Humanoid robot joint life monitoring method

Abstract

The invention provides a humanoid robot joint life monitoring method, and belongs to the technical field of life monitoring. The method comprises the steps that 1, a multi-mode sensor array is constructed to collect original joint motion data of a target robot; 2, standardizing the original joint movement data according to feature types to obtain standard feature data of all parameter types; 3, establishing a joint degradation model of the target robot based on the standard feature data; 4, inputting real-time motion data of the target robot into the joint degradation model, and outputting a joint residual life index; and 5, when the residual life index of the joint is lower than a preset threshold value, a self-adaptive maintenance strategy is triggered. Through real-time monitoring of joint health, early warning and optimization maintenance, the reliability of the robot is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of life monitoring, and in particular to a method for monitoring the life of joints of a humanoid robot. Background Art

[0002] With the widespread application of humanoid robots in industry, healthcare, and service sectors, the lifespan and health of robot joints, as one of their core components, directly impact the performance and reliability of robotic systems. Due to the high-frequency motion, loads, and environmental factors that robot joints experience during long-term operation, they are susceptible to wear and aging. However, monitoring the lifespan of humanoid robot joints typically relies solely on manual inspections after exercise, which prevents real-time analysis. This results in delayed joint lifespan analysis and potential errors.

[0003] Therefore, the present invention proposes a method for monitoring the joint life of a humanoid robot. Summary of the Invention

[0004] The present invention provides a method for monitoring the lifespan of humanoid robot joints, which is used to collect robot joint motion data through multimodal sensors, establish a degradation model after standardization, evaluate the remaining lifespan of joints in real time, trigger adaptive maintenance strategies, optimize robot maintenance, and improve reliability and safety.

[0005] In one aspect, the present invention provides a method for monitoring the lifespan of a humanoid robot's joints, comprising: Step 1: Construct a multimodal sensor array to collect the original joint motion data of the target robot; Step 2: Standardize the original joint motion data according to the feature type to obtain the standard feature data of all parameter types; Step 3: Establishing a joint degradation model of the target robot based on standard feature data; Step 4: Input the real-time motion data of the target robot into the joint degradation model and output the joint remaining life index; Step 5: When the remaining life index of the joint is lower than the preset threshold, the adaptive maintenance strategy is triggered.

[0006] On the other hand, a multimodal sensor array is constructed, including: Determine the target robot to be monitored and match the corresponding sensor type according to the preset acquisition parameter type; Obtain the initial structural parameter data of the target robot and divide the initial structure into three layers, including: surface layer, middle layer, and core layer; A three-layer embedded layout is used to mark the possible installation locations of sensors. A matching degree analysis is performed between any sensor and any possible installation location to obtain the matching coefficient. The matching coefficient is combined with the layer weight coefficient to generate a matching value, and for any sensor, the possible sensor installation position with the highest matching value is selected as the optimal sensor installation position; The optimal sensor installation positions are input into the three-layer embedded layout to generate a multimodal sensor array.

[0007] On the other hand, the original joint motion data of the target robot is collected, including: Starting the multimodal sensor array, and configuring a unique first identifier for a joint of the target robot according to the multimodal sensor array; Constructing a data acquisition module to synchronously acquire raw joint motion signal data of the joint from the first identified joint based on a preset acquisition frequency; The original joint motion signal data is combined with the demodulator for signal processing to obtain the original joint motion data of all joints of the target robot.

[0008] On the other hand, the original joint motion data is standardized according to the feature type to obtain standard feature data of all parameter types, including: Determine the feature category of the original joint motion data of the parameter type of any joint according to the parameter type-feature category mapping table; Screening the original joint motion data of the parameter type for missing values ​​and outliers, and performing standard replacement to obtain first joint motion data; The first joint motion data is normalized in combination with the feature categories.

