Robot joint module health state monitoring device and method
Patent Information
- Application Number
- CN202410900225.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-05
- Publication Date
- 2024-11-12
Smart Images

Figure CN118927306A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of humanoid robots, and in particular to a device and method for monitoring the health status of a robot joint module. Background Art
[0002] With the rapid advancement of science and technology, humanoid robots, as the product of the intersection of artificial intelligence and robotics, are attracting widespread attention around the world. With their human-like appearance and behavior, humanoid robots not only show great potential in the field of industrial production, but also gradually play an important role in daily life. Robot joints are the core components of robots. The service performance of joints determines the performance of the entire robot. Robot joints are mainly composed of motors, reducers, sensors and controllers. The service process is inevitably accompanied by wear and performance degradation of mechanical parts. The most common practice is to monitor the current and voltage of the motor, but due to the use of large transmission ratio reducers, the damage of mechanical parts is extremely weak in the motor current. Only major faults will be more obvious, and abnormal conditions in the operation of robot joint modules cannot be identified in time. Summary of the invention
[0003] The purpose of this application is to provide a robot joint module health status monitoring device and method, which can identify abnormal conditions in the operation of the robot joint module, so as to promptly find the source of abnormal vibration, so as to avoid a series of subsequent losses caused by untimely fault discovery.
[0004] To achieve the above objectives, this application provides the following solutions:
[0005] In a first aspect, the present application provides a robot joint module health status monitoring device, comprising:
[0006] A vibration temperature sensor, disposed on a housing of the robot joint module, for collecting vibration signals and temperature signals of the robot joint in real time when the robot joint module operates in a health detection mode;
[0007] Robot joint module driver board, used to collect current signals, voltage signals and position signals of robot joints;
[0008] A microcontroller connected to the vibration temperature sensor and the robot joint module driving board, and used for storing and transmitting signal data of the robot joint; the signal data includes vibration signal, temperature signal, current signal, voltage signal and position signal;
[0009] A host computer is connected to the microprocessor and is used to monitor the health status of the robot joint module according to the signal data.
[0010] In a second aspect, the present application provides a method for monitoring the health status of a robot joint module, the method being applied to the above-mentioned device for monitoring the health status of a robot joint module, the method comprising:
[0011] Collecting signal data of robot joints in the robot joint module; the signal data includes vibration signals, temperature signals, current signals, voltage signals and position signals;
[0012] Preprocessing the signal data and extracting characteristic parameters of the signal data; the characteristic parameters include frequency domain power spectrum characteristics, time-frequency domain wavelet energy characteristics and time domain statistical parameters;
[0013] The characteristic parameters are input into the diagnostic model to determine the state of the robot joint module; the state of the robot joint module includes a healthy state and a fault state; the diagnostic model is obtained by training the machine learning model with the characteristic parameters of the historical signal data in the normal mode and the fault mode.
[0014] According to the specific embodiments provided in this application, this application discloses the following technical effects:
[0015] In order to reduce power consumption, this application does not connect external sensors during the normal service of the robot joint module. Only when the robot joint module is regularly running in health monitoring mode, an external vibration temperature sensor is connected. At the same time, the current signal, voltage signal, and position signal of the robot joint are monitored, and these signals are uploaded to the host computer through the communication interface. The host computer will further analyze the collected signals to determine the health status of the robot joint module, thereby identifying abnormal conditions in the operation of the robot joint module and promptly troubleshooting the source of abnormal vibration to avoid a series of subsequent losses due to untimely fault detection. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0017] Figure 1 A schematic diagram of the structure of a robot joint module provided in an embodiment of the present application and equipped with a robot joint module health status monitoring device;
[0018] Figure 2 A front view of a robot joint module provided with a robot joint module health status monitoring device provided in an embodiment of the present application;
[0019] Figure 3A model diagram of a robot joint module driving board provided in one embodiment of the present application;
[0020] Figure 4 A flowchart of a method for monitoring the health status of a robot joint module provided in one embodiment of the present application. DETAILED DESCRIPTION
[0021] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0022] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below in conjunction with the accompanying drawings and specific implementation methods.
