Variable stiffness tactile hair-cutting paper piezoelectric film bionic sensor and system for ornithopter

Abstract

The application provides a variable-stiffness sensory hair-paper piezoelectric film bionic sensor and system for flapping-wing aircraft, the sensor comprising a piezoelectric film layer provided with an out-of-plane stiffness-reducing cut structure, a variable-stiffness sensory hair having a stiffness difference along an axial direction and connected to the center of the film at the bottom end, and a plurality of independent sensing channels in central symmetric distribution and electrical isolation; under external excitation, the sensory hair is deflected to drive the film to deform, and each channel synchronously outputs electrical signals with spatial and temporal distribution differences. The application also provides a sensing system comprising the above-mentioned sensor and an aircraft carrying the system, and through the combination of double gravimetric sensitization and multi-channel spatial coding with a deep learning network, high-precision independent decoupling sensing of flow field parameters and body attitude is realized.

Description

Technical Field

[0001] This invention relates to the field of flexible sensor and intelligent sensing technology, specifically to a variable stiffness-sensing piezoelectric thin film bionic sensor and system for flapping-wing aircraft. Background Technology

[0002] Ornithoptering aircraft, with their biomimetic advantages such as high maneuverability, miniaturization, and low power consumption, have broad application prospects in low-altitude environment monitoring, covert reconnaissance, and disaster relief. However, when ornithoptering aircraft fly autonomously in the complex, unstructured low-altitude atmosphere, they are highly susceptible to the effects of flow field disturbances such as gusts, and the high-frequency nonlinear vibrations caused by their flapping motion can also induce high-frequency changes in their attitude (yaw, roll, pitch). Therefore, real-time and accurate perception of flow field parameters and flight attitude is a prerequisite for achieving stable control of ornithoptering aircraft.

[0003] Currently, sensing for micro-aircraft mainly relies on traditional rigid sensors such as inertial measurement units (IMUs) or pitot tubes from microelectromechanical systems (MEMS). These sensors suffer from common problems such as large weight, high power consumption, and encroachment on payload. Furthermore, they are prone to severe signal drift and response delay under frequent small-amplitude attitude changes and nonlinear high-frequency flapping wing vibrations. In recent years, biomimetic sensors based on flexible materials have gradually emerged. However, existing flexible thin-film sensors mostly use continuous, non-porous substrates with high out-of-plane deflection stiffness, resulting in insufficient sensitivity to weak airflow and minute vibrations. More critically, when dealing with multidimensional complex excitations (such as simultaneous changes in wind direction and aircraft attitude), most sensors exhibit severe cross-coupling of signals from single-channel outputs, creating sensing blind spots and making it difficult to achieve independent decoupling between the multidimensional environment and the aircraft's state. Summary of the Invention

[0004] To address the shortcomings of existing technologies, the purpose of this invention is to provide a variable stiffness-sensing piezoelectric thin film bionic sensor and system for flapping-wing aircraft.

[0005] A biomimetic sensor for a flapping-wing aircraft, based on the present invention, comprising: A piezoelectric thin film layer, wherein the piezoelectric thin film layer is provided with an out-of-plane stiffness reduction shear structure; The variable stiffness sensitive hair has a stiffness difference along the axial direction, and the bottom end of the variable stiffness sensitive hair is connected to the central region of the piezoelectric thin film layer. Multiple independent sensing channels are centrally symmetrically distributed on the surface of the piezoelectric thin film layer, and each independent sensing channel is electrically isolated from the others. Under external excitation, the variable stiffness sensing hair deflects outward and causes the piezoelectric thin film layer to deform. The multiple independent sensing channels synchronously output electrical signals with spatiotemporal distribution differences based on the piezoelectric effect.

[0006] Preferably, the variable stiffness felt hair includes a first hair segment and a second hair segment distributed along the axial direction; The elastic modulus of the first segment is greater than that of the second segment; The bottom end of the second hair segment is connected to the central region of the piezoelectric thin film layer to serve as a flexible hinge structure.

