Loom networking data transmission optimization method and system based on distributed collaboration

Abstract

The invention relates to the technical field of textile wireless networking data transmission, in particular to a weaving machine networking data transmission optimization method and system based on distributed collaboration. The method comprises the following steps: firstly, predicting a loom data volume change rate and a loom health degree through an LSTM (Long Short Term Memory) model based on loom historical state data; the method comprises the following steps: dividing a loom wireless network into a plurality of sub-networks, and carrying out distributed networking; calculating the comprehensive cost of the path according to the node network state and the health degree of the loom, and selecting the path with the minimum comprehensive cost as a data transmission main path; after classification and redundant coding are carried out on the collected weaving machine data, three-level self-adaptive compression is carried out; the data transmission frequency is adjusted according to the loom data volume change rate and the minimum path comprehensive cost; and finally, according to the adjusted data transmission frequency, transmitting the compressed data to a central server through the path with the minimum comprehensive cost. According to the invention, high efficiency, high predictability, high reliability and real-time performance of weaving machine networking data transmission can be realized.

Description

Technical Field

[0001] The present invention relates to the technical field of textile wireless networking data transmission, and in particular to a method and system for optimizing loom networking data transmission based on distributed collaboration. Background Art

[0002] A loom is a machine that interweaves warp and weft yarns into fabric. In the modern textile industry, semi-automatic or fully automated loom production has been largely achieved to meet the needs of large-scale industrial weaving of various fabrics, including cotton, linen, and chemical fibers. As the modern textile industry transitions toward intelligent and digital technology, loom networking data transmission technology is key to achieving efficient management and precise control of textile production. By connecting loom equipment to the network, real-time information such as equipment operating status and production data can be collected, providing data support for production scheduling, equipment maintenance, and quality control. However, traditional loom networking data transmission suffers from numerous issues, severely hindering the development of intelligent textile industry.

[0003] On the one hand, the data generated during loom operation is dynamic and uncertain, and the amount of data can fluctuate significantly with changes in production conditions and equipment status. Traditional data transmission methods often use fixed transmission strategies that cannot adapt to dynamic changes in data volume, leading to network congestion or wasted bandwidth resources, affecting the timeliness and stability of data transmission. On the other hand, after the looms are connected to the network, data transmission faces problems such as network congestion, high latency, and poor reliability. When large amounts of loom data are transmitted simultaneously, traditional centralized data transmission methods are prone to network bottlenecks, affecting the real-time and accuracy of data, and failing to meet the needs of applications such as remote loom monitoring and intelligent scheduling. In addition, long-term operation of loom equipment may lead to health issues such as wear and tear and failure. Traditional transmission methods do not fully consider the impact of the loom's health status on data transmission. Once a loom fails, it may lead to data interruption or reduced transmission quality, which in turn affects the accuracy and timeliness of production decisions. Summary of the Invention

[0004] In order to address the deficiencies in the prior art, the present invention aims to provide a method and system for optimizing loom networking data transmission based on distributed collaboration. The technical solutions adopted are as follows: In a first aspect, a method for optimizing data transmission of a loom network based on distributed collaboration is provided, the method comprising: Step S1: Based on the historical status data of the loom, the LSTM model is used to predict the loom data volume change rate and loom health; Step S2: Divide the loom wireless network into multiple subnets, with each loom as a node. After all nodes are networked, initialize the neighbor list and receive neighbor node status information; Step S3: Calculate the comprehensive cost of the path based on the node network status and loom health, and select the path with the minimum comprehensive cost from the candidate paths as the main data transmission path; Step S4: After classifying and redundantly encoding the collected loom data, three-level adaptive compression is performed; Step S5: adjusting the data transmission frequency according to the loom data volume change rate and the minimum path comprehensive cost; Step S6: According to the adjusted data transmission frequency, the compressed data is transmitted to the central server via the path with the minimum comprehensive cost.

[0005] Furthermore, the step S1 specifically includes: Collect historical loom operation data, and perform linear interpolation filling, outlier replacement, and normalization processing on missing values in the historical operation data; According to the production variety switching frequency, variety switching complexity coefficient, and single switching benchmark data volume in the historical operation data of the loom, the loom data volume change rate in each historical period is obtained; Obtain the loom health status for each historical period based on the spindle amplitude, loom temperature, alarm frequency, and remaining time for the next maintenance in each period of the loom's historical operation data; The loom data volume change rate and loom health status in each historical period are constructed into input feature vectors according to time series, and an LSTM model is constructed for training to obtain the loom data volume change rate and loom health status predicted in the future period.

