Piezoelectric ultrasonic tactile sensor

Abstract

The invention discloses a piezoelectric ultrasonic tactile sensor, and relates to the technical field of sensors. The piezoelectric ultrasonic tactile sensor includes: a double-layer flexible printed circuit board; the piezoelectric micro-electro-mechanical transducer array is integrated in the double-layer flexible printed circuit board and is used for transmitting ultrasonic waves to the surface of a measured object and receiving reflection echo signals; the signal processing unit is integrated in the double-layer flexible printed circuit board and is used for sequentially carrying out amplification, filtering and analog-to-digital conversion processing on the reflection echo signal to generate an echo digital signal; the echo digital signal is used for identifying the surface characteristics of the measured object; the sensor provided by the invention can realize high-precision non-contact tactile perception in a complex environment, has flexible adaptation and anti-interference characteristics, and remarkably improves the stability and measurement reliability in a dynamic scene.

Description

Technical Field

[0001] The present invention relates to the technical field of sensors, and in particular to a piezoelectric ultrasonic tactile sensor. Background Art

[0002] As a core capability for robot-environment interaction, tactile sensing technology has evolved from single-purpose mechanical sensing to multi-dimensional information fusion. Tactile sensors are key components for robot perception of the environment. Their development began with simple contact force detection and has gradually expanded to multi-dimensional sensing based on resistance, capacitance, piezoelectricity, and optics.

[0003] Existing technologies, capacitive sensors rely on changes in the spacing between electrodes to measure force, but are difficult to adapt to irregular surfaces and limited in miniaturization. While visual-tactile sensors can sense shear force and texture, their bulk hinders widespread deployment. While magnetic sensors offer the advantage of non-contact measurement, their sensitivity and resolution remain bottlenecks in complex deformation scenarios. Furthermore, existing sensors exhibit poor stability in harsh environments such as high humidity and underwater, severely limiting their in-depth application in fields like industrial automation and medical procedures. Summary of the Invention

[0004] In order to solve the problems existing in the prior art, the present invention provides a piezoelectric ultrasonic tactile sensor that can achieve high-precision non-contact tactile perception in complex environments, has both flexible adaptation and anti-interference characteristics, and significantly improves stability and measurement reliability in dynamic scenarios.

[0005] To achieve the above objectives, the present invention provides a piezoelectric ultrasonic tactile sensor, comprising:

[0006] Double-layer flexible printed circuit board;

[0007] A piezoelectric micro-electromechanical transducer array is integrated into the double-layer flexible printed circuit board and is used to transmit ultrasonic waves to the surface of the object being measured and receive reflected echo signals;

[0008] A signal processing unit is integrated into the double-layer flexible printed circuit board and is used to amplify, filter and perform analog-to-digital conversion on the reflected echo signal in sequence to generate an echo digital signal; the echo digital signal is used to identify the surface characteristics of the object being measured.

[0009] Optionally, the double-layer flexible printed circuit board includes:

[0010] The upper circuit board is a polyimide layer with sound wave transmitting / receiving windows;

[0011] The lower circuit board is a polyimide layer, and an annealed copper patterned layer is provided on the surface facing the upper circuit board.

[0012] Optionally, the upper circuit board is a polyimide layer with a thickness of 200 microns, and the lower circuit board is a polyimide layer with a thickness of 100 microns; the thickness of the annealed copper patterned layer is 12 microns.

[0013] Optionally, the signal processing unit includes:

[0014] an amplifier circuit, configured to amplify the reflected echo signal to generate an amplified signal;

[0015] a filtering circuit, configured to filter the amplified signal to generate a filtered signal;

[0016] The analog-to-digital conversion module is used to perform analog-to-digital conversion on the filtered signal to generate an echo digital signal.

[0017] Optionally, the signal processing unit further includes:

[0018] a level raising circuit, provided between the filtering circuit and the analog-to-digital conversion module, for performing DC bias adjustment on the filtered signal to raise the signal level to an effective input voltage range of the analog-to-digital conversion module;

[0019] In terms of performing analog-to-digital conversion on the filtered signal, the analog-to-digital conversion module is specifically used to perform analog-to-digital conversion on the filtered signal after completing DC bias adjustment.

[0020] Optionally, the echo digital signal is transmitted to an identification unit that is communicatively connected to the piezoelectric ultrasonic tactile sensor; the identification unit inputs the echo digital signal into a pre-trained recognition model to obtain an identification result of the surface characteristics of the object being measured.

