data-driven strategy, utilizing neural networks to establish a mapping relationship between optical signals and multidimensional force vectors.

[0054] The data processing software runs on a computing terminal, and its core functions include data acquisition and calibration as well as real-time model solving.

[0055] First, execute the calibration procedure: control the loading device to apply a step-by-step scanning load to the optical elastomer 13 in both the horizontal and vertical directions. Set the horizontal displacement step size to... The vertical displacement step size is The system traverses all combined force states within its coverage range. It synchronously records the force values ​​measured by the force sensor under each loading condition. and the intensity values ​​of the three light sources , construct the training dataset.

[0056] A neural network model was then constructed, comprising an input layer, a feature extraction layer, a feature fusion layer, and an output layer. To address the varying sensitivity of the optical path to the force component across different detection directions, the feature extraction layer incorporated multiple parallel, specialized subnetworks, each independently processing the light intensity signal from a single channel. The feature fusion layer, acting as a joint parameter allocation layer, weighted and fused the features output from each subnetwork, ultimately outputting the predicted force component. The model was trained on a dataset using the backpropagation algorithm to learn the nonlinear mapping from the light intensity space to the force space. .

[0057] To verify the effectiveness of the system, the following two sets of implementation examples were tested.

[0058] Example 1: Static Response Characteristics and Model Training

[0059] This embodiment aims to verify the system's response to static forces. A green collimated light source with a center wavelength of 532 nm was used in the experiment, and the light spot was adjusted to a radius of 1 mm. The optical elastomer 13 is a cubic structure with dimensions of 10 mm × 10 mm × 10 mm, filled with magnetically induced oriented titanium phosphate nanosheets.

[0060] A static shear force ranging from 0 N to 1.2 N along the horizontal direction was applied to the optical elastomer 13 using a precision mechanical stage. Data processing software recorded the changes in the intensity of the three light sources, and the results are as follows: Figure 3 , Figure 4 and Figure 5 As shown:

[0061] In direction one ( Figure 3 The normalized light intensity increases nonlinearly and monotonically with increasing shear force, approaching saturation.

[0062] In direction two ( Figure 4 The light intensity also increases monotonically, but the slope of its curve is significantly different from that of the direction, reflecting the difference in the mechanism of the change in birefringence light intensity signal caused by the rotation of the nanosheet in different directions.

[0063] In direction three ( Figure 5 The light intensity decreases monotonically with increasing shear force, complementing the first two paths.

[0064] The combination of the three signals exhibits the characteristic of no multivaluedness, proving the effectiveness of the optical path configuration. Based on the acquired data, a dataset was constructed and a neural network was trained using an error threshold of 0.01N. Figure 6 and Figure 7 As shown, the model training loss decreased rapidly in the first few rounds, the test set accuracy converged at about 20 rounds, and the prediction accuracy of static shear force stabilized at over 99%, verifying the model's static solution capability.

[0065] Example 2: Real-time detection of dynamic multidimensional forces

[0066] This embodiment aims to verify the decoupling accuracy of the system under dynamic composite stress. To test the applicability to units of different sizes, the light source was replaced with 620nm red light, and the optical elastomer 13 was replaced with a structure with dimensions of 15mm × 15mm × 15mm.

[0067] First, high-density calibration was performed, with the force variation step set to 0.05N and the scanning range expanded to 0N to 5N. The model was then retrained. Dynamic testing followed: a resultant force of approximately 3N was applied to the optoelastomer at a 45° angle to the horizontal. Figure 8 , Figure 9 and Figure 10 As shown, at the moment of force application, the light intensity signals of the three detection channels change abruptly almost simultaneously, with a response delay of less than 10ms, demonstrating excellent dynamic capture capability.

[0068] The data processing software collects light intensity data in real time and inputs it into the model, outputting the prediction result: horizontal component. Vertical component force Vector synthesis calculations showed that the detected resultant force was 2.99 N and the direction angle was 44.05°. Compared with the actual applied value (resultant force 3 N, angle 45°), the resultant force error was controlled within ±0.01 N and the angle error within ±1°.

[0069] In addition, the torque detection function was also tested in this embodiment. A small torque (approximately 50 mN·m) was applied around the Z-axis, and the software interface successfully captured the characteristic fluctuations of the three light intensity signals (especially the phase difference changes in directions two and three), and output the torque prediction value of 48.5 mN·m on the interface. At the same time, the stress concentration phenomenon at the material edge was displayed in the visualized stress distribution map, verifying the system's ability to perceive and analyze complex mechanical types.

Claims

1. A multidimensional mechanical sensing and analysis system based on optical elastomers, characterized in that, It includes a collimated light source (1), a beam splitter (2), a three-dimensional detection optical path, an optical elastomer (13), a substrate (14), and a computing terminal; The beam output by the collimating light source (1) is split by the beam splitting device (2) and enters the detection optical paths of direction one, direction two and direction three respectively; Each of the aforementioned detection optical paths is sequentially provided with a polarizer, the optical elastomer (13), an analyzer, and a photodetector along the light propagation direction, and the optical elastomer (13) is located at the intersection of each detection optical path; The upper and lower surfaces of the optical elastomer (13) are connected to the substrate (14), and its interior contains layered nanomaterials that are oriented in a single direction and arranged in a coplanar manner. The signal output terminals of each photodetector are connected to the computing terminal running data processing software.

