Torque feedback control method for quick-release self-locking connector

Abstract

The invention discloses a torque feedback control method for a quick-release self-locking connector, and relates to the technical field of mechanical connection intelligent control, and the method comprises the following steps: deploying an electromagnetic sensitivity detection unit, dynamically capturing the electromagnetic abnormal response of any node in a torque feedback path of the connector, and building an interference pulse feature library; and based on the interference pulse feature library, constructing a multi-channel time domain and frequency spectrum cooperative sensing model, identifying amplitude fluctuation, frequency drift and phase mutation of the torque feedback signal, and generating an interference signal mark set. By dynamically adjusting the moment judgment threshold value, the force application rate and the stepping amplitude and combining force application response and signal authenticity evaluation, a self-adaptive adjustment mechanism of force application and feedback signals is established, the force application accuracy and safety are improved, structural damage and connection failure are effectively prevented, and the method is suitable for high-strength and high-precision assembly and has a wide application prospect. And the stability and the reliability of connection operation are obviously improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent control of mechanical connection, and particularly relates to a torque feedback control method for quick-release self-locking connectors. BACKGROUND

[0002] The torque feedback control of quick-release self-locking connectors refers to a mechanical connection device capable of quick assembly and disassembly, automatic locking and anti-loosening, and an intelligent control method based on real-time monitoring and dynamic feedback adjustment of torque. During use, the connector is automatically locked by the built-in self-locking mechanism (such as wedge locking, elastic buckle or thread anti-loosening structure) after connection is completed, preventing loosening or falling off caused by vibration, load impact or long-term use. At the same time, the control method performs real-time sensing and monitoring on the torque change during connection and disassembly, and dynamically compares it with the preset locking or disassembly torque threshold, forming a closed-loop feedback regulation to ensure reliable locking each time and not to damage the connecting parts, and to avoid component damage or jamming caused by overloading during disassembly. The scheme is widely used in industrial robot end replacement, aerospace component quick assembly, automotive intelligent manufacturing, medical device quick docking and other fields, significantly improving the connection efficiency, safety and intelligent level.

[0003] The prior art has the following disadvantages: In an industrial application environment with strong electromagnetic interference and complex wireless signals, factors such as strong magnetic field interference, lightning induction, pulse electromagnetic radiation and high-frequency electromagnetic wave interference exist externally, which can easily cause unpredictable occasional interference distortion to the torque sensor signal acquisition and transmission link in the connector torque feedback control system. When the feedback signal is distorted or deviates from the true value due to interference, the torque information received by the control system will not match the actual torque state, and thus the torque calculation and discrimination algorithm will produce a false judgment in the dynamic locking or disassembly process, and cannot accurately reflect the actual stress condition of the current connection or disassembly.

[0004] Under this false judgment condition, the control system will incorrectly judge that the connection has not reached the preset locking torque threshold, and thus continue to apply excessive torque to the connector, causing serious structural damage such as thread tearing, fitting part crushing, local stress concentration of the connection interface and material plastic deformation of the connecting parts. Especially in high-strength connection or precision assembly scenarios, such damage is hidden and irreversible, and once formed, the connection site will be damaged even if the locking is completed, and the subsequent disassembly process will be difficult to separate or completely damaged due to connection failure, jamming or material fatigue, and even the risk of permanent scrap of the connection site.

[0005] The above information disclosed in the background section is only used to enhance the understanding of the background of the present disclosure, and therefore it can include information that does not constitute prior art known to those of ordinary skill in the art. SUMMARY

[0006] The purpose of the present application is to provide a quick-release self-locking connector torque feedback control method, which dynamically adjusts the torque determination threshold, force application rate and step size, combines force application response and signal authenticity evaluation, establishes an adaptive adjustment mechanism for force application and feedback signal, improves force application accuracy and safety, effectively prevents structural damage and connection failure, is suitable for high-strength and high-precision assembly, significantly improves the stability and reliability of connection operation, and solves the problems in the background art.

[0007] In order to achieve the above purpose, the present application provides the following technical scheme: a quick-release self-locking connector torque feedback control method, comprising the following steps: S100: deploying an electromagnetic sensitivity detection unit to dynamically capture electromagnetic abnormal responses of any node in the connector torque feedback path, and establishing an interference pulse feature library; S200: based on the interference pulse feature library, constructing a multi-channel time domain and frequency spectrum collaborative perception model to identify the amplitude fluctuation, frequency drift and phase mutation of the torque feedback signal, and generating an interference signal marker set; S300: according to the interference signal marker set, performing synchronous cleaning and double reconstruction on the torque feedback signal, generating a authenticity evaluation index based on the difference residual of the signals before and after reconstruction, and excluding interference components; S400: based on the authenticity evaluation index, performing dynamic threshold correction to adjust the upper and lower limits of the torque determination threshold to avoid continuous overload force application; S500: according to the dynamic threshold correction result, scheduling the force application rate and step size, verifying the locking effect step by step, and recording the dynamic changes of the force application response sequence and the authenticity evaluation index; S600: based on the dynamic changes of the force application response sequence and the authenticity evaluation index, dynamically adjusting the force application increment and the force application rate, balancing the force application intensity and the stability of the torque feedback signal, continuously optimizing through the coupling relationship between force application and feedback signal, and realizing torque closed-loop control in the connection process.

[0008] Preferably, step S100 comprises: After deploying the electromagnetic sensitivity detection unit, a signal real-time acquisition mechanism is established to perform multi-channel synchronous acquisition on the electromagnetic abnormal responses of any node in the torque feedback path; Performing frequency spectrum analysis and transient pulse capture on the collected electromagnetic abnormal signals to extract frequency components, waveform features, amplitude size and time duration; Based on signal feature extraction and classification algorithm, extracting main frequency band range, peak amplitude, pulse rising edge and falling edge steepness, duration and periodicity characteristics, combining node position and physical environment, and classifying interference pulse sources; According to the classification result, an interference pulse feature library is constructed, and the feature library is continuously verified and updated through dynamic monitoring and similarity matching.

[0009] Preferably, the step S200 comprises: Based on the interference pulse feature library, a multi-channel signal parallel perception link is deployed to collect time domain waveform and frequency domain characteristics of the torque feedback signal; The collected signal is subjected to wavelet transform and frequency spectrum analysis to extract frequency variation and amplitude fluctuation, and Hilbert transform is used to extract the envelope and phase sequence of the signal; Based on phase difference analysis and sliding window detection, phase mutation is monitored to form a multi-dimensional feature vector of signal anomaly; According to the similarity of the feature vector and the interference template, a dynamic interference signal label set is generated.

[0010] Preferably, the step S300 comprises: According to the dynamic interference signal label set, the interference section of the torque feedback signal is divided, and adaptive filtering and band-stop filtering algorithms are applied to perform signal cleaning to form a cleaning result; The cleaning result is subjected to multi-scale reconstruction method of wavelet packet decomposition to repair the signal continuity in frequency and time dimensions to obtain a first reconstructed signal; Based on the first reconstructed signal, an intrinsic mode function reconstruction method of empirical mode decomposition is used to screen and recombine intrinsic mode components to obtain a second reconstructed signal; The difference residual is calculated between the original signal and the second reconstructed signal, and a signal authenticity evaluation index is generated based on the residual statistical characteristics.

[0011] Preferably, the step S400 comprises: Based on the signal authenticity evaluation index, a sensitivity mapping model with threshold dynamic correction is constructed to form a nonlinear mapping relationship between the authenticity index and the threshold correction amplitude; According to the mapping model, a dynamic threshold correction coefficient is calculated in real time, and combined with the original judgment standard, a dynamic torque judgment upper and lower limit is generated; The dynamic threshold is applied to each torque discrimination decision to continuously compare the force feedback signal with the dynamic threshold; The dynamic threshold adjustment history and authenticity index change are recorded during the force application process, and the threshold correction model parameters are dynamically optimized according to the cumulative data.

