System and method for monitoring real-time working state of intelligent sewing machine

Abstract

The invention discloses a system and a method for monitoring the real-time working state of an intelligent sewing machine, and relates to the technical field of sewing machine running state monitoring. By comparing an actual ratio with a task preset ratio, a generated synchronous error set can accurately reflect the stability state of a feeding action in a sewing task; and further extracted deviation level parameters, fluctuation amplitude parameters and fluctuation trend characteristics are counted through a sliding window, so that the numerical trend of synchronous deviation between a feeding system and a needle bar system can be dynamically quantified, and subsequent trend modeling of fluctuation behaviors can be supported. The stitch density change score obtained through calculation can evaluate the stitch density stability of the current sewing task section in real time, abnormal section recognition and dynamic marking are carried out in combination with density tolerance threshold parameters, and a finally formed abnormal marking result set enables the system to have traceable, quantifiable and linkable density abnormal state sensing capacity.

Description

Technical Field

[0001] This invention relates to the field of sewing machine operation status monitoring technology, specifically to a real-time monitoring system and method for the working status of an intelligent sewing machine. Background Technology

[0002] With the continuous integration of advanced manufacturing technologies and automated control methods, the level of intelligence in industrial equipment is constantly improving, and intelligent manufacturing has become an important development direction in the modern industrial system. Among them, the textile and apparel industry, as a typical representative of the upgrading of labor-intensive industries to automation and precision, has seen its core processing equipment, sewing machines, gradually become a key target for the intelligent transformation of automated equipment.

[0003] In current industrial sewing equipment, although some high-end sewing machines have functions such as electronically controlled feeding, servo drive, and electronic pattern control, the judgment of the synchronization status between the feeding system and the needle bar movement generally still remains at the level of coarse-grained rhythm matching or hardware limit monitoring. Especially in complex pattern sewing or multi-segment sewing speed changes, slight timing deviations can easily occur between the frequency of the feeding motor and the rhythm of the needle bar movement. This deviation will directly cause uneven distribution of stitch density per unit length, resulting in process problems such as visual pattern misalignment, and localized overly dense or sparse stitches.

[0004] Currently, most mainstream methods rely on preset fixed beat ratios or indirect judgment based on stitch sensor triggering. They lack real-time modeling and analysis of the actual difference between the feed frequency and the needle bar's up-and-down movement. Especially under conditions of fabric elasticity changes, tension disturbances, and slight speed drift, existing monitoring mechanisms cannot capture the dynamic process of stitch density shifts in a timely and accurate manner. This means that during normal sewing, the equipment may remain in a slightly abnormal state for extended periods without being detected, only revealing the problem during the finished product quality inspection stage, leading to the risk of production rework or rework. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a monitoring system and method for the real-time working status of an intelligent sewing machine, solving the problems mentioned in the background section.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for monitoring the real-time working status of an intelligent sewing machine, comprising the following steps: S1. Through the operation signal acquisition module in the main control system, the feeding frequency parameters output by the feeding controller and the needle bar cycle parameters acquired by the needle bar motion encoder during the sewing process are obtained to construct a basic synchronization data set. S2. Perform frame-by-frame pairing analysis on the synchronization data set, calculate the cycle ratio between the feed frequency parameter and the needle bar cycle parameter under each sewing cycle, form the actual cycle value, and derive the theoretical ratio based on the preset parameters of the sewing task to obtain the expected cycle value, calculate its synchronization difference parameter, and output the synchronization error set. S3. Perform sliding window statistical processing on the synchronization error set, and output the deviation level parameter, fluctuation amplitude parameter and fluctuation trend characteristics respectively to form the error feature set; S4. Based on the error feature set, perform fitting calculations on the deviation level parameter, fluctuation amplitude parameter, and fluctuation trend feature to generate a line density change score. S5. Compare the stitch density change score with the density tolerance threshold parameter set for the sewing task to generate an anomaly marker result set, and synchronously feed back the density status information of the current sewing task segment based on the anomaly marker result set.

[0007] Preferably, S1 includes S11 and S12; S11. The main control system calls and runs the signal acquisition module to sample the control signal output by the sewing feed controller in each sewing beat cycle, and extract the feed frequency change signal sequence. The feed frequency change signal sequence is processed by bandpass filtering for noise reduction and periodic rhythm segmentation recognition to obtain the frequency expression data in each beat segment. Based on the frequency expression data, the average frequency value of the cycle in the feeding frequency change signal sequence is calculated, and the average frequency value of the cycle is used as the feeding frequency parameter. S12. By synchronizing the signal with the needle bar motion encoder, the needle bar motion signal is received in real time during the sewing process, and the motion time period within the complete cycle is extracted. The time interval of the edge change of the needle bar motion signal is calculated, and the cycle time used to complete the up and down motion of the needle bar within the current beat is calculated based on the time interval to form the needle bar cycle parameter. The needle bar cycle parameter is used to represent the actual time required for the needle bar to complete one up-and-down reciprocating motion within a single sewing cycle. Using the cycle period of the needle bar cycle parameter as the time division benchmark, the feeding frequency parameter and the needle bar cycle parameter synchronously collected within the same cycle are paired one-to-one to form paired data pairs and construct a basic synchronous data set.

