Integrated joint module

Abstract

The application discloses a kind of feel drive integrated joint module, comprising: shell, for installing harmonic reducer, power module, power module back cover, brake module, fixed flange, drive module, control module, sensing module and main shaft;Main shaft, rotation support in shell and power module back cover;Power module, for the rotation of main shaft by the mode of driving electromagnetic excitation;Power module back cover, with wiring slot and with the gap between power module;Brake module, for generating brake force lock main shaft;Fixed flange, with wiring slot and positioning drive module and control module;Sensing module, for detecting the temperature, vibration and position signal of joint module, and transmit to control module;Control module, for controlling power module, brake module and realizing the regulation and control to joint module position, speed and torque;The stability of the application improves joint module, response speed and security.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of driving control devices, in particular to a sensing and driving integrated joint module. BACKGROUND

[0002] The sensing and driving integrated joint module is widely used in various intelligent robots, and undertakes comprehensive tasks such as motion generation, accurate positioning and driving control. In the fields of rescue robots, industrial robot arms and service robots, the module not only needs to integrate various functional elements in a limited space, but also needs to achieve efficient power transmission and accurate action execution to meet the diversified application requirements. Its performance has a crucial impact on the flexibility, load capacity and intelligent operation of the whole robot.

[0003] At present, the mainstream integrated joint module usually adopts the combination design of planetary gear reducer and brushless DC motor, and integrates the driver and controller inside to realize basic motion control. This kind of joint module provides stable reduction ratio and high transmission efficiency through the planetary gear reducer, and the brushless DC motor provides relatively stable output for motion driving.

[0004] Although the joint module based on planetary gear reducer and brushless DC motor has breakthroughs in performance, there are still obvious deficiencies. First of all, the overall structure of this kind of module is bulky, occupies a large space, and it is difficult to achieve efficient integration in a compact robot arm design. Secondly, its torque output capacity is limited, which cannot meet the high load task or instantaneous high torque demand. Thirdly, due to the low degree of modularity of the system, the cooperation efficiency and integration level between components are insufficient, making it difficult to realize soft control and intelligent operation. In addition, the dynamic response capability in complex environment is poor, which limits its application in high sensitivity and high flexibility scenes. These problems make it difficult for the existing technology to fully meet the diversified needs of modern robots in intelligence, flexibility and high performance.

[0005] Therefore, there is an urgent need for a technical solution that can effectively improve the power output performance, response speed and intelligent control ability of the joint module, to solve the problem of insufficient stability of the existing module under high load and complex working conditions, so as to meet the needs of smooth and reliable operation of robots in diversified scenes. SUMMARY

[0006] The purpose of the present application is to provide a sensing and driving integrated joint module to improve the stability, response speed and safety of the joint module, and significantly enhance the intelligent level through the real-time sensing of the module operating state by the sensing module, to meet the needs of high-performance equipment driving scenes.

[0007] The technical solution to achieve the purpose of the present application is:

[0008] A sense and drive integrated joint module, comprising:

[0009] A shell for mounting a harmonic reducer, a power module, a power module rear cover, a brake module, a fixing flange, a drive module, a control module, a sensing module and a main shaft;

[0010] A main shaft rotatably supported on the shell and the power module rear cover;

[0011] A power module for driving the rotation of the main shaft by means of electromagnetic excitation;

[0012] A power module rear cover provided with a wiring slot and having a gap with the power module;

[0013] A brake module for generating a braking force to lock the main shaft;

[0014] A fixing flange provided with a wiring slot and positioning the drive module and the control module;

[0015] A sensing module for detecting the temperature, vibration and position signals of the joint module and transmitting them to the control module;

[0016] A control module for controlling the power module, the brake module and realizing the regulation of the position, speed and torque of the joint module;

[0017] A harmonic reducer connected to the main shaft, which outputs the power module through the main shaft and increases the output torque.

[0018] Compared with the prior art, the present application has the following advantages:

[0019] Increase the torque output density: by adopting a high torque density harmonic reducer and a power module, efficient torque transmission and driving performance are realized, and the load capacity and motion efficiency of the joint module are effectively improved. This improvement significantly enhances the power output performance of the joint module, enabling it to meet the needs of high-load and high-dynamic performance scenarios.

[0020] Enhance the compactness of the structure: the harmonic reducer, power module, brake module, drive module, control module and sensing module are integrated into one, and an optimized spatial layout design is adopted, which greatly reduces the volume and weight of the joint module. The high compactness of the structure lays the foundation for the lightweight design of the joint module, improving the overall design freedom and adaptability.

[0021] Realize multi-functional modularization: through modular design, driving, control, sensing and other functions are integrated into a unified joint module, ensuring that each functional module operates independently and cooperatively. This multi-functional modular design significantly improves the expandability and versatility of the joint module, meeting the needs of diverse application scenarios.

[0022] Improve driving control precision: The perception module provides high-precision motion state feedback, combined with advanced algorithms in the control module and high-performance driving modules, to achieve precise regulation of joint position, speed and torque. High-precision driving control ensures stable operation of the module under high-performance equipment driving requirements, providing technical support for intelligent precision equipment.