[0012] Preferably, the step S500 comprises: According to the dynamic threshold correction result, the force application rate and step size are set, and the force application parameters are dynamically adjusted based on the feedback signal authenticity evaluation index; After each force application, the feedback torque is compared with the dynamic threshold upper and lower limits to verify whether the locking criterion is met, and the force application rate and step size are dynamically adjusted. The dynamic changes of the force application response sequence and the signal authenticity evaluation index are recorded in real time, and the force application control strategy is continuously optimized. When the feedback signal is within the dynamic threshold upper and lower limits, a fine force application mode is automatically triggered to reduce the force application rate and step size, and the locking threshold is approached step by step.

[0013] Preferably, when the feedback signal is within the dynamic threshold upper and lower limits, a fine force application mode is automatically triggered to reduce the force application rate and step size, and the locking threshold is approached step by step, specifically including: In the fine force application mode, the physical reasonableness of the torque increase and the signal stability after each force application are dynamically verified by increasing the feedback sampling frequency and signal authenticity evaluation. If the feedback signal continuously approaches the dynamic threshold upper and lower limits, the force application rate and step size are continuously tightened until the feedback torque is stable within the dynamic threshold range and continuously maintained for a set period, and the locking is determined to be completed.

[0014] Preferably, step S600 includes: The dynamic changes of the force application response sequence and the signal authenticity evaluation index are monitored in real time, and the authenticity index change rate, feedback signal volatility, force application and feedback correlation, and dynamic trend are analyzed. According to the dynamic change trend, the progression and slow-release rhythm of the force application are dynamically adjusted, and the force application increment and force application rate decrease amplitude are controlled. The influence of force application adjustment on the stability of the feedback signal is calculated in real time, and a dynamic mapping relationship between the force application and the feedback signal is established. By continuously learning the dynamic coupling relationship between the force application and the feedback signal, the force application adjustment strategy is optimized, and closed-loop intelligent adjustment of the torque force application process is realized.

[0015] In the above technical solutions, the present application provides technical effects and advantages: The present application dynamically adjusts the torque determination threshold, the force application rate and the step size, combines the dynamic evaluation of the force application response and the signal authenticity, forms a self-adaptive coupling adjustment mechanism between the force application and the feedback signal, effectively improves the force application accuracy and safety of the connection process, reduces the risk of structural damage, thread tearing or mating surface crushing of the connection part caused by excessive force application, realizes intelligent dynamic closed-loop control of the connection operation, and is particularly suitable for industrial application scenarios with high strength, high reliability and high precision assembly requirements, greatly improving the stability, reliability and service life of the connection operation. BRIEF DESCRIPTION OF DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the accompanying drawings needed in the embodiments will be briefly introduced as follows. Obviously, the accompanying drawings in the following description only represent some embodiments described in the present application, and for those skilled in the art, other drawings can also be obtained from these drawings.

[0017] Figure 1 Method flow chart of the torque feedback control method of the quick-release self-locking connector of the present application. DETAILED DESCRIPTION

[0018] Example implementations will now be described more fully with reference to the accompanying drawings. Example implementations may, however, be implemented in many different forms and should not be construed as limited to the examples set forth herein; rather, these example implementations are provided so that this disclosure will be thorough and complete, and will fully convey the scope of example implementations to those skilled in the art.

[0019] The present application provides a torque feedback control method for a quick-release self-locking connector as shown in Figure 1 The torque feedback control method for a quick-release self-locking connector includes the following steps: S100: Based on the torque feedback path of the connector, a composite electromagnetic sensitivity detection unit is deployed to dynamically capture the electromagnetic abnormal response amplitude and response frequency of any node in the torque sensor and signal transmission chain, and to establish an interference pulse feature library; In view of the problem that the torque feedback signal of the quick-release self-locking connector is affected by electromagnetic interference in complex industrial application environments, based on the torque feedback path of the connector, the method of deploying a composite electromagnetic sensitivity detection unit includes the following steps: According to the torque feedback path structure of the connector and the physical topology of the signal transmission chain, all key signal nodes between the torque sensor and the control signal receiving end are determined, including the signal output interface of the torque sensor, the input and output ends of the signal amplification and conditioning circuit, the interface of the digital-to-analog or analog-to-digital conversion circuit, and the physical connection points and distributed connection sections of the signal transmission line. Each of the above nodes is analyzed for electromagnetic sensitivity, and the sensitivity of each node to different types of electromagnetic interference is evaluated according to the interference characteristics of the strong magnetic field source, high-frequency electromagnetic wave source, lightning induction source and pulse radiation source in the node's surrounding environment. According to the sensitivity evaluation results, different types of electromagnetic perception sensors are configured for different nodes, including Hall elements sensitive to static and low-frequency magnetic fields, wideband antenna probes sensitive to radio frequency electromagnetic waves, and fast-response electric field sensing devices for pulse electromagnetic radiation, to ensure that all possible interference source types and characteristics of the signal nodes in the torque feedback path are covered.

[0020] For various types of electromagnetic sensing sensors that have been deployed, a signal real-time acquisition mechanism is established to continuously and dynamically capture the electromagnetic anomaly response at each signal node. The collected electromagnetic anomaly signals include but are not limited to magnetic field strength changes, electric field strength changes, frequency characteristics, response amplitude, and transient pulse morphology. During the acquisition process, to ensure the integrity and timing consistency of the data, a multi-channel synchronous acquisition method is used, and each sensor signal is marked with a timestamp to establish a time correlation of the full-link electromagnetic response. To further enhance the identification capability of the interference source, high-precision spectrum analysis and transient pulse capture algorithms are introduced in the signal acquisition stage, which can extract the frequency components, waveform characteristics, amplitude, and time persistence of the interference pulse from the collected signals.

[0021] The spectrum analysis and transient pulse capture algorithm is a means for analyzing the frequency, time domain morphology, and transient characteristics of the interference signal using mathematical and signal processing methods. In this embodiment, spectrum analysis is used to convert the collected raw electromagnetic signals from the time domain to the frequency domain, usually using the Fast Fourier Transform (FFT) method, to obtain the amplitude distribution and main frequency band interval of the interference signal at each frequency segment, thereby accurately determining the frequency components and energy concentration area of the interference pulse. The transient pulse capture algorithm focuses on the mutation and short-time changes of the interference signal in the time domain, and uses methods such as sliding window and wavelet transform to detect the instantaneous waveform mutation, rising edge and falling edge steepness of the signal in real time, and capture the start time, peak amplitude and duration of the pulse. The specific steps include: Bandwidth limitation and noise filtering of the original electromagnetic sensing signal to retain the signals of the key interference frequency band; Applying Fast Fourier Transform to analyze the spectrum of the filtered signal and obtaining the amplitude-frequency characteristics in the frequency domain; Using wavelet transform to analyze the time-frequency joint of the instantaneous energy change of the signal, to identify the transient components in the signal that change rapidly, and to capture the time position and instantaneous waveform characteristics of the pulse; Combining the extraction results of the frequency domain and the time domain to form a multi-dimensional feature set of the interference pulse, including frequency components, peak amplitude, pulse width, rising and falling edge rates, and duration, for subsequent interference identification and classification archiving, thereby fully depicting the dynamic characteristics of electromagnetic interference.