[0007] Preferably, the first segment is made of a high-modulus solid material, and the second segment is made of a low-modulus soft material; The high-modulus solid material includes a UV-curable resin, and the low-modulus soft material includes soft silicone.

[0008] Preferably, the out-of-plane stiffness reduction shear mark structure includes multiple concentrically distributed annular shear marks; An uncut area is retained between adjacent annular cuts, and the uncut area forms a flexible hinge, so that the piezoelectric thin film layer forms a structure in which rigid units and flexible hinges are arranged alternately.

[0009] Preferably, the number of the plurality of independent sensing channels is at least three; The multiple independent sensing channels are arranged in a cross or orthogonal or centrally symmetrical manner, dividing the surface of the piezoelectric thin film layer into multiple electrically isolated fan-shaped regions.

[0010] Preferably, the upper surface of the piezoelectric thin film layer is provided with an electrode layer, which is divided into multiple independent electrodes corresponding one-to-one with the plurality of sector regions, and each of the independent electrodes is connected to a wire extending radially outward; A common grounding electrode layer is provided on the lower surface of the piezoelectric thin film layer.

[0011] Preferably, the piezoelectric thin film layer is a polyvinylidene fluoride thin film; The biomimetic sensor also includes a substrate, on which the piezoelectric thin film layer is encapsulated. The substrate is a flexible polyimide substrate.

[0012] A biomimetic sensing system for flapping-wing aircraft based on a variable stiffness-sensing, paper-cutting piezoelectric thin film, according to the present invention, comprises: The aforementioned flapping-wing aircraft variable stiffness sensing hair-cutting piezoelectric film bionic sensor; The signal acquisition module is electrically connected to the multiple independent sensing channels and is used to synchronously acquire the electrical signals and perform preprocessing to generate multi-channel timing signals. The intelligent solution module has a built-in deep learning network model, which is used to receive the multi-channel time series signal, and perform feature extraction and solution on the multi-channel time series signal through the deep learning network model, and output flow field parameters and / or body attitude angle.

[0013] Preferably, the deep learning network model includes a flow field classification network and / or a multi-output pose regression network; The flow field classification network is configured to use an attention mechanism to assign weights to the feature maps of the multi-channel time-series signals in order to focus on specific frequency bands and output flow field classification results. The multi-output pose regression network includes fully connected layers and is configured to output continuous three-dimensional pose angles based on extracted features.

[0014] An aircraft according to the present invention includes an airframe and a flapping-wing aircraft variable stiffness-sensing hair-cutting piezoelectric film bionic sensing system; The bionic sensor is conformally attached to the surface of the body.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. Mechanical amplification and sensitivity leap: Macroscopically, torque amplification is achieved by using variable stiffness sensing hair, and microscopically, out-of-plane stiffness is reduced by using paper-cutting structure. The dual biomimetic design enables the sensor to have extremely high responsiveness and electromechanical conversion efficiency under extremely low load and micro airflow.

[0016] 2. Completely solve the blind spot of multidimensional sensing coupling: The four-channel symmetrical independent layout enables the excitation from any spatial orientation to generate a unique combination of time-series voltages, transforming the differences in physical space into matrix differences in electrical signals, providing a perfect high-dimensional data source for algorithm decoupling.

[0017] 3. Low-cost, high-precision intelligent solution: Combining CNN-Attention and PSLHS efficient sampling strategies, the time and experimental cost of exhaustive calibration of 3D pose are greatly reduced. With a relatively small network size, high-precision, low-latency continuous state solution with an average absolute error of less than 1.5° is achieved, which perfectly meets the needs of lightweight and low-power micro intelligent drones. Attached Figure Description

[0018] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings: Figure 1 This is a schematic diagram of the overall structure of the bionic sensor according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the out-of-plane stiffness reduction shear mark structure according to an embodiment of the present invention; Figure 3 This is a cross-sectional view of the mechanical transmission and deformation principle of an embodiment of the present invention; Figure 4 This is a graph showing the spatiotemporal distribution characteristics of multi-channel signals according to an embodiment of the present invention. Figure 5 This is a flowchart of the data processing of the intelligent calculation module in an embodiment of the present invention; Figure 6 This is a statistical analysis diagram of attitude calculation error in an embodiment of the present invention. Detailed Implementation

[0019] The present invention will now be described in detail with reference to specific embodiments. These embodiments will help those skilled in the art to further understand the present invention, but do not limit the invention in any way. It should be noted that those skilled in the art can make several changes and improvements without departing from the concept of the present invention. These all fall within the protection scope of the present invention.