2. The multidimensional mechanical sensing and analysis system based on an optical elastomer according to claim 1, characterized in that, The collimated light source (1) is configured to output collimated monochromatic light with a wavelength range covering 200nm to 1600nm, and the output spot size of the collimated light source (1) is adjustable.

3. The multidimensional mechanical sensing and analysis system based on optical elastomers according to claim 1, characterized in that, The polarizer and analyzer in the detection optical path of direction one are referred to as polarizer one (4) and analyzer one (5), both of which are linear polarizers and their transmission axes are orthogonal to each other. The angle between the light transmission axis of the polarizer (4) and the arrangement direction of the layered nanomaterials inside the optical elastomer (13) is set to 0 degrees, and the angle between the light transmission axis of the analyzer (5) and the arrangement direction of the layered nanomaterials inside the optical elastomer (13) is set to 90 degrees.

4. The multidimensional mechanical sensing and analysis system based on an optical elastomer according to claim 1, characterized in that, The polarizer and analyzer in the detection optical path of direction two are referred to as polarizer two (7) and analyzer two (8), both of which are linear polarizers and their transmission axes are orthogonal to each other. The light transmission axis of the polarizer 2 (7) and the light transmission axis of the analyzer 2 (8) are configured to form an angle of +45 degrees and -45 degrees with the vertical direction, respectively, so as to keep them orthogonal to each other.

5. The multidimensional mechanical sensing and analysis system based on an optical elastomer according to claim 1, characterized in that, The polarizer and analyzer in the detection optical path of direction three are referred to as polarizer three (11) and analyzer three (12), both of which are linear polarizers and their transmission axes are orthogonal to each other. The light transmission axis of the polarizer three (11) and the light transmission axis of the analyzer three (12) are respectively configured to form an angle of +45 degrees and -45 degrees with the arrangement direction of the layered nanomaterials inside the optical elastomer (13), thereby keeping them orthogonal to each other.

6. The multidimensional mechanical sensing and analysis system based on an optical elastomer according to claim 1, characterized in that, The system is capable of simultaneously sensing, analyzing and calculating at least two different types of mechanical information acting on the optical elastomer (13); The mechanical information includes basic mechanical quantities and biomechanical property parameters; The basic mechanical quantities are selected from at least one of shear force, compressive force, tensile force, torque, and bending moment; The biomechanical properties parameters are selected from at least one of tissue elastic modulus, biological tissue viscoelasticity, cell traction force, cell adhesion force, surface friction coefficient, and dynamic physiological load.

7. The multidimensional mechanical sensing and analysis system based on an optical elastomer according to claim 1, characterized in that, The optical elastomer (13) has a multifaceted structure; the surface area of ​​each of its light-transmitting surfaces is configured to completely cover the incident light spot, and the substrate is made of a high-transmittance material to ensure the stability of the optical path transmission.

8. The multidimensional mechanical sensing and analysis system based on an optical elastomer according to claim 1, characterized in that, The optical elastomer (13) is composed of a polymer matrix and layered nanomaterials dispersed in the polymer matrix; The polymer matrix is ​​selected from polyacrylamide hydrogel, polyvinyl alcohol hydrogel, polydimethylsiloxane, silicone rubber, or polyurethane; The layered nanomaterials are selected from titanium phosphate nanosheets, graphene oxide, layered silicate clay nanosheets, or transition metal carbonitrides.

9. A multidimensional mechanical sensing and analysis system based on an optical elastomer according to claim 8, characterized in that, The material hardness and elastic modulus of the optical elastomer (13) can be controlled by adjusting the ratio of monomers to crosslinking agents in its polymer matrix, so that the force sensitivity range of the system can cover 0.001N to 1000N.

10. A multidimensional mechanical sensing and analysis system based on an optical elastomer according to claim 8, characterized in that, The single-directional orientation and coplanar arrangement of the internal layered nanomaterials of the optical elastomer (13) are formed by magnetic field induction, electric field induction, rheological shear induction or mechanical stretching induction.

11. A multidimensional mechanical sensing and analysis system based on an optical elastomer according to claim 1, characterized in that, The data processing software running on the computing terminal includes a neural network model; The neural network model includes an input layer, a feature extraction layer, a feature fusion layer, and an output layer; The feature extraction layer includes multiple parallel specialized subnetworks, each of which is configured to independently receive light intensity signals from a single detection direction and extract features. The feature fusion layer is a parameter joint allocation layer, configured to perform weighted fusion of the features output by the specialized sub-network.

12. A multidimensional mechanical sensing and analysis system based on an optical elastomer according to claim 11, characterized in that, The data processing software includes a user interface for real-time visualization of at least one of the following information: the transmitted light intensity signal acquired by the photodetector in each detection optical path; and multidimensional mechanical parameters calculated by a calculation model based on the transmitted light intensity signal, including spatial resultant force, component forces in each direction, torque, and bending moment. The stress distribution diagram within the optical elastic body is reconstructed based on the mechanical parameters.

13. A multidimensional mechanical sensing and analysis system based on an optical elastomer according to claim 11, characterized in that, The data processing software is configured to execute data acquisition and calibration procedures: The control loading device applies a multidimensional combined load to the optical elastomer (13), the multidimensional combined load including a horizontal displacement component, a vertical displacement component and a rotation angle component that vary according to a preset step size, and performs step-by-step scanning loading; The data processing software synchronously records the mechanical component values ​​under each loading state, as well as the transmitted light intensity values ​​measured by each photodetector in the detection optical paths located in directions one, two, and three, respectively, to generate a dataset for training the neural network model.

Images