Robot health state authentication method and system based on multi-dimensional fusion

Abstract

The invention relates to the technical field of artificial intelligence and robots, and discloses a robot health state authentication method and system based on multi-dimensional fusion, and the method comprises the steps: collecting multi-source data, and carrying out the standardization preprocessing; 17 health dimensions are estimated based on an algorithm model; fusing the health dimensions to generate a comprehensive health index; authenticating the health state, and generating a health authentication report; reporting and coding, and storing to a block chain; the system comprises a multi-source data acquisition module, a dimension calculation engine, a correlation analysis module, a dynamic authentication generator, a block chain evidence storage module and an application interface layer. According to the method, a 17-dimensional health index system is constructed, hardware multiplexing is realized by applying an algorithm model based on current, speed, vision, network data and the like, an expensive physical sensor is effectively replaced, and more comprehensive health data can be obtained by matching with a chemical risk inversion model based on vision and network data to generate an environmental chemical index.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of artificial intelligence and robot technology, in particular to a robot health state authentication method and system based on multi-dimensional fusion. BACKGROUND

[0002] Humanoid robots show application potential in many fields such as industrial production, medical service, emergency rescue, etc. With the increase of its functional complexity and work intensity, it is of great significance to ensure the continuous healthy operation of the humanoid robot for its task reliability, service life and operation cost control.

[0003] The existing health state monitoring of humanoid robots or similar intelligent devices usually relies on direct measurement of specific operating parameters, such as monitoring the operating temperature of motors and joints through integrated temperature sensors, monitoring motor drive current through current sensors, or monitoring the angular position and speed of joints through encoders. The data directly obtained by these sensors can reflect the physical state of the device at a specific moment. Some systems will set fixed thresholds and trigger alarms when the monitoring data exceeds the preset range to indicate possible abnormal conditions.

[0004] However, the existing humanoid robot health state monitoring technology relies on single-dimensional data such as temperature and current, which is difficult to comprehensively and multi-dimensionally reflect the complex operating state and potential health risks of the humanoid robot. In order to obtain more comprehensive performance data, some solutions need to integrate special and expensive physical sensors such as high-precision torque sensors or infrared thermal imagers. The introduction of these additional hardware greatly increases the manufacturing cost and maintenance complexity of the robot, and most existing monitoring systems can only reflect the current instantaneous state of the device and mainly trigger alarms based on threshold values, which makes it difficult to provide early warning of potential failures, resulting in delayed maintenance response. Therefore, the present application provides a robot health state authentication method and system based on multi-dimensional fusion to solve the problems existing in the prior art. SUMMARY

[0005] In view of the deficiencies of the prior art, the present application provides a robot health state authentication method and system based on multi-dimensional fusion, which solves the problems of single-dimensional limitation, high hardware dependency, lack of prediction ability and lack of standardized authentication in the existing robot health state monitoring.

[0006] To achieve the above purpose, the present application is realized by the following technical scheme:

[0007] The first aspect of the present application provides a humanoid robot health state authentication method, comprising the following steps:

[0008] S1, collect multi-source data of the humanoid robot during operation. The data includes inertial measurement unit data, joint current signal, joint angular velocity, RGB-D camera data, microphone array data, battery management system data, controller log, network communication data, and external network air quality data. The collected data is standardized and pre-processed to ensure data consistency and availability.

[0009] S2, based on the collected multi-source data of the humanoid robot, use an algorithm model to realize hardware reuse, perceive and estimate 17 health dimensions of the humanoid robot. The 17 health dimensions include quantum behavior entropy, joint harmonic entropy, drive chain efficiency, visual perception entropy, sound field modeling strength, neural synchronization rate, energy management index, thermodynamic stability, battery health, neural efficiency, emotional service ability, environmental interaction, learning transfer rate, global collaboration degree, communication quality index, gait energy entropy, and environmental chemical index.

[0010] In the hardware reuse algorithm model, specifically includes:

[0011] Torque estimation model based on current and speed signals: This model is used to estimate the actual torque of the joints of the humanoid robot, so as to calculate the drive chain efficiency, the formula is: In the formula, k t represents the motor constant; I represents the joint current signal; J represents the moment of inertia; represents the angular acceleration; b represents the damping coefficient; ω represents the angular velocity. This model calculates the motor parameters, current signals, and angular velocity signals to replace the physical torque sensor.

[0012] Chemical risk inversion model based on visual analysis, network air quality data, and historical exposure records: This model is used to calculate the risk factor of the environmental chemical index. The formula is: risk_factor = 0.4S + 0.4A / 50 + 0.2C, where risk_factor represents the chemical risk factor; S represents the smoke concentration and dust visibility index obtained by the visual system; A represents the network air quality index; C represents the historical chemical contact record related index. This model integrates multi-source data to realize non-direct perception of environmental chemical risk.