[0009] On the other hand, the first joint motion data is normalized in combination with the feature category as follows: ; in, represents the normalized first joint motion data, Represents the feature category matrix function of the i-th time interval, Indicates the time window size, t indicates the initial time of time window cutting, n indicates that the time window is divided into n time intervals after the first joint motion data is cut. Represents the joint motion data within the current time window, represents the harmonic function, represents the imaginary unit, represents the time-integrated variable, represents the frequency coefficient, Represents the feature category matrix function of the preset initial time interval, represents the time splicing function; The standardized first joint motion data is the standard feature data of the joint, and the standard feature data is obtained by processing all joints of the target robot.

[0010] On the other hand, a joint degradation model of the target robot is established based on the standard feature data, including: According to the three-layer embedded layout, the original state space model of the target robot is defined, and the standard feature data of each joint is input to generate the joint state space model of the target robot: ; in, Represents standard feature data, Indicates the input control quantity, Represents a preset time-varying parameter, determined by the physical characteristics of the joint, represents the joint state variable, represents the joint state space model, ( ) represents a differential equation based on the laws of physics.

[0011] On the other hand, based on the law of mechanical wear, the degradation coefficient of the added joint is: ; in, represents the degradation coefficient, represents the friction function, Indicates the preset fatigue coefficient, represents the first material constant, represents the second material constant, Indicates the maximum lifespan of the product. The degradation coefficient of any joint is input into the joint state space model of the joint to generate a joint degradation model of the joint.

[0012] On the other hand, the real-time motion data of the target robot is input into the joint degradation model, and the joint remaining life index is output, including: docking the joint degeneration model to the corresponding joint in the multimodal sensor array according to the first identifier of any joint, and collecting real-time motion data of the joint in real time; The remaining lifespan index is calculated based on the joint degeneration model of the joint, specifically: ; in, represents the remaining life index, Indicates the The dynamic weight function of each time node, Indicates the Real-time motion data of the j1th parameter type at the time node, represents the degradation coefficient of the joint, represents the time decay function, Indicates the The time attenuation coefficient of the U-th real-time motion data of the j1-th parameter type at the time node, where U represents the number of working cycles of the joint.

[0013] On the other hand, a work recording chip is embedded in the joints of the target robot to cumulatively record the number of work cycles of the joints.

[0014] On the other hand, when the remaining life index of the joint falls below the preset threshold, the adaptive maintenance strategy is triggered, including: When the remaining life index of any joint is lower than the first threshold and greater than the second threshold, the motion planning optimization is triggered; when the remaining life index of the first joint is lower than the second threshold, the joint is forced to enter the maintenance state and the system is activated to upload the fault code; The motion planning optimization limits the peak torque of the joint based on the real-time motion data of the joint, and redistributes the joint load according to the Jacobian matrix transpose method to generate a dynamic planning result of the joint; An adaptive maintenance strategy is developed based on the log records of all joints of the target robot, and maintenance personnel are notified to check the faulty joints.

[0015] Compared with the prior art, the present invention has the following beneficial effects: The present invention provides a method for monitoring the lifespan of humanoid robot joints, which is used to collect robot joint motion data through multimodal sensors, establish a degradation model after standardization, evaluate the remaining lifespan of joints in real time, trigger adaptive maintenance strategies, optimize robot maintenance, and improve reliability and safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction is given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0017] Figure 1 The figure is a flow chart of a method for monitoring the life span of a humanoid robot joint provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0018] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0019] Example 1: like Figure 1 As shown, an embodiment of the present invention provides a method for monitoring the life of a humanoid robot joint, comprising: Step 1: Construct a multimodal sensor array to collect the original joint motion data of the target robot; Step 2: Standardize the original joint motion data according to the feature type to obtain the standard feature data of all parameter types; Step 3: Establishing a joint degradation model of the target robot based on standard feature data; Step 4: Input the real-time motion data of the target robot into the joint degradation model and output the joint remaining life index; Step 5: When the remaining life index of the joint is lower than the preset threshold, the adaptive maintenance strategy is triggered.

[0020] In this embodiment, the multimodal sensor array is an array composed of multiple types of sensors, which is used to collect different types of data simultaneously to more comprehensively monitor the status of the target object.