[0008] Preferably, S2 includes S21 and S22; S21. For each pair of paired data items in the basic synchronization data set, perform frame-by-frame extraction and sequentially read the corresponding feed frequency parameter and needle bar cycle parameter. Using the feed frequency parameter as the divisor and the needle bar cycle parameter as the reference variable, calculate the actual synchronization ratio and name the calculation result as the actual beat value. At the same time, extract the preset reference value of the sewing machine's corresponding sewing beat from the sewing task settings, process the preset reference value, and obtain the expected beat value. The expected beat value is used to represent the theoretical synchronization ratio that should exist between feed and needle movement under the current task parameters.

[0009] Preferably, in step S22, the periodic error difference between the actual value of the beat and the expected value of the beat is calculated with the needle bar movement cycle as the beat, to obtain the synchronization difference parameter; The synchronization difference parameters under all needle bar action cycles are accumulated and stored, and organized into structured sequence data according to the time sequence, and output as a synchronization error set.

[0010] Preferably, S3 includes S31 and S32; S31. Construct a fixed-length time sliding window structure based on the synchronization error set, and perform statistical analysis on the continuous synchronization difference parameters in each window of the set time sliding window to obtain the deviation level parameter, fluctuation amplitude parameter and fluctuation trend characteristics. The deviation level parameter is obtained by calculating the arithmetic mean of all synchronization difference parameters; The fluctuation amplitude parameter is obtained by the absolute difference between the maximum and minimum values ​​among all synchronization difference parameters within a statistical time sliding window. The fluctuation trend feature is based on the synchronous difference parameter sequence within the time sliding window, performs a linear trend fitting, and calculates the fluctuation trend feature based on the mean of the squared residuals of the fitted curve.

[0011] Preferably, in step S32, the difference level parameter, the fluctuation amplitude parameter, and the fluctuation trend feature are organized according to the time order of their respective windows to construct a structured statistical result set, named the error feature set; the error feature set, as a statistical set in the form of a time series, records the numerical state and trend characteristics of error changes within each window of the time sliding window.

[0012] Preferably, S4 includes S41; S41. Based on the error feature set, extract the deviation level parameter, fluctuation amplitude parameter and fluctuation trend feature in each window, perform normalization preprocessing, and eliminate the dimensional differences between different features. The nonlinear regression fitting method is then used to fit the normalized preprocessed deviation level parameter, fluctuation amplitude parameter and fluctuation trend characteristics to form a numerical output for representing the density fluctuation state, which is marked as the line trace density change score.

[0013] Preferably, S5 includes S51 and S52; S51. Compare the stitch density change score with the density tolerance threshold parameter preset for the sewing task, and construct an anomaly marker result set based on the judgment result. The anomaly marker result set is constructed using the following comparison and judgment method: When the stitch density change score is greater than the density tolerance threshold parameter, it is determined that there is a synchronous imbalance between the feeding frequency and the needle bar action within the window period, which causes the stitch density of the sewing machine to exceed the preset tolerance range and an abnormal density area appears. The window period mentioned in the stitch density change score is extracted synchronously and included in the abnormal marking result set. When the stitch density change score is less than or equal to the density tolerance threshold parameter, there is no synchronous imbalance between the feeding frequency and the needle bar action within the judgment window period, and it is not included in the abnormal marking result set.

[0014] Preferably, in step S52, the abnormal marking result set is used as the basis for state recognition. For the window time period recorded in the abnormal marking result set Tag, the sewing task segment corresponding to the system is determined to be in an abnormal density state; while the unmarked window time period is determined to be in a normal density state. All status information is organized by sewing machine sewing task segments and displayed graphically in the operation interface, thus forming a visual feedback result of the sewing task density status.

[0015] A real-time monitoring system for the working status of an intelligent sewing machine includes a running signal acquisition module, a data structuring and error analysis module, a feature extraction module, a fitting module, and an evaluation feedback module. The operation signal acquisition module collects the feeding frequency parameters output by the feeding controller and the needle bar cycle parameters collected by the needle bar motion encoder during the sewing process to construct a basic synchronization data set; The data structuring and error analysis module performs frame-by-frame pairing analysis on the synchronous data set, calculates the cycle ratio between the feed frequency parameter and the needle bar cycle parameter under each sewing cycle, constitutes the actual cycle value, and derives the theoretical ratio based on the preset parameters of the sewing task to obtain the expected cycle value, calculates its synchronization difference parameter, and outputs the synchronization error set. The feature extraction module performs sliding window statistical processing on the synchronization error set and outputs the deviation level parameter, fluctuation amplitude parameter and fluctuation trend feature respectively, which constitute the error feature set. The fitting module performs fitting operations on the deviation level parameter, fluctuation amplitude parameter and fluctuation trend feature based on the error feature set, and generates a line density change score. The evaluation feedback module compares the stitch density change score with the density tolerance threshold parameter set for the sewing task, generates an anomaly marker result set, and synchronously feeds back the density status information of the current sewing task segment based on the anomaly marker result set.