[0023] Optimize the layout scheme: Through the reasonable layout between modules and the design of wiring slots, the signal interference and cable redundancy are minimized, and the electrical reliability and maintenance convenience of the module are improved. The optimization of the line arrangement not only improves the overall operation efficiency of the system, but also provides technical support for the maintainability and durability of the module. BRIEF DESCRIPTION OF DRAWINGS

[0024] Figure 1 is a two-dimensional sectional view of the structure of the embodiment of the present application;

[0025] Figure 2 is a two-dimensional sectional view of the harmonic reducer of the embodiment of the present application;

[0026] Figure 3 is a three-dimensional structure diagram of the power module of the embodiment of the present application;

[0027] Figure 4 is a three-dimensional structure diagram of the rear cover of the power module of the embodiment of the present application;

[0028] Figure 5 is a three-dimensional structure diagram of the brake module of the embodiment of the present application;

[0029] Figure 6 is a three-dimensional structure diagram of the fixed flange of the embodiment of the present application;

[0030] Figure 7 is a working principle diagram of the long short-term memory network of the embodiment of the present application;

[0031] Figure 8 is a flow chart of the adaptive empirical mode decomposition algorithm of the embodiment of the present application;

[0032] Figure 9 is a three-dimensional structure diagram of the encoder of the embodiment of the present application;

[0033] Figure 10 is a working flow chart of the sense-drive control integrated joint module of the embodiment of the present application;

[0034] Figure 11 is a two-dimensional sectional view of the shell of the embodiment of the present application;

[0035] In the diagram, the markings are: 1-Harmonic reducer, 11-Wave generator, 111-Elliptical cam, 112-Flexible bearing, 12-Flexible wheel, 121-External gear, 122-Connecting flange, 13-Rigid wheel, 131-Rigid gear ring, 132-Rigid wheel mounting base, 2-Housing, 21-Joint housing, 22-Joint rear cover, 23-First bearing, 3-Power module, 31-Stator, 32-Rotor, 33-Integrated circuit board, 331-Power terminal block, 332-Hall sensor, 333-Thermistor, 4-Power... Module rear cover, 41-wiring groove, 42-second bearing, 43-fixed bracket, 5-brake module, 51-electromagnet, 511-coil, 512-spring, 52-armature, 53-friction plate, 54-brake seat, 55-brake housing, 6-fixed flange, 61-wiring groove, 7-drive module, 8-control module, 9-sensing module, 91-temperature sensor, 92-accelerometer, 93-encoder, 931-optical encoder disk, 932-signal processing unit, 933-interface module, 10-spindle. Detailed Implementation

[0036] The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0037] The present invention and its embodiments are described below. This description is not restrictive, and actual embodiments are not limited thereto. In short, if those skilled in the art are inspired by this description and, without departing from the spirit of the invention, design similar structures and embodiments to this technical solution, such designs should fall within the protection scope of the present invention.

[0038] like Figure 1 As shown, this embodiment discloses an integrated sensing, driving, and control joint module, including a harmonic reducer 1, a housing 2, a power module 3, a power module rear cover 4, a braking module 5, a fixing flange 6, a drive module 7, a control module 8, a sensing module 9, and a main shaft 10. The harmonic reducer 1 is fixed to the housing 2, and the main shaft 10 is disposed inside the housing 2. The power module 3, the power module rear cover 4, the braking module 5, and the fixing flange 6 are all sleeved with the main shaft 10.

[0039] like Figure 2As shown, the harmonic reducer 1 is composed of a wave generator 11, a flexspline 12 and a rigid spline 13, wherein the wave generator 11 is composed of an elliptical cam 111 and a flexible bearing 112, the flexspline 12 is composed of an external gear 121 and a connecting flange 122, and the rigid spline 13 includes a rigid gear ring 131 and a rigid spline fixing seat 132. The wave generator 11 is connected with the flexspline 12 through interference, and is stably connected with the main shaft 10 through a flat key to drive the power transmission between the flexspline 12 and the rigid spline 13. In a specific embodiment, the elliptical cam 111 is sleeved with the main shaft 10 and connected through a flat key, the elliptical cam 111 is supported in the rigid gear ring 131 through the flexible bearing 112, the rigid gear ring 131 is fixed in the rigid spline fixing seat 132, and the rigid spline fixing seat 132 is connected with the shell 2 through bolts; the elliptical cam 111 and the flexible bearing 112 are matched through clearance. The external gear 121 and the connecting flange 122 are fixed and rotatably supported in the rigid spline fixing seat 132 and used in cooperation with the rigid gear ring 131, and the engagement and disengagement between the external gear 121 of the flexspline 12 and the rigid gear ring 131 realizes differential transmission. The connecting flange 122 is connected with external loads through eight evenly distributed threaded holes to complete the torque output of the joint module. The rigid spline fixing seat 132 is firmly connected with the joint shell 21 through eight evenly distributed screws to ensure the stability of the overall structure.

[0040] In terms of working principle, when the main shaft 10 rotates, the wave generator 11 applies radial deformation to the flexible bearing 112 through the cooperation of the elliptical cam 111 and the flexible bearing 112, thereby driving the engagement and disengagement of the flexspline 12 and the rigid spline 13 at a specific position. The difference in tooth number between the flexspline 12 and the rigid spline 13 causes differential transmission, thereby realizing accurate torque transmission between the external gear 121 and the rigid gear ring 131. Power is input to the wave generator 11 through the main shaft 10 and torque is output through the connecting flange 122 to ensure efficient operation of the system.

[0041] The harmonic reducer 1 converts the high-speed, low-torque power module 3 output into a low-speed, high-torque output through the periodic engagement of the flexspline 12 and the rigid spline 13, and the transmission ratio is determined by the tooth number difference of the flexspline 12 and the rigid spline 13, and the specific calculation formula is:

[0042]

[0043] wherein Z g and Z f represent the tooth numbers of the rigid spline 12 and the flexspline 13, respectively. When the tooth number of the rigid spline 13 is greater than that of the flexspline 12, high transmission ratio and high precision control can be achieved. Through the principle of tooth difference, the harmonic reducer 1 not only gives the joint precise motion control ability, but also significantly improves the torque density of the joint, making it suitable for a variety of complex robot working scenarios.