[0021] Optionally, the piezoelectric ultrasonic tactile sensor further includes:

[0022] The recognition unit is used to input the echo digital signal into a pre-trained recognition model to obtain a recognition result of the surface characteristics of the object being measured.

[0023] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0024] The piezoelectric ultrasonic tactile sensor provided by the present invention uses a double-layer flexible printed circuit board (FPCB) to encapsulate a piezoelectric micro-electromechanical transducer (PMUT) array and a signal processing unit. The flexible printed circuit board has a low modulus characteristic. Compared with the existing rigid sensor housing, it can effectively eliminate redundant space, significantly compressing the overall volume of the sensor, thereby miniaturizing the sensor so that it can be widely laid out and can be adapted to robot fingers or complex curved surface installation scenarios, solving the installation limitations of traditional sensors due to their rigid structure. Furthermore, the tactile sensor provided by the present invention uses ultrasonic waves for non-contact detection, breaking through the limitations of traditional tactile sensors that rely on contact mechanical measurements. Even in complex deformation scenarios, it still has high sensitivity and reliability, and is particularly suitable for dynamic interaction or fragile object detection scenarios.

[0025] In addition, due to the low modulus characteristics of the flexible packaging structure and the ability of the piezoelectric element to resist environmental interference, the sensor can maintain stable performance in harsh environments such as high humidity and underwater, overcoming the measurement error problem caused by environmental sensitivity of traditional sensors and expanding its application potential in industrial automation, medical micro-manipulation, underwater robots and other fields. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The above and other objects, features and advantages of the present invention will become more apparent through a more detailed description of exemplary embodiments of the present invention with reference to the accompanying drawings, wherein like reference numerals generally represent like components throughout the exemplary embodiments of the present invention.

[0027] Figure 1 This is a schematic diagram of the hardware structure of a piezoelectric ultrasonic tactile sensor according to an embodiment of the present invention;

[0028] Figure 2 Schematic diagrams of sensor fabrication according to an embodiment of the present invention; Figures (I), (II), (III), (IV), (V), (VI), (VII), and (VIII) are schematic diagrams of silicon-on-insulator wafers, sputtering deposition of molybdenum / scandium aluminum nitride / molybdenum multilayer films, etching molybdenum electrodes, etching through holes, depositing gold wires, back cavity etching, solder bumping, and bonding to flexible circuit boards, respectively;

[0029] Figure 3 A schematic diagram of signal transmission processing according to an embodiment of the present invention;

[0030] Figure 4 Schematic diagrams of echo detection results shown in an embodiment of the present invention; wherein, Figures (4.1), (4.2), (4.3) and (4.4) are schematic diagrams of echo detection results of ceramic tiles, wooden boards, rubber and polypropylene respectively. DETAILED DESCRIPTION

[0031] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0032] See Figure 1 and Figure 3 , Figure 1 This is a schematic diagram of the hardware structure of the piezoelectric ultrasonic tactile sensor. Figure 3 The figure is a schematic diagram of signal transmission processing. The piezoelectric ultrasonic tactile sensor includes:

[0033] Double-layer flexible printed circuit board 1;

[0034] The piezoelectric micro-electromechanical transducer array 2 is integrated inside the double-layer flexible printed circuit board 1 and is used to transmit ultrasonic waves to the surface of the object under test 01 and receive reflected echo signals;

[0035] The signal processing unit 3 is integrated inside the double-layer flexible printed circuit board 1 and is used to amplify, filter and perform analog-to-digital conversion on the reflected echo signal in sequence to generate an echo digital signal; the echo digital signal is used to identify the surface characteristics of the object 01 under test.

[0036] Exemplarily, the surface characteristics include one or more of material, hardness, texture, and roughness.

[0037] Among them, the double-layer flexible printed circuit board includes:

[0038] The upper circuit board 11 is a polyimide layer with sound wave transmitting / receiving windows;

[0039] The lower circuit board 12 is a polyimide layer, and an annealed copper patterned layer 121 is provided on the surface facing the upper circuit board.

