System for controlling multiple shuttleless looms based on wireless network

Abstract

The invention relates to the technical field of shuttleless loom control, and discloses a system for controlling multiple shuttleless looms based on a wireless network, comprising a management computer used for global optimization and task scheduling; the wireless communication network comprises a management unit and a plurality of node units, and the management unit is connected with the management computer through an Ethernet interface; the node units are connected with the node control units through wireless radio frequency modules and used for achieving data transmission between the management computer and the node control units, and each node control unit is installed in a shuttleless loom control box. By adopting a wireless communication network and dual channels, an efficient and reliable data transmission effect is achieved. Compared with a single-channel communication scheme in the prior art, the method solves the defects of large delay jitter and untimely key instruction transmission under high load. And through dynamic switching of the main channel and the backup channel, the emergency instruction is ensured to be transmitted preferentially, and the response speed of the system is remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of shuttleless loom control, and specifically to a control system for multiple shuttleless looms based on a wireless network. Background Art

[0002] With the accelerating evolution of industrial automation towards intelligence, the communication architecture optimization and dynamic resource scheduling capabilities of multi-device collaborative control systems have become the core elements for improving manufacturing efficiency. In typical industrial scenarios such as textile manufacturing, the deep integration of wireless network technology and intelligent control algorithms provides a new path to break through the limitations of traditional wired control systems. Traditional wired systems rely on fixed topological structures and have defects such as high deployment costs and poor scalability, making it difficult to meet the needs of large-scale device networking and dynamic production tasks. Against this background, distributed control systems based on wireless networks have gradually become a key technical direction for industrial automation upgrades due to their advantages such as flexible networking and real-time response.

[0003] Currently, many shuttleless loom control systems rely on wired communication networks for data transmission and control instruction issuance. Existing control systems mostly adopt a centralized architecture, usually connecting a central computer to each node control unit. These systems generally use fixed channels for data transmission and transfer control instructions to the loom control unit through serial communication interfaces. However, with the increase in production tasks, these traditional systems will encounter problems such as excessive system load and severe delay jitter when facing large-scale concurrent communications. In addition, the control parameters of many existing systems are difficult to adjust in real time, resulting in the inability to adapt to dynamic changes in complex production environments and affecting the stability of the loom and the quality of the fabric surface.

[0004] In the prior art, the scheme of using a single channel for communication will have problems such as large delay jitter and untimely transmission of key instructions under high load conditions. Especially in case of emergencies, the response speed of the system often cannot meet the production requirements. At the same time, traditional static resource allocation mechanisms are prone to waste of communication resources and competition between nodes, increasing the conflict frequency of the system. Coupled with the fact that many existing technologies use fixed control parameters and cannot be flexibly adjusted according to changes in the actual production environment, the control accuracy is low and the loom stability is poor. Therefore, how to ensure real-time and efficient instruction transmission, optimize communication resource allocation, and achieve dynamic adjustment of control parameters in a high-load environment has become a major problem in the current technology. Summary of the Invention

[0005] Aiming at the deficiencies of the prior art, the present invention provides a control system for multiple shuttleless looms based on a wireless network, which solves the problems of large delay jitter caused by single-channel communication, untimely transmission of key instructions, communication conflicts caused by static resource allocation, and low control accuracy in the prior art in a high-load environment.

[0006] To achieve the above object, the present invention is realized by the following technical solutions: A control system for multiple shuttleless looms based on a wireless network, comprising: A management computer for global optimization and task scheduling; A wireless communication network, including a management unit and multiple node units, the management unit is connected to the management computer through an Ethernet interface; Multiple node control units, the node units are connected to the node control units through radio frequency modules, for realizing data transmission between the management computer and the node control units, each node control unit is installed in a control box of a shuttleless loom, and is connected to the control function unit of the shuttleless loom control box through a serial communication interface, for executing control instructions and collecting loom status data; An information geometry optimization module, connected to the management computer through a software interface, for optimizing control parameters based on the statistical characteristics of the system state distribution; A game theory scheduling module, connected to the management computer through a software interface, for optimizing communication resource allocation through a non-cooperative game model; A hardware acceleration and execution module, connected to the node control unit through an SPI bus, for realizing efficient calculation and execution of control instructions; A fault detection and self-healing module, connected to the node control unit through a GPIO interface, for real-time monitoring of the system state and realizing fault self-healing; Wherein, the management computer issues control instructions to the node control units through the wireless communication network, and the node control units distribute the control instructions to each control function unit of the shuttleless loom control box through the serial communication interface, and feedback the status data to the management computer.

[0007] Preferably, the wireless communication network includes a main channel and a backup channel, the main channel adopts a time slot allocation protocol, the time slot length is 10 milliseconds - 13 milliseconds, the backup channel adopts a pulse communication protocol, and when the instruction queue length of the node control unit exceeds a preset threshold, it automatically switches to the backup channel to transmit emergency instructions.

[0008] Preferably, the information geometry optimization module updates the control parameters by calculating the statistical characteristic matrix of the system state distribution and along the statistical manifold, specifically including: Calculating the statistical characteristic matrix of the system state distribution; Updating the control parameters based on the inverse matrix of the statistical characteristic matrix and the gradient of the loss function; Projecting the updated control parameters onto the feasible domain.

[0009] Preferably, the game theory scheduling module calculates the equilibrium strategy by defining a non-cooperative game model, specifically including: Define the user set as all node control units, the action set as transmission power and time slot selection, and the utility function as the weighted negative value of delay and energy consumption; Calculate the equilibrium strategy by iteratively updating the policy distribution; Use the equilibrium strategy result for communication resource allocation.