[0020] Example 1 like Figures 1 to 3 As shown in the figure, this embodiment provides a variable stiffness-sensing piezoelectric thin film bionic sensor for flapping-wing aircraft, and its fabrication process is as follows.

[0021] In this embodiment, the sensitive core is a flexible polyvinylidene fluoride (PVDF) piezoelectric thin film layer with a thickness of 30-50 micrometers. First, using a high-precision ultraviolet nanosecond laser marking machine, multiple concentric annular shear marks are etched onto the piezoelectric thin film layer to form an out-of-plane stiffness-reducing shear mark structure 5. For example... Figure 2 As shown, uncut areas are retained between adjacent annular cuts, forming flexible hinges 6, thus creating a structure in which rigid units 7 and flexible hinges 6 are alternately arranged in the piezoelectric thin film layer. The optimized key topology parameters can be set as follows: span angle of approximately 70° and radial spacing of 2.5 mm. This paper-cut-based topology design breaks the rigid constraints of continuous thin film materials at the geometric level.

[0022] When the sensor applies a force perpendicular to the membrane surface, the deformation energy is mainly concentrated at the flexible hinge 6, while the rigid unit 7 only undergoes rigid body displacement or slight deformation. This reduces the overall equivalent out-of-plane stiffness of the membrane by several orders of magnitude. Of course, those skilled in the art can adjust the shape (e.g., spiral, petal-shaped), number of turns, spacing, and span angle of the shear marks according to the required sensitivity and range, without departing from the scope of protection of this invention.

[0023] Secondly, a gold (Au) electrode layer is deposited on the surface of the etched piezoelectric thin film using magnetron sputtering, and then precisely divided into four independent orthogonal sector regions using a laser scribing machine, forming multiple independent sensing channels. For example... Figure 1 As shown, multiple independent sensing channels are centrally symmetrically distributed and electrically isolated from each other. Each independent electrode is connected to a radially outward-extending wire 11 to independently export the charge signal of each channel. The lower surface of the piezoelectric thin film layer retains an intact sputtered layer as a common ground electrode layer 10. This top-separated and bottom-common electrode architecture ensures the independence of signal acquisition while simplifying wiring complexity.

[0024] In actual fabrication, the electrode material can also be silver, copper, or a conductive polymer, and the slicing process can also employ photolithography or mask sputtering. It should be noted that the number of channels can also be three, six, or more, and the distribution pattern can be triangular symmetry, hexagonal symmetry, etc., as long as central symmetry and electrical isolation are ensured, multidimensional sensing functionality can be achieved.

[0025] Next, variable stiffness felt is fabricated. The variable stiffness felt is made using a "two-stage molding process": a high Young's modulus (e.g., several gigapascals) UV-curable resin is injected into the upper part of the mold, curing to form the first felt segment (high-stiffness felt shaft); a low Young's modulus (e.g., tens to hundreds of kilopascals) soft silicone material (e.g., Ecoflex) is injected into the lower part of the mold, curing to form the second felt segment (flexible hinge base). The elastic modulus of the first felt segment is greater than that of the second felt segment. For example... Figure 3 As shown, this segmented variable stiffness design constitutes a natural mechanical amplification lever: the first segment at the far end has high stiffness, which can maintain shape stability in the flow field, effectively capture the fluid drag force and transmit it downward as a concentrated load; while the second segment at the near end has low stiffness, which acts as a flexible hinge, and is more likely to produce large local deformation when receiving loads transmitted from the upper part, thereby converting small flow field forces into significant root bending moments.