[0022] Based on the collected full-link electromagnetic anomaly response data, the signal feature extraction and classification algorithm is used to extract the corresponding feature parameter set of different types of interference pulses, including the main frequency band range, peak amplitude, steepness of the pulse rising and falling edges, duration and periodicity characteristics, etc., and combined with the known conditions of node position and physical environment, the source of the interference pulse is further classified and identified. By summarizing and modeling the massive interference pulses, an interference pulse feature library is constructed, which is stored in layers according to the type, spectral characteristics, time domain response and potential impact on torque feedback signal of electromagnetic interference, forming a sustainable and dynamically expanding interference pulse data resource. The feature library not only contains the static templates of interference pulses, but also contains the dynamic response patterns of interference signals under different working conditions and different nodes to support the accuracy of subsequent interference identification and signal cleaning.

[0023] The signal feature extraction and classification algorithm refers to the algorithm method of automatically extracting the parameter set describing the essential features of the signal by analyzing the collected electromagnetic interference signals, and classifying and identifying the source of different types of interference pulses based on these features. In this scheme, its role is to convert complex and diverse electromagnetic interference signals into identifiable and distinguishable feature parameter sets, so as to accurately judge the interference type and possible interference source. This algorithm can be realized based on existing algorithms for feature extraction such as principal component analysis (PCA), wavelet packet decomposition (WPD), short-time Fourier transform (STFT), and classification algorithms such as support vector machine (SVM), K-nearest neighbor (KNN), random forest (RF) or convolutional neural network (CNN). The specific steps are as follows: For the captured interference pulse signals, wavelet packet decomposition and short-time Fourier transform are used to extract their frequency and time domain comprehensive features, such as main frequency band range, amplitude spectrum, energy density, etc. Based on the transient changes of the pulse in the time domain, the steepness of the waveform is extracted by calculating the change rate of the rising and falling edges of the pulse, and the duration and periodicity characteristics of the pulse are also calculated. Combined with the distribution characteristics of the signal at each node of the full link and the corresponding physical environment information, such as the node position close to the strong magnetic source or high frequency power supply, the joint feature vector of signal features and physical environment is formed. The above feature vector is input into a classification model such as support vector machine or random forest, and the source and type of the interference pulse are automatically classified and identified according to the trained interference sample library, forming a complete interference pulse feature label for subsequent archiving and dynamic updating of the feature library.

[0024] To ensure the dynamic and practicality of the interference pulse feature library, a verification and updating mechanism of the interference pulse feature library is established. By continuously monitoring the newly appeared electromagnetic abnormal response signals in the actual working process of the connector, and performing similarity matching with the pulse features in the existing feature library, if new interference features are found or the response amplitude, frequency, etc. of the existing features change significantly, the corresponding entries of the feature library are updated or new records are added. In this way, it is ensured that the interference pulse feature library can continuously reflect the latest changes and evolution trends of electromagnetic interference in industrial application environment, and enhance the full-link perception and protection capability of torque feedback signal interference in complex electromagnetic environment.

[0025] The role of this step is to provide a complete electromagnetic interference perception and feature modeling mechanism for the torque feedback control system for the full-link signal path, solving the torque feedback signal distortion and misjudgment problems caused by complex electromagnetic interference in industrial environment. In the actual application process of the connector, the signal path from the torque sensor to the control signal receiving end is often exposed to strong magnetic field, high frequency electromagnetic wave, lightning induction and pulse electromagnetic radiation, etc. These interference factors have randomness, concealment and time-varying, and often exist in multiple forms, superimposed, multi-band, nonlinear characteristics coexist, which can easily cause distortion, drift or even mutation of the feedback signal at any node of the signal chain, resulting in misjudgment of the torque feedback control algorithm for locking state or disassembly state, causing connection failure or structure damage. Therefore, by deploying a composite electromagnetic sensitivity detection unit in the full-link of the torque feedback path, the response of each node to different electromagnetic interference can be dynamically, comprehensively and real-time captured, forming accurate records of the amplitude, frequency, time characteristics, etc. of the interference signal. Further, by establishing an interference pulse feature library, it not only provides a basis support for subsequent interference signal identification, cleaning and signal authenticity evaluation, but also can be continuously updated and expanded, forming an electromagnetic interference knowledge system suitable for different industrial application scenarios, realizing the perception, identification and management of torque feedback signal full-link interference, and ensuring the stability and reliability of torque control.

[0026] S200: Based on the interference pulse feature library, a multi-channel time domain and frequency spectrum collaborative perception model of torque feedback signal is constructed, amplitude fluctuation, frequency drift and phase mutation in the torque feedback signal are identified, and a dynamic interference signal label set is generated; Based on the established interference pulse feature library, a multi-channel time domain and frequency spectrum collaborative perception model of torque feedback signal is constructed for the stability requirement of torque feedback signal in the working process of the connector, which includes the following steps: According to the multi-dimensional characteristic parameters of the main frequency band, amplitude range, pulse duration, rising edge and falling edge characteristics of various interference signals summarized in the interference pulse characteristic library, a multi-channel signal parallel sensing link is deployed for the collection path of the torque feedback signal. Each channel is configured with a high sampling rate data acquisition unit, which is used to record the time domain original waveform and frequency domain characteristics of the torque feedback signal respectively. The time domain channel continuously records the change trend of the signal amplitude with high time resolution, and captures the amplitude fluctuation and mutation of the signal in real time; the frequency domain channel synchronously performs fast Fourier transform processing on the collected signal to generate continuous frequency spectrum, and identifies whether there is a frequency drift or abnormal signal energy peak in the frequency band in the signal that matches the interference pulse characteristic library.

[0027] For the time domain signal and frequency domain signal collected by multiple channels, a time-frequency joint analysis mechanism is established, the frequency components of the torque feedback signal at different time scales are decomposed by wavelet transform, the frequency variation and energy distribution in a specific time period are accurately extracted, and the frequency drift is determined whether it exists and matches the interference characteristics recorded in the interference pulse characteristic library. For the time domain signal, envelope detection and variation coefficient calculation method are applied to identify the abnormal fluctuation amplitude and change rate of the signal amplitude in the local time period to detect the amplitude anomaly related to the interference pulse. Through time-frequency collaborative analysis, the change characteristics of time domain and frequency domain are one-to-one mapped to form the comprehensive perception results of the torque feedback signal in time and frequency dimensions.

[0028] On the basis of time-frequency joint analysis, the phase characteristics of the signal are further analyzed in detail, the Hilbert transform is adopted to extract the envelope and phase of the torque feedback signal, the continuity and abruptness of the signal phase in the time sequence are monitored, and the phase mutation characteristics caused by electromagnetic interference are identified. The detection results of amplitude fluctuation, frequency drift and phase mutation are summarized to form a multi-dimensional feature vector of signal anomaly detection. Through similarity calculation of each type of interference template in the interference pulse characteristic library, the matching degree of the current feedback signal abnormal characteristics and the known interference pulse is evaluated, whether the abnormal signal is caused by the interference pulse is judged, and the type and intensity level of the interference are identified.

