Humanoid robot finger magnetic touch sensor system with bimodal sensing function

Abstract

The invention discloses a humanoid robot finger magnetic tactile sensor system with a bimodal sensing function, and relates to the technical field of tactile sensor systems. A magnetic touch sensing module is included, and bimodal force sensing is achieved through a microstructure material; the signal conditioning module optimizes the signal by means of an adaptive neural network; the data acquisition and processing module, a multi-core heterogeneous processor and a deep learning algorithm are used for extracting features and calculating force information; the communication and interface module is used for transmitting data by adopting chaotic encryption and multiple protocols; the mechanical structure module uses a bionic material and a variable rigidity joint; and functional modules of self-calibration, temperature compensation and the like are also arranged. Positive pressure and tangential force are accurately sensed, signal processing is efficient, environmental interference resistance is high, and data transmission is safe and reliable; the mechanical structure flexibly adapts to various tasks, reliability and endurance are improved through fault diagnosis and energy management, and the robot is assisted to complete complex operation.

Description

Technical Field

[0001] The present invention relates to the technical field of tactile sensor systems, and particularly to a humanoid robot finger magneto-tactile sensor system with dual-modal perception function. Background Art

[0002] At present, with the rapid development of robotics, humanoid robots have shown great application potential in industrial production, service industries, medical care and other fields. Among them, the fingers of the robot, as the key parts directly interacting with the external environment, the strength of their perception ability directly affects the quality and efficiency of the tasks completed by the robot.

[0003] Most traditional robot finger sensors can only achieve single-modal perception, such as only being able to sense pressure or only being able to sense tactile texture. However, in actual application scenarios, the physical properties of objects are complex and diverse, and it is necessary to simultaneously obtain information in multiple aspects such as normal force and tangential force in order to more accurately achieve grasping, operation and recognition of objects. For example, on an industrial assembly line, a robot needs to accurately grasp parts of different shapes and materials, and a single-modal sensor cannot provide enough information to judge the grasping force and direction, which easily leads to the parts slipping or being damaged.

[0004] Although existing magneto-tactile sensors can sense force changes to a certain extent, they have deficiencies in terms of sensitivity, accuracy and the ability to sense complex forces. The magnetic materials and magnetosensitive element structures used are relatively simple, the response to small force changes is not obvious, and it is difficult to distinguish the combined action of normal force and tangential force. In addition, the signal processing and analysis methods of traditional sensors are relatively single, and the effective information in the sensor data cannot be fully exploited, resulting in inaccurate judgment of the physical properties of objects.

[0005] At the same time, the diversity of the robot working environment also poses higher requirements for the performance of the sensor. Environmental factors such as temperature and humidity will affect the performance of the sensor. Existing sensors lack an effective environmental compensation mechanism, resulting in large errors in the measurement results of the sensor under different environmental conditions. Moreover, the communication stability and data security between the sensor and the robot control system also need to be improved to ensure that the robot can obtain sensor data in real time and accurately and make correct decisions. Summary of the Invention

[0006] The humanoid robot finger magneto-tactile sensor system with dual-modal perception function proposed by the present invention is to solve the problems mentioned in the above prior art.

[0007] To achieve the above object, the present invention adopts the following technical solution: A humanoid robot finger magneto-tactile sensor system with dual-modal perception function, comprising:

[0008] Magnetic tactile sensing module: A microstructured magnetic elastomer material is used as the sensing medium. When an external force acts, the magnetic elastomer undergoes macroscopic deformation, and the microscopic orientation and spacing of the internal magnetosensitive fibers change accordingly, resulting in a change in the magnetic field distribution; for the normal force F n , the change in the electrical signal and F n satisfy the relationship where k 1n is the non-linear normal force sensitivity coefficient, α n is the normal force non-linearity exponent, b 1n is the normal force logarithmic adjustment coefficient; for the tangential force F t , the change in the electrical signal and F t satisfy k 2t is the non-linear tangential force sensitivity coefficient, α t is the tangential force non-linearity exponent, b 2t is the tangential force sine adjustment coefficient, ω t is the tangential force angular frequency; simultaneously sense the normal force and the tangential force to achieve dual-modal sensing and sense the dynamic change process of the force;