[0026] Of course, high-modulus solid materials can also be carbon fiber composites, rigid plastics, or metal wires, while low-modulus soft materials can be polyurethane elastomers, hydrogels, or other hyperelastic polymers. This invention does not limit these to a single material. It should be understood that although a two-segment structure is shown in the figure, in other embodiments, the variable stiffness sensor can also employ gradient variable stiffness materials or multi-segment stiffness combinations, as long as there is a stiffness difference along the axial direction to achieve torque amplification.

[0027] Finally, the bottom end of the variable stiffness sensitive hair (i.e., the bottom end of the second hair segment) is vertically and centrally bonded to the central region of the piezoelectric thin film layer using flexible conductive silver paste, and the entire assembly is encapsulated on a flexible polyimide (PI) substrate to complete the fabrication. The PI substrate has good heat resistance and insulation properties, and can adapt to the complex working conditions on the surface of aircraft. Of course, the piezoelectric material can also be replaced with PZT piezoelectric ceramic sheets, ZnO nanowire arrays, or piezoelectric composite materials, and the substrate material can also be selected from PET, PDMS, or flexible PCB boards, etc., as long as the functions of flexible encapsulation and signal transmission can be achieved, they are all within the protection scope of this invention.

[0028] like Figures 1 to 3As shown, the fabricated sensor includes a piezoelectric thin film layer, variable stiffness sensing hairs, and multiple independent sensing channels. During operation, when subjected to external airflow or vibration excitation, the variable stiffness sensing hairs undergo out-of-plane deflection, causing the piezoelectric thin film layer to deform. Based on the piezoelectric effect, the multiple independent sensing channels synchronously output electrical signals with spatiotemporal distribution differences. These signals contain multidimensional information such as the direction, magnitude, and frequency of the external excitation.

[0029] Example 2 When the sensor prepared in Example 1 is mounted on a flapping-wing aircraft, it encounters crosswinds or changes in aircraft attitude that induce inertial forces. The first segment of the variable-stiffness sensor is subjected to force and acts as a rigid lever, transferring the load to the second segment at the bottom in a bent rectangular manner. Thanks to the low modulus of the second segment and the dual flexibility of the out-of-plane stiffness reduction shear structure 5 of the piezoelectric thin film layer, the piezoelectric thin film layer undergoes significant localized asymmetric out-of-plane warping. For example, as... Figure 4 As shown, when the force direction is directly above an independent sensing channel, that channel experiences compressive stress, the opposite channel experiences tensile stress, and the other two channels are in a stress transition zone. Due to the characteristics of piezoelectric polarization, the four independent sensing channels synchronously output AC voltage signals with different amplitude polarities, phase differences, and amplitude ratios. This direct mapping from mechanics to electricity enables parallel encoding of spatially continuous vectors, completely overcoming the signal coupling and orientation blind spots faced by single-channel sensors.

[0030] Example 3 This embodiment provides a variable stiffness hair-cutting piezoelectric thin film biomimetic sensing system for flapping-wing aircraft. The system includes the variable stiffness hair-cutting piezoelectric thin film biomimetic sensor for flapping-wing aircraft as described in Embodiment 1, a signal acquisition module, and an intelligent calculation module. The signal acquisition module is electrically connected to multiple independent sensing channels to synchronously acquire electrical signals and perform preprocessing to generate multi-channel time-series signals. The intelligent calculation module incorporates a deep learning network model to receive the multi-channel time-series signals and perform feature extraction and calculation on the signals using the deep learning network model, outputting flow field parameters and / or body attitude angles.

[0031] Specifically, such as Figure 5 As shown, the sensing system hardware includes a sensor array, a multi-channel charge amplifier, a high-speed data acquisition card (such as an NI-DAQ card with a sampling rate set to 1000Hz), and an onboard edge computing unit. The signal acquisition module not only performs analog-to-digital conversion but also integrates impedance matching circuitry and a Butterworth low-pass filter to eliminate high-frequency electromagnetic noise and adapt to the high internal resistance characteristics of the piezoelectric sensor. After hardware filtering, the signal is divided into a sliding window sequence with a fixed time step and normalized to form a standardized tensor input, providing a high-quality data foundation for subsequent intelligent computation.