[0023] In an exemplary embodiment, Figure 1-Figure 2 As shown, a robot joint module health status monitoring device is provided, including a vibration temperature sensor 1, a robot joint module driving board 3 and a microcontroller 4.
[0024] Among them, the vibration temperature sensor 1 is arranged on the housing 2 of the robot joint module, and is used to collect the vibration signal and temperature signal of the robot joint in real time when the robot joint module operates in the health detection mode.
[0025] Specifically, the shell 2 of the robot joint module is cylindrical as a whole, and three platforms 6 are processed along the circumferential direction near the reducer end of the robot joint module, and each platform 6 is 120° apart. A magnetic base 5 is adsorbed on the platform. The vibration temperature sensor 1 is connected to the magnetic base 5 by a hexagon socket screw, and the vibration temperature sensor 1 is fixed on the processed platform by the magnetic force of the magnetic base 5. There is a cylindrical knob 7 with a thread at the tail of the vibration temperature sensor 1, and the middle of the cylindrical knob 7 is connected, and the signal line and power line of the vibration temperature sensor 1 extend from this hollow knob. During installation, the wiring harness 8 of the three vibration temperature sensors 1 extends along the busbar direction of the shell 2 of the robot joint module to the robot joint module drive board 3, and is connected to the wiring terminal on the robot joint module drive board 3, and the wiring harness is fixed to the shell 2 of the robot joint module through the card seat 9. The wiring harness 8 has no relative rotation, and the wiring is simple and reliable, which avoids unfavorable factors such as inconvenient assembly operation, unreliable connection, and easy interference. The cylindrical knob 7 has a large inner hole, which is beneficial to the arrangement of the wiring harness 8 and the trachea without movement and twisting.
[0026] Among them, the robot joint module driving board 3 is used to collect current signals, voltage signals and position signals of the robot joints.
[0027] Specifically, the robot joint module driving board 3 has built-in current detection circuit, voltage detection circuit and position detection circuit. The current detection circuit uses a Hall current sensor to convert the current of the motor winding into a voltage analog output. The voltage detection circuit directly samples the power supply voltage. Position detection relies on a photoelectric encoder or an absolute encoder to obtain the angular displacement information of the motor rotor.
[0028] Among them, the microcontroller 4 is connected to the vibration temperature sensor 1 and the robot joint module driving board 3, and the microcontroller 4 is used to store and transmit signal data of the robot joint; the signal data includes vibration signals, temperature signals, current signals, voltage signals and position signals.
[0029] Specifically, the microcontroller 4 and the robot joint module driving board 3 are connected via the pin header on the microcontroller 4 and the female header on the robot joint module driving board 3. Figure 4 As shown, three positioning holes 10 are provided on the microcontroller 4 and the robot joint module driving board 3 along the circumferential direction, and screws and nuts are inserted into the positioning holes to fix the robot joint module motor control board 2 and the robot joint module driving board 3. The coaxiality of the microcontroller 4 and the robot joint module driving board 3 is ensured to prevent movement twisting during movement.
[0030] Wherein, the host computer is connected to the microprocessor 4 and is used to monitor the health status of the robot joint module according to the signal data.
[0031] Specifically, the control program is burned into the microcontroller 4. When the upper computer issues an order for the robot joint module to operate in the health monitoring mode, the robot joint will perform reciprocating sinusoidal motion for a long time. The long-term motion is to keep the temperature of the robot joint module constant. At this time, if the robot joint module has abnormal heating, it will be clearly observed. The reciprocating sinusoidal motion is because the vibration of the robot joint module under the reciprocating sinusoidal motion is periodic and regular, and the abnormal vibration signal can be clearly captured at this time. When operating in the health monitoring mode, the microcontroller 4 will periodically read the digital vibration signal and temperature signal in the memory at a fixed acquisition frequency, and the current, voltage, and position acquisition circuit built into the robot joint module drive board 3 will collect the current signal, voltage signal, and position signal of the joint itself. The acquisition timing of these three signals is synchronized, that is, in the same control cycle, the current, voltage, and position signals are all collected once. All the above signals are converted into digital quantities through ADC (Analog-to-Digital Converter) and stored in the memory of microcontroller 4. Microcontroller 4 is connected to and communicates with the host computer through a serial port to USB data cable, and the collected vibration signal, temperature signal, current signal, voltage signal, and position signal are transmitted to the host computer through the serial port.