[0009] Signal conditioning module: Amplify, filter, and linearize the weak electrical signal output by the magnetic tactile sensing module. Use a gain amplifier based on an adaptive neural network for signal amplification. The neural network automatically adjusts the gain G according to the characteristics of the input signal. The gain adjustment formula is where G0 is the initial gain, V ref is the reference voltage, V in is the input signal voltage, V target is the target signal voltage, and β is the gain adjustment coefficient; filter out the frequency noise through an adaptive notch filter, and use a polynomial fitting method with fuzzy logic control to linearize the signal. Dynamically adjust the order and coefficients of the polynomial according to the dynamic range and noise level of the input signal;

[0010] Data acquisition and processing module: Use a data acquisition card to collect the conditioned signal. The sampling frequency f s satisfies f s = max(2f max , f dynamic ), where f max is the highest frequency component of the signal, and f dynamic is the dynamic sampling frequency adjusted in real time according to the dynamic change rate of the force; the collected data is transmitted to the microprocessor, and a feature extraction algorithm based on deep learning is used for data processing. Combine the convolutional neural network CNN and the recurrent neural network RNN to extract the features of the force signal from different time scales and spatial scales;

[0011] Communication and Interface Module: It supports SPI, I2C, USB, and CAN communication protocols to achieve data transmission with the control system of the humanoid robot; it uses an adaptive data encryption mechanism based on a chaotic encryption algorithm to encrypt the transmitted data. The encryption key K is generated by a chaotic system, and the iterative formula is x n+1 = μ·x n ·(1 - x n ), where μ is the chaotic parameter, and x n is the nth iteration value of the chaotic system; the encryption formula is C = E(M, K)·exp(-γ·t), where M is the original data, C is the encrypted data, E is the encryption function, γ is the time decay coefficient, and t is the transmission time;

[0012] Mechanical Structure Module: Design the mechanical structure of the humanoid robot's finger, integrate the magnetic tactile perception module into the fingertip part of the finger, wrap the sensor with a material with shape memory characteristics, and the finger joints adopt a bionic variable stiffness structure to adjust the stiffness of the joints in real time according to the requirements of the grasping task; the rotation angle θ of the joint and the applied force F satisfy the relationship where a i is the polynomial coefficient, is the angle fluctuation amplitude, ω is the fluctuation angular frequency, and t is the time.

[0013] Furthermore, it also includes:

[0014] Self-Calibration Module: Perform self-calibration operations when the sensor system starts or regularly. Apply known normal and tangential forces with dynamic characteristics through the built-in standard force source, collect the output signals of the sensors, compare them with the theoretical values, and calculate the error values; according to the error values, correct the sensitivity coefficients k 1n , k 2t , and the correction formula is where Δk 1n , Δk 2t are the error values of the sensitivity coefficients, and β n , β t are the correction exponents of the normal and tangential forces; at the same time, automatically calibrate the gain and filter parameters of the signal conditioning module.

[0015] Temperature Compensation Module: Considering the comprehensive influence of temperature on the performance of magnetic sensitive elements, magneto-elastic materials, and signal conditioning circuits, integrate a temperature sensor inside the sensor to monitor the temperature change T i of different parts in real time; according to the multivariable coupling relationship between temperature and sensitivity coefficients, gain, and filter parameters, establish a temperature compensation model based on support vector regression (SVR) to compensate the electrical signals output by the magnetic tactile perception module and the parameters of the signal conditioning module; the compensated electrical signal V compensated is calculated by the formula Realize eliminating the influence of temperature on the measurement result, where V original is the original electrical signal, w i is the temperature weight coefficient, T 0i is the reference temperature, ∈ is the temperature fluctuation adjustment coefficient, ω T is the temperature angular frequency.

[0016] Furthermore, in the magnetic tactile perception module, the magnetosensitive element adopts a composite magnetosensitive sensor combining the anisotropic magnetoresistance (AMR) effect and the tunneling magnetoresistance (TMR) effect. By optimizing the microstructure of the magnetoelastic material, nano-scale magnetic particles and quantum dots are introduced to form a quantum-classical hybrid magnetic sensitive system; at the same time, the magnetosensitive elements are arranged in a three-dimensional staggered array layout to sense the distribution of force. By calculating the difference and ratio of the electrical signals of the magnetosensitive elements in different layers, the acting position, direction and force gradient change of the force are judged in combination with the spatial geometry algorithm.