[0044] AsFigure 3 As shown, the power module 3 is a high-efficiency drive device with various forms, and here an inner rotor frameless torque motor is taken as an example for illustration. The module mainly consists of a stator 31, a rotor 32 and an integrated circuit board 33. The integrated circuit board 33 includes power connection terminals 331, Hall sensors 332 and thermal sensors 333. The stator 31 is firmly fixed in the inner hole of the joint shell 21 through adhesive or mechanical assembly. The rotor 32 is placed inside the stator 31 and is firmly fixed on the main shaft 10 through adhesive or mechanical assembly. The axial positioning of the rotor 32 is based on the shoulder surface of the main shaft 10, while maintaining a spacing of 3-5 mm between the left end surface of the power module rear cover 4, thereby avoiding the leakage phenomenon of the power module 3. The integrated circuit board 33 is adhered to the right end surface of the stator 31. The power connection terminals 331, Hall sensors 332 and thermal sensors 333 are all welded on the integrated circuit board 33. In terms of working principle, the stator 31 is responsible for generating a rotating magnetic field to drive the rotor 32 to rotate. The rotor 32 interacts with the stator 31 through electromagnetic induction and generates torque under the magnetic field action of the stator 31. The integrated circuit board 33 provides power control and signal processing functions for the power module 3. The integrated circuit board 33 is provided with power connection terminals 331 for connecting with external power supply to provide power supply. At the same time, the integrated circuit board 33 also contains Hall sensors 332 and thermal sensors 333, which are used to monitor the position of the rotor 32 and the temperature of the power module 3 in real time, respectively. The Hall sensors 332 provide accurate rotor 32 position information by detecting magnetic field changes, so as to realize closed-loop control; the thermal sensors 333 monitor the working temperature of the power module 3 in real time and feed back the temperature data to the control module 8. Once the temperature exceeds the safety threshold, the control module 8 will automatically adjust the power output or start the cooling mechanism, so as to ensure that the power module 3 always operates within the safe temperature range and effectively avoids faults or performance degradation caused by overheating. The design structure of the power module 3 is simple and compact, which can effectively improve the transmission efficiency and control performance.

[0045] As Figure 4As shown, the power module rear cover 4 includes a wiring groove 41, a second bearing 42, and a mounting bracket 43. The wiring groove 41 is located on the end face of the mounting bracket 43. The second bearing 42 is interference-fitted with the inner hole at the center of the mounting bracket 43 to support the spindle 10. The mounting bracket 43 is securely fixed to the joint housing 21 by three circumferentially distributed protrusions on its outer circumference. The wiring groove 41 inside the power module rear cover 4 optimizes the wiring arrangement between the power module 3 and the temperature sensor 91, avoiding wiring crossing, interference, and wear, and further improving the stability and durability of the system. In addition, the second bearing 42 is also installed inside the power module rear cover 4. This bearing effectively distributes the radial and axial loads generated by the power module 3 during high-speed operation, reducing vibration and friction caused by rotation, thereby improving the operational stability of the power module 3 and the reliability of power output. While improving the system accuracy, the second bearing 42 also ensures the smooth movement and long-term stable operation of the sensor-driven control integrated joint module.

[0046] like Figure 5 As shown, the braking module 5 consists of an electromagnet 51, an armature 52, a friction plate 53, a brake seat 54, and a brake housing 55. The electromagnet 51 includes a coil 511 and a spring 512. The electromagnet 51 is fixed to the right end face of the power module rear cover 4 by three evenly distributed screws and is connected to the drive module 7 via two shielded communication lines, one connected to the positive terminal of the drive module 7 and the other to the negative terminal. The spring 512 is bonded to the right end face of the coil 511 with adhesive. The left end face of the armature 52 is welded to the spring 512. A friction plate 53 is placed between the armature 52 and the brake housing 55. The square outer wall of the brake seat 54 and the square inner hole of the friction plate 53 are interference-fitted to ensure synchronous start and stop of the friction plate 53 and the brake seat 54. The inner hole of the brake seat 54 is reliably connected to the main shaft 10 via a flat key, and axial positioning is achieved using two left and right shaft retaining rings. The brake housing 55 is fixed to the right end face of the electromagnet 51 by three circumferentially distributed screws. When the power is on, the coil 511 generates magnetic force, attracting the armature 52 and compressing the spring 512. At this time, the friction plate 53 and the brake seat 54 rotate with the main shaft 10. When the power is off, the coil 511 loses its electromagnetic force, and the armature 52, pushed by the spring 512, presses the friction plate 53 against the brake housing 55, generating braking force through friction, quickly locking the main shaft 10, and realizing the braking of the power module 3. The advantage of the braking module 5 is that it responds quickly and can achieve smooth start and stop, ensuring the dynamic safety and operational stability of the joint module.

[0047] like Figure 6As shown, the fixing flange 6 includes eight circumferentially distributed wiring grooves 61. The fixing flange 6 is fixed to the inner hole of the joint shell 21 through a hole fitting retainer, thereby ensuring its stable positioning and preventing axial movement. The mounting mode of the flange provides reliable structural support, ensuring that the encoder 93 always maintains the correct position during operation. The fixing flange 6 is internally provided with wiring grooves 61, which not only optimize the arrangement of the power module 3 and temperature sensor 91 wiring, but also effectively avoid interference with other components, ensuring stable and clean electrical connections of the entire system, reducing interference, and thereby improving the overall performance and reliability of the joint module.