[0021] In this embodiment, the target robot refers to the robot that is monitored, evaluated, and maintained in this technical solution.

[0022] In this embodiment, the raw joint motion data refers to the raw data related to joint motion collected in real time by the multimodal sensor array during the operation of the target robot, such as joint position, joint temperature, joint torque, joint vibration, etc.

[0023] In this embodiment, the feature type refers to various types of data information extracted from the original joint motion data, which is used to describe and analyze the motion state and health status of the target robot joint, including: time domain features, frequency domain features, vibration features, etc.

[0024] In this embodiment, standardization is used to adjust the range and distribution of data to make it conform to certain standards.

[0025] In this embodiment, the parameter type refers to different data variable types related to the robot joint motion, such as joint position, joint temperature, joint torque, joint vibration, etc.

[0026] In this embodiment, the standard feature data refers to various data features obtained through standardization processing and used to describe the joint motion state of the target robot.

[0027] In this embodiment, the joint degeneration model is a mathematical calculation model used to predict the lifespan and health status of each joint in the robot. Based on the motion data and environmental data of the robot joints, the remaining service life of the joints is estimated by analyzing factors such as joint wear, fatigue, and load.

[0028] In this embodiment, real-time motion data refers to data related to the motion of each joint of the robot collected in real time during the operation of the target robot, including dynamic parameters such as the robot's position, speed, acceleration, load, and temperature.

[0029] In this embodiment, the joint remaining life index is an indicator used to represent the current expected life of the robot joint.

[0030] In this embodiment, the preset threshold is a value set according to the robot's usage environment, joint characteristics and working conditions. When the joint remaining life index drops below the threshold, the system will automatically trigger the predetermined maintenance measures.

[0031] In this embodiment, the adaptive maintenance strategy is a maintenance strategy that is dynamically adjusted based on the target robot joint remaining life index and actual operating data, ensuring that the robot is properly maintained before a failure is about to occur, thereby avoiding sudden failures or downtime and improving the robot's reliability and work efficiency.

[0032] The working principle and beneficial effects of the above technical solution are as follows: multimodal sensors collect and standardize joint motion data in real time, establish degradation models, and evaluate joint lifespan. Through continuous monitoring and adaptive maintenance, the robot's operational reliability is improved, failures are avoided, and service life is extended.

[0033] Example 2: Based on the above embodiment 1, a multimodal sensor array is constructed, including: Determine the target robot to be monitored and match the corresponding sensor type according to the preset acquisition parameter type; Obtain the initial structural parameter data of the target robot and divide the initial structure into three layers, including: surface layer, middle layer, and core layer; A three-layer embedded layout is used to mark the possible installation locations of sensors. A matching degree analysis is performed between any sensor and any possible installation location to obtain the matching coefficient. The matching coefficient is combined with the layer weight coefficient to generate a matching value, and for any sensor, the possible sensor installation position with the highest matching value is selected as the optimal sensor installation position; The optimal sensor installation positions are input into the three-layer embedded layout to generate a multimodal sensor array.

[0034] In this embodiment, the preset acquisition parameter types refer to various data parameters that need to be monitored and collected during the operation of the robot, including: joint position, joint temperature, joint torque, joint vibration, etc.

[0035] In this embodiment, the sensor is a device that can sense environmental changes or object states and convert them into measurable signals, and monitor changes in the measured object or environment in real time through physical, chemical or biological methods, including: position, speed, acceleration, load, temperature and other types.

[0036] In this embodiment, the initial structural parameter data refers to various basic parameters of the target robot in the initial design or construction stage, including but not limited to the robot's size, weight, shape, number of joints, range of motion, load capacity, power requirements, etc.

[0037] In this embodiment, the surface layer refers to the outer area of ​​the robot structure, the part that directly contacts the external environment, including the shell, surface or external parts of the robot.

[0038] In this embodiment, the middle layer refers to the area between the surface layer and the core layer, which plays an intermediary role between the internal system of the robot and the external environment.

[0039] In this embodiment, the core layer refers to the innermost part, which carries the key computing, control and data processing functions of the robot system.