[0029] Hilbert transform is a mathematical method for signal analysis, which can convert real signals into complex signals, construct the analytical form of the signal to contain the amplitude and phase information of the original signal, and then extract the envelope and instantaneous phase of the signal. In this step, the role of Hilbert transform is to dynamically monitor the phase change of the torque feedback signal in the time domain to identify the signal phase mutation caused by electromagnetic interference, so as to judge the influence of interference on the continuity of the feedback signal. The specific steps include: A Hilbert transform is applied to the collected original torque feedback signal to convert the signal from a real number domain to a complex number domain to form an analytic signal including a real part of the original signal and an imaginary part with a phase shift of 90 degrees; Based on the analytic signal, an envelope thereof is calculated, that is, an instantaneous amplitude variation of the signal at each time point is obtained by square root of a sum of squares of the real part and the imaginary part, so as to facilitate observation of an energy profile of the signal under interference; An instantaneous phase is calculated by using the analytic signal, specifically, a phase sequence of the signal on a time sequence is obtained by taking an inverse tangent function of the imaginary part and the real part, to form a complete phase trajectory; Based on the phase sequence, a phase difference analysis and a sliding window detection method are used to dynamically monitor continuity of the phase with time, to identify abnormal mutation or discontinuous jump of the phase in a short time, to judge whether the signal is affected by electromagnetic interference in combination with envelope variation of the mutation point and a characteristic parameter of a known interference pulse, and to classify a type and intensity of the interference. In this way, not only is the signal amplitude and the phase double-dimension accurately analyzed, but also the phase stability destruction of the torque feedback signal when the signal is interfered is effectively captured, to provide a key basis for subsequent signal authenticity evaluation and interference cleaning.

[0030] Based on the above feature vector and the interference discrimination result, a dynamic interference signal marking set is generated. The marking set records a time of occurrence, a duration, an influence amplitude and a corresponding interference type identification of amplitude fluctuation, frequency drift and phase mutation in each torque feedback signal, and simultaneously associates a connector working state and a torque change when the signal anomaly occurs, to form a complete dynamic interference marking archive. The dynamic interference signal marking set is not only used as an input basis for subsequent signal cleaning and reconstruction steps, but also updates and iterates in real time, to gradually optimize accuracy and sensitivity of interference discrimination, to realize accurate monitoring and dynamic adaptation of the torque feedback signal in a complex electromagnetic interference environment, and to thereby guarantee reliability and safety of the connector in various working scenarios.

[0031] The step is to construct a multi-channel time domain and spectrum collaborative perception model of torque feedback signal, to realize the accurate identification of various abnormal features in the torque feedback signal affected by electromagnetic interference, and to mark and archive these abnormalities in the form of dynamic interference signal marker set, laying a foundation for subsequent signal cleaning, reconstruction and control threshold dynamic adjustment. In complex industrial application environment, torque feedback signal often causes signal amplitude fluctuation, frequency drift and phase mutation due to multi-source electromagnetic interference, and single time domain or frequency domain analysis method is difficult to fully reveal the influence of interference on the full dimension of the signal. By introducing a multi-channel perception mechanism, the amplitude change of torque feedback signal in time domain, the frequency component and energy distribution in frequency domain, and the phase continuity and abruptness in phase domain are perceived and analyzed in parallel to form a full-dimensional monitoring of time-frequency-phase. Combined with the established interference pulse feature library, the perception model can extract and compare the feature parameters of each captured signal abnormal feature, such as the change rate of abnormal amplitude, the drift amplitude and position of frequency in spectrum, and the mutation gradient and amplitude of phase change, to judge whether the abnormality conforms to the known interference pulse feature mode. When it is confirmed that the abnormality has high similarity with the interference feature library, the corresponding dynamic interference signal marker is generated in the torque feedback signal, and the occurrence time, influence duration, signal node position and interference type of the interference are recorded. Through this step, not only the full perception and identification of torque feedback signal in electromagnetic interference environment are realized, but also the dynamic updating and traceable interference marker data are formed, providing key perception and data support for the signal dynamic cleaning, control parameter adaptive adjustment and connection safety guarantee of the connector torque control system.

[0032] S300: According to the dynamic interference signal marker set, the torque feedback signal is executed with synchronous cleaning and double reconstruction processing, and the signal authenticity evaluation index is generated based on the difference residual of the signals before and after reconstruction, and the interference distortion components in the torque feedback signal are screened out; According to the dynamic interference signal marker set, the torque feedback signal is executed with synchronous cleaning and double reconstruction processing, and the signal authenticity evaluation index is generated based on the difference residual of the signals before and after reconstruction, and the interference distortion components in the torque feedback signal are screened out, which specifically includes the following steps: Based on the dynamic interference signal marker set generated in the previous stage, the complete time sequence and frequency components of the torque feedback signal are segmented and divided, and the time position, duration, corresponding frequency range and phase mutation characteristics of each disturbed signal segment indicated by the marker set are clearly marked. For each marked interference signal segment, an adaptive filtering method and a band-stop filtering algorithm are applied to specifically filter and suppress the frequency drift, amplitude anomaly and phase mutation in the marked segment, and remove the distorted signal components corresponding to the interference characteristic frequency band, amplitude mutation and phase jump, forming the first-stage signal cleaning result. The goal of this stage is to reduce noise and suppress interference in the signal segment affected by known interference through signal preprocessing, ensuring that the undisturbed part of the original torque feedback signal is not affected.

[0033] Based on the signal after signal cleaning, the first signal reconstruction processing is performed. The first reconstruction uses a multi-scale reconstruction method based on wavelet packet decomposition to decompose the cleaned signal into multiple wavelet subspaces, and by reconstructing the sub-signals at different scales, the disturbed signal segment is informationally completed in both time and frequency dimensions, restoring the continuity and integrity of the signal in low and high frequency components. This process realizes the continuity repair of the waveform, frequency and phase of the interference segment by combining the characteristics of the adjacent signal segments that are not disturbed in the multi-layer wavelet decomposition and reconstruction of the signal, improving the overall stability and restoration of the signal.

[0034] The multi-scale reconstruction method of wavelet packet decomposition is a signal processing technology that aims to realize detailed analysis and restoration of signals in full-band and multiple levels through decomposition and reconstruction of signals at different frequency scales, especially suitable for processing complex and non-stationary signals. In this step, the method is used to comprehensively analyze the frequency and time of the cleaned torque feedback signal, effectively repair the signal components lost in each frequency level due to filtering or cleaning in the disturbed segment, and restore the continuity and physical integrity of the signal in the low and high frequency domains. The specific steps are as follows: Apply wavelet packet decomposition method to the cleaned torque feedback signal, select a mother wavelet function that adapts to the signal characteristics, and set the decomposition level to decompose the signal into multiple wavelet subspaces layer by layer, each subspace corresponding to a different frequency range, realizing multi-level division of the full-band signal. The key to selecting the mother wavelet function that matches the signal characteristics lies in the mathematical characteristics of the wavelet function that need to match the time-frequency characteristics, variation rules and local mutation characteristics of the signal to be analyzed. For the torque feedback signal, it usually contains low-frequency components that grow smoothly, high-frequency oscillations when disturbed, and local waveform mutations caused by transient interference. Therefore, when selecting the mother wavelet function, priority should be given to its good time-frequency localization ability, moderate orthogonality and sensitivity to signal mutations. Specifically, if the signal is strong in smoothness and changes gently, a high-order function in the Daubechies (db) wavelet, such as db8 or db10, can be selected to ensure high-fidelity capture of smooth changes in the signal; if the signal has frequent local mutations or transient interference, Symlet (sym) or Coiflet (coif) wavelets are preferred, as these wavelet functions are more sensitive to signal edges and mutation points. In addition, for cases where both wide frequency characteristics and non-stationary components exist in the signal, the adaptability of the mother wavelet can be evaluated by calculating the energy distribution, reconstruction error and signal-to-noise ratio of the original signal after decomposition by different mother wavelets, and the optimal mother wavelet function can be determined through simulation testing and cross-validation to ensure that the wavelet packet decomposition process retains the main characteristics of the signal while removing the interference components to the greatest extent.