[0016] This invention provides a system and method for monitoring the real-time working status of an intelligent sewing machine, which has the following beneficial effects: (1) Based on the feeding frequency parameter Fed output by the feeding controller and the needle bar cycle parameter Ncy collected by the needle bar motion encoder, a basic synchronization data set Dat is constructed. On this basis, the frame-by-frame pairing analysis and dynamic error identification of the feeding and needle movement coordination state during the sewing process are completed. By comparing the actual ratio with the task preset ratio, the generated synchronization error set Dif can accurately reflect the stability state of the feeding action in the sewing task. On this basis, the deviation level parameter Dev, fluctuation amplitude parameter Vol, and fluctuation trend feature Res are further extracted through sliding window statistics. This not only dynamically quantifies the numerical trend of the synchronization offset between the feeding system and the needle bar system, but also supports the subsequent trend modeling of fluctuation behavior. The stitch density change score Qsd calculated from this can evaluate the stitch density stability of the current sewing task segment in real time. Combined with the density tolerance threshold parameter Lim, abnormal segments are identified and dynamically marked. The final abnormal marking result set Tag enables the system to have traceable, quantifiable, and linkage density abnormal state perception capabilities.

[0017] (2) By introducing a time sliding window structure based on the time series of the synchronization difference parameter Dlt, not only is the ability to identify local fluctuations in error information improved, but also a multi-dimensional quantitative expression of error change states is realized. Taking each sliding window as a unit, the deviation level parameter Dev, fluctuation amplitude parameter Vol, and fluctuation trend feature Res are calculated respectively, which can be used to identify the error directional shift, error numerical amplitude intensity, and error evolution trend within that time period, thereby establishing a three-dimensional modeling framework for dynamic error behavior in sewing tasks. All statistical parameters are uniformly organized into an error feature set Sta in the time dimension, forming a statistical result set with trend judgment ability and structural continuity. This effectively breaks through the technical bottleneck of traditional methods that can only make judgments based on instantaneous error values ​​and cannot see the error evolution path. It has the ability to continuously track the evolution trend of synchronization deviation on the time axis, which is the key foundation for subsequent scoring function construction and dynamic feedback control.

[0018] (3) A unified scoring function model is constructed based on the error feature set Sta. Furthermore, by normalizing and preprocessing the deviation level parameter Dev, fluctuation amplitude parameter Vol, and fluctuation trend feature Res, and performing nonlinear regression fitting, a stitch density change score Qsd reflecting the density stability state is accurately generated. By comparing the stitch density change score Qsd with the density tolerance threshold parameter Lim set for the sewing task in real time, abnormal density fluctuation segments can be accurately identified, and an abnormal label result set Tag can be dynamically constructed. This label set can not only be refined to specific window time periods, but can also be updated and called in real time during task operation, forming a highly time-series density anomaly identification channel. Furthermore, using the abnormal label result set Tag as the core state identification basis, the density fluctuation situation can be displayed graphically as a visual feedback result of the sewing task density state, improving the operator's intuitive control over the density change trend. Based on the label, frequency conversion compensation or density parameter adjustment commands are triggered, thereby establishing a complete closed-loop structure of abnormal score identification, density state visualization, and adaptive control linkage. Attached Figure Description

[0019] Figure 1 This is a schematic diagram illustrating the steps of a method for monitoring the real-time working status of an intelligent sewing machine according to the present invention. Figure 2 This is a flowchart illustrating a real-time monitoring system for the working status of an intelligent sewing machine according to the present invention. Detailed Implementation

[0020] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0021] Example 1 This invention provides a method for monitoring the real-time working status of an intelligent sewing machine. Please refer to [link / reference]. Figure 1 This includes the following steps: S1. Through the operation signal acquisition module of the main control system, obtain the feeding frequency parameter Fed output by the feeding controller during the sewing process and the needle bar cycle parameter Ncy acquired by the needle bar motion encoder, and construct the basic synchronization data set Dat; S2. Perform frame-by-frame pairing analysis on the synchronization data set Dat, calculate the cycle ratio between the feed frequency parameter Fed and the needle bar cycle parameter Ncy under each sewing cycle, form the actual cycle value Rac, and derive the theoretical ratio based on the preset parameters of the sewing task to obtain the expected cycle value Rex, calculate its synchronization difference parameter Dlt, and output the synchronization error set Dif. S3. Perform sliding window statistical processing on the synchronization error set Dif, and output the deviation level parameter Dev, fluctuation amplitude parameter Vol and fluctuation trend feature Res respectively to form the error feature set Sta. S4. Based on the error feature set Sta, the deviation level parameter Dev, the fluctuation amplitude parameter Vol, and the fluctuation trend feature Res are fitted to generate the line density change score Qsd. S5. Compare the stitch density change score Qsd with the density tolerance threshold parameter Lim set for the sewing task to generate an anomaly marker result set Tag. Based on the anomaly marker result set Tag, synchronously feed back the density status information of the current sewing task segment.