[0013] Component temperature prediction model based on thermodynamic equation: This model is used to replace the physical thermal imaging device to predict the temperature distribution or temperature rise of the key components of the humanoid robot to evaluate the thermodynamic stability.

[0014] S3, fuse the 17 health dimensions of the humanoid robot perceived and estimated to generate the comprehensive health index HHV of the humanoid robot. The comprehensive health index is obtained by using a multi-dimensional fusion algorithm, and its calculation formula is:

[0015]

[0016] In the formula, D i is the core health dimension; w i is the core dimension weight; is the core dimension geometric mean; 0.5 is the square root transformation; D j is the auxiliary health dimension; is the auxiliary dimension arithmetic mean; count(D k <5.0) is the number of dimensions of health severe abnormalities; is the health attenuation factor. The core dimension weight is dynamically adjusted according to the application scenario of the humanoid robot, for example, when the humanoid robot is applied to an industrial carrying scene, the weight of the drive chain efficiency is increased; when applied to a service robot scene, the weight of the neural efficiency is increased; when applied to a medical robot scene, the weight of the environmental chemical index is increased. The generation steps of the environmental chemical index specifically include: collecting smoke concentration and dust visibility data obtained by a vision system; obtaining a regional air quality index provided by an external network interface; analyzing chemical exposure history data recorded in a robot task log; calculating a chemical risk factor based on the collected data, obtained data and analyzed data; and generating an environmental chemical index according to the risk factor and recognition delay.

[0017] S4, according to the comprehensive health index and the specific scores of the 17 humanoid robot health dimensions, the current health status of the humanoid robot is authenticated, and a health authentication report is generated. When the score of the drive chain efficiency or the thermodynamic stability is lower than a preset threshold, a predictive maintenance warning of the humanoid robot is triggered, and a maintenance work order is automatically generated according to the warning information, which can indicate that component replacement or system calibration is performed.

[0018] S5, the generated health authentication report and the identity code associated with the humanoid robot are stored to a blockchain, so as to ensure the authenticity, non-tamperability and traceability of the authentication information.

[0019] The second aspect of the present application provides a robot health state authentication system based on multi-dimensional fusion, which is used to realize the above-mentioned method. The system comprises:

[0020] A multi-source data acquisition module: this module is used for acquiring and preprocessing multi-source data of the humanoid robot during operation, and providing basic data for subsequent health state evaluation.

[0021] A dimension calculation engine: this module contains parallel calculation functions of 17 authentication dimensions, which is used for perceiving and estimating 17 humanoid robot health dimensions based on a multi-source data algorithm model, and realizing quantification of different aspects of the robot performance.

[0022] Correlation analysis module: This module is used for cross-dimensional analysis based on dynamic correlation coefficient matrix, analyzing the mutual influence and potential correlation between different health dimensions, so as to reveal deeper health problems or potential risks.

[0023] Dynamic authentication generator: This module is used for authenticating the current health status of the humanoid robot and generating a unique identity code and health authentication report of the humanoid robot. The identity code is designed to follow specific rules, aiming to provide a unique identification and health status overview of the robot.

[0024] Blockchain storage module: This module is used to store the unique identity code and health authentication report of the humanoid robot to the blockchain storage, ensuring the security, transparency and credibility of the authentication data.

[0025] Application interface layer: This module is used to provide an interface for data interaction with external systems and users, so that the health authentication results can be widely applied, such as for maintenance decision support, second-hand transaction evaluation, etc.

[0026] The present application provides a robot health status authentication method and system based on multi-dimensional fusion. It has the following beneficial effects:

[0027] 1. The present application constructs a 17-dimensional health index system, and uses algorithm models based on current, speed, vision and network data to realize hardware reuse, effectively replacing expensive physical sensors. Through torque estimation model to calculate the driving chain efficiency, and chemical risk inversion model based on vision and network data to generate environmental chemical index, more comprehensive health data can be obtained without additional hardware investment.

[0028] 2. The present application integrates 17-dimensional health data through a comprehensive health index formula, and introduces dynamic weight distribution and health decay factor. The geometric mean of the core dimension can highlight the system bottleneck, and when the score of the core dimension or auxiliary dimension is lower than the preset threshold, predictive maintenance warning can be triggered and maintenance work order can be automatically generated, prompting component replacement or system calibration, thereby improving maintenance response speed.

[0029] 3. The present application stores the generated health authentication report and humanoid robot identity code to the blockchain, uses the decentralized and tamper-proof characteristics of the blockchain, and ensures the authenticity and traceability of the health authentication data. BRIEF DESCRIPTION OF DRAWINGS

[0030] Figure 1 The method flowchart of the present application is shown in the figure;

[0031] Figure 2 The hardware reuse flowchart of the present application is shown in the figure;

[0032] Figure 3 System architecture diagram of the present application;

[0033] Figure 4 17-dimensional index formula specification diagram of the present application. DETAILED DESCRIPTION

[0034] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all the other embodiments obtained by those skilled in the art without creative work fall within the protection scope of the present application.