[0048] The drive module 7, control module 8, and encoder 93 are fixed to each other by three circumferentially distributed studs. This design not only exhibits high structural compactness in terms of reducing the axial size of the joint module and optimizing wiring arrangement, but also ensures accurate alignment and mechanical stability of the modules during assembly through evenly distributed studs, thereby effectively reducing vibration and signal delay caused by installation errors. Further, the integration of the drive module 7 and control module 8 significantly shortens the signal transmission path, reduces redundant wiring and electromagnetic interference, and improves system response speed and control accuracy, providing a solid technical guarantee for the optimization of overall performance.

[0049] The core function of the drive module 7 is to control the electromagnetic excitation of the power module 3, achieving precise output and dynamic adjustment of the torque of the power module 3. The drive module 7 provides a stable power driving basis for the joint module with high-efficiency power transmission and high-precision signal processing capability. Through integrated design, the drive module 7 optimizes energy efficiency while reducing modular size, providing a guarantee for overall performance improvement.

[0050] The control module 8 is responsible for analyzing data from the drive module 7 and sensing module 9, and executing high-precision closed-loop control strategies. Through real-time calculation and signal processing, the control module 8 adjusts the speed and position of the power module 3, ensuring that the motion of the joint module conforms to the expected trajectory and achieves precise and rapid response under complex working conditions, further enhancing the intelligent level of the system.

[0051] The perception module 9 is used to realize the omnidirectional state perception of the joint module, which is composed of a temperature sensor 91, an acceleration sensor 92 and an encoder 93. The encoder 93 includes an optical encoding disc 931, a signal processing unit 932 and an interface module 933. The temperature sensor 91 is firmly bonded to the left end face of the power module rear cover 4 by an adhesive, and the acceleration sensor 92 is pasted on the right end face of the fixed flange 6. The optical encoding disc 931 is stably fixed on the main shaft 10 by three evenly distributed tightening screws in the circumference, and is also fixed on the right end face of the fixed flange 6 by three screws in the axial direction. The temperature sensor 91 monitors the temperature of the joint module in real time to prevent performance degradation or damage to critical components due to abnormal temperature. The acceleration sensor 92 uses micro-electro-mechanical system (MEMS) technology to realize high-precision acceleration signal acquisition, providing important data for motion state analysis. The encoder 93 provides high-precision position and speed feedback to ensure the accuracy and real-time performance of motion control. Through the cooperative work of the three, the perception module 9 can accurately and comprehensively reflect the running state of the joint module, ensuring the stability and reliability of the joint module operation.

[0052] The temperature sensor 91 has various forms, and the invention takes a patch type temperature sensor as an example, and other types of temperature sensors can also be selected according to actual needs. Since the working environment of the power module 3 is usually a closed space, when the temperature is too low, the system may have problems such as material embrittlement and lubricating liquid solidification, which can cause mechanical moving parts to work abnormally; when the temperature is too high, it may cause electrical components to overheat, even causing short circuit or fire. Therefore, through the real-time monitoring of the temperature sensor 91, the system can detect abnormal fluctuations in temperature in time, automatically trigger the protection mechanism, such as adjusting the working load or stopping running, to prevent component damage or performance degradation caused by overheating, thereby effectively improving the safety and stability of the system. The patch type temperature sensor is based on the characteristic that the resistance changes with temperature, and reflects the temperature change by measuring the resistance change, which has the characteristics of high precision, low noise and wide measurement range, and is especially suitable for precise temperature monitoring in high temperature environment. The temperature sensor 91 adopts a three-wire circuit connection mode, in which two shielded communication lines are used for current input, and the third shielded communication line feedbacks resistance value signal to ensure high precision and low temperature drift performance.

[0053] The temperature sensor 91 is used for temperature measurement of the integrated joint module. The method is as follows. First, the temperature sensor 91 samples the real-time temperature of the left end face of the power module rear cover 4. Then, the collected temperature data is input to the control module to generate corresponding weight feature sequences. Finally, these feature sequences are passed to the LSTM (Long Short-Term Memory Network) for reconstructing the temperature distribution of the left end face of the power module rear cover 4. LSTM has a significant advantage in processing temperature data. LSTM effectively overcomes the limitations of traditional RNN in processing long-time dependence by introducing a gating mechanism, avoiding the problem of gradient vanishing or explosion. This feature enables LSTM to capture more rich long-term dynamic features in the time series data of temperature changes, thus providing stronger time-dependent modeling capability and higher accuracy in temperature monitoring.

[0054] Figure 7 The working principle diagram of LSTM (Long Short-Term Memory Network) is shown in the figure. LSTM neural network significantly enhances its ability to process long-term dependencies in feature sequences by introducing four key mechanisms: forget gate, input gate, output gate and cell state.

[0055] In the joint module, the cell state C t represents the long-term memory of all historical temperature data, which stores the temperature evolution of the joint module and adds the information of the current temperature change at each time step update. In this way, the LSTM network can dynamically adjust the temperature control of the joint module according to the current input and long-term temperature memory.

[0056] The hidden state h t is the output of each time step, which contains the temperature change information of the joint module at the current time and integrates the cell state of the previous time. In the joint module, the hidden state is equivalent to the immediate feedback of the current temperature change of the joint module, helping the system to respond to the current temperature change and provide necessary input in the next step calculation. The hidden state and the cell state work together through the gating mechanism to ensure that LSTM can capture the dynamic characteristics of temperature changes over a long time span while responding to short-term temperature fluctuations in a timely manner.

[0057] The Sigmoid activation function is a smooth, non-linear function whose output value ranges from 0 to 1. It compresses the weighted sum of the inputs, allowing the output value of the forget gate to be adjusted between 0 (complete forgetting) and 1 (complete retention). In temperature monitoring, this function helps determine which temperature data is most important for the current decision and which can be ignored. For example, in the temperature monitoring of a joint module, some historical temperature data (such as temperature changes at distant time steps) may be irrelevant to the current decision. This can be "forgotten" by the forget gate, while retaining more relevant temperature data, thereby optimizing the system's response to temperature changes.