[0032] The robot joint module health status monitoring device disclosed in the embodiments of the present application has the following advantages:
[0033] (1) It is possible to detect robot joint module failures or abnormal working conditions in a timely manner, so that engineers can immediately troubleshoot the source of the failure or replace the joint. This application installs a vibration temperature sensor near the harmonic reducer end of the robot joint module to directly measure the vibration temperature signal at the output end of the robot joint module. A health monitoring mode is also designed. In this mode, the robot joint module performs reciprocating sinusoidal motion for a long time, and abnormal signals generated by the movement of the robot joint module will be discovered more quickly.
[0034] (2) It can more accurately reflect the failure of the robot joint module. This application collects vibration signals and temperature signals while monitoring the current signal, voltage signal, and position signal of the joint, and uploads these signals to the host computer for processing and analysis. The mutual support of vibration, temperature, current signal, voltage signal, and position signal can more accurately judge and predict the failure of the robot joint module.
[0035] (3) Low power consumption. The present application does not connect external sensors during the normal service period of the joint module, and only connects external vibration and temperature sensors when the robot joint module is regularly running in health monitoring mode.
[0036] In an exemplary embodiment, Figure 3As shown, a method for monitoring the health status of a robot joint module is provided, comprising steps S1 to S3. Among them:
[0037] S1: Collect signal data of robot joints in the robot joint module; the signal data includes vibration signals, temperature signals, current signals, voltage signals and position signals.
[0038] Vibration and temperature sensors 1 are provided on the robot joint module to collect vibration and temperature signals of the robot joint in real time when the robot joint module operates in a health detection mode.
[0039] The current signal, voltage signal and position signal of the robot joint are collected through the robot joint module driving board 3.
[0040] S2: Preprocess the signal data and extract characteristic parameters of the signal data; the characteristic parameters include frequency domain power spectrum characteristics, time-frequency domain wavelet energy characteristics and time domain statistical parameters.
[0041] Specifically, a low-pass filter is used to filter and reduce noise on the signal data; a fast Fourier transform algorithm is used to extract frequency domain power spectrum features of the filtered and denoised signal data, a wavelet analysis algorithm is used to extract time-frequency domain wavelet energy features, and time domain statistical parameters are extracted.
[0042] Since most noise signals are high-frequency signals, a low-pass filter is used to filter and reduce noise on the signal data, and the preprocessed data is obtained after removing the high-frequency noise signal. The preprocessed data is used to extract the frequency domain power spectrum features using FFT (Fast Fourier Transform), and the wavelet energy features in the time-frequency domain are extracted using wavelet analysis, and the time domain statistical parameters such as peak value and root mean square are extracted as feature parameters.
[0043] S3: Input the characteristic parameters into the diagnostic model to determine the state of the robot joint module; the state of the robot joint module includes a healthy mode and a fault mode; the diagnostic model is obtained by training the machine learning model with the characteristic parameters of the historical signal data in the normal mode and different fault modes.
[0044] Specifically, a joint health data set containing normal operation and failure conditions is established. The data set contains historical signal data of robot joint modules in normal mode and failure mode under various working conditions.
[0045] The host computer runs an intelligent diagnosis algorithm based on machine learning and other technologies to perform online analysis and diagnosis on the received joint sensor data. The algorithm is trained in advance using historical signal data to establish a diagnosis model. During online diagnosis, the characteristic parameters of the real-time received signal data are input into the diagnosis model, which determines whether the robot joint module is in normal mode or fault mode based on the current data. If it is a fault mode, the fault type will also be determined and an early warning will be issued.
[0046] When the state of the robot joint module is in normal mode, the joint health index score is calculated, and the lower the score, the worse the health condition.
[0047] Specifically, the score corresponding to each characteristic parameter of different signal inputs is calculated according to the degree to which the characteristic parameter of the normal state deviates from the characteristic parameter. After obtaining the scores corresponding to the characteristic parameters of vibration, temperature, voltage, current, and position signals, different weights are assigned to each score according to the different effects of each signal on the health status of the joint, and finally the joint health index score is calculated by adding them up.