[0035] Refer to the accompanying drawings Figure 1 , Figure 1 It is a health state authentication method flow chart of a humanoid robot according to an embodiment of the present application. The present application provides a health state authentication method of a humanoid robot, comprising the following steps:

[0036] S1, collect multi-source data of the humanoid robot in the running process: the multi-source data includes inertial measurement unit data, joint current signal, joint angular velocity, RGB-D camera data, microphone array data, battery management system data, controller log, network communication data and external network air quality data. The data is collected by a multi-source data acquisition module and standardized input and preliminary pretreatment is performed.

[0037] S2, hardware multiplexing based on the collected multi-source data algorithm model of the humanoid robot, perceiving and estimating 17 health dimensions of the humanoid robot: the 17 health dimensions of the humanoid robot include quantum behavioral entropy, joint harmonic entropy, drive chain efficiency, visual perception entropy, acoustic modeling strength, neural synchronization rate, energy management index, thermodynamic stability degree, battery health index, neural efficiency index, emotionalized service capacity, environmental interaction force, learning transfer rate, global collaboration degree, communication quality index, gait energy entropy, and environmental chemical index. The perception and estimation process is performed by a dimension calculation engine.

[0038] In the hardware multiplexing algorithm model, the following models are included:

[0039] A torque estimation model based on current and speed signals is used to estimate the actual torque of the joints of the humanoid robot, thereby calculating the drive chain efficiency DCE.

[0040] A chemical risk inversion model based on visual analysis, network air quality data, and historical exposure records is used to calculate the risk factor of the environmental chemical index ECI.

[0041] A component temperature prediction model based on thermodynamic equations is used to replace physical thermal imaging devices to predict the temperature distribution or temperature rise of key components of the humanoid robot to evaluate the thermodynamic stability TSD.

[0042] S3, the 17 humanoid robot health dimensions obtained by perception and estimation are fused to generate a comprehensive health index HHV of the humanoid robot: the comprehensive health index is generated by the correlation analysis module and the dynamic authentication generator.

[0043] S4, according to the comprehensive health index and the specific scores of the 17 humanoid robot health dimensions, the current health status of the humanoid robot is authenticated, and a health authentication report and a unique identity code of the humanoid robot are generated: the identity code is generated by the dynamic authentication generator. When the score of the drive chain efficiency DCE or the thermodynamic stability TSD is lower than the preset threshold, the system triggers a predictive maintenance warning of the humanoid robot, and automatically generates a maintenance work order according to the warning information, which can indicate component replacement or system calibration.

[0044] S5, the generated health authentication report and the identity code associated with the humanoid robot are stored in the blockchain: the storage process is completed by the blockchain storage module 500, which ensures the authenticity, non-tamperability and traceability of the authentication results.

[0045] For S1, the multi-source data acquisition module is used to acquire various data of the humanoid robot during operation. The data acquisition includes the following aspects:

[0046] Inertial measurement unit IMU data: through the IMU sensor integrated on the humanoid robot body, the three-axis acceleration, angular velocity and attitude information of the robot are collected in real time. These data are transmitted to the multi-source data acquisition module in the form of digital signals.

[0047] Joint current signal and joint angular velocity: through the current sensor and encoder connected to the robot joint motor, the real-time current value of the motor and the real-time rotation speed of the joint are collected respectively. These electrical signals and digital signals are converted by A / D or directly transmitted in digital form.

[0048] RGB-D camera data: through the RGB-D camera installed on the robot, color image RGB and depth image Depth data are obtained, which contain visual information and spatial geometric information of the environment, such as smoke concentration, dust visibility and feature point position, etc.

[0049] Microphone array data: through the microphone array carried by the robot, environmental sound field data and robot's own running sound are collected, which are used for voiceprint feature analysis and noise detection.

[0050] Battery management system BMS data: the key parameters such as voltage, current, temperature, internal resistance and charge-discharge cycle number of the battery are obtained from the BMS of the humanoid robot.

[0051] Controller log: Obtain system running log, task execution status, preset action time, actual execution time, etc. from the main controller or each sub-controller of the robot.

[0052] Network communication data: Through the communication module of the robot, obtain real-time signal-to-noise ratio SNR, packet loss rate, actual communication bandwidth, etc. Network communication quality parameters.

[0053] External network air quality data: Through the network interface, obtain the regional air quality index AQI and other environmental chemical information provided by the external air quality monitoring system.