[0017] Furthermore, in the signal conditioning module, a composite filtering algorithm based on wavelet transform and Kalman filtering is used for filtering; at the same time, an adaptive linearization algorithm optimized by genetic algorithm is adopted to adjust the parameters of the linearization curve in real time according to the statistical characteristics and dynamic changes of the input signal, so as to accurately realize the linear relationship between the output signal and the actual force value.

[0018] Furthermore, in the data acquisition and processing module, a heterogeneous computing architecture based on system on chip (SoC) is adopted, and the CPU, GPU and FPGA are integrated on the same chip to realize parallel computing data processing; at the same time, the collected data is analyzed and decision-making in real time in combination with the reinforcement learning algorithm, and the working mode and parameters of the sensor are automatically adjusted according to different task requirements.

[0019] Furthermore, in the communication and interface module, a hybrid wireless communication method combining millimeter wave communication technology and visible light communication technology is adopted; at the same time, the sensor fusion communication protocol is supported to fuse and transmit the magnetic tactile sensor data and the auxiliary sensor data.

[0020] Furthermore, in the mechanical structure module, the finger joint adopts a composite drive structure based on shape memory alloy (SMA) and piezoelectric ceramics; at the same time, a bionic adsorption layer with micro-nano structure is designed on the finger surface to simulate the adsorption principle of insect feet, and adsorption and grasping are realized on objects with different materials and surface topographies.

[0021] Furthermore, it also includes:

[0022] Fault diagnosis module: It monitors the working state of the sensor system in real time. By analyzing the output signals of the magnetic tactile perception module, the gain and filtering parameters of the signal conditioning module, the calculation results of the data acquisition and processing module, and the communication quality information of the communication and interface module, it uses a fault diagnosis algorithm based on Bayesian network to determine whether the sensor system has a fault. When a fault is detected, the system automatically triggers an alarm mechanism. At the same time, it uses a fault prediction algorithm to predict the probability and time of fault occurrence based on historical fault data and real-time monitoring information.

[0023] Energy management module: It adopts a composite energy recovery technology based on triboelectric nanogenerator (TENG) and thermoelectric devices. The mechanical energy and thermal energy generated during the movement of the robot finger are converted into electrical energy and stored in a hybrid energy storage device composed of a capacitor and a lithium battery. By optimizing the structure and materials of the TENG, the efficiency η of converting mechanical energy into electrical energy is improved. mech , and the efficiency calculation formula is where E mech-elec is the electrical energy converted from mechanical energy, and E total-mech is the total mechanical energy generated by the movement of the robot finger. By optimizing the thermoelectric conversion materials and structure of the thermoelectric device, the efficiency η of converting thermal energy into electrical energy is improved. therm , and the efficiency calculation formula is where E therm-elec is the electrical energy converted from thermal energy, and E total-therm is the total thermal energy generated by the robot finger. At the same time, it manages the charging and discharging process of the hybrid energy storage device and uses a fuzzy control algorithm to adjust the charging and discharging strategy in real time according to the energy demand and energy storage state.

[0024] Compared with the existing technologies, the beneficial effects of the present invention are as follows:

[0025] In terms of perception ability, the system realizes bimodal perception of normal force and tangential force, and can obtain the physical properties of objects more comprehensively and accurately. By using a magnetic elastomer material with intelligent microstructures and advanced magnetic sensitive elements, the sensitivity and accuracy of the sensor are greatly improved, and it can sense the change of tiny forces, providing richer information for the robot and making it more precise and stable when grasping and operating objects. For example, when grasping fragile items, the grasping force can be accurately controlled according to the magnitude and direction of the sensed force to avoid damage to the items.

[0026] In terms of signal processing and data acquisition, the system adopts advanced algorithms and technologies, which can efficiently process and analyze the weak signals output by the sensors. The signal conditioning based on adaptive neural network and the data processing method based on deep learning improve the accuracy and stability of the signals, and at the same time can extract useful information from a large amount of data, providing strong support for the decision-making of the robot.

[0027] In terms of environmental adaptability, the temperature compensation module and self-calibration module of the system can effectively eliminate the influence of environmental factors on the performance of the sensor, ensuring stable measurement accuracy under different temperatures and working conditions. This enables the robot to work in a wider range of environments and expands its application scope.