[0006] Furthermore, the frequency of product switching, the complexity coefficient of product switching, and the benchmark data volume of a single switching are all positively correlated with the changes in the amount of loom data in each historical period; The spindle amplitude, alarm frequency, and remaining time for the next maintenance in each period of the historical operation data are positively correlated with the changes in the loom health in each historical period, and the higher or lower the loom temperature, the worse the loom health.

[0007] Furthermore, step S2 specifically includes: configuring a wireless communication module for each loom, dividing the loom wireless network into multiple subnets, each loom as a node, and connecting adjacent subnets through a gateway to form a Mesh topology; each node accesses the designated subnet through a wireless access point, and each node periodically sends broadcast packets. After receiving the broadcast, the neighboring node parses the information and adds it to the local neighbor list.

[0008] Furthermore, the mathematical calculation formula for the comprehensive cost of the loom data transmission path constructed in step S3 is expressed as: Where, represents the comprehensive cost of the path, represents the network cost, represents the network cost weight, Indicates the current health of the loom. represents the predicted loom health in the future period, Indicates the time interval between the future period and the current period. Indicates the weight of the loom health change rate.

[0009] Furthermore, the classification and redundant coding of the collected loom data in step S4 specifically includes: The loom controller divides the collected loom data into three categories according to their importance: control alarm data, status monitoring data, and general log data; The highest redundancy encoding, medium redundancy encoding, and lowest redundancy encoding are performed for control alarm data, status monitoring data, and general log data respectively.

[0010] Furthermore, the three-level adaptive compression in step S4 specifically includes: For control alarm data, lossless compression is used, using the LZ77 algorithm to encode phrases in the forward buffer into corresponding tags through a sliding window, thereby achieving the purpose of compression; For status monitoring data, two-level dynamic lossy compression is adopted, and efficient compression is achieved through the dual-level strategy of "dynamic quantization + differential compression"; For general log data, the Transformer model is combined with the loom production cycle to achieve dynamic adjustment of the compression cycle. The compression frequency is increased during peak data periods to avoid data backlogs; conversely, the compression cycle is extended during low data periods.

[0011] Furthermore, the data transmission frequency formula after adjustment in step S5 can be expressed as: Where, Indicates the adjusted data transmission frequency, represents the basic transmission frequency, Indicates the rate of change of loom data volume, represents the frequency adjustment coefficient, represents the minimum comprehensive path cost, Indicates the maximum comprehensive path cost.

[0012] Furthermore, after step S6, the following steps are further included: Key loom data is copied and stored in different subnet nodes. The loom with worse health, the higher the service type weight, the node with the largest proportion of control alarm data in the loom data, and the higher the average network latency, the higher the data synchronization frequency. If there is a minor data transmission failure, adjust the loom node position and / or add wireless communication nodes and / or remove obstacles and / or detect and reduce radio interference; if there is a serious data transmission failure at the loom node, check the power off status of the loom node wireless communication module, check the distance between the wireless nodes, and confirm whether there is any line of sight obstruction between the nodes.

[0013] In another aspect, the present invention provides a distributed collaborative loom networking data transmission optimization system, the system comprising: The invention comprises a loom system controller, wherein the loom system controller integrates an operating program of any one of the above-mentioned methods for optimizing data transmission of a loom network based on distributed collaboration, and further comprises: The data prediction module uses the LSTM model to predict the loom data volume change rate and loom health based on the loom's historical status data; The network division and initialization module is used to divide the loom wireless network into multiple subnets, with each loom as a node. After all nodes are networked, the neighbor list is initialized and the status information of neighbor nodes is received; The path selection module is used to calculate the comprehensive cost of the path based on the node network status and loom health, and select the path with the minimum comprehensive cost from the candidate paths as the main path for data transmission; The data processing module is used to classify and encode the collected loom data, and then perform three-level adaptive compression; The transmission frequency adjustment module is used to adjust the data transmission frequency according to the loom data volume change rate and the minimum path comprehensive cost; The data transmission module is used to transmit the compressed data to the central server through the path with the minimum comprehensive cost according to the adjusted data transmission frequency.

[0014] The present invention has the following beneficial effects: It uses an LSTM model to dynamically predict the rate of change and health of loom data volume. Combined with a distributed subnet collaboration mechanism, it establishes a three-dimensional prediction system combining "data features, equipment status, and network status." This distributed collaboration mechanism enables real-time evaluation and dynamic optimization of network-wide paths. Compared to existing path planning methods that rely solely on network-layer metrics, this solution effectively mitigates the risk of transmission node failure due to loom health deterioration while balancing network load. The present invention fully considers the potential health issues of loom equipment, such as wear and failure, that may arise over long-term operation. Conventional transmission methods fail to fully account for the impact of loom health on data transmission. Failures can lead to data interruptions or degradation, impacting the accuracy and timeliness of production decisions. The present invention classifies and encodes collected loom data, then applies three-level adaptive compression and adjusts the data transmission frequency, improving the efficiency, reliability, and real-time performance of networked loom data transmission. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0016] Figure 1 This is a flow chart of a method for optimizing loom networking data transmission based on distributed collaboration provided by one embodiment of the present invention.