[0040] In this embodiment, the three-layer embedded layout refers to dividing the structure of the robot system into three main layers, and performing layout design at these layers based on functions, performance requirements, and sensor installation requirements.

[0041] In this embodiment, the possible sensor installation position refers to a specific position in the robot structure where the sensor can be installed.

[0042] In this embodiment, the matching analysis refers to evaluating and comparing the selected sensor and the sensor installation location to determine the matching degree between them.

[0043] In this embodiment, the matching coefficient is a first numerical indicator used to measure the degree of adaptation between the sensor and the installation location. For example, the peripheral distance between the sensor and the possible installation location of the sensor is calculated using the Pearson correlation coefficient, and the matching coefficient between the sensor and the possible installation location of the sensor is obtained by superposition.

[0044] In this embodiment, the layer weight coefficient refers to the weight coefficient set in advance for each layer, for example, the surface weight is 0.5, the middle weight is 0.3, and the core weight is 0.2.

[0045] In this embodiment, the matching value is a value that quantifies the sensor and the sensor position. Assuming that the surface layer weights are 0.5 for the surface layer, 0.3 for the middle layer, and 0.2 for the core layer, and the matching coefficient is 0.7, the surface layer matching value = 0.7×0.5 = 0.35.

[0046] In this embodiment, the optimal sensor installation position is the final position calculated according to the following steps, which is intended to ensure that the sensor can perform its function in the best manner on the target robot.

[0047] The working principle and beneficial effects of this technical solution are as follows: Through a three-layer embedded layout and matching analysis, the sensor installation position is optimized, improving the compatibility of the sensors with the target robot structure. By selecting the optimal installation location, the accuracy and monitoring effect of the sensor array are improved, thereby enhancing the performance and reliability of the robot.

[0048] Example 3: Based on the above embodiment 1, the original joint motion data of the target robot is collected, including: Starting the multimodal sensor array, and configuring a unique first identifier for a joint of the target robot according to the multimodal sensor array; Constructing a data acquisition module to synchronously acquire raw joint motion signal data of the joint from the first identified joint based on a preset acquisition frequency; The original joint motion signal data is combined with the demodulator for signal processing to obtain the original joint motion data of all joints of the target robot.

[0049] In this embodiment, the first identifier refers to an identifier used to uniquely identify a target robot joint.

[0050] In this embodiment, the data acquisition module is a system component for collecting raw motion signal data from the joints of the target robot and transmitting the data to a subsequent processing unit for analysis and processing.

[0051] In this embodiment, the preset acquisition frequency refers to a fixed time interval set by the system for the acquisition module during the data acquisition process.

[0052] In this embodiment, the raw joint motion signal data refers to the raw data directly collected on the robot joint through the sensor array without any signal processing or demodulation.

[0053] In this embodiment, the demodulator is a device used to convert the modulated signal back to the original signal.

[0054] In this embodiment, signal processing refers to the process of performing operations such as analysis, conversion, enhancement, and denoising on the collected original signal to extract useful information or optimize signal quality.

[0055] The working principle and beneficial effects of the above technical solution are as follows: by activating a multimodal sensor array, uniquely identifying the target robot joint, synchronously collecting joint motion data in real time, and extracting effective information through signal processing, the accuracy of data collection is improved, ensuring accurate monitoring and analysis of joint status.

[0056] Example 4: Based on the above embodiment 2, the original joint motion data is standardized according to the feature type to obtain standard feature data of all parameter types, including: Determine the feature category of the original joint motion data of the parameter type of any joint according to the parameter type-feature category mapping table; Screening the original joint motion data of the parameter type for missing values ​​and outliers, and performing standard replacement to obtain first joint motion data; Normalize the first joint motion data combined with the feature category: ; in, represents the normalized first joint motion data, Represents the feature category matrix function of the i-th time interval, Indicates the time window size, t indicates the initial time of time window cutting, n indicates that the time window is divided into n time intervals after the first joint motion data is cut. Represents the joint motion data within the current time window, represents the harmonic function, represents the imaginary unit, represents the time-integrated variable, represents the frequency coefficient, Represents the feature category matrix function of the preset initial time interval, represents the time splicing function; The standardized first joint motion data is the standard feature data of the joint, and the standard feature data is obtained by processing all joints of the target robot.