[0040] The piezoelectric ultrasonic tactile sensor provided by the present invention utilizes a double-layer flexible printed circuit board (FPCB) 1 as its core packaging structure. Its upper circuit board 11 is made of flexible polyimide material, with acoustic wave transmitting and receiving windows on its surface, enabling efficient transmission of ultrasonic waves and accurate reception of reflected echoes. A lower circuit board 12, through a patterned conductive layer, forms a coordinated circuit layout with the upper circuit board, providing stable electrical connections for the internal functional units. The signal processing unit 3 is located within the lower circuit board 12, specifically beneath the annealed copper patterned layer 121. A piezoelectric microelectromechanical transducer (PMUT) array 2 is integrated within the FPCB 1 through a precision lamination process. It utilizes the inverse piezoelectric effect of the piezoelectric material to transmit ultrasonic waves and the direct piezoelectric effect to receive echo signals reflected from the surface of the measured object (01). The PMUT array 2's miniaturized design, combined with the low modulus of the FPCB 1, gives the sensor excellent mechanical adaptability and compactness, enabling it to conform to robot fingers and meet the requirements of installation on complex curved surfaces, avoiding the mechanical adaptation limitations imposed by traditional rigid packaging structures.

[0041] In applications, the thickness and material parameters of the dual-layer flexible printed circuit board 1 can be optimized to achieve a balanced performance. For example, the upper circuit board 11 can utilize a 200-micron-thick polyimide layer, whose mechanical strength and flexibility meet the requirements for efficient transmission of the acoustic window while ensuring stability under dynamic deformation. The lower circuit board 12 can utilize a 100-micron-thick polyimide layer, whose thinner thickness reduces the overall package height of the sensor. A 12-micron-thick annealed copper patterned layer 121 on the surface of the lower circuit board 12 provides a high-precision conductive path.

[0042] The annealed copper patterned layer 121 undergoes a special annealing process, resulting in both low resistivity and high ductility, ensuring reliable electrical connections even when the flexible printed circuit board is deformed. The upper and lower layers of polyimide material are formed into a compact laminate structure through a heat-pressing process, providing mechanical protection for the internal piezoelectric MEMS transducer array 2. The differentiated thickness design also balances package rigidity and flexibility, enabling the sensor to be mounted on curved surfaces while avoiding device damage caused by stress concentration.

[0043] See also Figure 2 , Figure 2 Create a schematic diagram for sensor processing; Figure 2 In the figure, (Ⅰ), (Ⅱ), (Ⅲ), (Ⅳ), (Ⅴ), (Ⅵ), (Ⅶ) and (Ⅷ) are schematic diagrams of silicon wafer on insulator, sputtering deposition of molybdenum / scandium aluminum nitride / molybdenum multilayer film, etching of molybdenum electrode, etching of through hole, deposition of gold wire, back cavity etching, solder ball planting and flexible circuit board bonding.

[0044] The following is an exemplary description of the sensor manufacturing method, wherein the manufacturing and packaging of the piezoelectric micro-electromechanical transducer array 2 includes the following steps:

[0045] Cleaning of silicon-on-insulator (SOI) wafers; SOI wafers include a silicon device layer, a buried silicon dioxide layer, and a silicon substrate;

[0046] A scandium aluminum nitride seed layer is formed on the surface of the SOI wafer using an atomic layer deposition process;

[0047] forming a molybdenum layer, a scandium aluminum nitride layer, and a molybdenum layer in sequence on the scandium aluminum nitride seed layer by physical vapor deposition to obtain a first structure;

[0048] Using an inductively coupled plasma etching process, a molybdenum electrode pattern is formed in the top molybdenum layer of the first structure, and an electrode lead-out hole is etched in the bottom molybdenum layer to obtain a second structure;

[0049] depositing a metal layer in the electrode lead-out hole region of the second structure by a magnetron sputtering process, and forming metal leads and pads by a patterning process to obtain a third structure;

[0050] Deep reactive ion etching is performed on the back side of the silicon substrate of the SOI wafer of the third structure to release the cavity and form a thin film vibration structure to obtain a piezoelectric micro-electromechanical transducer;

[0051] A plurality of piezoelectric micro-electromechanical transducers are fixed inside the double-layer flexible printed circuit board 1 through ball planting and flip-chip bonding processes to form a piezoelectric micro-electromechanical transducer array 2, thereby completing the packaging.