[0017] Figure 2 This is a block diagram of a distributed collaborative loom networking data transmission optimization system provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0018] To further illustrate the technical means and effectiveness of the present invention in achieving its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, describes in detail the specific implementation, structure, features, and effectiveness of the distributed collaborative loom networking data transmission optimization method and system proposed by the present invention. In the following description, different references to "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this invention belongs.

[0019] The scenario targeted by this invention is: under the wireless networking of large-scale loom equipment, the whole process of loom data prediction, path planning, data processing to transmission strategy is realized through a distributed collaborative mechanism, so as to achieve high efficiency, stability and adaptability of loom network data transmission.

[0020] The specific scheme of the method and system for optimizing data transmission of a loom network based on distributed collaboration provided by the present invention will be described in detail below with reference to the accompanying drawings.

[0021] First, see Figure 1 , which shows a flow chart of a method for optimizing data transmission of a loom network based on distributed collaboration provided by an embodiment of the present invention, the method comprising the following steps: Step S1: Based on the historical status data of the loom, the LSTM model is used to predict the loom data volume change rate and loom health.

[0022] Among them, step S1 specifically includes: collecting historical operation data of the loom, performing linear interpolation filling, outlier replacement, and normalization processing on missing values in the historical operation data; obtaining the loom data volume change rate of each historical period according to the production variety switching frequency, variety switching complexity coefficient, and single switching benchmark data volume in the loom historical operation data; obtaining the loom health of each historical period according to the spindle amplitude, loom temperature, alarm frequency, and remaining time for the next maintenance in each period in the loom historical operation data; constructing the loom data volume change rate and loom health of each historical period into an input feature vector according to a time series, constructing an LSTM model for training, and obtaining the loom data volume change rate and loom health predicted for future periods.

[0023] More specifically, the historical operating data of the loom is obtained through various sensors and data storage devices built into the loom, such as speed, output, warp tension, loom energy consumption, fault code, product type, temperature, air pump pressure, alarm frequency, product type switching frequency, main shaft amplitude, remaining time for the next maintenance, etc. Through time series analysis, the missing time points in the data are marked, and the missing numerical data (such as loom speed and output) are filled using linear interpolation of the two known data points before and after. Then, the data is detected for outliers based on the Z-score (standard score), and outliers are filtered according to the loom operation logic. For example, the loom speed cannot be negative or exceed the maximum design value. Then, the sliding window mean or median is used to replace the outliers, and all the data are normalized.

[0024] Furthermore, based on the product switching frequency, product switching complexity coefficient, and single switching benchmark data volume in the historical loom operation data, the loom data volume change rate for each historical period is obtained. In this embodiment, the mathematical calculation formula for the loom data volume change rate is constructed as follows: Where, Indicates the rate of change of loom data volume, Indicates the switching frequency of production types within the time interval, Indicates the complexity coefficient of product switching within the time interval, Indicates the amount of benchmark data for a single switch. Indicates the stable data volume when there is no production switching within the time interval.

[0025] In the mathematical calculation formula for the loom data volume change rate constructed above, the higher the switching frequency of production varieties within a time interval, the greater the amount of loom data that needs to be collected and reported due to the change in production varieties; the greater the complexity coefficient of variety switching within a time interval, the more complex the process of the switched fabric variety, the more data that needs to be collected, and the greater the amount of loom data. It represents the data value of the cumulative complexity of the loom production switching, which is different from the stable data volume when there is no production switching in the time interval. The ratio can be expressed as the change in loom data volume. Therefore, the frequency of product switching, the complexity coefficient of product switching, and the benchmark data volume of a single switching are all positively correlated with the change in loom data volume in each historical period.

[0026] Furthermore, the health of the loom in each historical period is obtained based on the spindle amplitude, loom temperature, alarm frequency, and remaining time for the next maintenance in each period of the loom's historical operation data. In this embodiment, the mathematical calculation formula for constructing the loom health can be expressed as: Where, For loom health, represents the loom temperature score, express The weight of represents the alarm frequency score, express The weight of represents the spindle amplitude score, express The weight of Indicates the maximum maintenance cycle, Indicates the remaining time until maintenance.