[0058] The forget gate in LSTM is used to determine which temperature information should be forgotten from the cell state. This mechanism allows LSTM to effectively suppress the accumulation of unnecessary historical information, thereby enhancing the network's stability and performance when processing long-term data. In the temperature monitoring application of joint modules, the forget gate uses a sigmoid activation function based on the current temperature input x. t and the hidden state h from the previous moment t-1 The calculation outputs a value between 0 and 1. This value represents the forgetting gate for cell state C. t The degree to which temperature information is retained is calculated using the following formula.

[0059] f t =σ(W f ·[h t-1 ,x t ]+b f (2)

[0060] Among them, f t This indicates the amount of temperature information to retain; 0 means discarding all information, and 1 means retaining all information. σ represents the sigmoid activation function. W f The weight matrix representing the forget gate determines how the system adjusts its dependence on historical temperature data based on previous temperature states and changes in the current input temperature. t-1 and x t This represents the hidden state at the previous moment and the temperature input at the current moment. (b) f The bias term representing the forget gate provides adjustment for the forget gate output, allowing the forget gate to further optimize its temperature data processing logic according to specific application requirements.

[0061] In LSTM, the input gate controls the degree to which the current input information influences the cell state update. In the temperature monitoring application of the joint module, the input gate uses a sigmoid activation function to process the current temperature input x. t and the hidden state h from the previous moment t-1The weighted sum produces a value between 0 and 1, which determines which temperature information will be written into the cell state C t , i.e., the degree of influence of the current temperature data on the long-term temperature memory of the joint module. Specifically, a higher value means more temperature information is retained, and a lower value means the current temperature data contributes less to the system state.

[0062] Subsequently, this value is multiplied by a candidate cell state that has been processed by a tanh activation function. The candidate cell state represents new temperature information extracted from the current input and the previous state, which is a prediction of future temperature trends. Through the multiplication with the sigmoid output, the LSTM decides how much new information to add to the cell state C t , thereby affecting the long-term memory of the system and dynamically adjusting the temperature control strategy. This process ensures the effective updating of temperature data, enabling the joint module to make reasonable temperature adjustment decisions based on current and historical data.

[0063]

[0064] where i t represents the activation value of the input gate, determining the degree of influence of the current input temperature x t on the update of the cell state C t . In the temperature monitoring of the joint module, the input gate controls the weight of the input temperature information in the long-term temperature memory, ensuring the effective transmission of temperature information. W i represents the weight matrix of the input gate, used to convert the previous hidden state h t-1 and the current temperature input x t into weighted values for the input gate, determining the contribution of each temperature sensor signal to the current state of the system. b i is the bias term of the input gate, added to the weighted sum result, used to adjust the model's response to the input gate, thereby making appropriate adjustments when the temperature data fluctuates greatly. represents the candidate value of the cell state, which is calculated by weighting and summing the current input temperature information x t and the previous hidden state h t-1 and then passing it through a tanh activation function. This candidate value provides potential update information about the temperature change at the current time, determining which new temperature information will be written into the current cell state to update the long-term temperature memory. The tanh activation function performs a nonlinear transformation on the input, resulting in an output value between -1 and 1, controlling the flow of information in the LSTM and ensuring smooth transition of temperature information. W C represents the weight matrix of the cell state candidate, similar to W iUsed to hide the previous time step h t-1 and the current temperature input x t Convert to candidate cell state values. C The bias term for the cell state candidates is added to the weighted summation result to adjust the model's response to the cell state candidate values, ensuring adaptability and accuracy to temperature changes.

[0065] In an LSTM, the output gate is responsible for determining the hidden state h at the current time step. t What information should be output? In the temperature monitoring application of the joint module, the output gate adjusts the cell state C. t and the hidden state h from the previous moment t-1 This enables the filtering and transmission of temperature information, as shown in the following formula. Specifically, the output gate generates the hidden state h. t At this time, it acts as a filter and selector, ensuring that only the effective temperature information from the cell state is passed to the next layer or used as the final system output. In this way, the output gate ensures that only the temperature data most critical to temperature control can influence the next decision, enhancing the system's responsiveness and stability to temperature changes.

[0066]

[0067] Among them, o t Indicates the hidden state h at the current moment. t The amount of information. In the temperature monitoring application of joint modules, the output gate's role is to regulate the cell state C. t The temperature information within ensures that only data meaningful for temperature control is passed to the next layer or output. o This represents the weight matrix of the output gate, which controls the hidden state h from the previous time step. t-1 And the current input temperature data x t Impact on the output gate. Through weighted summation, the weight matrix determines the importance of historical temperature information and current temperature input in adjusting the output. o This represents the bias term of the output gate, which adjusts the output of the gate to ensure that the system can better fit actual temperature data changes during network training. This adjustment process helps the model optimize its response to temperature data, ensuring accurate capture of temperature changes in the system.

[0068] The fixed flange 6 is far away from the output end. Since the joint module often generates a large dynamic impact and vibration when running at high speed, the acceleration sensor 92 is pasted at this position, which can more accurately perceive these mechanical changes and timely feedback the dynamic state of the joint module. The acceleration sensor 92 is based on micro-electro-mechanical system (MEMS) technology, uses the capacitive principle, the deformation of the cantilever beam causes the resistance value to change, thereby changing the analog voltage, realizing the acceleration signal acquisition of the fixed flange 6. The acceleration signal is transmitted to the control module 8 through the digital interface, which is used for real-time monitoring of the dynamic motion state of the joint and the external vibration condition. The acceleration sensor 92 is connected with the control module 8 through I2C or SPI protocol, and transmits the acceleration signal to the control module 8. The acceleration sensor 92 can accurately perceive the impact and vibration, and help the control system quickly adjust the operation strategy through the feedback signal, reduce the influence of external disturbance on the system, thereby significantly improving the dynamic response ability and operation stability of the module.