[0028] The communication and interface module adopts advanced encryption and communication technologies, ensuring the security and reliability of data transmission and achieving seamless docking between the sensor system and the robot control system. At the same time, the innovative design of the mechanical structure module, such as the bionic multi-degree-of-freedom variable stiffness structure and flexible materials with shape memory characteristics, improves the flexibility and adaptability of the robot's fingers, enabling it to better complete complex tasks.

[0029] The addition of the fault diagnosis and energy management module improves the reliability and endurance of the system. The fault diagnosis module can detect and handle faults in a timely manner, reducing the downtime of the robot; the energy management module extends the endurance time of the sensor system and reduces the operating cost through composite energy recovery technology. Brief Description of the Drawings

[0030] Figure 1 It is a schematic block diagram of the humanoid robot finger magnetic tactile sensor system with dual-modal sensing function proposed by the present invention;

[0031] Figure 2 It is a schematic block diagram of the humanoid robot finger magnetic tactile sensor system with dual-modal sensing function for comparing the sensing accuracy of different sensors;

[0032] Figure 3 It is a schematic block diagram of the signal processing speed of the humanoid robot finger magnetic tactile sensor system with dual-modal sensing function proposed by the present invention changing with time. Detailed Embodiments

[0033] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0034] In the description of the present invention, it should be understood that the orientation or positional relationship indicated by the terms "center", "longitudinal", "transverse", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise", etc. is based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation to the present invention.

[0035] In addition, the terms "first" and "second" are only used for descriptive purposes and should not be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of the said features. In the description of the present invention, the meaning of "a plurality" is two or more unless otherwise specifically defined. In addition, the terms "mounted", "connected", and "coupled" should be understood in a broad sense. For example, it may be a fixed connection, a detachable connection, or an integral connection; it may be a mechanical connection or an electrical connection; it may be directly connected or indirectly connected through an intermediate medium, and it may be the communication inside two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances. The present invention will be further described in detail below with reference to the drawings.

[0036] Refer to Figure 1 and Figure 3 : A specific implementation manner of a humanoid robot finger magneto-tactile sensor system with dual-modal perception function.

[0037] Magneto-tactile perception module: This module is the core to achieve dual-modal perception. A magnetic elastomer material with intelligent microstructures is selected. During the material preparation process, microscopic magnetic sensitive fibers are evenly embedded therein to form a three-dimensional magnetic sensitive structure. When subjected to an external force, the magnetic elastomer not only undergoes macroscopic deformation, but also the microscopic orientation and spacing of the internal magnetic sensitive fibers change, thereby causing a complex and precise change in the magnetic field distribution.

[0038] For the normal pressure F n , the change amount of the electrical signal and F n satisfy the relationship where the non-linear normal pressure sensitivity coefficient k 1n , the non-linear exponent α of the normal pressure n and the logarithmic adjustment coefficient b of the normal pressure 1nIt needs to be determined through experimental calibration. For example, in a laboratory environment, multiple measurements of normal pressures of different magnitudes are carried out, and then the values of these coefficients are obtained by fitting using the least squares method.

[0039] For the tangential force F t , the change in the electrical signal and F t satisfy Similarly, the non - linear tangential force sensitivity coefficient k 2t , the tangential force non - linear index α t , the tangential force sine adjustment coefficient b 2t and the tangential force angular frequency ω t also need to be determined through experiments. By measuring and fitting tangential forces of different directions and magnitudes, accurate coefficient values are obtained.

[0040] The magnetosensitive element uses a composite magnetosensitive sensor that combines the anisotropic magnetoresistance (AMR) effect and the tunneling magnetoresistance (TMR) effect. This type of sensor has high sensitivity and fast response speed. Arranging the magnetosensitive elements in a three - dimensional staggered array layout can perceive the force distribution comprehensively and with high precision. By calculating the difference and ratio of the electrical signals of magnetosensitive elements in different layers and combining with spatial geometry algorithms, the acting position, direction of the force and the gradient change of the force can be judged more accurately.

[0041] Signal conditioning module: This module processes the weak electrical signals output by the magnetic tactile perception module. A programmable gain amplifier based on an adaptive neural network is used for signal amplification. The input of the neural network is the characteristics of the input signal, such as the amplitude, frequency, etc. of the signal, and the output is the gain G. The gain adjustment formula is where G0 is the initial gain, V ref is the reference voltage, V in is the input signal voltage, V target is the target signal voltage, and β is the gain adjustment coefficient. By continuously adjusting the value of β, the amplifier can automatically adjust the gain according to the intensity of the input signal, improving the signal amplification effect.