[0048] For each signal characteristic parameter x i , the score of the signal characteristic parameter is calculated using the linear scoring method:
[0049]
[0050] Among them, Score i is the score corresponding to different signal characteristic parameters, C i is the center value of the normal range, D i is the distance between the normal range and the fault range, and these two values are available in the established joint health dataset containing normal operation and fault conditions.
[0051] Since the characteristic parameters of different signals have different effects on joint health, vibration signals and temperature signals can better reflect the health of joints than voltage, current and position signals. Therefore, the scores corresponding to each signal characteristic parameter are calculated and weights are added artificially:
[0052] Vibration signal characteristic score: a%
[0053] Temperature signal characteristic fraction: b%
[0054] Current signal characteristic fraction: c%
[0055] Voltage signal characteristic fraction: d%
[0056] Position signal characteristic score: e%
[0057] Among them, a%+b%+c%+d%+e%=1, and the values of a, b, c, d, and e need to be determined through experimental data based on the specific parameters of the joint. Finally, the score corresponding to each signal is multiplied by the corresponding weight and added together to obtain the joint health index score.
[0058] The following is a detailed explanation of the robot joint module health status monitoring method provided by the present application through examples:
[0059] (1) After the robot joint module has been working for a period of time, click the button in the host computer to put the robot joint module into health monitoring mode. In this mode, the robot joint module performs reciprocating sinusoidal motion. The vibration temperature sensor 1 installed on the robot joint module collects the vibration signal and temperature signal of the joint. At the same time, the robot joint module drive board 3 collects the current signal, voltage signal, and position signal of the joint. The sampling frequency is 25600Hz. These signals are uploaded to the host computer through serial communication.
[0060] (2) The host computer receives the transmitted original signal x(n), and first uses a zero-phase filter to filter the original signal to obtain the filtered preprocessed data xf(n). The zero-phase filter parameters are as follows: Butterworth low-pass filter, 4th order, 0.1Hz cutoff frequency.
[0061] (3) Use FFT to extract the frequency domain power spectrum features of the preprocessed data xf(n). Apply FFT transformation to xf(n):
[0062]
[0063] Among them, X(k) is the frequency domain sequence after FFT transformation; n = 0, 1, ..., N-1, a total of N sampling points; e^(-j2πkn / N) is the basis function of FFT; k = 0, 1, ..., N-1, is the sampling index of the frequency domain, a total of N points.
[0064] Calculate the power spectrum P(k):
[0065] P(k)=|X(k)| 2 / N
[0066] The power spectrum peak frequency Fp and peak amplitude Ap are extracted as frequency domain power spectrum features.
[0067] (4) Wavelet analysis to extract wavelet energy features
[0068] Select Daubechies wavelet base db4, decomposition scale J = 5 and perform 5-layer wavelet decomposition on xf(n):
[0069] xf(n)=a5(n)+d5(n)+d4(n)+d3(n)+d2(n)+d1(n)
[0070] where a5(n) is the approximate component of the fifth layer, d j (n)(j=1,2,3…,5) is the detail component of the jth layer.
[0071] For each layer’s detail component d j (n) and the approximate component a5(n) of the fifth layer to calculate the energy:
[0072] Energy of detail component E j =∑ n |d j (n)| 2 ;
[0073] Energy E of the approximate component a5 =Σ n |a5(n)| 2 .
[0074] (5) Calculate the time domain statistical parameters such as the root mean square (RMS) and peak factor (CF) of xf(n).
[0075]
[0076] CF=(max|xf(n)|) / RMS
[0077] The maximum absolute value of xf(n) is taken as the peak value of the signal, and RMS is the root mean square value obtained by the above calculation. Finally, all the extracted feature parameters: Fp, Ap, E j 、E a5 , RMS, and CF form the feature vector X, which is input into the trained machine learning diagnostic model.
[0078] (6) The extracted feature vector X is input into the trained diagnostic model f(X), and the diagnostic model outputs a category judgment result y', where y' = 0 indicates that the current sensor data corresponds to the normal mode, and y' = 1, 2, 3... indicates different fault types.