[0022] In this embodiment, a basic synchronization data set Dat is constructed based on the feeding frequency parameter Fed output by the feeding controller and the needle bar cycle parameter Ncy collected by the needle bar motion encoder. Based on this, frame-by-frame pairing analysis and dynamic error identification of the feeding and needle movement coordination states during sewing are performed. By comparing the actual ratio with the task-preset ratio, the generated synchronization error set Dif accurately reflects the stability of the feeding action in the sewing task. Furthermore, by using a sliding window to statistically extract the deviation level parameter Dev, fluctuation amplitude parameter Vol, and fluctuation trend feature Res, not only can the numerical trend of the synchronization offset between the feeding system and the needle bar system be dynamically quantified, but it can also support subsequent trend modeling of fluctuation behavior. The calculated stitch density change score Qsd can evaluate the stitch density stability of the current sewing task segment in real time. Combined with the density tolerance threshold parameter Lim, abnormal segments are identified and dynamically marked. The final abnormal marking result set Tag enables the system to have traceable, quantifiable, and interactive density anomaly perception capabilities.

[0023] The overall approach not only effectively solves the technical problem of existing sewing monitoring systems being unable to distinguish "stitch density anomalies caused by non-human factors," but also achieves a closed-loop feedback capability throughout the entire process from error analysis to density status judgment through parameter scoring models and dynamic fluctuation modeling. This significantly improves the stability of pattern density and process consistency in the sewing process, while also providing accurate and efficient decision-making basis for subsequent automated density adjustment optimization and adaptive control strategies.

[0024] Example 2 This embodiment is an explanation based on Embodiment 1. Please refer to it. Figure 1 Specifically: S1 includes S11 and S12; S11. The main control system calls and runs the signal acquisition module to sample the control signal output by the sewing feed controller in each sewing beat cycle, and extract the feed frequency change signal sequence. The feed frequency change signal sequence is processed by bandpass filtering for noise reduction and periodic rhythm segmentation recognition to obtain the frequency expression data in each beat segment. Based on the frequency expression data, the average frequency value of the feed frequency change signal sequence is calculated, and the average frequency value of the feed frequency is used as the feed frequency parameter Fed; The feeding frequency parameter Fed is used to represent the drive execution frequency characteristics of the feeding actuator in the current cycle, and is one of the basic input parameters for subsequent cycle synchronization ratio analysis. S12. By synchronizing the signal with the needle bar motion encoder, the needle bar motion signal is received in real time during the sewing process, and the motion time period within the complete cycle is extracted. The time interval of the edge change of the needle bar motion signal is calculated, and the cycle time used to complete the up and down motion of the needle bar within the current beat is calculated based on the time interval, forming the needle bar cycle parameter Ncy. The needle bar cycle parameter Ncy is used to represent the actual time required for the needle bar to complete one up-and-down reciprocating motion within a single sewing beat, and is one of the paired input data for beat comparison analysis; Using the cycle period of the needle bar period parameter Ncy as the time division benchmark, the feeding frequency parameter Fed and the needle bar period parameter Ncy synchronously collected within the same cycle are paired one-to-one to form paired data pairs and construct the basic synchronous data set Dat. Each data pair represents the coordination between the feeding state and the needle movement state under one cycle. All paired data pairs are organized according to the cycle sequence to form the basic synchronization data set Dat, which provides structured raw data support for subsequent synchronization analysis.

[0025] It should be noted that the defined actual value of the cycle time, Rac, is not actually a unit to directly express the stitch density, but rather to reflect whether the feeding frequency and the needle bar movement are coordinated and synchronized. This value is essentially a ratio, a means of judging the degree of coordination between the two movements, rather than an absolute quantity in a physical sense.

[0026] Although the "unit" mathematically appears to be "times per second" 2 However, it doesn't focus on the physical unit, but rather on whether this ratio remains within the normal range when the feeding frequency increases or the needle bar movement decreases. This comparison is valid as long as the theoretical value (Rex) used for subsequent comparisons is calculated using the same logic. It's a suitable method for constructing a synchronization index to express whether the rhythm between feeding and needle movement is synchronized.

[0027] In this embodiment, high-precision synchronous acquisition of the feeding frequency change signal sequence and the needle bar action signal is achieved. Furthermore, through periodic segmentation identification and filtering, the acquired signals possess a high degree of structural integrity in both the time and frequency characteristics dimensions. The feeding frequency parameter Fed, calculated based on the feeding frequency change signal sequence, accurately characterizes the execution rhythm of the feeding system within the current cycle. Meanwhile, the needle bar cycle parameter Ncy, calculated from the edge changes of the needle bar action signal, precisely describes the cycle time required for the needle bar to complete one complete action. By using the needle bar cycle parameter Ncy as the cycle alignment benchmark, a one-to-one pairing relationship is established between the two types of parameters within the same cycle and organized into a basic synchronization data set Dat. This not only effectively avoids the data drift problem caused by cycle alignment ambiguity or mismatch in traditional signal acquisition but also significantly improves the timing accuracy and physical correlation strength of subsequent synchronization error calculations. This method ensures that the entire system establishes a stable, traceable, and structurally consistent data foundation from the source, laying a reliable foundation for the analysis of cycle coordination states.