[0042] In terms of filtering, a multi - order adaptive notch filter is used to filter out noise of specific frequencies. According to the frequency characteristics of the input signal, the parameters of the filter, such as the cut - off frequency, order, etc., are automatically adjusted to better adapt to signals in different frequency ranges. A polynomial fitting method controlled by fuzzy logic is used for signal linearization, and the order and coefficients of the polynomial are dynamically adjusted according to the dynamic range and noise level of the input signal. For example, when the dynamic range of the input signal is large, the order of the polynomial is increased to improve the accuracy of linearization.

[0043] Data acquisition and processing module: A high - speed data acquisition card is used to collect the conditioned signals, and the sampling frequency fs Satisfy f s = max(2f max , f dynamic ), where f max is the highest frequency component of the signal, and f dynamic is the dynamic sampling frequency adjusted in real time according to the dynamic change rate of the force. By monitoring the change rate of the force in real time and dynamically adjusting the sampling frequency, it is ensured that the change information of the signal can be accurately collected.

[0044] The collected data is transmitted to a multi-core heterogeneous microprocessor, and a multi-scale feature extraction algorithm based on deep learning is used for data processing. Combining a convolutional neural network (CNN) and a recurrent neural network (RNN), the features of the force signal are extracted from different time scales and spatial scales. CNN is used to extract the spatial features of the signal, and RNN is used to process the time series features of the signal. The actual force value is calculated through the established multi-physical field coupling mathematical model, and the decomposition and synthesis of the force are carried out to obtain the magnitude, direction, action point and dynamic change trend information of the force. At the same time, this module can also predict the motion state and physical properties of the object according to the change of the force. For example, by analyzing the change trend of the force, it is predicted whether the object will slide or fall.

[0045] Communication and interface module: This module supports multiple communication protocols such as SPI, I2C, USB and CAN, which is convenient for data transmission with the control system of the humanoid robot. An adaptive data encryption mechanism based on a chaotic encryption algorithm is used to encrypt the transmitted data. The encryption key K is generated by a chaotic system, and its iterative formula is x n+1 = μ·x n ·(1 - x n ), where μ is the chaotic parameter and x n is the nth iteration value of the chaotic system. The encryption formula is C = E(M, K)·exp(-γ·t), where M is the original data, C is the encrypted data, E is the encryption function, γ is the time decay coefficient, and t is the transmission time. By continuously adjusting the values of μ and γ, the encryption intensity and transmission rate are adaptively adjusted according to the change of the communication environment to ensure the security and reliability of data transmission.

[0046] Mechanical structure module: Integrate the magnetic tactile perception module into the fingertip part of the finger, and wrap the sensor with a flexible material with shape memory characteristics. After being deformed by an external force, this material can return to its original shape under specific conditions (such as temperature change), simulating the elasticity and self-healing ability of the human finger skin. The finger joint adopts a bionic multi-degree-of-freedom variable stiffness structure, and the rotation angle θ of the joint and the applied force F satisfy the relationship where a i is the polynomial coefficient, is the angular fluctuation amplitude, ω is the fluctuation angular frequency, and t is the time. By monitoring the applied force in real time, the stiffness of the joint is adjusted to adapt to different grasping tasks. For example, when grasping a heavy object, the stiffness of the joint is increased to improve the stability of the grasp.

[0047] Self-calibration module: This module performs self-calibration operations when the sensor system is started or periodically. By applying known normal forces and tangential forces with complex dynamic characteristics through a built-in multi-dimensional standard force source, the output signals of the sensors are collected, compared with the theoretical values, and the error values are calculated. According to the error values, the sensitivity coefficients k 1n , k 2t of the magnetic tactile perception module are corrected, and the correction formula is where Δk 1n , Δk 2t are the error values of the sensitivity coefficients, and β n , β t are the correction exponents of the normal force and the tangential force. At the same time, this module can also automatically calibrate the gain and filtering parameters of the signal conditioning module to improve the measurement accuracy and stability of the sensor.