[0027] In the mathematical calculation formula of the loom health constructed above, the larger the main shaft amplitude, the higher or lower the loom temperature, and the more frequent the alarm frequency, the worse the loom health. Indicates the basic health of the loom's operating parameters. Since vibration has the greatest impact on the loom's structure, temperature affects the performance of loom components, and alarms reflect the real-time failure risk of the loom, the weight can be 、 、 Assign values of 0.4, 0.3, and 0.3 respectively; Indicates the loom maintenance urgency correction coefficient, when M< When it is not the maintenance period, the correction coefficient is , when M= When M=0, maintenance has just been completed and the health level is unaffected. When M=0, immediate maintenance is required and the health level is reduced by 20%, forcing a maintenance urgency. Therefore, the spindle amplitude, alarm frequency, and remaining time to next maintenance for each period in the historical operation data are positively correlated with the changes in loom health level during each historical period. Furthermore, the higher or lower the loom temperature, the worse the loom health level.

[0028] It should be noted that in the above formula, the loom temperature scoring formula can be expressed as: Where, represents the loom temperature score, Indicates the loom temperature. The preset high temperature threshold is 40℃ and the low temperature threshold is 20℃. Since the risk of high temperature is higher than that of low temperature, 4 points will be deducted for every 1℃ above the high temperature and 2 points for every 1℃ below the low temperature.

[0029] The scoring formula for loom alarm frequency can be expressed as: Where, represents the alarm frequency score, Indicates the number of loom alarms. Since alarms directly reflect faults and require strong penalties, the preset base value for alarms is 2 times / period, and 40 points will be deducted for each additional alarm.

[0030] The loom spindle amplitude scoring formula can be expressed as: Where, represents the loom amplitude score, Indicates the spindle amplitude. Exceeding the amplitude limit will directly affect the mechanical life of the loom, so the preset amplitude threshold is 50μm. 30 points will be deducted for every 10μm exceeding the threshold.

[0031] Furthermore, the loom data volume change rate and loom health of each historical period are obtained through historical relevant data and the above-mentioned loom data volume change rate formula and loom health mathematical calculation formula. Then, the loom data volume change rate and loom health of each historical period are constructed into input feature vectors according to time series, and an LSTM model is constructed for training. The processed training data set is input into the LSTM model, and backpropagation training is performed using a suitable loss function and optimizer. The model parameters are iteratively updated until the loss function converges or the preset number of training times is reached. Finally, the processed test data set is input into the trained LSTM model to obtain the loom data volume change rate and loom health prediction values for future periods.

[0032] Step S2: Divide the loom wireless network into multiple subnets, with each loom as a node. After all nodes are networked, initialize the neighbor list and receive neighbor node status information.

[0033] Among them, step S2 specifically includes: configuring a wireless communication module on each loom, dividing the loom wireless network into multiple subnets, each loom as a node, and connecting adjacent subnets through a gateway to form a mesh topology; each node accesses the designated subnet through a wireless access point, and each node periodically sends broadcast packets. After receiving the broadcast, the neighboring node parses the information and adds it to the local neighbor list.

[0034] More specifically, each loom is first equipped with a low-power wide area network (LPWAN) communication module (such as LoRa or Thread) that supports mesh protocols. This module balances long-distance transmission with low energy consumption, ensuring signal coverage and penetration in complex industrial environments. Furthermore, it supports multi-band switching to avoid frequency interference. Then, dynamic or static subnetting is implemented based on the physical distribution of the looms, production area divisions, and data traffic characteristics. For example, subnets can be divided by workshop area, with looms on the same production line assigned to the same subnet. Alternatively, subnets can be dynamically adjusted based on data transmission priority, with high-priority nodes preferentially connected to independent subnets to ensure critical data transmission. After each node accesses a designated subnet through a wireless access point, it periodically sends broadcast packets containing information such as node ID, signal strength, and subnet affiliation. The broadcast period is dynamically adjusted based on network load, such as 10 seconds for light load and 30 seconds for heavy load. Upon receiving the broadcast, neighboring nodes filter valid neighbors based on signal strength, removing nodes with weak signals. A neighbor list aging timer is set, deleting records that have not been updated for 120 seconds to reduce redundant information storage. Furthermore, distributed routing algorithms, such as OLSR (Optimized Link State Routing Protocol), are introduced to enable automatic discovery of network nodes and route establishment. When a node fails or a link is interrupted, the network automatically triggers a path recalculation mechanism, forwarding data through neighboring nodes to ensure network connectivity. Regular network-wide topology updates are performed to optimize routing paths and reduce transmission latency. The cloud platform monitors subnet load change rates in real time and re-subnets when the load change rate exceeds a preset threshold, such as a load change rate of >20% or <-15%. If an anomaly is detected, the central system triggers a subnet re-partitioning algorithm and generates a new subnetting solution, such as adjusting node ownership or switching gateways.