[0028] Example 3 This embodiment is an explanation of Embodiment 1. Specifically, S2 includes S21 and S22. S21. For each pair of paired data items in the basic synchronization data set Dat, perform frame-by-frame extraction and sequentially read the corresponding feed frequency parameter Fed and needle bar cycle parameter Ncy. Using the feed frequency parameter Fed as the divisor and the needle bar cycle parameter Ncy as the reference variable, calculate the actual synchronization ratio and name the calculation result as the actual rhythm value Rac. At the same time, extract the preset reference value of the sewing rhythm corresponding to the sewing machine from the sewing task setting, and after processing the preset reference value, obtain the expected rhythm value Rex. The expected rhythm value Rex is used to represent the theoretical synchronization ratio that should exist between feed and needle movement under the current task parameters. The actual beat value Rac and the expected beat value Rex constitute the basic ratio structure for synchronicity analysis.

[0029] S22. Perform periodic error difference calculation on the actual beat value Rac and the expected beat value Rex with the needle bar action cycle as the beat to obtain the synchronization difference parameter Dlt; The synchronization difference parameter Dlt under all needle bar action cycles is accumulated and stored, and organized into structured sequence data according to the time sequence, and output as a synchronization error set Dif; the synchronization error set Dif serves as the core foundation set for subsequent density fluctuation feature extraction and scoring modeling, and is used to reflect the synchronization stability level of the feeding system and the needle action system during the sewing process.

[0030] In this embodiment, a standardized synchronization comparison structure is constructed between the actual beat value Rac and the expected beat value Rex. Using the feed frequency parameter Fed and the needle bar period parameter Ncy as input variables, the actual beat value Rac, calculated through unified mathematical logic, accurately reflects the true ratio of feed and needle movement coordination during sewing. The expected beat value Rex, extracted from the sewing task, serves as the reference baseline for the theoretical synchronization model. The difference calculation between the two not only constructs a periodic error extraction mechanism but also outputs a synchronization difference parameter Dlt and organizes it into a continuous time series synchronization error set Dif, providing stable, dynamic, and coherent basic data support for subsequent multidimensional error trend analysis and density scoring modeling. This step significantly improves the microscopic detection accuracy of abnormal feed and needle movement coordination, enabling the system to perceive changes in beat synchronization at a higher resolution. It is a key supporting link in solving the problem that traditional methods cannot quantify and dynamically track beat error fluctuations.

[0031] Example 4 This embodiment is an explanation of Embodiment 1. Specifically, S3 includes S31 and S32. S31. Construct a fixed-length time sliding window structure based on the synchronization error set Dif, and perform statistical analysis on the continuous synchronization difference parameter Dlt in each window of the set time sliding window to obtain the deviation level parameter Dev, the fluctuation amplitude parameter Vol, and the fluctuation trend feature Res. The deviation level parameter Dev is obtained by calculating the arithmetic mean of all synchronization difference parameters Dlt, and is used to describe the overall offset direction of the error during this period. The fluctuation amplitude parameter Vol is obtained by the absolute difference between the maximum and minimum values ​​of all synchronization difference parameters Dlt within the statistical time sliding window, and is used to reflect the fluctuation intensity of the error during that period. The fluctuation trend feature Res is based on the synchronous difference parameter Dlt sequence within the time sliding window, performs a linear trend fitting, and calculates the fluctuation trend feature Res based on the mean of the squared residuals of the fitted curve, which is used to identify the directional trend of error changes.

[0032] It should be noted that the "fluctuation trend feature Res" here is not used to determine whether the trend is "up" or "down", but rather to determine whether the error data in the current period has a "clear trend", that is, whether the fluctuation is developing in one direction.

[0033] for example: If the error variation is chaotic, the fitted straight line will not fit well and the residual will be large; If the error changes are very stable and show a clear trend (whether it is rising or falling), the fitted line will be very close and the residual will be small.

[0034] Therefore, the mean of the fitted residuals is used to determine whether the fluctuation trend is obvious during this period, rather than to determine the direction.

[0035] S32. The difference level parameter Dev, the fluctuation amplitude parameter Vol, and the fluctuation trend feature Res are organized according to the time order of their respective windows to construct a structured statistical result set, named the error feature set Sta. The error feature set Sta, as a statistical set in the form of a time series, records the numerical state and trend characteristics of error changes in each window within the time sliding window.