[0035] For each subspace signal, identify the time window corresponding to the cleaned interference section, and use the adjacent undisturbed signal segment to locally repair the missing or distorted signal segment through interpolation completion, smooth fitting and other methods, ensuring the time continuity and spectral consistency of the signal at each scale; The repaired signals of each wavelet subspace are transformed back to the time domain according to the reconstruction algorithm of the wavelet packet, and the complete torque feedback signal curve is synthesized, realizing seamless fusion and restoration of different frequency components; Smooth and trend fit the reconstructed signal as a whole to further eliminate the phase shift or amplitude mutation that may be introduced during the reconstruction process, ensuring the continuity and consistency of the signal in low and high frequency components. Through this method, the signal discontinuity or frequency information loss caused by interference cleaning can be repaired without losing important signal details, realizing high-fidelity restoration of the torque feedback signal and ensuring the accuracy and stability of the subsequent control logic.

[0036] Based on the first re-reconstruction result, further perform the second re-signal reconstruction processing, adopt the intrinsic mode function reconstruction method based on empirical mode decomposition, decompose the cleaned signal into several intrinsic mode components again, and screen and recombine these components according to the physical characteristics and dynamic characteristics of the signal, so as to reconstruct the signal curve closer to the real moment feedback change rule. Through the second re-reconstruction, not only the reconstruction error and local distortion in the first re-reconstruction are corrected, but also the adaptability of the signal to nonlinear and non-stationary interference is enhanced. After completing the double reconstruction, the original signal and the double reconstructed signal are compared, and the difference residual of the two in the whole time sequence is calculated. The smaller the residual is, the higher the authenticity of the original signal is, and vice versa.

[0037] The intrinsic mode function reconstruction method of empirical mode decomposition is an adaptive signal decomposition and reconstruction technology for nonlinear and non-stationary signals. Its core is to decompose the original signal through iterative local characteristic scale decomposition, extract a group of intrinsic mode function components (IMF) with different frequency characteristics and physical significance, and then realize signal noise reduction and feature enhancement by screening and recombining these modal components, and restore the signal curve more consistent with the actual physical process. In this step, the function of this method is to perform second deep feature extraction and repair on the moment feedback signal after preliminary cleaning and wavelet packet reconstruction. By decomposing the intrinsic mode components reflecting the different dynamic characteristics of the signal, removing the distortion components introduced by residual interference or cleaning error, and further improving the authenticity and physical continuity of the signal. The specific steps include: Apply the empirical mode decomposition algorithm to the cleaned moment feedback signal, iteratively find the upper and lower envelope lines according to the local extreme points of the signal, calculate the local average value of the signal and continuously iterate, and decompose the intrinsic mode component sequence arranged from high frequency to low frequency layer by layer, each modal component represents the characteristic change of the signal at a certain time scale; According to the physical characteristics and dynamic change rule of the signal, combined with the frequency range, amplitude level and change stability that the moment feedback signal should show in the actual physical process, the modal components with too high frequency, abnormal amplitude or violent change are screened out, which are often non-physical components introduced by residual interference or cleaning error; Recombine and superimpose the screened modal components in the order from high frequency to low frequency, gradually restore the composite signal curve consistent with the actual dynamics of the moment feedback signal, and perform trend smoothing and boundary correction on the reconstructed signal to ensure its consistency with the real physical process in time, frequency and phase change; The reconstructed signal is compared with the original cleaning signal and the first double reconstructed signal, and through analysis of amplitude recovery degree, frequency stability and phase continuity and other indicators, the authenticity and integrity of the signal are verified, and the reconstructed signal is taken as the basis for subsequent torque feedback control and dynamic threshold adjustment. Through this method, the lost nonlinear characteristics and dynamic information in the cleaning and initial reconstruction process can be effectively compensated, and the response accuracy and reliability of the torque feedback signal to the actual connection or disconnection process changes can be improved.

[0038] Based on the statistical characteristics of the difference residual, such as mean square error, maximum residual amplitude and coefficient of variation of residual, a signal authenticity evaluation index is established to comprehensively evaluate whether the signal at each time point is authentic and reliable. This evaluation index not only makes a quantitative judgment on the authenticity of the current signal, but also is used in the subsequent control logic to dynamically adjust the trust weight of the signal, ensuring that the torque feedback signal is only used as the basis for control command generation and torque adjustment when the signal authenticity reaches a certain threshold.

[0039] The purpose of this step is to address the distortion, drift and distortion of the torque feedback signal in the electromagnetic interference environment. Through dynamic interference signal marking set, the contaminated signal segment is accurately cleaned and double reconstructed, thereby restoring the authenticity and integrity of the signal. The authenticity and reliability of the signal are objectively evaluated through the difference residual between the reconstructed signal and the original signal, thereby effectively filtering out the interference components and providing reliable data support for subsequent torque control. In complex industrial environments, torque feedback signals often produce irregular signal abnormalities such as amplitude mutation, frequency shift or phase instability due to transient electromagnetic interference or frequency interference. If these abnormalities are not processed and directly involved in control decisions, it is easy to misjudge the connection state or miscontrol the force, causing irreversible damage such as connector thread tearing, local crushing or material fatigue. Therefore, this step first uses the dynamic interference signal marking set to accurately locate the signal segment affected by interference and implement targeted synchronous cleaning to remove the distorted signal components in the marked segment. On the basis of cleaning, the signal is double reconstructed through multi-scale reconstruction of wavelet packet decomposition and intrinsic mode function reconstruction method of empirical mode decomposition, respectively restoring and completing the signal from the frequency scale and dynamic characteristic level, and maximizing the restoration of the physical nature and dynamic change characteristics of the signal. Through difference residual calculation of the original signal and the double reconstructed signal, the authenticity of each time point signal can be quantified, forming a signal authenticity evaluation index to ensure that the signal is only used for feedback control when the signal authenticity reaches a certain threshold. This step effectively solves the problems of distortion compensation and information loss after signal cleaning, ensuring that the torque feedback signal still has high fidelity and high reliability under interference, providing a solid data foundation and judgment basis for intelligent dynamic closed-loop control of connectors.

[0040] S400: Based on the signal authenticity evaluation index, the dynamic threshold correction of the torque feedback control is performed, the upper and lower limits of the torque determination threshold are adjusted in real time, the torque discrimination standard is corrected, and the continuous overload force is avoided due to interference distortion; Based on the signal authenticity evaluation index, the dynamic threshold correction of the torque feedback control is performed, and the upper and lower limits of the torque determination threshold are adjusted in real time to correct the torque discrimination standard, prevent continuous overload force due to interference distortion, and specifically include the following steps: According to the signal authenticity evaluation index generated after the double reconstruction processing in the previous stage, a sensitivity mapping model for threshold dynamic correction is constructed. The mapping model forms a nonlinear mapping relationship between the authenticity index and the threshold correction amplitude by modeling the correlation between the signal authenticity evaluation index and the historical torque feedback signal under different interference levels. Through learning of historical signal samples, the model can identify the amplitude and direction of the torque determination threshold that should be dynamically adjusted when the authenticity evaluation index is in different confidence intervals. For example, when the authenticity index is close to the full score, indicating that the signal interference is extremely low, the torque determination threshold can be strictly limited within the set standard upper and lower limits; on the contrary, if the authenticity index is below the warning threshold, the upper and lower limits of the torque determination threshold will be automatically relaxed to improve the tolerance to signal uncertainty to prevent misjudgment overload caused by signal distortion.