[0054] All collected raw data are transmitted to the multi-source data acquisition module for standardized input and preliminary preprocessing. Standardized input includes data format unification, unit conversion, etc. to ensure the compatibility of data from different sources. Preliminary preprocessing includes data denoising, outlier detection, missing value filling, etc. to improve data quality and provide accurate data basis for subsequent dimension calculation.

[0055] For S2, the dimension calculation engine perceives and estimates 17 health dimensions of human-shaped robots based on the preprocessed data provided by the multi-source data acquisition module.

[0056] Quantum behavior entropy QBE is used to evaluate the uncertainty and abnormality of robot behavior, and the calculation is based on IMU sensor data and task log:

[0057] QBE = 10-2F f -5D d +2πh∑lnp i ;

[0058] In the formula, F f represents the number of falls per thousand hours, which is obtained by analyzing the frequency of falls detected by the IMU sensor; D d represents the task delay rate, which is obtained by calculating the ratio of the number of delayed tasks to the total number of tasks, and its range is [0, 1]; p i represents the state probability distribution, including normal state, delay state, and fall state, which is obtained by quantum noise analysis of the state data during the robot running process; h represents the Planck constant scaling factor, which is used to simulate the quantum tunneling effect, and its value range is 0.01-0.05.

[0059] Joint harmonic entropy JHE is used to evaluate the distortion degree of robot joint current waveform, reflecting the smoothness of joint operation. This dimension is calculated through joint current sensor data.

[0060] JHE = 10-min(THD × 0.8, 8);

[0061] In the formula, THD represents the total harmonic distortion rate, and its calculation formula is I n represents the nth harmonic component; I1 represents the fundamental component. The larger the THD value, the more serious the current waveform distortion, and the worse the smoothness of joint operation.

[0062] Drive chain efficiency DCE is used to evaluate the efficiency and loss of the robot power system. With hardware multiplexing mechanism, torque is estimated by current and speed signals:

[0063]

[0064] where τ i represents the actual torque, obtained by current and angular velocity estimation; represents the desired torque, i.e. the target value output by the control system; τ max represents the maximum output torque capacity of the joint; n represents the number of sampling points.

[0065] The calculation of drive chain efficiency DCE relies on the torque estimation model based on current and speed signals. This model estimates the real-time torque τ est of the joint through parameters such as motor constant, moment of inertia, damping coefficient, joint current signal and angular acceleration.

[0066]

[0067] where k t represents the motor constant; I represents the joint current signal; J represents the moment of inertia; represents the angular acceleration; b represents the damping coefficient; ω represents the angular velocity.

[0068] Neural synchronization rate NSR is used to evaluate the consistency between the instructions of the robot control system and the actual executed actions.

[0069] Neural efficiency NEX is used to evaluate the precision and energy efficiency of the robot in executing specific tasks.

[0070]

[0071] where t p represents the preset action time, i.e. the planned theoretical execution time; t e represents the actual execution time; I σ represents the current standard deviation during execution, reflecting the current fluctuation.

[0072] Environmental chemical index ECI is used to evaluate the adaptability of the robot to chemical risks in a specific environment. This dimension is realized through a chemical risk inversion model based on visual analysis, network air quality data and historical exposure records.

[0073]

[0074] wherein risk_factor represents the chemical risk factor; t response represents the recognition delay time of the robot to the environmental changes.

[0075] The formula for calculating the chemical risk factor risk_factor is:

[0076] risk_factor = 0.4S + 0.4A / 50 + 0.2C;

[0077] wherein S represents the smoke concentration and dust visibility index obtained by the visual system, with a value range of [0, 1]; A represents the network air quality index, with a value range of [0, 500]; C represents the historical chemical exposure record related index, with a value range of [0, 1]. The value of S is obtained by analyzing the image through the RGB-D camera to detect the visual features of smoke and dust; the value of A is obtained from the external air quality monitoring platform through the network interface; and the value of C is obtained by analyzing the chemical exposure type and duration recorded in the robot task log.

[0078] The visual perception entropy VPE is used to evaluate the stability of the robot visual system in recognizing environmental feature points; the acoustic field modeling strength AMS is used to evaluate the identification and modeling ability of the robot to environmental acoustic features; the energy management index EMI is used to evaluate the stability of the robot power consumption and the efficiency of energy consumption management; the thermodynamic stability TSD is used to evaluate the uniformity of heat dissipation and temperature stability of the internal components of the robot, which is replaced by a component temperature prediction model based on thermodynamic equations instead of physical thermal imaging equipment to predict the temperature distribution or temperature rise of the key components of the humanoid robot; the battery health BHI is used to evaluate the current performance and expected life of the robot battery; the emotional service ability ECS is used to evaluate the emotional recognition and feedback ability of the service robot in human-robot interaction; the environmental interaction force EIF is used to evaluate the perception, recognition and response ability of the robot to environmental changes; the learning transfer rate LTR is used to evaluate the efficiency of the robot in learning from existing task experience and adapting to new tasks; the global collaboration degree GCD is used to evaluate the task allocation, data exchange and consensus reaching level among robots in a multi-robot collaboration system; the communication quality index CQI is used to evaluate the reliability and efficiency of the communication link between the robot and the external system; the gait energy entropy GEE is used to evaluate the stability and balance of the gait of the humanoid robot, reflecting the abnormal vibration in the movement process.