[0036] In this embodiment, by introducing a time-sliding window structure based on the time series of the synchronization difference parameter Dlt, not only is the ability to identify local fluctuations in error information improved, but also a multi-dimensional quantitative expression of error change states is achieved. Using each sliding window as a unit, the resulting deviation level parameter Dev, fluctuation amplitude parameter Vol, and fluctuation trend feature Res are calculated respectively. These can be used to identify the directional shift of error, the amplitude intensity of error values, and the error evolution trend within that time period, thereby establishing a three-dimensional modeling framework for dynamic error behavior in sewing tasks. All statistical parameters are uniformly organized into an error feature set Sta in the time dimension, forming a statistical result set with trend judgment capabilities and structural continuity. This effectively overcomes the technical bottleneck of traditional methods that can only judge based on instantaneous error values ​​and cannot perceive the error evolution path. It possesses the ability to continuously track the evolution trend of synchronization deviation on the time axis, which is a key foundation for subsequent scoring function construction and dynamic feedback control.

[0037] Example 5 This embodiment is an explanation of Embodiment 1. Specifically, S4 includes S41; S41. Based on the error feature set Sta, extract the deviation level parameter Dev, fluctuation amplitude parameter Vol and fluctuation trend feature Res in each window, perform normalization preprocessing, and eliminate the dimensional differences between different features. Then, a nonlinear regression fitting method is used to fit the normalized preprocessed deviation level parameter Dev, fluctuation amplitude parameter Vol, and fluctuation trend feature Res to form a numerical output for representing the density fluctuation state, which is labeled as the line trace density change score Qsd. The line density change score Qsd numerically reflects the intensity of line density fluctuation caused by the accumulation of synchronization error between the feeding frequency and the needle movement rhythm over time. The higher the score, the more significant the synchronization deviation and the more unstable the density.

[0038] S5 includes S51 and S52; S51. Compare the stitch density change score Qsd with the density tolerance threshold parameter Lim preset for the sewing task, and construct an anomaly marker result set Tag based on the judgment result. The anomaly tagging result set is constructed using the following comparison and judgment method: When the stitch density change score Qsd > the density tolerance threshold parameter Lim, it is determined that there is a synchronous imbalance between the feeding frequency and the needle bar action within the window period, which causes the stitch density of the sewing machine to exceed the preset tolerance range and an abnormal density area appears. The window period mentioned in the stitch density change score Qsd is extracted synchronously and included in the abnormal label result set Tag. When the stitch density change score Qsd ≤ density tolerance threshold parameter Lim, there is no synchronous imbalance between the feeding frequency and the needle bar action within the judgment window period, and it is not included in the abnormal labeling result set Tag. The anomaly tag result set is used to clearly indicate which window periods during the sewing process have been identified as density anomaly areas, which can be used for subsequent density adjustment and optimization, equipment operation strategy correction and operation data backtracking.

[0039] It should be noted that the designed line density variation score Qsd is a score value calculated by a set of normalization and fitting algorithms. It is not a physical unit (such as "millimeters / needle"), but an abstract score used to reflect how severe the density fluctuation is.

[0040] Similarly, the density tolerance threshold Lim is not a direct data point in physical units, but rather an "acceptable range" set within the same scoring system as Qsd.

[0041] In other words, these two values ​​are inherently part of the same scoring system. The units are not the issue; it's just that a numerical system is being used to determine whether density fluctuations exceed acceptable limits.

[0042] S52. Using the anomaly labeling result set Tag as the basis for state recognition, for the window time period recorded in the anomaly labeling result set Tag, the sewing task segment corresponding to the system is determined to be in a density abnormal state; while the unlabeled window time period is determined to be in a density normal state. All status information is organized by sewing machine sewing task segment and displayed graphically in the operation interface, thus forming a visual feedback result of the sewing task density status, allowing the operator to view the density stability performance during the sewing process. While completing the density state visualization, the main control system uses the anomaly tag result set as the core trigger condition, and calls the control model interface in conjunction with it to execute adaptive control strategies for the corresponding anomaly window period, including: Send parameter fine-tuning instructions to the pattern density adjustment module to correct the fabric feeding rate or needle density factor, so as to alleviate local density fluctuations. Output rhythm synchronization compensation command to the frequency converter control module to dynamically adjust the feeding frequency or needle bar action response time in real time, so as to improve the matching accuracy of subsequent cycle time. This feedback and control process is dynamically and cyclically executed during the task's operation cycle, ensuring that the system can achieve adaptive and stable control of the trace density state without relying on manual intervention.

[0043] In this embodiment, a unified scoring function model is constructed based on the error feature set Sta. Furthermore, by normalizing the deviation level parameter Dev, the fluctuation amplitude parameter Vol, and the fluctuation trend feature Res, and performing nonlinear regression fitting, a stitch density change score Qsd reflecting the stable density state is accurately generated. By comparing the stitch density change score Qsd with the density tolerance threshold parameter Lim set for the sewing task in real time, abnormal density fluctuation segments can be accurately identified, and an abnormality marker result set Tag can be dynamically constructed. This marker set can not only be refined to specific window periods but can also be updated and called in real time during task execution, forming a highly time-seriesd density anomaly identification channel. Further, using the abnormality marker result set Tag as the core state identification basis, the density fluctuation situation can be graphically displayed as a visual feedback result of the sewing task density state, improving the operator's intuitive control over density change trends. Simultaneously, this method also realizes the direct linkage of abnormal states to the control strategy module, triggering frequency conversion compensation or density parameter adjustment commands based on the markers, thereby establishing a complete closed-loop structure of abnormality score identification—density state visualization—adaptive control linkage.