[0041] According to the mapping model, the dynamic correction coefficient of the determination threshold under the current torque feedback signal is calculated in real time, and the original torque locking and dismounting determination standard is combined to generate dynamic torque determination upper and lower limits that take effect immediately. This process applies the correction coefficient to the amplitude and interval of the original determination threshold, so that the dynamic threshold not only adjusts according to the authenticity index of the current signal, but also responds to the change trend and fluctuation intensity of the torque signal in real time. For example, when the torque signal fluctuates intensively in a short time and the authenticity index is low, the determination upper limit will be increased simultaneously to prevent premature determination of locking completion caused by false signal peaks or continuous force caused by false signal valleys.

[0042] The corrected dynamic threshold is applied to the torque feedback control process, and the dynamic comparison is performed for each torque discrimination decision. Specifically, in each step cycle of torque application, the real-time torque feedback signal and the dynamically adjusted upper and lower limit thresholds are determined to determine whether the currently applied torque meets the optimal conditions for locking or dismounting. If the feedback signal fluctuates within the dynamic threshold, the current force application rhythm is maintained, if the feedback signal is stable within the dynamic threshold interval for a plurality of cycles, it is determined that locking or dismounting is completed, the force application is stopped, and the final torque value and force application timing are recorded; if the feedback signal continuously deviates from the threshold range and the authenticity index is low, the threshold will be further adjusted and the force application rate will be reduced to prevent overload.

[0043] The adjustment history of the dynamic threshold, the real-time index change of each force feedback, and the discrimination result are continuously recorded during the force application process, and the parameters of the threshold correction model are dynamically optimized according to the accumulated data. Through the closed-loop iteration of continuous force application and feedback, a set of threshold adjustment mechanism that evolves adaptively with the working environment and interference state is formed, ensuring that under different interference intensities and different connection working conditions, the upper limit of the force and the judgment standard can always be accurately controlled based on the cooperation of signal authenticity and dynamic threshold correction, completely avoiding the overloading of force, the damage of connection or the failure of disassembly caused by signal interference distortion.

[0044] The role of this step is to dynamically adjust the upper and lower limits of the judgment threshold in the torque feedback control by introducing the signal authenticity evaluation index, ensuring that when the feedback signal is interfered or distorted, the control system can intelligently adapt to the credibility of the feedback signal, avoiding the risk of force misjudgment and continuous overload caused by signal abnormalities. In the torque feedback control process, the traditional judgment method usually relies on a fixed torque judgment threshold. When the feedback signal appears error or noise interference, the fixed threshold cannot adapt to the change of signal authenticity, which is easy to cause the force to be continuously increased without reaching the torque requirement of locking or disassembly, and finally cause irreversible damage such as thread tearing, mating surface crushing or material plastic deformation of the connected parts. Through this step, the torque judgment standard is dynamically corrected in real time according to the signal authenticity evaluation index, and the upper and lower limits of the judgment threshold are flexibly adjusted according to the authenticity of the feedback signal. When the signal authenticity evaluation index is high, it means that the feedback signal is close to the real state, at this time the standard judgment threshold is maintained to ensure the accuracy of locking or disassembly discrimination; when the authenticity index is low, the feedback signal has distortion risk, the judgment threshold range is automatically expanded, and the force rate and control force amplitude are reduced to prevent the occurrence of overload force. The introduction of dynamic threshold makes the torque feedback control have real-time adaptive ability, which is no longer dependent on static threshold judgment, but flexibly discriminates according to the dynamic credibility of the feedback signal, effectively ensuring the safety of force application and the reliability of connection under strong electromagnetic interference or complex working environment, and further improving the intelligent control level of the connector and the practicality of industrial application.

[0045] S500: According to the dynamic threshold correction result, the force rate and step amplitude of the torque feedback control are scheduled in real time, the locking effect is verified step by step, the force stability of the connection process is enhanced, and the dynamic changes of the force response sequence and the signal authenticity evaluation index are recorded; According to the dynamic threshold correction result, the force rate and step amplitude of the torque feedback control are scheduled in real time, the locking effect is verified step by step, the force stability of the connection process is enhanced, and the dynamic changes of the force response sequence and the signal authenticity evaluation index are recorded; According to the latest results of dynamic threshold correction, the initial rate and step size of the current force are set in real time. The force rate refers to the rate of change of the torque applied per unit time, and the step size refers to the incremental torque applied per force application cycle. The initial setting of the force parameters is determined according to the authenticity evaluation index of the feedback signal. If the authenticity evaluation index of the current signal is high, it indicates that the signal is reliable, and the force rate and step size are executed according to the preset standard larger value to improve the connection efficiency; if the authenticity evaluation index is low, it indicates that the signal may be disturbed or distorted, and the force rate is automatically adjusted downward, and the step size is reduced, so that the torque is applied in a small progressive manner to prevent the connection from being overloaded or damaged due to the fast force caused by the untruthful signal.

[0046] After each force step is completed, the feedback torque signal is immediately compared with the dynamically corrected upper and lower thresholds to verify whether the current force has approached or reached the locking effect determination standard. During the verification process, not only is it detected whether the current feedback torque enters or exceeds the dynamic threshold range, but also the current signal authenticity evaluation index and the torque growth trend of the previous and next two cycles are combined for joint determination. If the feedback torque grows steadily and the authenticity index does not fluctuate significantly, it is determined that the signal is true, and the current force rate and step size are continued to execute the next cycle of force. If the feedback signal is close to the threshold, but the authenticity index decreases or the signal fluctuation increases, the dynamic adjustment of the force rate and step size is triggered immediately, the force increment rate is reduced, and the torque progressive size is reduced to ensure that the feedback signal is stable and true before the force is applied. The blind force causes local fatigue or structural deformation of the connection.

[0047] In each cycle of force and verification, the force response sequence is recorded in real time, including the applied torque value, corresponding feedback torque value, force rate, step size, dynamic threshold upper and lower limits used for determination, authenticity evaluation index of the feedback signal, and time stamp corresponding to the force. The sequence not only reflects the dynamic change of the torque during the force application process, but also retains the historical trajectory of the signal quality evaluation and control parameters during the entire process. Through continuous accumulation of the force response sequence, the torque force characteristics curve of the connection under specific environmental and interference conditions can be formed in subsequent operations, providing a data basis for future force strategy optimization and model iteration.

[0048] According to the dynamic change trend of the force response sequence and the signal authenticity evaluation index, the force control strategy is continuously adjusted. When the connection process approaches the locking judgment critical value or the feedback signal appears near the edge state of the dynamic threshold, the fine force mode is automatically triggered, further reducing the force rate and step size, and using small, gradual and slow force to accurately approach the locking threshold. This way avoids the sudden force when the locking is close to completion, preventing overlocking or damage. At the same time, after the entire connection process is completed, a complete force response sequence and authenticity evaluation record is formed for subsequent review and performance optimization.

[0049] The implementation of the fine force mode is based on the results of dynamic monitoring and threshold judgment of the torque feedback signal in the connection process. When it is detected that the feedback torque has approached the upper and lower limits of the dynamic judgment threshold or the signal is in the judgment critical state, the fine force mode is automatically switched to. In this mode, the force rate and step size are greatly reduced by a preset reduction ratio, for example, 30% or less of the original force rate as the new force speed, and the step size is reduced to 20% or less of the normal mode, to achieve more detailed torque increment. After each force, the physical reasonableness of the current torque growth and signal stability are verified through higher frequency feedback sampling and signal authenticity evaluation, and the force amplitude of the next step is dynamically adjusted. If the feedback signal is still approaching the threshold, the force parameters are continuously maintained or further tightened until the torque is stable in the dynamic judgment threshold range and continuously maintained for a set period, and finally the locking is completed. Through the progressive, small-amplitude force of the fine force mode and high-frequency feedback verification, it is ensured that before the locking is completed, the overload, plastic deformation of the connection material or assembly precision decline caused by excessive force or speed is avoided, and the mechanical integrity and use reliability of the connector are guaranteed.