[0057] In this embodiment, the parameter type-feature category mapping table is a table used to associate different types of sensor data with their corresponding feature categories.

[0058] In this embodiment, the feature category is the type of feature extracted from the original joint motion data, such as mean, standard deviation, maximum value, etc.

[0059] In this embodiment, missing values ​​refer to the situation in which certain data points in a data set are not recorded or cannot be obtained, and thus the data position has no value.

[0060] In this embodiment, outliers refer to data points in a data set that deviate significantly from other data points, which may be caused by erroneous measurements or sensor failures.

[0061] In this embodiment, standard replacement refers to replacing missing or abnormal data points with reasonable values ​​using a mean replacement method when processing missing values ​​or abnormal values.

[0062] In this embodiment, the first joint motion data is data obtained after standard replacement.

[0063] In this embodiment, the feature category matrix function is a matrix used to describe the feature performance of joint motion data in different time intervals. The matrix is ​​a multidimensional array used to represent the feature value at each moment in the time window.

[0064] In this embodiment, the time window refers to the concept of segmenting data, which cuts the original joint motion data into several small intervals in the time dimension to facilitate more fine-grained analysis or processing of the data.

[0065] In this embodiment, the harmonic function refers to a mathematical function based on sine and cosine functions, which is used to describe a phenomenon of periodic change and to analyze and represent periodic components in a signal or data.

[0066] In this embodiment, the time splicing function is a function for merging or splicing data of multiple time intervals into a continuous time series.

[0067] In this embodiment, ; in, Frequency domain feature extraction, followed by weighted fusion and normalization, converts the original joint signals into more robust standardized data, which is suitable for high-precision motion control or fault diagnosis scenarios.

[0068] The working principle and beneficial effects of the above technical solution are as follows: by filtering, standardizing, and processing joint motion data, and optimizing it in combination with feature categories, standard feature data is obtained. This ensures data accuracy and consistency, improves the analysis accuracy of the target robot's joint motion, and provides a reliable data foundation for subsequent analysis.

[0069] Example 5: On the basis of the above-mentioned embodiment 1, a joint degradation model of the target robot is established based on standard feature data, including: According to the three-layer embedded layout, the original state space model of the target robot is defined, and the standard feature data of each joint is input to generate the joint state space model of the target robot: ; in, Represents standard feature data, Indicates the input control quantity, Represents a preset time-varying parameter, determined by the physical characteristics of the joint, represents the joint state variable, represents the joint state space model, ( ) represents a differential equation based on physical laws; Based on the law of mechanical wear, the degradation coefficient of the added joint is: ; in, represents the degradation coefficient, represents the friction function, Indicates the preset fatigue coefficient, represents the first material constant, represents the second material constant, Indicates the maximum lifespan of the product. The degradation coefficient of any joint is input into the joint state space model of the joint to generate a joint degradation model of the joint.

[0070] In this embodiment, the original state space model is a framework of a mathematical model that originally describes the states of the target robot joints and how they change over time.

[0071] In this embodiment, the joint state space model is a mathematical model used to describe the dynamic behavior of each joint of the target robot.

[0072] In this embodiment, the input control variable refers to an external signal or instruction that controls the action of the robot.

[0073] In this embodiment, the preset time-varying parameters refer to parameters that change with time, and these parameters are usually determined by factors such as the physical characteristics of the robot joints, environmental factors, and control strategies.

[0074] In this embodiment, ; Among them, the differential equations are based on the physical laws of the joint (such as dynamics, thermodynamics), ensuring that the model matches the actual system and the time-varying parameters The model can dynamically respond to environmental changes (such as load mutation, mechanical wear), state equation For controller design, if a new sensor is added, only the expansion No need to reconstruct the kinetic model.