[0069] In the process of collecting acceleration signals by the acceleration sensor 92, in order to improve the sampling accuracy of the signal, first, the control module 8 decomposes the original acceleration signal by using adaptive empirical mode decomposition (EMD). The original signal refers to the vibration data collected by the acceleration sensor during the operation of the joint module. These signals reflect the vibration conditions of the components of the joint module, including impact, friction and other factors. In the decomposition process, the EMD algorithm decomposes the acceleration signal into several intrinsic mode functions (IMF), each IMF represents the local vibration mode of the joint module at different time scales. For example, rapid impact vibration may occur when components collide or are operated violently, while continuous low-frequency vibration may be caused by long-term operation or load. Through these IMFs, different working states of the components of the joint module and potential abnormal vibrations can be effectively identified.

[0070] EMD decomposes the input signal into several intrinsic mode functions and a residual, i.e. formula 5.

[0071]

[0072] Wherein, X(t) represents the input vibration signal, IMF m (t) represents the intrinsic mode function of the mth input vibration signal, r M (t) represents the residual, and M represents the total number of input vibration signals. The process of extracting IMF is as shown in Figure 8 .

[0073] (1) Signal preprocessing: The raw acceleration signals collected from the acceleration sensors are denoised and normalized to ensure the signal quality is suitable for subsequent analysis. In the joint module, this process removes irrelevant information caused by sensor noise or external environmental interference. Common preprocessing methods include high-pass filtering to remove low-frequency noise (such as device drift); low-pass filtering to remove high-frequency noise (such as electrical interference); and wavelet denoising to remove noise in transient impacts or short vibrations.

[0074] (2) Local extreme point extraction: Extract all local maximum and minimum points from the preprocessed acceleration signal, which represent key peaks in the joint module vibration signal, such as the start or end position of the vibration impact. To ensure accurate extraction of these extreme values, cubic spline interpolation can be used to smooth the local extreme points in the signal, further enhancing the accuracy and stability of the signal. By extracting extreme points, abnormal vibrations or impacts of the joint module can be effectively identified, helping to diagnose potential faults or abnormal operating conditions.

[0075] (3) Envelope construction: Construct the upper envelope and lower envelope of the signal by performing spline interpolation on the local maximum points e upper (t) and local minimum points e lower (t), respectively. In the acceleration signal processing of the joint module, the upper envelope and lower envelope reflect the upper and lower limits of the oscillation amplitude of the signal. Through the construction of the envelope, the vibration trend of the signal can be clearly determined, and the change range of the vibration amplitude in the module can be identified. The upper envelope and lower envelope provide clear boundaries for subsequent signal analysis, helping to monitor the dynamic response of the joint module, especially for detecting changes in vibration intensity and sudden impacts.

[0076] (4) Calculate envelope average m(t): Calculate the average of the upper envelope and lower envelope to obtain the local average envelope of the signal, as shown in the formula. In the acceleration signal processing of the joint module, the local average envelope reflects the low-frequency trend of the acceleration signal, helping to capture the main change pattern of the joint module in normal operation. By calculating the envelope average, the overall fluctuation trend of the signal and the long-term dynamic response of the system can be identified.

[0077]

[0078] (5) Remove envelope: Subtract the local average envelope from the original acceleration signal to obtain the intermediate signal h(t), which is the high-frequency part of the signal, as shown in formula 7. The process of removing the envelope helps to remove the low-frequency trend component in the signal, highlighting the transient vibration characteristics in the joint module. This process allows the high-frequency vibration signal to be clearly displayed, which helps to analyze the dynamic changes caused by impacts or rapid vibrations.

[0079] h(t) = X(t) - m(t) (7)

[0080] where X(t) represents the raw acceleration data collected by the acceleration sensor, m(t) is the mean signal obtained by calculating the envelope average, reflecting the low-frequency trend of the signal, and h(t) is the intermediate signal after removing the trend, highlighting the high-frequency vibration components of the joint module, helping to identify rapidly changing vibration patterns or impacts.

[0081] (6) Iteration: If the intermediate signal h(t) does not meet the definition of IMF, h(t) is taken as the new input signal g(t), and steps (2) to (5) are repeated. If the intermediate signal h(t) meets the definition of IMF, h(t) is extracted as an IMF component. Then X(t)-h(t) is taken as the new input signal g(t), and steps (2) to (5) are repeated. If the intermediate signal h(t) is a monotonic function or a signal below the set frequency, h(t) is assigned to r M (t), and the iteration process ends.

[0082] (7) Decomposition results: After multiple decompositions, the original acceleration signal is finally decomposed into multiple intrinsic mode functions (IMF) and a residual term. Each IMF represents the oscillation component of the acceleration signal in different frequency bands, reflecting the vibration patterns of the joint module under different working conditions. For example, low-frequency IMFs may correspond to the persistent low-frequency vibration of the joint module, while high-frequency IMFs may correspond to high-frequency vibration caused by impacts or sudden events. The residual term represents the components of the signal that cannot be attributed to any IMF, usually containing long-term trends or noise in the signal. By decomposing these IMFs, the dynamic characteristics of the joint module can be effectively analyzed and fault diagnosis can be performed.