[0050] The role of this step is to dynamically schedule the force application rate and step size in the torque feedback control process by modifying the results according to the dynamic threshold, and to verify the locking effect step by step to ensure the stability and safety of the force application process during connection, and to realize the dynamic recording and tracking of the response data and signal authenticity throughout the process. In actual connection or locking process, torque feedback signals are often affected by electromagnetic interference or mechanical noise, resulting in fluctuations in signal stability and authenticity. At this time, if the force application rate or step size is too large, it is easy to produce overload force in the case of uncertain feedback signals, which may not only cause excessive stress on the threads or mating parts of the connector, but also cause local material damage, plastic deformation or assembly failure. Therefore, through dynamic threshold correction, real-time judgment results are obtained to intelligently schedule the force application rate and step size. When the signal authenticity is high, the force is quickly advanced; when the authenticity is low or the signal is close to the dynamic judgment threshold, the force application rate and step size are automatically reduced, and the force is applied in a small, gradual and slow manner to ensure that each force is realized within the safety range. At the same time, through the verification of the locking effect after each force, it can be dynamically detected whether the feedback signal after the force meets the threshold condition, and the force strategy is corrected in real time to avoid the risk of excessive force at one time. In addition, the dynamic changes of force response sequence and signal authenticity evaluation index throughout the process not only help to trace and verify the force process, but also form a dynamic database of force and feedback, providing support for control strategy optimization and model training under different connection conditions and different interference intensities, thereby improving the adaptability and intelligent level of control in complex application environments.

[0051] S600: Based on the dynamic change trend of the force response sequence and the signal authenticity evaluation index, dynamically adjust the progress and slow-release rhythm of the force, control the decrement amplitude of the force increment and the force rate in real time, continuously balance the force intensity and the stability of the torque feedback signal, and continuously learn and optimize the dynamic coupling relationship between the force and the feedback signal to realize intelligent dynamic closed-loop control of torque in the connection process; Based on the dynamic change trend of the force response sequence and the signal authenticity evaluation index, dynamically adjust the progress and slow-release rhythm of the force, control the decrement amplitude of the force increment and the force rate in real time, continuously balance the force intensity and the stability of the torque feedback signal, and continuously learn and optimize the dynamic coupling relationship between the force and the feedback signal to realize intelligent dynamic closed-loop control of torque in the connection process, comprising the following steps: In the process of connecting force, the dynamic change trend of force response sequence and signal authenticity evaluation index is continuously monitored in real time. The force response sequence records the torque value, force rate, feedback torque, authenticity score of feedback signal and change process of dynamic threshold correction at each force progression. The signal authenticity evaluation index reflects the reliability and interference degree of the current feedback signal in real time. Through continuous analysis of these data, the change rate of authenticity index and the stability response of feedback signal caused by force are captured to form the dynamic change trend between force behavior and feedback signal, and to determine whether the current force is in the positive interval of stable torque rise or in the risk interval of feedback signal authenticity fluctuation and signal stability decline caused by too fast force.

[0052] In this process, the analysis of data mainly includes the following categories: change rate analysis of authenticity index, fluctuation analysis of feedback signal, correlation analysis of force-feedback and dynamic trend discrimination analysis. The change rate analysis of authenticity index identifies whether the signal authenticity is continuously improved or deteriorated by calculating the increase and decrease amplitude and change gradient of signal authenticity score in continuous multiple force cycles. The fluctuation analysis of feedback signal judges whether the feedback signal is in a stable interval or produces abnormal fluctuations based on the amplitude standard deviation, fluctuation frequency and phase continuity of feedback torque. The correlation analysis of force-feedback measures the direct effect of force progression on feedback by statistical response ratio of force increment and feedback torque change, and identifies the sensitivity and hysteresis between force and feedback. The dynamic trend discrimination analysis dynamically draws the three-dimensional change curve between force intensity, feedback torque and authenticity index based on sliding window and time series fitting algorithm, and discriminates whether the current force stage is in the positive interval of stable torque rise or in the risk interval of feedback instability due to too fast force or signal quality decline. Through these four types of data analysis, the safety of force process and the reliability of feedback signal can be dynamically perceived to guide the intelligent adjustment of subsequent force parameters.

[0053] According to the above dynamic change trend, the dynamic adjustment mechanism of force progression and slow release rhythm is started. When it is detected that the authenticity index of feedback signal continuously rises and the feedback torque growth caused by force progression is in a steady state, the force increment and force rate are appropriately increased to speed up the force rhythm and improve the efficiency of locking process. On the contrary, when the authenticity index decreases or the feedback signal produces severe fluctuations after force, or the signal is close to the edge of dynamic judgment threshold, the force rate is immediately reduced, the force increment is reduced, and a small amplitude torque slow release operation is inserted, that is, a small amplitude attenuation of force amplitude is inserted in the force progression rhythm, so that the connection part obtains a short buffer of stress release in the force process, avoiding local fatigue or material strain hardening of the connecting part caused by continuous unidirectional force accumulation, thereby ensuring the stability and continuity of feedback signal.

[0054] While dynamically adjusting the progression and slow-release rhythm of force application, the influence degree of force adjustment on feedback signal stability is calculated in real time, and the response relationship between force increment, force rate decrement amplitude and feedback signal change is mapped. Through this mapping relationship, the dynamic coupling model between force and feedback signal is continuously optimized, and the sensitivity and response law of force adjustment on feedback signal stability under different materials, different connection conditions and different interference environments are identified. According to the authenticity recovery degree and stability level of the feedback signal after each force adjustment, the force control parameters are dynamically fine-tuned, and a force adjustment strategy that adapts to the dynamic changes of the feedback signal is formed.

[0055] Through continuous learning and iterative optimization of the dynamic coupling relationship between the foregoing force and feedback signal, a set of adaptive force and feedback dynamic control strategy is gradually formed, realizing the closed-loop intelligent adjustment of the torque force process. This adaptive adjustment strategy not only can automatically adjust the force rhythm according to the real-time feedback, but also can gradually accumulate experience data in multiple connection operations. Through long-term learning of the force response sequence and authenticity index, the balance point of force strength and signal stability is optimized, ensuring that the connection process always maintains high coordination and dynamic stability between force and feedback while achieving the goal of locking or disassembling. Ultimately, the intelligent dynamic closed-loop control of torque in the connection process is realized, effectively improving the safety, accuracy and stability of the connection operation.

[0056] The role of this step is to dynamically adjust the progress and slow-release rhythm in the force application process through real-time analysis of the dynamic change trend of the force response sequence and the signal authenticity evaluation index, accurately control the increment of force and the decrement amplitude of force rate, realize the dynamic balance between force intensity and feedback signal stability, and ensure the continuity, safety and authenticity of the feedback signal in the connection process. In the actual connection process, there is a complex dynamic coupling relationship between the force application process and the feedback signal. If the force rhythm is too fast and the increment is too large, it is easy to cause stress concentration or material hardening in the mechanical structure of the connection part, which causes the feedback signal to fluctuate sharply or even distort, and affects the authenticity evaluation index of the signal, causing the control system to misjudge the force state and produce the risk of overloading or insufficient force. Through this step, based on the feedback signal change caused by each torque progression in the force response sequence, combined with the rising, falling or fluctuation trend of the signal authenticity index, the force increment and decrement rhythm can be intelligently adjusted to avoid signal distortion caused by continuous force, and at the same time, the force can be accelerated in the stage of stable and high authenticity feedback signal to improve the efficiency of locking or disassembly. By continuously learning the dynamic coupling relationship between force and feedback signal, it can adapt to the differences of different connection materials, structural stiffness and electromagnetic interference environment, optimize the force strategy, realize the self-adaptation and closed-loop adjustment of the torque control process, and ensure that each step of the force is based on the reliable signal, thereby improving the accuracy, stability and safety of the connection operation.