[0044] Example 6 A real-time monitoring system for the operating status of an intelligent sewing machine; please refer to [reference needed]. Figure 2 Specifically, it includes a signal acquisition module, a data structuring and error analysis module, a feature extraction module, a fitting module, and an evaluation feedback module; The operation signal acquisition module collects the feeding frequency parameter Fed output by the feeding controller during the sewing process and the needle bar cycle parameter Ncy collected by the needle bar motion encoder to construct the basic synchronization data set Dat; The data structuring and error analysis module performs frame-by-frame pairing analysis on the synchronization data set Dat, calculates the cycle ratio between the feed frequency parameter Fed and the needle bar cycle parameter Ncy under each sewing cycle, forms the actual cycle value Rac, and derives the theoretical ratio based on the preset parameters of the sewing task to obtain the expected cycle value Rex, calculates its synchronization difference parameter Dlt, and outputs the synchronization error set Dif. The feature extraction module performs sliding window statistical processing on the synchronization error set Dif, and outputs the deviation level parameter Dev, the fluctuation amplitude parameter Vol, and the fluctuation trend feature Res respectively, which constitute the error feature set Sta. The fitting module performs fitting operations on the deviation level parameter Dev, fluctuation amplitude parameter Vol, and fluctuation trend feature Res based on the error feature set Sta, and generates a line density change score Qsd. The evaluation feedback module compares the stitch density change score Qsd with the density tolerance threshold parameter Lim set for the sewing task, generates an anomaly marker result set Tag, and synchronously feeds back the density status information of the current sewing task segment based on the anomaly marker result set Tag.

[0045] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for monitoring the real-time working status of an intelligent sewing machine, characterized in that: Includes the following steps: S1. Through the operation signal acquisition module in the main control system, the feeding frequency parameters output by the feeding controller and the needle bar cycle parameters acquired by the needle bar motion encoder during the sewing process are obtained to construct a basic synchronization data set. S2. Perform frame-by-frame pairing analysis on the synchronization data set, calculate the cycle ratio between the feed frequency parameter and the needle bar cycle parameter under each sewing cycle, form the actual cycle value, and derive the theoretical ratio based on the preset parameters of the sewing task to obtain the expected cycle value, calculate its synchronization difference parameter, and output the synchronization error set. S3. Perform sliding window statistical processing on the synchronization error set, and output the deviation level parameter, fluctuation amplitude parameter and fluctuation trend characteristics respectively to form the error feature set; S4. Based on the error feature set, perform fitting calculations on the deviation level parameter, fluctuation amplitude parameter, and fluctuation trend feature to generate a line density change score. S5. Compare the stitch density change score with the density tolerance threshold parameter set for the sewing task to generate an anomaly marker result set, and synchronously feed back the density status information of the current sewing task segment based on the anomaly marker result set.

2. The method for monitoring the real-time working status of an intelligent sewing machine according to claim 1, characterized in that: S1 includes S11 and S12; S11. The main control system calls and runs the signal acquisition module to sample the control signal output by the sewing feed controller in each sewing beat cycle, and extract the feed frequency change signal sequence. The feed frequency change signal sequence is processed by bandpass filtering for noise reduction and periodic rhythm segmentation recognition to obtain the frequency expression data in each beat segment. Based on the frequency expression data, the average frequency value of the cycle in the feeding frequency change signal sequence is calculated, and the average frequency value of the cycle is used as the feeding frequency parameter. S12. By synchronizing the signal with the needle bar motion encoder, the needle bar motion signal is received in real time during the sewing process, and the motion time period within the complete cycle is extracted. The time interval of the edge change of the needle bar motion signal is calculated, and the cycle time used to complete the up and down motion of the needle bar within the current beat is calculated based on the time interval to form the needle bar cycle parameter. The needle bar cycle parameter is used to represent the actual time required for the needle bar to complete one up-and-down reciprocating motion within a single sewing cycle. Using the cycle period of the needle bar cycle parameter as the time division benchmark, the feeding frequency parameter and the needle bar cycle parameter synchronously collected within the same cycle are paired one-to-one to form paired data pairs and construct a basic synchronous data set.

3. The method for monitoring the real-time working status of an intelligent sewing machine according to claim 2, characterized in that: S2 includes S21 and S22; S21. For each pair of paired data items in the basic synchronization data set, perform frame-by-frame extraction and sequentially read the corresponding feed frequency parameter and needle bar cycle parameter. Using the feed frequency parameter as the divisor and the needle bar cycle parameter as the reference variable, calculate the actual synchronization ratio and name the calculation result as the actual beat value. At the same time, extract the preset reference value of the sewing machine's corresponding sewing beat from the sewing task settings, process the preset reference value, and obtain the expected beat value. The expected beat value is used to represent the theoretical synchronization ratio that should exist between feed and needle movement under the current task parameters.