[0035] Step S3: Calculate the comprehensive cost of the path based on the node network status and loom health, and select the path with the minimum comprehensive cost from the candidate paths as the main path for data transmission.

[0036] Because network status factors such as bandwidth, latency, and packet loss rate, as well as equipment status factors such as loom health and failure risk, route selection requires a comprehensive assessment of these indicators. Avoiding loom nodes with low health (prone to failure) or high-load subnets reduces transmission interruptions and data loss, ensuring real-time transmission of production data. Costs must be quantified to prioritize cost-effective routes, avoid network congestion, and improve overall transmission efficiency.

[0037] In this embodiment, the mathematical calculation formula for constructing the comprehensive cost of the loom data transmission path is expressed as: Where, represents the comprehensive cost of the path, represents the network cost, represents the network cost weight, Indicates the current health of the loom. represents the predicted loom health in the future period, Indicates the time interval between the future period and the current period. Indicates the weight of the loom health change rate.

[0038] In the mathematical calculation formula for the comprehensive cost of the loom data transmission path constructed above, the network cost It represents the node network cost under the comprehensive combination of delay, bandwidth and packet loss rate of the node network. It is the most important component of the data transmission path. The higher the delay, the smaller the bandwidth and the greater the packet loss rate, the higher the network cost. The larger the value, the greater the path comprehensive cost. The larger the loom's current health The lower the value, the more likely it is that the data is being transmitted from a device with a high risk of failure, and there is a high probability of losing critical data. Therefore, the current health of the loom is Combined cost with path negative correlation; represents the rate of change of loom health, , indicating that the loom health is stable or improved, the path comprehensive cost It will get smaller. , indicating that the health of the loom deteriorates, which will increase the comprehensive cost of the path In the above formula, the network cost weight The value can be 0.6, the loom health change rate The weight value can be 0.3.

[0039] Then traverse the path comprehensive cost TPC of the candidate path to obtain the candidate path set In the process, the path with the smallest comprehensive path cost TPC is selected as the main path for data transmission.

[0040] In this embodiment, the minimum path comprehensive cost calculation formula is constructed as follows: Where, represents the minimum comprehensive path cost, Indicates the candidate paths, represents the comprehensive cost of the path, Represents a set of candidate paths.

[0041] Furthermore, the three paths with the smallest comprehensive cost ranking are pre-stored as data transmission paths. Among them, the path with the smallest comprehensive cost is used as the main transmission path, and the rest are two redundant paths. When the path with the smallest comprehensive cost fails, it automatically switches to the redundant path for data transmission.

[0042] Step S4: After the collected loom data is classified and redundantly encoded, three-level adaptive compression is performed.

[0043] Specifically, the classification and redundant coding of the collected loom data in step S4 specifically includes: The loom controller divides the collected loom data into three categories according to their importance: control alarm data, status monitoring data, and general log data. The control alarm data includes loom health, emergency stop signal, fault code, etc. The status monitoring data includes speed, tension, temperature, energy consumption, pump air pressure, spindle amplitude, etc. The general log data includes historical production, operation records, general operation logs, etc. Then, the highest redundancy coding, medium redundancy coding, and lowest redundancy coding are performed for the control alarm data, status monitoring data, and general log data respectively. More specifically, when the highest redundancy coding is performed, the forward error correction coding + repeated transmission double redundancy strategy is adopted. When the forward error correction coding is performed, a large number of check bits are inserted into the original data. For example, 5 check bytes are generated for every 10 data bytes, so that the receiving end can restore the original information even if some data is lost. In the repeated transmission strategy, the same data is sent at intervals of 3 times. The receiving end takes the earliest correctly received data as the basis, reducing the risk of transmission failure due to instantaneous interference. In actual applications, when the loom triggers a high temperature alarm, the alarm information is encoded with high redundancy and takes priority in occupying the network bandwidth for transmission to ensure timely response from the control center; when the medium redundancy coding is performed, the low-density parity check code + selective retransmission mechanism is adopted. The low-density parity check strategy achieves efficient error correction with a low redundancy overhead, such as a 15% redundancy rate, and is suitable for medium-importance data. In actual application, data such as loom speed and tension are sent periodically to ensure basic data integrity while reducing bandwidth occupancy. When encoding with minimum redundancy, a lightweight check + batch transmission method is adopted, and only a small number of check bits are added to the log data. The focus relies on the storage verification mechanism at the receiving end. Data is cached to a certain amount, such as 1MB, and then sent in batches. Delayed transmission is allowed during network congestion. In actual application, daily production logs are transmitted in batches during non-production periods, such as in the early morning. Even if a small amount of data is lost, it can be supplemented and analyzed through data from other periods.