[0075] In this embodiment, the mechanical wear law is a law that describes the degradation of component performance in a mechanical system due to friction, wear, fatigue, and other factors.

[0076] In this embodiment, the degradation coefficient is a measure that describes the gradual decline in performance of the robot joint due to factors such as friction, wear, and fatigue during use.

[0077] In this embodiment, the preset fatigue coefficient is a parameter used to describe the fatigue damage rate of the joint caused by cyclic stress of repeated loading and unloading during long-term use.

[0078] In this embodiment, the material constant is a basic parameter that describes the behavior of a material under specific physical conditions, and is related to properties such as elasticity, hardness, and fatigue strength.

[0079] In this embodiment, the maximum factory life refers to the longest usable life of a mechanical component (such as a joint, part, equipment, etc.) from the time of production to its expected stoppage or failure.

[0080] In this embodiment, ; Among them, fatigue damage Will change the surface properties of the material, causing friction ( ) nonlinear growth, the two accelerate degradation together, this formula only needs to monitor the fatigue accumulation and known , you can evaluate the health status in real time without complicated online testing.

[0081] The working principle and beneficial effects of this technical solution are as follows: By defining a joint state-space model for the target robot and combining joint degradation coefficients with the laws of mechanical wear, a joint degradation model is generated. This model can accurately predict joint degradation processes, providing data support for robot maintenance and performance optimization, and improving system reliability and lifespan.

[0082] Example 6: Based on the above embodiment 1, the real-time motion data of the target robot is input into the joint degradation model, and the joint remaining life index is output, including: docking the joint degeneration model to the corresponding joint in the multimodal sensor array according to the first identifier of any joint, and collecting real-time motion data of the joint in real time; The remaining lifespan index is calculated based on the joint degeneration model of the joint, specifically: ; in, represents the remaining life index, Indicates the The dynamic weight function of each time node, Indicates the Real-time motion data of the j1th parameter type at the time node, represents the degradation coefficient of the joint, represents the time decay function, Indicates the The time attenuation coefficient of the U-th real-time motion data of the j1-th parameter type at the time node, where U represents the number of working cycles of the joint.

[0083] In this embodiment, the dynamic weight function in the joint degeneration model is a function used to describe the importance or contribution of motion data at different time nodes, and its weight is adjusted according to factors such as the working condition, usage environment, and load conditions of the joint.

[0084] In this embodiment, the time decay function is used in the joint degeneration model to describe that the influence of certain factors (such as joint motion data or certain characteristics in the degeneration process) gradually weakens over time.

[0085] In this embodiment, ; This formula quantifies the complex degradation process into an intuitive remaining life index through multi-parameter time series fusion and dynamic attenuation weighting, and the dynamic weight can be adaptively adjusted according to the operating conditions.

[0086] The working principle and beneficial effects of this technical solution are as follows: By combining a joint degradation model with a multimodal sensor array, real-time joint motion data is collected and the remaining life index is calculated based on the degradation model. This accurately predicts the remaining life of the joint, assisting in maintenance decisions, extending equipment life, and improving the reliability and efficiency of the robotic system.

[0087] Example 7: On the basis of the above-mentioned embodiment 6, a work recording chip is embedded in the joint of the target robot to cumulatively record the number of work cycles of the joint.

[0088] In this embodiment, the work recording chip is an electronic component embedded in the robot's joints. Its primary function is to record and track the joint's work cycles. It accumulates real-time operational data for each joint, particularly key parameters such as the number of joint movements, load, and time.

[0089] The working principle and beneficial effects of this technical solution are as follows: By embedding a work-recording chip within the robot joint, the number of joint work cycles is accumulated in real time, enhancing the monitoring capability of joint usage status. This enables more accurate assessment of joint degradation, improving prediction accuracy, optimizing maintenance strategies, and extending the service life of the equipment.