[0052] In the application, a silicon-on-insulator (SOI) wafer (6N600-1-5N, Okmetic) is first cleaned. The SOI wafer can be composed of a 5-micron-thick silicon device layer, a 1-micron-thick buried silicon dioxide layer, and a 400-micron-thick silicon substrate. An atomic layer deposition (ALD) process (NLD-4000 ALD equipment, Nano-master) is used to grow a 100-nanometer ScAlN (scandium aluminum nitride) seed layer on the wafer surface to reduce the surface roughness of subsequent thin film deposition and improve the structure and morphology of molybdenum (Mo) and c-axis oriented ScAlN. A physical vapor deposition (PVD) process (Sigma PVD equipment, SPTS) is used to sequentially deposit a 0.3-micron-thick Mo layer, a 1-micron-thick ScAlN layer, and a 0.3-micron-thick Mo layer on the ScAlN seed layer. The top Mo electrode pattern was formed using an inductively coupled plasma (ICP) process (GSEC200 ICP equipment, NMC), followed by etching of ScAlN to form the bottom electrode lead-out holes. A 200-nanometer-thick gold (Au) layer was deposited using magnetron sputtering (MS150X-L magnetron sputtering equipment, FHR), and metal leads and pads were formed by lift-off. Deep reactive ion etching (DRIE) was performed on the backside of the SOI wafer (Omega LPX Rapier DRIE equipment, SPTS) to release the cavity and form the thin-film vibration structure. Finally, the PMUT array 2 was mounted on the inside of a double-layer flexible printed circuit board 1 through ball placement and flip-chip bonding, achieving electrical connection and mechanical fixation.

[0053] See Figure 3 , Figure 3 Schematic diagram of signal transmission and processing; in application, an excitation signal can be sent to the signal processing unit 3 through the field programmable gate array FPGA connected to the sensor or the FPGA built into the sensor; the signal processing unit 3 converts the excitation signal, a digital signal, into an analog signal (D / A processing) and transmits it to the PMUT array 2, thereby turning on the transmission mode. The PMUT array 2 transmits an ultrasonic wave to the object under test 01 according to the received analog signal and receives the corresponding reflected echo signal; the PMUT array 2 transmits the reflected echo signal to the signal processing unit 3, which amplifies, filters, samples, and processes it, and finally converts the analog signal into a digital signal (A / D processing) to obtain an echo digital signal; the signal processing unit 3 transmits the echo digital signal to the FPGA, which performs algorithm processing to identify the surface characteristics of the object under test 01.

[0054] In one embodiment, the signal processing unit 3 includes:

[0055] an amplifier circuit, used for amplifying the reflected echo signal to generate an amplified signal;

[0056] A filter circuit, used for filtering the amplified signal to generate a filtered signal;

[0057] The analog-to-digital conversion module is used to perform analog-to-digital conversion on the filtered signal to generate an echo digital signal.

[0058] In the application, the signal processing unit 3 is integrated inside the double-layer flexible printed circuit board 1, and a modular design is adopted to realize the full-process processing of the echo signal: the reflected echo signal passes through the amplification circuit in turn to improve the signal-to-noise ratio, eliminates environmental interference through the filtering circuit, and is finally converted into a digital signal by the analog-to-digital conversion module.

[0059] In one embodiment, the signal processing unit 3 further includes:

[0060] The level raising circuit is provided between the filtering circuit and the analog-to-digital conversion module, and is used to adjust the DC bias of the filtered signal to raise the signal level to the effective input voltage range of the analog-to-digital conversion module;

[0061] In terms of performing analog-to-digital conversion on the filtered signal, the analog-to-digital conversion module is specifically used to perform analog-to-digital conversion on the filtered signal after completing DC bias adjustment.

[0062] In applications, the level-lifting circuit can be set between the filtering circuit and the analog-to-digital conversion module. The level-lifting circuit dynamically adjusts the DC bias of the filtered signal to raise the overall signal level to the effective input voltage window of the analog-to-digital conversion module, thereby avoiding quantization errors caused by signal amplitude being too low or out of range. After the level-adjusted filtered signal enters the analog-to-digital conversion module, its voltage range completely matches the linear operating range of the converter, significantly improving the accuracy and linearity of signal sampling. This design effectively solves the problem of dynamic range adaptation of weak echo signals in complex environments. Combined with the low-noise amplification and filtering modules integrated on the flexible circuit board, the sensor can still stably output high-quality echo digital signals under complex working conditions (such as underwater).

[0063] In one embodiment, the echo digital signal is transmitted to a recognition unit that is in communication with the piezoelectric ultrasonic tactile sensor; the recognition unit inputs the echo digital signal into a pre-trained recognition model to obtain a recognition result of the surface characteristics of the object 01 being measured.