[0044] Furthermore, three-level adaptive compression is performed on the classified and redundantly encoded data. Specifically, for control alarm data, lossless compression is applied using the LZ77 algorithm. Sentences in the forward buffer are encoded into corresponding tokens using a sliding window, achieving compression with a compression ratio between 1.2 and 1.5. For condition monitoring data, two-level dynamic lossy compression is employed, utilizing a two-stage strategy of "dynamic quantization + differential compression" to achieve efficient compression. Specifically, in the first stage of compression, the data is first divided into blocks and then subjected to a discrete cosine transform (DCT) to convert the time-domain data into the frequency domain, concentrating energy on low-frequency coefficients. The quantization factor is then dynamically calculated based on data volatility (standard deviation) and network bandwidth utilization. The quantization step size is adjusted to process the DCT coefficients, preserving low-frequency data and drastically processing high-frequency data, completing the initial compression. In the second stage of compression, the rate of change between the compressed and original data volumes is calculated. When the change exceeds 30%, differential compression is initiated. Differences between adjacent data blocks are calculated and encoded into a sequence of these differences to further reduce data redundancy. The receiving end then reverse-decompresses the data based on the quantization factor and compression flag, restoring the data. This solution balances compression efficiency and data accuracy, flexibly adjusting to the data and network status to meet the dynamic transmission needs of loom data. For general log data, the Transformer model is combined with the loom production cycle to achieve dynamic adjustment of the compression cycle to improve data management efficiency and resource utilization. By training on historical data, the model can predict the log data volume in the next production cycle. The compression cycle calculation formula is constructed as follows: Where, Indicates the general log data compression period; It is the average time it takes for a loom to complete a complete production process, which serves as the basic cycle; Indicates the amount of predicted data; Indicates the peak value of historical log data volume. Represents the normalized prediction data volume , when the amount of prediction data The larger the compression cycle The shorter it is, the more compression frequency is increased during peak data volume periods to avoid data backlogs; conversely, the compression cycle is extended during low data volume periods to reduce unnecessary resource consumption.

[0045] Step S5: adjusting the data transmission frequency according to the loom data volume change rate and the minimum path comprehensive cost.

[0046] In this embodiment, the constructed adjusted data transmission frequency formula can be expressed as: Where, Indicates the adjusted data transmission frequency, represents the basic transmission frequency, Indicates the rate of change of loom data volume, represents the frequency adjustment coefficient, represents the minimum comprehensive path cost, Indicates the maximum comprehensive path cost.

[0047] In the above-constructed adjusted data transmission frequency formula, the basic transmission frequency It is the default transmission frequency of the loom under stable working conditions, which is the reference value of data transmission frequency; the loom data volume changes It is positively correlated with the adjusted data transmission frequency. When the amount of data increases, the transmission frequency needs to be increased to avoid data backlog. The corresponding factor If the value is greater than 1, when the amount of data decreases, the frequency can be reduced to save resources. In this case, the factor may be less than 1. For the path transmission cost, the higher the value, the higher the transmission path delay and packet loss rate, the lower the data transmission reliability, and the more retransmission opportunities need to be increased by increasing the frequency. Indicates the skill-based normalization processing of the selected minimum path comprehensive cost, Indicates the adjustment factor for adjusting the data transmission frequency corresponding to the path cost.

[0048] Step S6: According to the adjusted data transmission frequency, the compressed data is transmitted to the central server via the path with the minimum comprehensive cost.

[0049] Furthermore, after step S6, the following steps are further included: key loom data is copied to different subnet nodes. The looms with poorer health, the greater the service type weight, the nodes with the largest proportion of control alarm data in the loom data, and the greater the average network latency, the higher the data synchronization frequency. Specifically, taking health as an example, real-time monitoring is performed to monitor changes in the loom's health status and dynamically adjust the synchronization frequency. When the loom health score falls below a threshold, such as below 60 points, the synchronization frequency is automatically increased by 50%-100% to ensure data backup is completed before equipment failure. This adaptive mechanism continuously optimizes through machine learning models, predicts potential equipment failure risks, and initiates high-frequency synchronization mode in advance.

[0050] Furthermore, if there is a minor failure in data transmission, adjust the loom node position and / or add wireless communication nodes and / or remove obstacles and / or detect and reduce radio interference; if there is a serious failure in data transmission of the loom node, check the power off status of the loom node wireless communication module, check the distance between the wireless nodes, and confirm whether there is any line of sight obstruction between the nodes.