[0090] Example 8: Based on the above embodiment 1, when the joint remaining life index is lower than the preset threshold, the adaptive maintenance strategy is triggered, including: When the remaining life index of any joint is lower than the first threshold and greater than the second threshold, the motion planning optimization is triggered; when the remaining life index of the first joint is lower than the second threshold, the joint is forced to enter the maintenance state and the system is activated to upload the fault code; The motion planning optimization limits the peak torque of the joint based on the real-time motion data of the joint, and redistributes the joint load according to the Jacobian matrix transpose method to generate a dynamic planning result of the joint; An adaptive maintenance strategy is developed based on the log records of all joints of the target robot, and maintenance personnel are notified to check the faulty joints.

[0091] In this embodiment, the first threshold is a set value, which is used to indicate that when the remaining life index of the joint is lower than this value, the system will trigger motion planning optimization.

[0092] In this embodiment, the second threshold refers to a lower value, which means that when the remaining life index of the joint drops below this value, the system will force the joint to enter the maintenance state and activate the mechanism of uploading the fault code.

[0093] In this embodiment, motion planning optimization refers to adjusting the motion process of joints in a robot or mechanical system through an optimization algorithm to extend their service life or improve work efficiency.

[0094] In this embodiment, the fault code refers to an identifier code automatically generated and recorded by the system when the remaining life index of a joint in the robot system is lower than a set threshold value, which is used to indicate the fault state of the joint.

[0095] In this embodiment, peak torque refers to the maximum torque that a joint or component can withstand during the movement of the robot joint. Torque is a physical quantity that represents the torsional effect generated by force on the rotating shaft, and is represented by the product of force and lever arm.

[0096] In this embodiment, the Jacobian matrix transpose method calculates and adjusts the movement of the robot joints based on the transpose of the Jacobian matrix. By utilizing the transpose of the Jacobian matrix, the solution process of the motion instructions is simplified, and the robot joints can move more smoothly and accurately.

[0097] In this embodiment, joint load refers to the force or torque that each joint of the robot bears when performing movement. The magnitude of the load depends on multiple factors, such as the force requirement of the end effector, the angle of the joint, the movement speed of the robot, and the influence of the external environment.

[0098] The working principle and beneficial effects of this technical solution are as follows: by real-time monitoring of the remaining life index of joints, motion planning optimization or maintenance is triggered to ensure safe operation of the robot. Motion planning optimization improves operational efficiency by limiting peak torque and redistributing loads. Furthermore, logging and adaptive maintenance strategies enhance the timeliness and accuracy of fault detection and maintenance.

[0099] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for monitoring the life of a humanoid robot joint, characterized in that: include: Step 1: Construct a multimodal sensor array to collect the original joint motion data of the target robot; Step 2: Standardize the original joint motion data according to the feature type to obtain the standard feature data of all parameter types; Step 3: Establishing a joint degradation model of the target robot based on standard feature data; Step 4: Input the real-time motion data of the target robot into the joint degradation model and output the joint remaining life index; Step 5: When the remaining life index of the joint is lower than the preset threshold, the adaptive maintenance strategy is triggered.

2. The method for monitoring the life of a humanoid robot joint according to claim 1, characterized in that: Build a multimodal sensor array, including: Determine the target robot to be monitored and match the corresponding sensor type according to the preset acquisition parameter type; Obtain the initial structural parameter data of the target robot and divide the initial structure into three layers, including: surface layer, middle layer, and core layer; A three-layer embedded layout is used to mark the possible installation locations of sensors. A matching degree analysis is performed between any sensor and any possible installation location to obtain the matching coefficient. The matching coefficient is combined with the layer weight coefficient to generate a matching value, and for any sensor, the possible sensor installation position with the highest matching value is selected as the optimal sensor installation position; The optimal sensor installation positions are input into the three-layer embedded layout to generate a multimodal sensor array.

3. The method for monitoring the life of a humanoid robot joint according to claim 1, characterized in that: Collect the target robot's raw joint motion data, including: Starting the multimodal sensor array, and configuring a unique first identifier for a joint of the target robot according to the multimodal sensor array; Constructing a data acquisition module to synchronously acquire raw joint motion signal data of the joint from the first identified joint based on a preset acquisition frequency; The original joint motion signal data is combined with the demodulator for signal processing to obtain the original joint motion data of all joints of the target robot.