[0057] Through the torque feedback control method of the quick-release self-locking connector described above, the whole process monitoring, dynamic interference identification and cleaning reconstruction of the torque feedback signal of the connector in the industrial environment with strong electromagnetic interference and complex wireless signal can be realized, the authenticity and stability of the feedback signal are ensured, and the force misjudgment and continuous overload caused by signal distortion are avoided. The method dynamically adjusts the torque determination threshold, the force rate and the step amplitude, combines the dynamic evaluation of the force response and the signal authenticity, forms an adaptive coupling adjustment mechanism between the force and the feedback signal, effectively improves the force accuracy and safety of the connection process, reduces the risk of structural damage, thread tearing or mating surface crushing of the connection part caused by overload force, realizes the intelligent dynamic closed-loop control of the connection operation, and is especially suitable for industrial application scenarios with high strength, high reliability and high precision assembly requirements, greatly improving the stability, reliability and service life of the connection operation.

[0058] The above only describes certain exemplary embodiments of the present application by way of illustration, and it is self-evident that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present application. Therefore, the above figures and description are illustrative in nature and should not be construed as limiting the scope of the claims of the present application.

Claims

1. A torque feedback control method for quick-release self-locking connectors, characterized in that, The method comprises the following steps: S100: deploying an electromagnetic sensitivity detection unit, dynamically capturing electromagnetic abnormal responses of any node in the connector torque feedback path, and establishing an interference pulse feature library; S200: based on the interference pulse feature library, constructing a multi-channel time domain and frequency spectrum collaborative perception model, identifying amplitude fluctuations, frequency drifts and phase mutations of the torque feedback signal, and generating an interference signal marker set; S300: according to the interference signal marker set, performing synchronous cleaning and double reconstruction on the torque feedback signal, generating a reality evaluation index based on the difference residual of the signal before and after reconstruction, and filtering out interference components; S400: based on the reality evaluation index, performing dynamic threshold correction, and adjusting the upper and lower limits of the torque determination threshold; S500: according to the dynamic threshold correction result, scheduling the force applying rate and step size, verifying the locking effect step by step, and recording the dynamic changes of the force applying response sequence and the reality evaluation index; S600: based on the dynamic changes of the force applying response sequence and the signal reality evaluation index, dynamically adjusting the force applying increment and the force applying rate, balancing the force applying strength and the stability of the torque feedback signal, continuously optimizing through the coupling relationship between the force applying and the feedback signal, and realizing the torque closed-loop control of the connection process.

2. The torque feedback control method for quick-release self-locking connector according to claim 1, wherein Step S100 comprises: After deploying the electromagnetic sensitivity detection unit, a signal real-time acquisition mechanism is established, and the electromagnetic abnormal responses of any node in the torque feedback path are synchronously acquired in multiple channels; Performing frequency spectrum analysis and transient pulse capture on the acquired electromagnetic abnormal signals, extracting frequency components, waveform features, amplitude size and time continuity; Based on the signal feature extraction and classification algorithm, the main frequency band range, peak amplitude, pulse rising edge and falling edge steepness, duration and periodicity characteristics are extracted, combined with the node position and physical environment, and the interference pulse source is classified; According to the classification result, an interference pulse feature library is constructed, and the feature library is continuously verified and updated through dynamic monitoring and similarity matching.

3. The torque feedback control method for quick-release self-locking connector according to claim 2, wherein Step S200 comprises: Based on the interference pulse feature library, a multi-channel signal parallel perception link is deployed to acquire the time domain waveform and frequency domain characteristics of the torque feedback signal; Performing wavelet transform and frequency spectrum analysis on the acquired signal, extracting frequency variation and amplitude fluctuation, and combining Hilbert transform to extract the envelope and phase sequence of the signal; Based on phase difference analysis and sliding window detection, the phase mutation is monitored to form a multi-dimensional feature vector of signal anomaly; According to the similarity of the feature vector and the interference template, a dynamic interference signal marker set is generated.

4. The torque feedback control method for quick-release self-locking connector according to claim 1, wherein Step S300 comprises: According to the dynamic interference signal marker set, the disturbed section of the torque feedback signal is divided, the adaptive filtering and band-stop filtering algorithm is applied to perform signal cleaning, and the cleaning result is formed; A multi-scale reconstruction method of wavelet packet decomposition is applied to the cleaning result to repair the signal continuity in frequency and time dimensions, and a first reconstructed signal is obtained; Based on the first reconstructed signal, an intrinsic mode function reconstruction method of empirical mode decomposition is adopted to select and recombine the intrinsic mode components, and a second reconstructed signal is obtained; The difference residual between the original signal and the second reconstructed signal is calculated, and a signal reality evaluation index is generated based on the residual statistical characteristics.

5. The torque feedback control method for quick-release self-locking connector according to claim 1, wherein, Step S400 comprises: Based on the signal authenticity evaluation index, a sensitivity mapping model with threshold dynamic correction is constructed to form a nonlinear mapping relationship between the authenticity index and the threshold correction amplitude; According to the mapping model, the dynamic threshold correction coefficient is calculated in real time, and the dynamic torque judgment upper and lower limits are generated combined with the original judgment standard; Apply the dynamic threshold to each torque discrimination decision, and continuously compare the force feedback signal with the dynamic threshold; Record the dynamic threshold adjustment history and authenticity index changes during the force application process, and dynamically optimize the threshold correction model parameters based on the cumulative data.

6. The torque feedback control method of quick release self-locking connector according to claim 5, wherein, Step S500 includes: According to the dynamic threshold correction result, set the force application rate and step size, and dynamically adjust the force application parameters based on the feedback signal authenticity evaluation index; After each force application, compare the feedback torque with the dynamic threshold upper and lower limits to verify whether the locking judgment standard is reached, and dynamically adjust the force application rate and step size; Real-time record the dynamic changes of force response sequence and signal authenticity evaluation index, and continuously optimize the force control strategy; When the feedback signal is within the dynamic threshold upper and lower limits, automatically trigger the fine force application mode, reduce the force application rate and step size, and gradually approach the locking threshold.

7. The torque feedback control method for quick-release self-locking connector according to claim 6, wherein, When the feedback signal is within the dynamic threshold upper and lower limits, automatically trigger the fine force application mode, reduce the force application rate and step size, and gradually approach the locking threshold, which specifically includes: In the fine force application mode, after each force application, increase the feedback sampling frequency and signal authenticity evaluation to dynamically verify the physical reasonableness and signal stability of the torque growth after force application; If the feedback signal continuously approaches the dynamic threshold upper and lower limits, continuously tighten the force application rate and step size until the feedback torque is stable within the dynamic threshold range and continuously maintains for a set period, and the locking is determined to be completed.

8. The torque feedback control method for quick-release self-locking connector according to claim 7, wherein, Step S600 includes: Real-time monitor the dynamic changes of force response sequence and signal authenticity evaluation index, analyze the authenticity index change rate, feedback signal volatility, force and feedback correlation, and dynamic trend; According to the dynamic change trend, dynamically adjust the progression and slow-release rhythm of the force, control the force increment and force rate decrease amplitude; Real-time calculate the influence of force adjustment on feedback signal stability, establish the dynamic mapping relationship between force and feedback signal; Through continuous learning of the dynamic coupling relationship between force and feedback signal, optimize the force adjustment strategy, and realize the closed-loop intelligent adjustment of the torque force application process.

Images