[0079] For S3, the correlation analysis module and the dynamic authentication generator cooperate to fuse the 17 health dimension scores output by the dimension calculation engine, generating the comprehensive health index HHV of the humanoid robot. The fusion process adopts a multi-dimension fusion algorithm, and the formula is:

[0080]

[0081] wherein D i is the core health dimension; w i is the core dimension weight; is the core dimension geometric mean; 0.5 is the square root transformation; D j is the auxiliary health dimension; is the auxiliary dimension arithmetic mean; count(D k <5.0) is the number of dimensions with severe abnormal health; and is the health decay factor. The core dimension geometric mean G is calculated as:

[0082]

[0083] By taking the square root of the weighted product of the core dimensions, the aim is to highlight the impact of the dimensions that perform worse on the overall health status. When a core dimension has a low score, the final geometric mean is significantly affected even if other dimensions have high scores.

[0084] D j is the auxiliary health dimension, including visual perception entropy VPE, sound field modeling strength AMS, energy management index EMI, thermodynamic stability TSD, battery health index BHI, emotional service capability ECS, environmental interaction force EIF, learning transfer rate LTR, global collaboration degree GCD, communication quality index CQI, and gait energy entropy GEE, a total of 11 dimensions, whose value range is [0, 10]. represents the arithmetic mean of the auxiliary dimensions, reflecting the overall performance of the secondary performance indicators of the robot.

[0085] count(D k <5.0) represents the number of dimensions with severe abnormal health, i.e., the number of dimensions with scores lower than the preset threshold 5.0 among all 17 health dimensions, whose value range is [0, 17]. is the health decay factor, which is used to punish the dimensions with severe abnormal health. The more abnormal dimensions there are, the smaller the health decay factor, resulting in a lower final HHV value, which objectively reflects the degree of deterioration of the overall health status of the robot.

[0086] The core dimension weight w i is dynamically adjusted according to the specific application scenario of the humanoid robot, to adapt to different requirements of different scenarios for the performance emphasis of the robot. For example:

[0087] When the humanoid robot is applied to an industrial carrying scenario, the weight of the drive chain efficiency DCE is increased to more accurately reflect its health status in terms of power output and load capacity.

[0088] When the humanoid robot is applied to the service robot scenario, the weight of the neural efficiency NEX increases to highlight its health status in terms of task execution efficiency and precision.

[0089] When the humanoid robot is applied to the medical robot scenario, the weight of the environmental chemical index ECI increases to emphasize its adaptability and avoidance ability to chemical pollution in complex medical environments.

[0090] For S4, the dynamic authentication generator generates the current health authentication report of the humanoid robot according to the comprehensive health index HHV and the specific scores of the 17 humanoid robot health dimensions, and creates a unique identity code for the robot.

[0091] The health authentication report contains the following information:

[0092] Comprehensive health index HHV: Quantitative overall health score of the robot.

[0093] Detailed scores of each dimension: Specific scores of the 17 health dimensions, revealing the performance of the robot in each aspect.

[0094] Health status rating: Grade the health status of the robot according to HHV and dimension scores, such as excellent, good, average, attention needed, failure, etc.

[0095] Abnormal dimension identification: Clearly list the health dimensions with scores below the preset threshold, indicating potential failure points or performance bottlenecks.

[0096] Maintenance recommendations: Provide specific maintenance or calibration recommendations for detected abnormalities or potential risks.

[0097] The dynamic authentication generator is responsible for generating a unique 17-bit identity code for the humanoid robot, which is combined through specific rules to ensure its uniqueness, resolvability and information carrying capacity. The specific generation method is:

[0098] Authority stamp: Fixed value ZK, occupying the first 2 bits of the code.

[0099] Scene key: Encode according to the application scenario of the humanoid robot, occupying 1 bit. For example, H represents the industrial handling scenario, C represents the service robot scenario, and M represents the medical robot scenario.

[0100] Enterprise chain: Consists of 4 bits, combining enterprise ID and robot model ID, represented by 62 binary compression code. This field uniquely identifies the manufacturer and specific model of the robot.

[0101] HHV core: consists of 3 bits, representing the comprehensive health value HHV. The numerical value of HHV is mapped to a 3-bit coding space through a nonlinear mapping. For example, the HHV value can be converted to the code through the following nonlinear mapping function:

[0102]

[0103] In the formula, HHV encoded represents the encoded comprehensive health value; the value range of HHV is [0, 10], represents the floor. This mapping ensures the range and accuracy of the representation of HHV in the code.