[0083] (8) Subsequent analysis: Each IMF can be further processed through time-frequency analysis (such as Hilbert transform) to extract the frequency characteristics, energy distribution, and time-frequency localization information of the acceleration signal. In the application of joint modules, the Hilbert transform can calculate the instantaneous frequency and amplitude of each IMF, revealing the frequency variation and energy distribution of the vibration signal. This analysis provides more detailed time-frequency characteristics for impact vibration diagnosis, analysis, and monitoring, enabling the system to accurately identify abnormal states in the joint module caused by impacts or vibrations, and further optimizing fault diagnosis and early warning mechanisms.

[0084] For example, Figure 9As shown, encoder 93 employs the absolute value encoding principle, using an optical encoder disk 931 to sample rotational displacement in real time. The optical encoder disk 931 works in conjunction with the signal processing unit 932 to convert the acquired optical signals into digital signals, achieving high-precision detection of the rotation angle. This signal is then transmitted to the control module 8 via the interface module 933, ensuring accurate position information feedback.

[0085] like Figure 10 As shown, after the external power supply is connected, the control module 8 is powered on and issues control commands to the drive module 7. If the drive current output by the drive module 7 is not zero, the current drives the power module 3 to rotate. If the drive current output by the drive module 7 is zero, the braking module 5 receives a power-off command and then achieves precise braking of the power module 3 through an electromagnetic engagement mechanism. The power module 3 outputs mechanical motion to the harmonic reducer 1 through the main shaft 10, achieving high torque output and ensuring stable operation of the system under high load conditions.

[0086] The driving module 7, the control module 8, and the perception module 9 work together precisely to achieve the efficient operation of the integrated joint module. The control module 8 plays a core role in this process, analyzing and adjusting based on real-time feedback data. By integrating the LSTM algorithm, the control module 8 can predict and analyze the long-term dependence of temperature data, accurately controlling the dynamic response of the joint module. At the same time, combined with the adaptive EMD algorithm, the control module can effectively decompose the acceleration signal, identify the key vibration mode in the signal, and adjust the output of the driving module in real time to respond to rapid changes in the system. Through these advanced algorithms, the control module can more accurately adjust the behavior of the driving module 7, achieving efficient and stable motion control, ensuring the optimal performance of the joint module in various working environments. The driving module 7 accurately controls the output and torque transmission of the power module 3 according to the instructions of the control module 8, thereby optimizing the operating efficiency of the joint module and ensuring efficient and stable operation under different working conditions. The perception module 9 includes temperature sensors 91, acceleration sensors 92, and encoders 93, which respectively monitor the temperature changes, vibration conditions, and motion states of the joint module in real time, providing key feedback information. The temperature sensor 91 is responsible for monitoring the temperature changes in the entire joint module space and providing real-time feedback on the temperature conditions of the system to prevent performance degradation or damage to key components due to abnormal temperatures. Unlike the temperature sensor 91, the thermal sensor 333 integrated in the power module 3 is specifically designed to monitor the temperature of the stator 31 coil and provide data feedback to the control module 8, providing important evidence for local thermal management. These two temperature sensors work together to ensure the thermal stability of the joint module globally and locally. The acceleration sensor 92 is used to monitor abnormal vibrations of the joint module and provide real-time impact and vibration data to ensure that the system is not disturbed by external interference during operation, ensuring the stability of the system. The encoder 93 provides high-precision position and speed feedback to ensure the accuracy of the motion trajectory of the joint module and meet the predetermined motion requirements.

[0087] Through this precise cooperation, the perception module 9 provides real-time data support for the control module 8, which adjusts the output of the driving module 7 based on these data, thereby achieving high-precision control, dynamic response, and intelligent operation of the joint module under complex working conditions. The efficient cooperation of each module significantly improves the response speed, stability, and intelligence level of the system, ensuring the accuracy and safety of the joint module under high load and high-speed operation.

[0088] As Figure 11As shown, the shell 2 is composed of a joint shell 21, a joint rear cover 22 and a first bearing 23. Among them, the joint shell 21 is located on the left side, the joint rear cover 22 is located on the right side, and the two are tightly connected by four evenly distributed screws. The first bearing 23 is connected with the joint shell 21 by interference. The joint shell 21 and the joint rear cover 22 form an integral support structure, which ensures that the integrated control module has sufficient strength and rigidity to withstand mechanical loads in various working environments. An observation hole is provided on the joint rear cover 22, which facilitates users to check the running status of the joint module in real time, timely discovers and deals with potential problems, and further enhances the reliability and safety of the overall system. On the left side of the joint shell 21, a first bearing 23 is arranged near the wave generator 11, which effectively supports the main shaft 10 and ensures that the power is transmitted to the harmonic reducer 1 stably and accurately.

[0089] Each functional module is reasonably configured according to the principles of compact structure, performance optimization and convenient maintenance. The harmonic reducer 1 is located at the leftmost end of the module, and high-precision torque transmission is achieved through the meshing of the flexspline 12 and the rigid gear 13. The stator 31 of the power module is bonded to the inner wall of the joint shell 21, and the rotor 32 is fixed on the main shaft 10 and closely linked with the harmonic reducer 1, ensuring efficient and stable power output. The brake module 5 is installed to the right of the power module 3, and precise stop control is achieved through electromagnetic braking. The driving module 7 and the control module 8 are coaxially arranged, effectively shortening the axial size and optimizing the utilization rate of the internal space. The encoder 93 in the sensing module 9 is located at the rear end of the main shaft 10, which is used to provide high-precision position feedback; the temperature sensor 91 is bonded to the left end face of the power module rear cover 4, which is used to monitor the operating temperature in real time; the acceleration sensor 92 is fixed on the right end face of the fixed flange 6 to monitor the impact and vibration. The circuit of each module adopts a layered wiring scheme, and the signal lines and power lines are arranged separately, which reduces signal interference and improves electrical reliability. Through the design of the wiring groove 41, the reasonable distribution and efficient management of the circuit are ensured, thereby realizing the organic combination between the modules and the optimal cooperation of the functions.