[0079] (7) Finally, if the host computer classifies the model as "normal mode", the health index score H = 80 is displayed on the UI interface, indicating that the joint module used in this example is in good health and does not require maintenance.
[0080] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0081] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.
Claims
1. A robot joint module health status monitoring device, characterized in that: include: A vibration temperature sensor, disposed on a housing of the robot joint module, for collecting vibration signals and temperature signals of the robot joint in real time when the robot joint module operates in a health detection mode; Robot joint module driver board, used to collect current signals, voltage signals and position signals of robot joints; A microcontroller connected to the vibration temperature sensor and the robot joint module driving board, and used to store and transmit signal data of the robot joint; the signal data includes vibration signal, temperature signal, current signal, voltage signal and position signal; A host computer is connected to the microprocessor and is used to monitor the health status of the robot joint module according to the signal data.
2. The robot joint module driver according to claim 1, characterized in that: A plurality of platforms are arranged at equal intervals along the circumferential direction at the reducer end close to the robot joint module, a magnetic base is arranged on the platform, and the vibration temperature sensor is connected to the magnetic base via a hexagon socket screw.
3. The robot joint module driver according to claim 1, characterized in that: The wiring harness of the vibration sensor extends along the busbar direction of the shell of the robot joint module to the robot joint module driving board, and is connected to the wiring terminals on the robot joint module driving board. The wiring harness is fixed to the shell of the robot joint module through a socket.
4. The robot joint module driver according to claim 1, characterized in that: The pin header on the microcontroller is connected to the female header on the robot joint module driving board.
5. The robot joint module driver according to claim 1, characterized in that: Positioning holes are provided on the microcontroller and the robot joint module driving plates along the circumferential direction, and the microcontroller and the robot joint module are fixed through the positioning holes.
6. The robot joint module driver according to claim 1, characterized in that: The robot joint module driving board has built-in current detection circuit, voltage detection circuit and position detection circuit.
7. A method for monitoring the health status of a robot joint module, the method being applied to a device for monitoring the health status of a robot joint module according to any one of claims 1 to 6, characterized in that: The method comprises: Collecting signal data of robot joints in the robot joint module; the signal data includes vibration signals, temperature signals, current signals, voltage signals and position signals; Preprocessing the signal data and extracting characteristic parameters of the signal data; the characteristic parameters include frequency domain power spectrum characteristics, time-frequency domain wavelet energy characteristics and time domain statistical parameters; The characteristic parameters are input into the diagnostic model to determine the state of the robot joint module; the state of the robot joint module includes a normal mode and a fault mode; the diagnostic model is obtained by training a machine learning model with the characteristic parameters of the historical signal data in the normal mode and the fault mode.
8. The method for monitoring the health status of a robot joint module according to claim 7, characterized in that: Preprocessing the signal data and extracting characteristic parameters of the signal data specifically includes: Using a low-pass filter to perform filtering and noise reduction processing on the signal data; For the signal data processed by filtering and denoising, the fast Fourier transform algorithm is used to extract the frequency domain power spectrum characteristics, the wavelet analysis algorithm is used to extract the time-frequency domain wavelet energy characteristics, and the time domain statistical parameters are extracted.
9. The robot joint module health status monitoring method according to claim 7, characterized in that: After inputting the characteristic parameters into the diagnosis model to determine the state of the robot joint module, the method further includes: When the state of the robot joint module is in normal mode, the joint health index score is calculated.
10. The robot joint module health status monitoring method according to claim 7, characterized in that: After inputting the characteristic parameters into the diagnosis model to determine the state of the robot joint module, the method further includes: When the state of the robot joint module is in a fault mode, the fault type is determined and an early warning is issued.
Citation Information
Patent Citations
Driving and control integrated type intelligent integration joint
CN104552329A
Method and device for open-loop / closed-loop control of an actuator-driven robot joint
CN107645980A
Real-time abnormity monitoring method and system for industrial robot
CN108638128A
Industrial robot operation fault determination method and system
CN111283731A
High-speed rail gear box bearing testing machine
CN111855213A
Cited By
Multi-modal fusion robot joint module health state monitoring method
CN122100237A
Humanoid robot joint module health state monitoring equipment
CN224509747U