[0104] Time anchor: consists of 2 bits, representing the century epoch code, using a continuous serial number system. For example, 2023 is mapped to "A0", 2050 is mapped to "C7", and 2 is mapped to "M9". This field provides timestamp information for the generation of the health report.

[0105] Data key: consists of 5 bits, representing the dimension index. This field is generated by taking the first 5 bits of the hash value of the scores of all 17 health dimensions. This field can be used as a unique identifier or index for the health report data, supporting subsequent data retrieval and verification.

[0106] When the score of the drive chain efficiency DCE or the thermodynamic stability TSD is lower than the preset threshold (for example, 5.0), the dynamic authentication generator will trigger a humanoid robot predictive maintenance warning. According to the warning information, the system automatically generates a maintenance work order. The maintenance work order contains specific maintenance instructions, such as indicating the replacement of a specific component, system calibration, or suggesting in-depth diagnosis.

[0107] For S5, the generated health authentication report and the identity code associated with the humanoid robot are stored in the blockchain: the storage process is completed by the blockchain evidence module 500, ensuring the authenticity, non-tamperability and traceability of the authentication result

[0108] Please refer to the attached Figure 2 -attached Figure 4 This part will further elaborate the execution details and technical effects of the method of the present application through specific examples.

[0109] Example 1:

[0110] This example describes the authentication process of the drive chain efficiency of the humanoid robot during operation.

[0111] The multi-source data acquisition module acquires the operation data of the humanoid robot joints in real time, and the operation data of the humanoid robot joints includes:

[0112] Joint current signal: acquired through current sensors within joint motor drivers;

[0113] Joint angular velocity: acquired through joint motor encoders.

[0114] The dimension calculation engine estimates the actual torque of the joint based on the collected joint current signal and joint angular velocity, combined with pre-set motor parameters. By calculating motor parameters, current signals, and angular velocity signals, it replaces physical torque sensors and achieves hardware reuse.

[0115] The dimension calculation engine calculates the drive chain efficiency DCE based on the estimated torque and the target torque output by the control system. By quantifying the squared sum of the deviation between the actual torque and the expected torque and performing normalization processing, it evaluates the actual working efficiency of the drive chain. The higher the drive chain efficiency, the smaller the loss. When the DCE score is below the pre-set threshold, the dynamic authentication generator will trigger a predictive maintenance warning.

[0116] Example 2:

[0117] This example describes the generation process of the environmental chemical index of a humanoid robot in a specific environment.

[0118] The multi-source data acquisition module obtains environment-related information from multiple data sources:

[0119] Vision system: acquires smoke concentration and dust visibility indicators through RGB-D cameras, such as analyzing color and texture changes in specific pixel regions through image processing algorithms;

[0120] Network interface: obtains the regional air quality index AQI through external network interfaces;

[0121] Task log: analyzes the chemical exposure history recorded in the robot's historical task log, such as recording the type and duration of chemical exposure.

[0122] The dimension calculation engine calculates the risk factors of the environmental chemical index based on the collected multi-source data, and quantifies the chemical risk level of the environment where the robot is located by integrating environmental data from different sources.

[0123] The dimension calculation engine generates the environmental chemical index ECI based on the calculated chemical risk factors and the robot's recognition delay time for environmental changes. It considers the size of the chemical risk factor and the impact of the robot's response speed on health. The higher the ECI score, the better the robot's health status in that chemical environment.

[0124] Example 3:

[0125] This embodiment describes the process of fusing 17 health dimensions to generate the humanoid robot comprehensive health index HHV. The process is completed by the correlation analysis module and the dynamic authentication generator.

[0126] The core dimension weight is dynamically adjusted according to the application scenario of the humanoid robot. The scores of the six core health dimensions (QBE, JHE, DCE, NSR, NEX, and ECI) are geometrically averaged and square rooted, that is, a low score of any core dimension will significantly lower the overall health assessment. The geometric mean G of the core dimensions is calculated as follows:

[0127]

[0128] In the formula, D i is the core dimension score; w i is the corresponding weight.

[0129] The scores of the 11 auxiliary health dimensions (VPE, AMS, EMI, TSD, BHI, ECS, EIF, LTR, GCD, CQI, and GEE) are arithmetically averaged to reflect the overall performance of the secondary performance indicators of the robot. The arithmetic mean A of the auxiliary dimensions is calculated as follows:

[0130]

[0131] In the formula, D j is the auxiliary dimension score.

[0132] The health attenuation factor F is used to quantify the penalty effect of the health condition with a serious abnormal dimension. When there is a dimension score below 5.0, the factor will be less than 1, thereby reducing the final HHV. The calculation formula of the health attenuation factor F is as follows:

[0133]

[0134] In the formula, count(D k <5.0) represents the number of dimensions with scores below 5.0 among the 17 dimensions.