[0090] As described above, the present application can be well implemented, and the above-mentioned embodiments are only preferred embodiments of the present application, but not used to limit the implementation range of the present application; that is, all equivalent changes and modifications made according to the content of the present application are covered by the scope of the claims of the present application.

Claims

1. A sensor-driven and control integrated joint module, characterized in that, Comprise: A shell for mounting a harmonic reducer, a power module, a power module rear cover, a brake module, a fixed flange, a drive module, a control module, a sensing module and a main shaft; The main shaft is rotatably supported on the shell and the power module rear cover; The power module is used to drive the rotation of the main shaft by means of electromagnetic excitation; The power module rear cover is provided with a wiring slot and has a gap with the power module; The brake module is used to generate a braking force to lock the main shaft; The fixed flange is provided with a wiring slot and is used to position the drive module and the control module; The sensing module is used to detect the temperature, vibration and position signal of the joint module and transmit it to the control module; The control module is used to control the power module, the brake module and realize the regulation of the position, speed and torque of the joint module; The harmonic reducer is connected with the main shaft, and the power module is output through the main shaft and the increased output torque; The harmonic reducer is composed of a wave generator, a flexible gear and a rigid gear; the rigid gear is fixed with the shell, the wave generator is connected with the main shaft and supported in the rigid gear; the flexible gear is supported in the rigid gear and cooperates with the rigid gear; The wave generator is composed of an elliptical cam and a flexible bearing; the flexible gear is composed of an outer gear and a connecting flange, the rigid gear includes a rigid gear ring and a rigid gear fixed seat; the elliptical cam is sleeved with the main shaft, the elliptical cam is supported in the rigid gear ring through the flexible bearing, the rigid gear ring is fixed in the rigid gear fixed seat, the rigid gear fixed seat is connected with the shell, the outer gear is fixed with the connecting flange and cooperates with the rigid gear ring; the elliptical cam and the flexible bearing adopt gap fit; The power module rear cover includes a second bearing and a fixed frame; the fixed frame is provided with a wiring slot, the second bearing is connected with the inner hole of the center of the fixed frame by interference, and is used to support the main shaft; The brake module is composed of an electromagnet, an armature, a friction plate, a brake seat and a brake housing; the electromagnet is fixed with the power module rear cover and connected to the drive module through two shielded communication lines; the electromagnet includes a coil and a spring, the armature is connected with the spring, the friction plate is placed between the armature and the brake housing; the square outer wall of the brake seat and the square inner hole of the friction plate adopt interference fit, and the inner hole of the brake seat is connected with the main shaft; The sensing module is composed of a temperature sensor, an acceleration sensor and an encoder; the temperature sensor is fixed on the end face of the power module rear cover, the control module processes temperature data based on LSTM neural network algorithm and real-time feedback joint module temperature change.

2. The force and motion integrated joint module of claim 1, wherein, The power module is composed of a stator, a rotor and an integrated circuit board; the stator is fixed with the shell, and the rotor is fixed with the main shaft; the integrated circuit board provides power control and signal processing function for the power module, and is provided with a Hall sensor and a thermal sensor; respectively used to monitor the position of the rotor and the temperature of the power module in real time.

3. The force and motion integrated joint module of claim 1, wherein, The LSTM neural network comprises: A forgetting gate for deciding which temperature information should be forgotten from the cell state; An input gate for controlling the influence degree of the current input information when updating the cell state; An output gate for deciding which information of the current time hidden state should be outputted; The acceleration sensor is fixed on the end face of the fixed flange to detect impact and vibration data of the joint module. The control module uses an empirical mode decomposition algorithm to analyze the vibration signal, specifically including: (1) Signal preprocessing: denoising and normalization processing of the original acceleration signal collected from the acceleration sensor; (2) Local extreme point extraction: extracting all local maximum and minimum points from the preprocessed acceleration signal; (3) Constructing Envelope: Through spline interpolation on local maximum points e upper ( t ) and local minimum points e lower ( t ), the upper envelope and the lower envelope of the signal are constructed; (4) Calculate the average value m of the envelope t : Calculate the average value of the upper envelope and the lower envelope to obtain the local average envelope of the signal; (5) Envelope removal: subtract the local mean envelope from the original acceleration signal to obtain the intermediate signal h ( t ) (6) if the intermediate signal h ( t ) does not satisfy the definition of IMF, then h ( t ) is taken as a new input signal g ( t ) and steps (2) to (5) are repeated; If the intermediate signal h ( t ) satisfies the definition of IMF, it is extracted as an IMF component, and the h ( t )- X ( t )- h ( t ) is taken as a new input signal g ( t ), and steps (2) to (5) are repeated. If the intermediate signal h ( t ) is a monotonic function or a signal below a set frequency, then h ( t ) is assigned to the residual, while the iteration process is ended. (7) Decomposition result: the original acceleration signal is finally decomposed into multiple IMFs and a residual term, each IMF represents the oscillation component of the acceleration signal in different frequency bands, reflecting the vibration mode of the joint module under different working conditions; (8) Each IMF can be further processed through time-frequency analysis to extract the frequency characteristics, energy distribution and time-frequency localization information of the acceleration signal; The encoder is fixed with the main shaft and is responsible for measuring the position and speed of the main shaft.

4. The force and motion integrated joint module of claim 1, wherein, The driving module, the control module and the encoder are fixed with each other through three evenly distributed studs in the circumferential direction.

Images