[0051] Specifically, if the system monitors data transmission metrics in real time and finds a packet loss rate <10%, latency fluctuation <50%, or intermittent transmission interruptions, this situation does not impact core business continuity. However, adaptive adjustments, such as switching to alternative data transmission paths or neighboring node assistance, still cannot resolve these issues. In this case, the network topology map and node status information can be used to quickly locate the faulty node, adjust the loom node position, add wireless communication nodes, remove obstacles, and / or detect and reduce radio interference. Alternatively, if the system monitors data transmission packet loss rates ≥ 10%, continuous disconnections lasting more than 5 minutes, or critical data transmission failures, potentially leading to production interruptions, adaptive adjustments may not resolve the data transmission issue. In this case, the node can be remotely controlled to reboot and the power module voltage and current can be checked for normal operation. UWB (ultra-wideband) positioning technology can be used to measure the actual distance between nodes and compare it to the theoretical optimal spacing. For nodes whose distance is too far, causing signal attenuation, telescopic brackets can be used to adjust their positions or signal amplifiers can be added. Combining three-dimensional maps with signal propagation models, signal paths are simulated and obstructions are identified. For fixed obstructions that cannot be removed, reflectors or refraction devices are deployed to change the direction of signal propagation and ensure smooth communication links.

[0052] Secondly, this embodiment provides a distributed collaborative loom networking data transmission optimization system, see Figure 2 , which shows a block diagram of a loom networking data transmission optimization system based on distributed collaboration provided by one embodiment of the present invention, the system includes a loom system controller, and the loom system controller integrates the operating program of the loom networking data transmission optimization method based on distributed collaboration mentioned in the first aspect of this embodiment; The invention also includes: a data prediction module 101, which predicts the loom data volume change rate and loom health status through an LSTM model based on the loom historical status data; The network division and initialization module 102 is used to divide the loom wireless network into multiple subnets, with each loom as a node. After all nodes are networked, the neighbor list is initialized and the neighbor node status information is received; The path selection module 103 is used to calculate the comprehensive cost of the path according to the node network status and the health of the loom, and select the path with the minimum comprehensive cost from the candidate paths as the main path for data transmission; The data processing module 104 is used to classify and encode the collected loom data, and then perform three-level adaptive compression; The transmission frequency adjustment module 105 is used to adjust the data transmission frequency according to the loom data volume change rate and the minimum path comprehensive cost; The data transmission module 106 is configured to transmit the compressed data to the central server via a path with the minimum comprehensive cost according to the adjusted data transmission frequency.

[0053] This embodiment provides a distributed collaborative loom network data transmission optimization method and system. Based on the loom's historical status data, an LSTM model is used to predict the loom data volume change rate and loom health. The loom wireless network is divided into multiple subnets, with each loom acting as a node. After all nodes are networked, a neighbor list is initialized and neighbor node status information is received. Based on the node network status and loom health, the path cost is calculated, and the path with the lowest cost is selected from the candidate paths as the primary data transmission path. The collected loom data is classified and redundantly encoded, followed by three-level adaptive compression. The data transmission frequency is adjusted based on the loom data volume change rate and the lowest path cost. The compressed data is then transmitted to a central server via the path with the lowest cost according to the adjusted data transmission frequency. This achieves high efficiency, high reliability, and real-time performance for loom network data transmission.

[0054] It should be noted that the order in which the embodiments of the present invention are described above is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0055] The above-described embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present application, and should all be included in the scope of protection of the present application.

Claims

1. A method for optimizing loom networking data transmission based on distributed collaboration, characterized in that: The method comprises: Step S1: Based on the historical status data of the loom, the LSTM model is used to predict the loom data volume change rate and loom health; Step S2: Divide the loom wireless network into multiple subnets, with each loom as a node. After all nodes are networked, initialize the neighbor list and receive neighbor node status information; Step S3: Calculate the comprehensive cost of the path based on the node network status and loom health, and select the path with the minimum comprehensive cost from the candidate paths as the main data transmission path; Step S4: After classifying and redundantly encoding the collected loom data, three-level adaptive compression is performed; Step S5: adjusting the data transmission frequency according to the loom data volume change rate and the minimum path comprehensive cost; Step S6: According to the adjusted data transmission frequency, the compressed data is transmitted to the central server via the path with the minimum comprehensive cost.

2. The method for optimizing data transmission of a loom network based on distributed collaboration according to claim 1 is characterized in that: The step S1 specifically includes: Collect historical loom operation data, and perform linear interpolation filling, outlier replacement, and normalization processing on missing values in the historical operation data; According to the production variety switching frequency, variety switching complexity coefficient, and single switching benchmark data volume in the historical operation data of the loom, the loom data volume change rate in each historical period is obtained; Obtain the loom health status for each historical period based on the spindle amplitude, loom temperature, alarm frequency, and remaining time for the next maintenance in each period of the loom's historical operation data; The loom data volume change rate and loom health status in each historical period are constructed into input feature vectors according to time series, and an LSTM model is constructed for training to obtain the loom data volume change rate and loom health status predicted in the future period.