[0135] The core dimension geometric mean, the auxiliary dimension arithmetic mean, and the health attenuation factor are multiplied to obtain the final comprehensive health index. The calculation formula of the final comprehensive health index is as follows.

[0136] HHV = G × A × F;

[0137] The finally calculated HHV will be used as a quantitative indicator of the current health status of the humanoid robot and by the dynamic authentication generator to generate a health authentication report.

[0138] Please refer to the attached Figure 3The robot health state authentication system based on multi-dimensional fusion of the application comprises:

[0139] A multi-source data acquisition module is responsible for collecting various raw data of the humanoid robot during operation, including an inertial measurement unit (IMU), a joint current sensor, a joint encoder, an RGB-D camera, a microphone array, a battery management system (BMS) interface, a controller log interface, a network communication module, and an external network data interface.

[0140] Each sensor and interface transmits the collected raw data (such as three-axis acceleration, angular velocity, joint current, joint angular velocity, image data, audio data, battery parameters, system logs, network communication parameters, and external environment data) to the module. The module performs preliminary processing on the received data, including data format unification, unit conversion, data denoising, outlier detection, and missing value filling, to ensure data quality and consistency, providing clean and standardized input for subsequent calculations.

[0141] A dimension calculation engine estimates 17 health dimensions of the humanoid robot based on the preprocessed data provided by the multi-source data acquisition module through various algorithm models, including multiple algorithm submodules such as a torque estimation model, a chemical risk inversion model, and a thermodynamic equation temperature prediction model.

[0142] The module receives standardized multi-source data, calculates and analyzes specific data according to preset algorithm models, and outputs quantified scores of 17 health dimensions. For example, using joint current and angular velocity data, the torque estimation model calculates the drive chain efficiency (DCE) score; using visual data, network air quality data, and task log data, the chemical risk inversion model calculates the environmental chemical index (ECI) score. The module realizes hardware reuse and reduces dependence on expensive physical sensors.

[0143] A correlation analysis module performs preliminary correlation analysis and fusion on the 17 health dimension scores output by the dimension calculation engine, including a data storage unit and a fusion algorithm execution unit.

[0144] The module receives 17 health dimension scores and performs weighted processing on core dimensions according to preset weight distribution rules. The module provides intermediate results for the calculation of the comprehensive health index (HHV), such as calculating the geometric mean of core dimensions and the arithmetic mean of auxiliary dimensions.

[0145] A dynamic authentication generator generates the comprehensive health index (HHV) of the humanoid robot based on the results of the correlation analysis module, authenticates its current health state, generates a unique identity code and health authentication report, and is responsible for triggering predictive maintenance warnings and generating maintenance work orders, including an HHV calculation unit, an authentication logic unit, an identity code generation unit, a report generation unit, and a warning management unit.

[0146] According to the preset HHV calculation formula, the core dimension geometric mean, the auxiliary dimension arithmetic mean and the health attenuation factor are fused to obtain the final HHV. Based on the HHV and the specific scores of each dimension, the health state level of the robot is evaluated. At the same time, the unique identity code of the robot is generated according to the 17-bit coding rule. When the score of a specific dimension (such as DCE, TSD) is lower than the preset threshold, the module will trigger the early warning mechanism and automatically generate a work order containing specific maintenance suggestions.

[0147] The blockchain storage module is responsible for storing the generated health authentication report and the unique identity code of the humanoid robot into the blockchain, including the blockchain interface, data encryption unit and transaction chaining unit.

[0148] The health authentication report and the identity code received by the dynamic authentication generator are hashed and encrypted, and then uploaded to the specified blockchain as transaction records through the blockchain network interface. By utilizing the decentralization and tamper-proofing features of the blockchain, the authenticity, transparency and traceability of the health authentication data are ensured.

[0149] The application interface layer provides an interface for external systems to interact with the health state authentication system, supporting a variety of application scenarios, including API interface, data transmission protocol and identity verification mechanism.

[0150] The standardized application programming interface API is provided to authorized third-party applications (such as robot predictive maintenance platform, second-hand trading platform, asset evaluation agency, etc.) to query the health authentication report and identity code of the robot. This module ensures secure data transmission and access control, supporting health state-based maintenance scheduling, asset value assessment and transaction process optimization.