[0048] Temperature compensation module: Multiple high-precision temperature sensors are integrated inside the sensor to monitor the temperature changes T i at different locations in real time. According to the multi-variable coupling relationship between temperature and sensitivity coefficients, gain, filtering parameters, etc., a temperature compensation model based on support vector regression (SVR) is established. The compensated electrical signal V compensated is calculated by the formula where V original is the original electrical signal, w i is the temperature weight coefficient, T 0i is the reference temperature, ∈ is the temperature fluctuation adjustment coefficient, ω T is the temperature angular frequency. By continuously adjusting the values of w i , ∈ and ω T , the electrical signals output by the magnetic tactile perception module and the parameters of the signal conditioning module are compensated to eliminate the influence of temperature on the measurement results.

[0049] Fault diagnosis module: This module monitors the working state of the sensor system in real time. By analyzing multi-source information such as the output signal of the magnetic tactile perception module, the gain and filtering parameters of the signal conditioning module, the calculation results of the data acquisition and processing module, and the communication quality of the communication and interface module, it uses a fault diagnosis algorithm based on Bayesian network to determine whether the sensor system has a fault. When a fault is detected, the system automatically triggers a multi-level alarm mechanism, adopts different alarm methods according to the severity of the fault, such as audible and visual alarms, SMS alarms, etc., and sends the fault information and diagnosis results to the robot control system through the communication and interface module. At the same time, a fault prediction algorithm is adopted to predict the probability and time of fault occurrence based on historical fault data and real-time monitoring information, and maintenance and replacement are carried out in advance to reduce the downtime of the robot.

[0050] Energy management module: Adopting a composite energy recovery technology based on triboelectric nanogenerator (TENG) and thermoelectric devices, it converts the mechanical energy and heat energy generated during the movement of the robot finger into electrical energy, which is stored in a hybrid energy storage device composed of a supercapacitor and a lithium battery. The efficiency η of converting mechanical energy into electrical energy mech The calculation formula is By optimizing the structure and materials of TENG, the efficiency of converting mechanical energy into electrical energy is improved. The efficiency η of converting heat energy into electrical energy therm The calculation formula is By optimizing the thermoelectric conversion materials and structure of the thermoelectric device, the efficiency of converting heat energy into electrical energy is improved. A fuzzy control algorithm is used to adjust the charge and discharge strategy in real time according to the energy demand and energy storage state, reasonably manage the charge and discharge process of the hybrid energy storage device, and extend the battery life of the sensor system.

[0051] Data representation and interpretation

[0052] Evaluation Index Traditional Sensor System System of the Present Application Improvement Effect Force Sensing Accuracy (%) 70 95 Improved by 25 percentage points Signal Processing Speed (ms) 100 20 Shortened by 80 ms Environmental Adaptability (Temperature Range) -20°C to -60°C -40°C to -80°C Broadened by 40°C Data Transmission Security (Encryption Strength) Low High Significantly Improved Battery Life (Hours) 8 24 Extended by 16 hours

[0053] It can be seen from the tabular data that the sensor system of this application has significant advantages in many aspects. The substantial improvement in force perception accuracy enables the robot to complete grasping and operating tasks more accurately, improving work efficiency and quality. The shortening of signal processing speed allows the robot to respond to external changes faster, enhancing real-time performance. The broadening of environmental adaptability enables the robot to work in harsher environments, expanding the application range. The significant improvement in data transmission security ensures the reliable transmission of sensor data and avoids data leakage and interference. The extension of battery life reduces the charging times of the robot and improves the continuous working ability of the robot. These advantages make the sensor system of this application have broad application prospects in the field of humanoid robots.

[0054] The above are only the preferred specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution and inventive concept of the present invention, making equivalent replacements or changes should be covered within the protection scope of the present invention.