4. The method for monitoring the real-time working status of an intelligent sewing machine according to claim 3, characterized in that: S22. Perform periodic error difference calculation on the actual value of the beat and the expected value of the beat, with the needle bar action cycle as the beat, to obtain the synchronization difference parameter; The synchronization difference parameters under all needle bar action cycles are accumulated and stored, and organized into structured sequence data according to the time sequence, and output as a synchronization error set.

5. The method for monitoring the real-time working status of an intelligent sewing machine according to claim 4, characterized in that: S3 includes S31 and S32; S31. Construct a fixed-length time sliding window structure based on the synchronization error set, and perform statistical analysis on the continuous synchronization difference parameters in each window of the set time sliding window to obtain the deviation level parameter, fluctuation amplitude parameter and fluctuation trend characteristics. The deviation level parameter is obtained by calculating the arithmetic mean of all synchronization difference parameters; The fluctuation amplitude parameter is obtained by the absolute difference between the maximum and minimum values ​​among all synchronization difference parameters within a statistical time sliding window. The fluctuation trend feature is based on the synchronous difference parameter sequence within the time sliding window, performs a linear trend fitting, and calculates the fluctuation trend feature based on the mean of the squared residuals of the fitted curve.

6. The method for monitoring the real-time working status of an intelligent sewing machine according to claim 5, characterized in that: S32. The difference level parameter, the fluctuation amplitude parameter, and the fluctuation trend feature are organized according to the time order of their respective windows to construct a structured statistical result set, named the error feature set; the error feature set, as a statistical set in the form of a time series, records the numerical state and trend characteristics of error changes in each window within the time sliding window.

7. The method for monitoring the real-time working status of an intelligent sewing machine according to claim 6, characterized in that: S4 includes S41; S41. Based on the error feature set, extract the deviation level parameter, fluctuation amplitude parameter and fluctuation trend feature in each window, perform normalization preprocessing, and eliminate the dimensional differences between different features. The nonlinear regression fitting method is then used to fit the normalized preprocessed deviation level parameter, fluctuation amplitude parameter and fluctuation trend characteristics to form a numerical output for representing the density fluctuation state, which is marked as the line trace density change score.

8. The method for monitoring the real-time working status of an intelligent sewing machine according to claim 7, characterized in that: S5 includes S51 and S52; S51. Compare the stitch density change score with the density tolerance threshold parameter preset for the sewing task, and construct an anomaly marker result set based on the judgment result. The anomaly marker result set is constructed using the following comparison and judgment method: When the stitch density change score is greater than the density tolerance threshold parameter, it is determined that there is a synchronous imbalance between the feeding frequency and the needle bar action within the window period, which causes the stitch density of the sewing machine to exceed the preset tolerance range and an abnormal density area appears. The window period mentioned in the stitch density change score is extracted synchronously and included in the abnormal marking result set. When the stitch density change score is less than or equal to the density tolerance threshold parameter, there is no synchronous imbalance between the feeding frequency and the needle bar action within the judgment window period, and it is not included in the abnormal marking result set.

9. The method for monitoring the real-time working status of an intelligent sewing machine according to claim 8, characterized in that: S52. Using the anomaly marking result set as the basis for state recognition, for the window time period recorded in the anomaly marking result set, determine that the sewing task segment corresponding to the system is in a density anomaly state. Unmarked window periods were judged to be in a normal density state; All status information is organized by sewing machine sewing task segments and displayed graphically in the operation interface, thus forming a visual feedback result of the sewing task density status.

10. A monitoring system for the real-time working status of an intelligent sewing machine, comprising the monitoring method for the real-time working status of an intelligent sewing machine as described in any one of claims 1 to 9, characterized in that: It includes a signal acquisition module, a data structuring and error analysis module, a feature extraction module, a fitting module, and an evaluation and feedback module; The operation signal acquisition module collects the feeding frequency parameters output by the feeding controller and the needle bar cycle parameters collected by the needle bar motion encoder during the sewing process to construct a basic synchronization data set; The data structuring and error analysis module performs frame-by-frame pairing analysis on the synchronous data set, calculates the cycle ratio between the feed frequency parameter and the needle bar cycle parameter under each sewing cycle, constitutes the actual cycle value, and derives the theoretical ratio based on the preset parameters of the sewing task to obtain the expected cycle value, calculates its synchronization difference parameter, and outputs the synchronization error set. The feature extraction module performs sliding window statistical processing on the synchronization error set and outputs the deviation level parameter, fluctuation amplitude parameter and fluctuation trend feature respectively, which constitute the error feature set. The fitting module performs fitting operations on the deviation level parameter, fluctuation amplitude parameter and fluctuation trend feature based on the error feature set, and generates a line density change score. The evaluation feedback module compares the stitch density change score with the density tolerance threshold parameter set for the sewing task, generates an anomaly marker result set, and synchronously feeds back the density status information of the current sewing task segment based on the anomaly marker result set.

Images