[0151] Although embodiments of the present application have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made therein without departing from the principles and spirit of the application, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A robot health state authentication method based on multi-dimensional fusion, characterized in that, The method comprises the following steps: S1, collecting multi-source data of the humanoid robot during operation, including inertial measurement unit data, joint current signal, joint angular velocity, RGB-D camera data, microphone array data, battery management system data, controller log, network communication data and external network air quality data, and standardizing the collected data and performing preliminary preprocessing; S2, hardware multiplexing based on the multi-source data algorithm model of the humanoid robot, perceiving and estimating 17 health dimensions of the humanoid robot; S3, fusing the 17 health dimensions of the humanoid robot perceived and estimated to generate a comprehensive health index of the humanoid robot; S4, authenticating the current health status of the humanoid robot according to the comprehensive health index and the specific scores of the 17 health dimensions of the humanoid robot, and generating a health authentication report; S5, storing the generated health authentication report and the identity code associated with the humanoid robot to the blockchain. 2.The multi-dimension fusion based robot health state authentication method of claim 1, wherein, In step S2, the 17 health dimensions of the humanoid robot include quantum behavior entropy, joint harmonic entropy, drive chain efficiency, visual perception entropy, sound field modeling strength, neural synchronization rate, energy management index, thermodynamic stability, battery health, neural efficiency, emotional service ability, environmental interaction, learning transfer rate, global collaboration degree, communication quality index, gait energy entropy and environmental chemical index. 3.The multi-dimension fusion based robot health state authentication method of claim 1, wherein, In step S2, the multi-source data algorithm model of the humanoid robot comprises: A torque estimation model based on current and speed signals is used to estimate the actual torque of the joints of the humanoid robot, so as to calculate the drive chain efficiency, and the formula of the torque estimation model is: where k t is a motor constant, I is a joint current signal, J is a moment of inertia, is an angular acceleration, b is a damping coefficient, and ω represents an angular velocity; A chemical risk inversion model based on visual analysis, network air quality data and historical exposure records is used to calculate the risk factor of the environmental chemical index; A component temperature prediction model based on thermodynamic equations is used to replace physical thermal imaging devices to predict the temperature distribution or temperature rise of key components of the humanoid robot and evaluate the thermodynamic stability. 4.The multi-dimension fusion based robot health state authentication method of claim 3, wherein, The formula of the chemical risk inversion model based on visual analysis, network air quality data and historical exposure records is: risk_factor=0.4S+0.4A / 50+0.2C; In the formula, risk_factor is the chemical risk factor; S is the smoke concentration and dust visibility index obtained by the vision system; A is the network air quality index; C is the historical chemical contact record related index. 5.The multi-dimension fusion based robot health state authentication method of claim 1, wherein, In step S3, the comprehensive health index is obtained by using a multi-dimensional fusion algorithm, and the calculation formula of the multi-dimensional fusion algorithm is: where D i is the core health dimension, w i is the core dimension weight, is the core dimension geometric mean, () 0.5 is the square root transformation, D j is the auxiliary health dimension, is the auxiliary dimension arithmetic mean, count(D k <5.0) is the number of dimensions with severe health abnormalities, is the health decay factor. 6.The multi-dimension fusion based robot health state authentication method of claim 4, wherein, The core dimension weight is dynamically adjusted according to the application scene of the humanoid robot, including: When the humanoid robot is applied to an industrial carrying scene, the weight of the drive chain efficiency increases; When applied to a service robot scene, the weight of neural efficiency increases; When applied to a medical robot scene, the weight of the environmental chemical index increases. 7.The multi-dimension fusion based robot health state authentication method of claim 2, wherein, The generation step of the environmental chemical index comprises: Collecting the smoke concentration and dust visibility data obtained by the vision system; Obtaining the regional air quality index provided by the external network interface; Analyzing the chemical contact history data recorded in the robot task log; calculating a chemical risk factor based on the collected data, acquired data, and analyzed data; generating an environmental chemical index based on the risk factor and the identified delay. 8.The multi-dimension fusion based robot health state authentication method of claim 1, wherein, When the score of the driving chain efficiency or thermodynamic stability is lower than a preset threshold, a humanoid robot predictive maintenance warning is triggered, and a maintenance work order is automatically generated according to the warning information, which can indicate component replacement or system calibration.

9. The robot health state authentication system based on multi-dimensional fusion, applied to the robot health state authentication method based on multi-dimensional fusion according to any one of claims 1-7, characterized in that, Comprise: A multi-source data acquisition module for collecting and preprocessing multi-source data of the humanoid robot during operation; A dimension calculation engine including parallel calculation functions of 17 authentication dimensions, for perceiving and estimating 17 health dimensions of the humanoid robot based on a multi-source data algorithm model; A correlation analysis module for cross-dimension analysis based on a dynamic correlation coefficient matrix to analyze the mutual influence and potential correlation between different health dimensions; A dynamic authentication generator for authenticating the current health status of the humanoid robot and generating a unique identity code and health authentication report of the humanoid robot; A blockchain storage module for storing the unique identity code and health authentication report of the humanoid robot to the blockchain storage; An application interface layer for providing an interface for data interaction with external systems and users.

Images