[0032] Regarding the specific execution logic of the intelligent solution module, the deep learning network model includes a flow field classification network and / or a multi-output pose regression network.

[0033] 1) Decoupling of flow field parameters: Four-channel time-series segments are input into a CNN-Attention network that combines channel and spatial attention. The network front end uses a one-dimensional convolutional layer (1D-CNN) to extract local phase difference features and high-frequency vibration features from multiple channels. The attention module dynamically assigns weights based on the importance of the current feature map, strengthening the frequency bands that reflect the real flow field information and suppressing high-order harmonic noise from the airborne motor. Finally, the Softmax layer classifies and outputs discrete wind speed levels (e.g., 1 m / s to 5 m / s) and two-dimensional spatial wind vector angles.

[0034] 2) Flight Attitude Regression: The roll, pitch, and yaw angles of the flapping wing constitute a vast and continuous three-dimensional state space. To reduce experimental calibration costs, the system employs the Progressive Sliced ​​Latin Hypercube Sampling (PSLHS) algorithm to generate a small number of evenly distributed slice sample points within a predefined attitude envelope. The pre-training (source domain) of the Multiple Input Multiple Output (MIMO) regression network is then performed on this sparse dataset. During actual flight, only a very small amount of calibration flight test data needs to be collected. The weights of the network's pre-feature extraction layers are frozen, and only the fully connected layers are fine-tuned through transfer learning. This allows for rapid generalization to the real flight distribution, outputting the aircraft's continuous three-dimensional attitude angles in real time with high precision. It should be understood that although this embodiment describes a specific network architecture, in other implementations, recurrent neural networks, Transformers, or other model architectures suitable for temporal signal processing can also be used, as long as they possess the capability to extract and compute multi-channel spatiotemporal features.

[0035] Furthermore, this embodiment also provides an aircraft, including a fuselage and a variable stiffness-sensing piezoelectric thin film bionic sensing system for flapping-wing aircraft as described above, wherein the variable stiffness-sensing piezoelectric thin film bionic sensor for flapping-wing aircraft is conformally attached to the surface of the fuselage. Figure 6As shown, in actual flight tests, thanks to the high-sensitivity mechanical response brought by the variable stiffness-sensing hair and out-of-plane stiffness reduction shear structure 5 in the fabrication of the variable stiffness-sensing hair-paper-cutting piezoelectric thin film bionic sensor in Example 1, and the efficient decoupling capability of the intelligent calculation module for multi-dimensional signals in this embodiment, the system exhibits excellent sensing accuracy in complex dynamic environments. Verification shows that the average absolute error of the aircraft's three-axis attitude angles remains stably within 1.5°, and the error distribution is bell-shaped and concentrated near zero, indicating that the system has good robustness and consistency. It should be noted that although this embodiment uses a flapping-wing aircraft as an example for verification, this variable stiffness-sensing hair-paper-cutting piezoelectric thin film bionic sensing system for flapping-wing aircraft is also applicable to other intelligent equipment requiring multi-dimensional tactile or flow field sensing, such as fixed-wing UAVs, rotary-wing aircraft, underwater robots, or flexible robotic arms. The scope of protection of this invention should not be limited to this specific application scenario.

[0036] Specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art can make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. Unless otherwise specified, the embodiments and features described in this application can be arbitrarily combined with each other.

Claims

1. A variable stiffness-sensing piezoelectric thin film bionic sensor for flapping-wing aircraft, characterized in that, include: A piezoelectric thin film layer, wherein the piezoelectric thin film layer is provided with an out-of-plane stiffness reduction shear structure; The variable stiffness sensitive hair has a stiffness difference along the axial direction, and the bottom end of the variable stiffness sensitive hair is connected to the central region of the piezoelectric thin film layer. Multiple independent sensing channels are centrally symmetrically distributed on the surface of the piezoelectric thin film layer, and each independent sensing channel is electrically isolated from the others. Under external excitation, the variable stiffness sensing hair deflects outward and causes the piezoelectric thin film layer to deform. The multiple independent sensing channels synchronously output electrical signals with spatiotemporal distribution differences based on the piezoelectric effect.