3. The method for optimizing loom networking data transmission based on distributed collaboration according to claim 2 is characterized in that: The frequency of product switching, the complexity coefficient of product switching, and the benchmark data volume of a single switching are all positively correlated with the changes in the amount of loom data in each historical period; The spindle amplitude, alarm frequency, and remaining time for the next maintenance in each period of the historical operation data are positively correlated with the changes in the loom health in each historical period, and the higher or lower the loom temperature, the worse the loom health.

4. The method for optimizing data transmission of a loom network based on distributed collaboration according to claim 1, characterized in that: The step S2 specifically includes: configuring a wireless communication module on each loom, dividing the loom wireless network into multiple subnets, with each loom acting as a node, and adjacent subnets connected via a gateway to form a mesh topology; each node accesses a designated subnet via a wireless access point, and each node periodically sends broadcast packets. After receiving the broadcast, the neighboring node parses the information and adds it to the local neighbor list.

5. The method for optimizing loom networking data transmission based on distributed collaboration according to claim 1 is characterized in that: The mathematical calculation formula for the comprehensive cost of the loom data transmission path constructed in step S3 is expressed as: Where, represents the comprehensive cost of the path, represents the network cost, represents the network cost weight, Indicates the current health of the loom. represents the predicted loom health in the future period, Indicates the time interval between the future period and the current period. Indicates the weight of the loom health change rate.

6. The method for optimizing data transmission of a loom network based on distributed collaboration according to claim 1, characterized in that: The classification and redundant coding of the collected loom data in step S4 specifically includes: The loom controller divides the collected loom data into three categories according to their importance: control alarm data, status monitoring data, and general log data; The highest redundancy encoding, medium redundancy encoding, and lowest redundancy encoding are performed for control alarm data, status monitoring data, and general log data respectively.

7. The method for optimizing loom networking data transmission based on distributed collaboration according to claim 1 is characterized in that: The three-level adaptive compression in step S4 specifically includes: For control alarm data, lossless compression is used, using the LZ77 algorithm to encode phrases in the forward buffer into corresponding tags through a sliding window, thereby achieving the purpose of compression; For status monitoring data, two-level dynamic lossy compression is adopted, and efficient compression is achieved through a two-level strategy of "dynamic quantization + differential compression"; For general log data, the Transformer model is combined with the loom production cycle to achieve dynamic adjustment of the compression cycle. The compression frequency is increased during peak data periods to avoid data backlogs; conversely, the compression cycle is extended during low data periods.

8. The method for optimizing loom networking data transmission based on distributed collaboration according to claim 1 is characterized in that: The data transmission frequency formula after adjustment in step S5 can be expressed as: Where, Indicates the adjusted data transmission frequency, represents the basic transmission frequency, Indicates the rate of change of loom data volume, represents the frequency adjustment coefficient, represents the minimum comprehensive path cost, Indicates the maximum comprehensive path cost.

9. The method for optimizing data transmission of a loom network based on distributed collaboration according to claim 1, characterized in that: After step S6, the following steps are also included: Key loom data is copied and stored in different subnet nodes. The loom with worse health, the higher the service type weight, the node with the largest proportion of control alarm data in the loom data, and the higher the average network latency, the higher the data synchronization frequency. If there is a minor data transmission failure, adjust the loom node position and / or add wireless communication nodes and / or remove obstacles and / or detect and reduce radio interference; if there is a serious data transmission failure at the loom node, check the power off status of the loom node wireless communication module, check the distance between the wireless nodes, and confirm whether there is any line of sight obstruction between the nodes.

10. A distributed collaborative loom networking data transmission optimization system, characterized in that: The invention comprises a loom system controller, wherein the loom system controller integrates an operating program of the method for optimizing the data transmission of a loom network based on distributed collaboration according to any one of claims 1 to 9, and further comprises: The data prediction module uses the LSTM model to predict the loom data volume change rate and loom health based on the loom's historical status data; The network division and initialization module is used to divide the loom wireless network into multiple subnets, with each loom as a node. After all nodes are networked, the neighbor list is initialized and the status information of neighbor nodes is received; The path selection module is used to calculate the comprehensive cost of the path based on the node network status and loom health, and select the path with the minimum comprehensive cost from the candidate paths as the main path for data transmission; The data processing module is used to classify and encode the collected loom data, and then perform three-level adaptive compression; The transmission frequency adjustment module is used to adjust the data transmission frequency according to the loom data volume change rate and the minimum path comprehensive cost; The data transmission module is used to transmit the compressed data to the central server through the path with the minimum comprehensive cost according to the adjusted data transmission frequency.

Images