4. The method for monitoring the life of a humanoid robot joint according to claim 2, characterized in that: The original joint motion data is standardized according to the feature type to obtain standard feature data of all parameter types, including: Determine the feature category of the original joint motion data of the parameter type of any joint according to the parameter type-feature category mapping table; Screening the original joint motion data of the parameter type for missing values ​​and outliers, and performing standard replacement to obtain first joint motion data; The first joint motion data is normalized in combination with the feature categories.

5. The method for monitoring the life of a humanoid robot joint according to claim 4, characterized in that: The first joint motion data is normalized in combination with the feature categories as follows: ； in, represents the normalized first joint motion data, Represents the feature category matrix function of the i-th time interval, Indicates the time window size, t indicates the initial time of time window cutting, n indicates that the time window is divided into n time intervals after the first joint motion data is cut. Represents the joint motion data within the current time window, represents the harmonic function, represents the imaginary unit, represents the time-integrated variable, represents the frequency coefficient, Represents the feature category matrix function of the preset initial time interval, represents the time splicing function; The standardized first joint motion data is the standard feature data of the joint, and the standard feature data is obtained by processing all joints of the target robot.

6. The method for monitoring the life span of a humanoid robot joint according to claim 1, characterized in that: Establishing a joint degradation model of the target robot based on standard feature data includes: According to the three-layer embedded layout, the original state space model of the target robot is defined, and the standard feature data of each joint is input to generate the joint state space model of the target robot: ； in, Represents standard feature data, Indicates the input control quantity, Represents a preset time-varying parameter, determined by the physical characteristics of the joint, represents the joint state variable, represents the joint state space model, ( ) represents a differential equation based on the laws of physics.

7. The method for monitoring the life span of a humanoid robot joint according to claim 1, characterized in that: Based on the law of mechanical wear, the degradation coefficient of the added joint is: ； in, represents the degradation coefficient, represents the friction function, Indicates the preset fatigue coefficient, represents the first material constant, represents the second material constant, Indicates the maximum lifespan of the product. The degradation coefficient of any joint is input into the joint state space model of the joint to generate a joint degradation model of the joint.

8. The method for monitoring the life span of a humanoid robot joint according to claim 1, characterized in that: The real-time motion data of the target robot is input into the joint degradation model, and the joint remaining life index is output, including: docking the joint degeneration model to the corresponding joint in the multimodal sensor array according to the first identifier of any joint, and collecting real-time motion data of the joint in real time; The remaining lifespan index is calculated based on the joint degeneration model of the joint, specifically: ； in, represents the remaining life index, Indicates the The dynamic weight function of each time node, Indicates the Real-time motion data of the j1th parameter type at the time node, represents the degradation coefficient of the joint, represents the time decay function, Indicates the The time attenuation coefficient of the U-th real-time motion data of the j1-th parameter type at the time node, where U represents the number of working cycles of the joint; Represents the total number of parameter types; q represents the total number of time points.

9. The method for monitoring the life span of a humanoid robot joint according to claim 6, characterized in that: A work recording chip is embedded in the joints of the target robot to cumulatively record the number of work cycles of the joints.

10. The method for monitoring the life of a humanoid robot joint according to claim 1, characterized in that: When the remaining life index of a joint falls below a preset threshold, an adaptive maintenance strategy is triggered, including: When the remaining life index of any joint is lower than the first threshold and greater than the second threshold, the motion planning optimization is triggered; when the remaining life index of the first joint is lower than the second threshold, the joint is forced to enter the maintenance state and the system is activated to upload the fault code; The motion planning optimization limits the peak torque of the joint based on the real-time motion data of the joint, and redistributes the joint load according to the Jacobian matrix transpose method to generate a dynamic planning result of the joint; An adaptive maintenance strategy is developed based on the log records of all joints of the target robot, and maintenance personnel are notified to check the faulty joints.

Images