2. The flapping-wing aircraft variable stiffness-sensing hair-cutting piezoelectric thin film bionic sensor according to claim 1, characterized in that, The variable stiffness sensitive hair includes a first hair segment and a second hair segment distributed along the axial direction. The elastic modulus of the first segment is greater than that of the second segment; The bottom end of the second hair segment is connected to the central region of the piezoelectric thin film layer to serve as a flexible hinge structure.

3. The flapping-wing aircraft variable stiffness-sensing hair-cutting piezoelectric thin film bionic sensor according to claim 2, characterized in that, The first segment is made of a high-modulus solid material, and the second segment is made of a low-modulus soft material; The high-modulus solid material includes a UV-curable resin, and the low-modulus soft material includes soft silicone.

4. The flapping-wing aircraft variable stiffness-sensing hair-cutting piezoelectric thin film bionic sensor according to claim 1, characterized in that, The out-of-plane stiffness reduction shear mark structure includes multiple concentric ring-shaped shear marks; An uncut area is retained between adjacent annular cuts, and the uncut area forms a flexible hinge, so that the piezoelectric thin film layer forms a structure in which rigid units and flexible hinges are arranged alternately.

5. The flapping-wing aircraft variable stiffness sensing hair-cutting piezoelectric thin film bionic sensor according to claim 1, characterized in that, The number of the multiple independent sensing channels is at least three; The multiple independent sensing channels are arranged in a cross or orthogonal or centrally symmetrical manner, dividing the surface of the piezoelectric thin film layer into multiple electrically isolated fan-shaped regions.

6. The flapping-wing aircraft variable stiffness-sensing hair-cutting piezoelectric thin film bionic sensor according to claim 5, characterized in that, An electrode layer is provided on the upper surface of the piezoelectric thin film layer. The electrode layer is divided into multiple independent electrodes that correspond one-to-one with the multiple sector regions. Each of the independent electrodes is connected to a wire extending radially outward. A common grounding electrode layer is provided on the lower surface of the piezoelectric thin film layer.

7. The flapping-wing aircraft variable stiffness sensing hair-cutting piezoelectric thin film bionic sensor according to claim 1, characterized in that, The piezoelectric thin film layer is a polyvinylidene fluoride film; The biomimetic sensor also includes a substrate, on which the piezoelectric thin film layer is encapsulated. The substrate is a flexible polyimide substrate.

8. A biomimetic sensing system for a flapping-wing aircraft, characterized in that: include: The flapping-wing aircraft variable stiffness sensing hair-cutting piezoelectric thin film bionic sensor as described in claim 1; The signal acquisition module is electrically connected to the multiple independent sensing channels and is used to synchronously acquire the electrical signals and perform preprocessing to generate multi-channel timing signals. The intelligent solution module has a built-in deep learning network model, which is used to receive the multi-channel time series signal, and perform feature extraction and solution on the multi-channel time series signal through the deep learning network model, and output flow field parameters and / or body attitude angle.

9. The flapping-wing aircraft variable stiffness sensing hair-cutting piezoelectric thin film bionic sensing system according to claim 8, characterized in that, The deep learning network model includes a flow field classification network and / or a multi-output pose regression network; The flow field classification network is configured to use an attention mechanism to assign weights to the feature maps of the multi-channel time-series signals in order to focus on specific frequency bands and output flow field classification results. The multi-output pose regression network includes fully connected layers and is configured to output continuous three-dimensional pose angles based on extracted features.

10. An aircraft, characterized in that, Includes the airframe and the flapping-wing aircraft variable stiffness-sensing piezoelectric thin film bionic sensing system as described in claim 8; The bionic sensor is conformally attached to the surface of the body.

Images