Claims

1. A humanoid robot finger magnetic tactile sensor system with dual-modal perception function, characterized in that, Including: Magnetic tactile perception module: A microstructured magnetic elastomer material is used as the sensing medium. When an external force acts, the magnetic elastomer undergoes macroscopic deformation, and the microscopic orientation and spacing of the internal magnetosensitive fibers change accordingly, resulting in a change in the magnetic field distribution; for the normal force F n , the change in the electrical signal and F n satisfy the relationship where k 1n is the non-linear normal force sensitivity coefficient, α n is the normal force non-linearity exponent, b 1n is the normal force logarithmic adjustment coefficient; for the tangential force F t , the change in the electrical signal and F t satisfy k 2t is the non-linear tangential force sensitivity coefficient, α t is the tangential force non-linearity exponent, b 2t is the tangential force sine adjustment coefficient, ω t is the tangential force angular frequency; simultaneously sense the normal force and the tangential force to achieve dual-modal perception and sense the dynamic change process of the force; Signal conditioning module: Amplify, filter, and linearize the weak electrical signals output by the magnetic tactile sensing module. The neural network automatically adjusts the gain G according to the characteristics of the input signal. The gain adjustment formula is where G0 is the initial gain, V ref is the reference voltage, V in is the input signal voltage, V target is the target signal voltage, β is the gain adjustment coefficient; Filter out the noise of the frequency through an adaptive notch filter, and use the polynomial fitting method of fuzzy logic control to linearize the signal, and dynamically adjust the order and coefficients of the polynomial according to the dynamic range and noise level of the input signal; Data acquisition and processing module: Using a data acquisition card to collect the conditioned signals, with a sampling frequency f s satisfying f s = max(2f max , f aynamic ), where f max is the highest frequency component of the signal, and f dynamic is the dynamic sampling frequency adjusted in real time according to the dynamic change rate of the force; The collected data is transmitted to the microprocessor, and a feature extraction algorithm based on deep learning is used for data processing. Combining the convolutional neural network CNN and the recurrent neural network RNN to extract the features of the force signal from different time scales and spatial scales.

2. The anthropomorphic robot finger magneto-tactile sensor system with dual-modal perception function according to claim 1, characterized in that Also including: Communication and Interface Module: Supports SPI, I2C, USB, and CAN communication protocols to achieve data transmission with the control system of the humanoid robot. Adopt an adaptive data encryption mechanism based on a chaotic encryption algorithm to encrypt the transmitted data. The encryption key K is generated by a chaotic system, and the iterative formula is x n+1 = μ·x n .(1 - x n ), where μ is the chaotic parameter, and x n is the nth iteration value of the chaotic system; the encryption formula is C = E(M, K)·exp(-γ·t), where M is the original data, C is the encrypted data, E is the encryption function, γ is the time decay coefficient, and t is the transmission time; Mechanical Structure Module: Design the mechanical structure of the humanoid robot finger, integrate the magnetic tactile sensing module into the fingertip part of the finger, use a material with shape memory characteristics to wrap the sensor, and adopt a bionic variable stiffness structure for the finger joints to adjust the stiffness of the joints in real time according to the requirements of the grasping task; the rotation angle θ of the joint and the applied force F satisfy the relationship where a i is the polynomial coefficient, is the angle fluctuation amplitude, ω is the fluctuation angular frequency, and t is the time; Self - calibration module: When the sensor system starts or performs self - calibration regularly, a known normal pressure and tangential force with dynamic characteristics are applied through a built - in standard force source, the output signal of the sensor is collected and compared with the theoretical value, and the error value is calculated; according to the error value, the sensitivity coefficients k 1n , k 2t of the magnetic tactile perception module are corrected, and the correction formula is where Δk 1n , Δk 2t are the error values of the sensitivity coefficients, and β n , β t are the correction exponents of the normal pressure and tangential force; at the same time, the gain and filter parameters of the signal conditioning module are automatically calibrated.

3. The humanoid robot finger magneto-tactile sensor system with dual-modal perception function according to claim 1, characterized in that, Also including: Temperature compensation module: Considering the influence of temperature on the performance of the magnetic sensor element, magnetoelastic material, and signal conditioning circuit, a temperature sensor is integrated inside the sensor to monitor the temperature change T at different parts in real time i ; According to the multivariable coupling relationship between temperature and sensitivity coefficient, gain, and filtering parameters, a temperature compensation model based on support vector regression (SVR) is established to compensate the electrical signal output by the magnetic tactile perception module and the parameters of the signal conditioning module; the compensated electrical signal V compensated The calculation formula is to eliminate the influence of temperature on the measurement result, where V original is the original electrical signal, w i is the temperature weight coefficient, T 0i is the reference temperature, ∈ is the temperature fluctuation adjustment coefficient, ω T is the temperature angular frequency.