[0064] In applications, the echo digital signal can be sent to an external recognition unit via wireless transmission. Based on a pre-trained machine learning model, the recognition unit performs a multi-dimensional analysis of the timing characteristics, amplitude changes, and attenuation characteristics of the echo signal, and accurately determines the surface material, hardness, texture, and / or roughness of the object being measured 01 through feature extraction and pattern matching. The independent design of the external recognition unit allows the sensor body to maintain its lightweight and low power consumption characteristics, while supporting centralized processing and collaborative analysis of multi-sensor data, suitable for application scenarios requiring high computing resources or complex decision-making. For example, in a robot's multi-finger collaborative grasping task, the echo data of multiple sensors can be aggregated into the same recognition unit for global optimization, improving the robustness of object attribute recognition and system response efficiency.

[0065] Exemplarily, the training method of the recognition model may include the following steps:

[0066] Acquire the echo digital signal collected by the piezoelectric ultrasonic tactile sensor;

[0067] extracting Mel-frequency cepstral coefficients from the echo digital signal as surface tactile features;

[0068] Based on a transfer learning method of a convolutional neural network, feature learning is performed on the surface tactile features to construct multiple classification models;

[0069] Performing cross-domain conversion on the surface tactile features through a recurrent generative adversarial network to generate conversion features that are adapted to different data distributions;

[0070] Taking the surface tactile features and the conversion features as input data, and inputting the input data into various classification models for collaborative prediction;

[0071] If the prediction results of at least two classification models are consistent, the reliability of each classification model is verified based on the consistent prediction results, and the verified classification models are combined into the final recognition model; otherwise, the echo digital signal is reacquired and the above steps are repeated.

[0072] In applications, before extracting surface tactile features, the acquired echo digital signal can also be preprocessed, such as data segmentation, fast Fourier transform, normalization, removal of DC interference, and principal component analysis dimensionality reduction operations.

[0073] Before collaborative prediction, the data distribution difference can be eliminated by minimizing the loss between the surface tactile features and the conversion features. Afterwards, the surface tactile features and the conversion features after eliminating the data distribution difference are used as input data.

[0074] Specifically, the application preprocesses the acquired digital echo signals and extracts Mel-Frequency Cepstral Coefficients (MFCCs) as surface tactile features. Using transfer learning methods with convolutional neural networks (CNNs), ultrasound imaging data, including amplitude, time of flight, and surface texture, is analyzed. A pretrained CNN model learns classification criteria from labeled data to construct multiple classification models.

[0075] To address the mismatch problem between tactile data of different materials, a cyclic generative adversarial network (CycleGAN) is introduced. Through domain adaptation technology (such as minimizing the loss between the original feature matrix and the restored feature matrix), cross-domain feature conversion is achieved, and conversion features that adapt to different data distributions are generated, thereby reducing dependence on data consistency.

[0076] During the model verification phase, surface tactile features and conversion features are used as input data and fed into multiple classification models, such as Gaussian kernel support vector machines (SVMs), linear kernel SVMs, and multilayer perceptrons (MLPs), for collaborative prediction. If the prediction results of at least two classification models are consistent, the reliability of each classification model is verified based on the consistent prediction results, and the verified classification models are combined into the final recognition model; otherwise, the echo digital signal must be reacquired and the above steps repeated. This method improves classification reliability through multi-model collaborative verification and uses feature conversion technology to reduce the need for repeated data collection, adapting to dynamic perception scenarios in complex environments.

[0077] In one embodiment, the piezoelectric ultrasonic tactile sensor further comprises:

[0078] The recognition unit is used to input the echo digital signal into a pre-trained recognition model to obtain a recognition result of the surface characteristics of the object under test 01.

[0079] In applications, the recognition unit can be directly integrated into the piezoelectric ultrasonic tactile sensor, such as within a built-in FPGA. The recognition unit is located within a double-layer flexible printed circuit board. The recognition unit uses an embedded processor to run a lightweight machine learning model, processing the echo digital signal in real time and outputting surface feature recognition results. This design significantly reduces signal transmission and processing latency, meeting the needs of demanding real-time scenarios, such as the instant tactile feedback of a robotic arm during minimally invasive medical surgery or rapid sorting operations on industrial assembly lines. The built-in recognition unit allows for identification and detection without relying on external equipment, significantly improving the sensor's autonomy and deployment flexibility.

[0080] See also Figure 4 , Figure 4Figures (4.1), (4.2), (4.3) and (4.4) are schematic diagrams of the echo detection results for tiles, wood boards, rubber and polypropylene, respectively.

[0081] The test object in Figure (4.1) is ceramic tile. The test process is as follows: Figure 4 .1(a), Figure 4 The red curve in Figure 1(b) is the echo signal after shielding the blind area, while the blue curve is the excitation signal, which serves as a time comparison and calibration. As can be seen, the tile echo's flight time is approximately 76.37 μs, the maximum echo amplitude is approximately 2.27 V, and the echo decay time from reception to disappearance is approximately 55.88 μs, indicating high signal amplitude and long duration.

[0082] The testing process of the wood board is as follows Figure 4 .2(a), Figure 4 .2(b) shows the corresponding echo signal. It can be seen that the echo flight time of the wooden board is about 75.54 μs, the maximum amplitude of the echo is about 1.73 V, and the decay time of the echo from reception to disappearance is about 39.6 μs.

[0083] The testing process of rubber floor mats is as follows Figure 4 .3(a), Figure 4 Figure 3(b) shows the corresponding echo signal. It can be seen that the echo flight time of the rubber mat is approximately 72.73 μs, the maximum amplitude of the echo is approximately 0.71 V, and the decay time from reception to disappearance of the echo is approximately 21.3 μs.

[0084] The testing process of polypropylene floor mats is as follows Figure 4 .4(a), Figure 4 Figure 4(b) shows the corresponding echo signal. It can be seen that the echo flight time of the polypropylene mat is approximately 70.59 μs, the maximum echo amplitude is approximately 0.47 V, and the echo decay time from reception to disappearance is approximately 16.8 μs.

[0085] After multiple experiments, we were able to determine the flight time, maximum echo amplitude, and signal decay time of the echo signals for different objects under test. Specific values ​​are listed in Table 1. As can be seen, the echo signal strength and decay time for each material are unique, and their ranges do not overlap. Therefore, they can be used as a basis for material identification with high accuracy.

[0086] Table 1

[0087]

[0088] While various embodiments of the present invention have been described above, the above descriptions are intended to be illustrative, non-exhaustive, and not limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is selected to best explain the principles of the embodiments, their practical applications, or improvements to existing technologies, or to enable others skilled in the art to understand the embodiments disclosed herein.

Claims

1. A piezoelectric ultrasonic tactile sensor, characterized in that: include: Double-layer flexible printed circuit board; A piezoelectric micro-electromechanical transducer array is integrated into the double-layer flexible printed circuit board and is used to transmit ultrasonic waves to the surface of the object being measured and receive reflected echo signals; A signal processing unit is integrated into the double-layer flexible printed circuit board and is used to amplify, filter and perform analog-to-digital conversion on the reflected echo signal in sequence to generate an echo digital signal; the echo digital signal is used to identify the surface characteristics of the object being measured.

2. The piezoelectric ultrasonic tactile sensor according to claim 1, characterized in that: The double-layer flexible printed circuit board comprises: The upper circuit board is a polyimide layer with sound wave transmitting / receiving windows; The lower circuit board is a polyimide layer, and an annealed copper patterned layer is provided on the surface facing the upper circuit board.

3. The piezoelectric ultrasonic tactile sensor according to claim 2, characterized in that: The upper circuit board is a polyimide layer with a thickness of 200 microns, and the lower circuit board is a polyimide layer with a thickness of 100 microns; the thickness of the annealed copper patterned layer is 12 microns.

4. The piezoelectric ultrasonic tactile sensor according to claim 1, wherein: The signal processing unit includes: an amplifier circuit, configured to amplify the reflected echo signal to generate an amplified signal; a filtering circuit, configured to filter the amplified signal to generate a filtered signal; The analog-to-digital conversion module is used to perform analog-to-digital conversion on the filtered signal to generate an echo digital signal.

5. The piezoelectric ultrasonic tactile sensor according to claim 4, characterized in that: The signal processing unit further includes: a level raising circuit, provided between the filtering circuit and the analog-to-digital conversion module, for performing DC bias adjustment on the filtered signal to raise the signal level to an effective input voltage range of the analog-to-digital conversion module; In terms of performing analog-to-digital conversion on the filtered signal, the analog-to-digital conversion module is specifically used to perform analog-to-digital conversion on the filtered signal after completing DC bias adjustment.

6. The piezoelectric ultrasonic tactile sensor according to claim 1, characterized in that: The echo digital signal is transmitted to an identification unit that is in communication with the piezoelectric ultrasonic tactile sensor; the identification unit inputs the echo digital signal into a pre-trained identification model to obtain an identification result of the surface characteristics of the object being measured.

7. The piezoelectric ultrasonic tactile sensor according to claim 1, characterized in that: Also includes: The recognition unit is used to input the echo digital signal into a pre-trained recognition model to obtain a recognition result of the surface characteristics of the object being measured.

Images