4. The humanoid robot finger magneto-tactile sensor system with dual-modal perception function according to claim 1, characterized in that, In the magnetic tactile sensing module, the magnetic sensitive element uses a composite magnetic sensitive sensor that combines the anisotropic magnetoresistance (AMR) effect and the tunneling magnetoresistance (TMR) effect. By optimizing the microstructure of the magnetoelastic material, magnetic particles and quantum dots are introduced to form a quantum-classical hybrid magnetic sensitive system. At the same time, the magnetic sensitive elements are arranged in a three-dimensional staggered array layout to sense the distribution of force. By calculating the difference and ratio of the electrical signals of the magnetic sensitive elements in different layers, combined with the spatial geometry algorithm, the acting position, direction, and force gradient change of the force are judged.

5. The humanoid robot finger magneto-tactile sensor system with dual-modal perception function according to claim 1, characterized in that, In the signal conditioning module, a composite filtering algorithm based on wavelet transform and Kalman filter is used for filtering. At the same time, an adaptive linearization algorithm optimized by genetic algorithm is used to adjust the parameters of the linearization curve in real time according to the statistical characteristics and dynamic changes of the input signal, so as to accurately achieve the linear relationship between the output signal and the actual force value.

6. The humanoid robot finger magneto-tactile sensor system with dual-modal perception function according to claim 1, wherein, In the data acquisition and processing module, a heterogeneous computing architecture based on system-on-chip (SoC) is adopted, integrating CPU, GPU, and FPGA on the same chip to achieve parallel computing data processing. At the same time, combined with the reinforcement learning algorithm, the collected data is analyzed and decision-making in real time, and the working mode and parameters of the sensor are automatically adjusted according to different task requirements.

7. The humanoid robot finger magneto-tactile sensor system with dual-modal perception function according to claim 2, characterized in that, In the communication and interface module, a hybrid wireless communication method combining millimeter-wave communication technology and visible light communication technology is adopted. At the same time, it supports the sensor fusion communication protocol to fuse and transmit the magnetic tactile sensor data and auxiliary sensor data.

8. The humanoid robot finger magneto-tactile sensor system with dual-modal perception function according to claim 2, characterized in that, In the mechanical structure module, the finger joints adopt a composite drive structure based on shape memory alloy (SMA) and piezoelectric ceramics. At the same time, a bionic adsorption layer with micro-nano structure is designed on the finger surface to simulate the adsorption principle of insect feet and achieve adsorption and grasping on objects with different materials and surface morphologies.

9. The anthropomorphic robot finger magneto-tactile sensor system with dual-modal perception function according to claim 1, characterized in that, Also including: Fault Diagnosis Module: Real-time monitors the working state of the sensor system by analyzing the output signal of the magnetic tactile sensing module, the gain and filtering parameters of the signal conditioning module, the calculation results of the data acquisition and processing module, and the communication quality information of the communication and interface module. Adopts a fault diagnosis algorithm based on Bayesian network to judge whether the sensor system fails. When a fault is detected, the system automatically triggers an alarm mechanism. At the same time, a fault prediction algorithm is adopted to predict the probability and time of fault occurrence according to historical fault data and real-time monitoring information.

10. The humanoid robot finger magnetic tactile sensor system with dual-modal perception function according to claim 1, characterized in that, Also including: Energy Management Module: Adopts a composite energy recovery technology based on triboelectric nanogenerator (TENG) and thermoelectric devices to convert the mechanical energy and heat energy generated during the movement of the robot finger into electrical energy and store it in a hybrid energy storage device composed of a capacitor and a lithium battery. By optimizing the structure and materials of the TENG, the efficiency η of converting mechanical energy into electrical energy is improved. mech , and the efficiency calculation formula is where E mech-elec is the electrical energy converted from mechanical energy, and E total-mech is the total mechanical energy generated by the movement of the robot finger; by optimizing the thermoelectric conversion materials and structure of the thermoelectric device, the efficiency η of converting thermal energy into electrical energy is improved. therm , and the efficiency calculation formula is where E therm-elec is the electrical energy converted from thermal energy, and E total-therm is the total thermal energy generated by the robot finger; at the same time, manage the charging and discharging process of the hybrid energy storage device, and use a fuzzy control algorithm to adjust the charging and discharging strategy in real time according to the energy demand and the energy storage state.

Images