[0010] Preferably, the hardware acceleration and execution module includes: A statistical characteristic matrix calculation unit for parallel calculating the statistical characteristic matrix of the system state distribution and sending the calculation result to the information geometry optimization module; A game theory scheduling engine for quickly calculating the equilibrium strategy and sending the calculation result to the game theory scheduling module; A dynamic voltage and frequency regulation unit for dynamically adjusting the calculation frequency according to the system load and sending the adjustment result to the node control unit.

[0011] Preferably, the fault detection and self-healing module includes: A broken yarn detection unit for detecting broken yarn events through a photoelectric sensor array and a high-speed analog-to-digital converter and sending the detection result to the node control unit; A delay compensation unit for compensating communication delay through a prediction observer and sending the compensation result to the node control unit; A self-healing control unit for automatically starting a backup control strategy when a fault is detected and sending the fault information to the management computer.

[0012] Preferably, the management computer sends control instructions to the node control unit through a wireless communication network. The control instructions include communication rate, calculation frequency, and control parameters. The node control unit distributes the control instructions to each control function unit of the shuttleless loom control box through a serial communication interface and feeds back the execution result to the management computer.

[0013] Preferably, the node control unit includes a main control chip, a communication interface, and a sensor interface. The main control chip uses a multi-core processor. The communication interface supports a serial bus and a controller area network bus. The sensor interface is used to connect the photoelectric sensor array.

[0014] Preferably, the information geometry optimization module and the game theory scheduling module are called by the management computer. The management computer dynamically adjusts the optimization objectives and constraints according to the state data of the node control unit. The optimization objectives include the weighted sum of delay, energy consumption, and control error.

[0015] Preferably, the fault detection and self-healing module collects sensor data in real time through the node control unit. When a broken yarn or communication delay limit is detected, it automatically triggers the self-healing process and reports the fault information to the management computer. The management computer recalculates the optimization strategy according to the fault information and sends it down.

[0016] The present invention provides a control system for multiple shuttleless looms based on a wireless network, having the following beneficial effects: 1. By adopting a wireless communication network and a dual-channel, the present invention achieves efficient and reliable data transmission. Compared with the single-channel communication solution in the prior art, it solves the problems of large delay jitter and untimely transmission of key instructions under high load. Through the dynamic switching between the main channel and the backup channel, it ensures the priority transmission of emergency instructions and significantly improves the system response speed.

[0017] 2. Through the information geometry optimization module, the present invention realizes the dynamic parameter optimization under multi-objective constraints. Compared with the fixed-parameter control solution in the prior art, it solves the problems of inability to adapt to complex working conditions and low control accuracy. Through the optimization algorithm of statistical manifold and Fisher information matrix, the system can adjust the control parameters in real time and significantly improve the stability of loom operation and the quality of the fabric surface.

[0018] 3. By optimizing the communication resource allocation through the game theory scheduling module, the present invention achieves the maximum utilization rate of resources. Compared with the static resource allocation solution in the prior art, it solves the problems of resource waste and communication conflicts caused by node competition. Through the non-cooperative game model and Nash equilibrium solution, the communication resources are dynamically allocated and the overall efficiency of the system is greatly improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 is the overall architecture diagram of the system of the present invention; Figure 2 is the flowchart of the wireless communication network of the present invention; Figure 3 is the flowchart of fault detection and self-healing of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0020] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative work shall fall within the protection scope of the present invention.

[0021] Please refer to the attached Figure 1 - attached Figure 3 , the embodiments of the present invention provide a control system for multiple shuttleless looms based on a wireless network, including: The management computer serves as the core control center of the present invention, undertaking the hub functions of global optimization, task scheduling, and multi-module coordination. It establishes a two-way data link with each node control unit through a wireless communication network, receives loom status information in real time, and issues optimization instructions. At the same time, it integrates an information geometry optimization module and a game theory scheduling module to realize the generation of dynamic strategies under multi-objective constraints. The implementation method of the management computer is described in detail below in combination with the technical solutions.

[0022] In this embodiment, the management computer is deployed on an industrial-grade server platform, runs a real-time operating system based on the Linux kernel, and configures a multi-threaded task scheduling framework. Its software architecture includes a data acquisition layer, an optimization calculation layer, and an instruction distribution layer. The data acquisition layer receives loom status data from each node control unit through the management unit of the wireless communication network, including but not limited to the main shaft speed, weft yarn tension, yarn breakage signal, and communication link quality index. The optimization calculation layer calls the information geometry optimization module and the game theory scheduling module to generate control parameters and resource allocation strategies based on real-time data. The instruction distribution layer encapsulates the optimization results into control instruction packets and issues them to the corresponding node control units through the management unit.

[0023] The hardware configuration of the management computer includes: Processor: Intel Xeon E5-2678 v3 twelve-core processor, with a main frequency of 2.5 GHz, supporting AVX2 instruction set acceleration for matrix operations; Memory: 64 GB DDR4 ECC memory, with a data throughput rate of 2133 MT / s; Storage: 1 TB NVMe SSD for storing historical status data and optimization logs; Network interface: Dual gigabit Ethernet ports are bonded as an aggregated link to connect to the wireless communication network management unit.

[0024] The wireless communication network is responsible for data transmission between the management computer and each node control unit. Its design aims to ensure the real-time performance and reliability of control instructions, while optimizing the utilization rate of communication resources. The wireless communication network adopts a dual-channel architecture, combined with a dynamic resource allocation and priority scheduling mechanism, to effectively address the challenges brought by multi-node concurrent communication. The implementation method of the wireless communication network is described in detail below in combination with the technical solutions.

[0025] In this embodiment, the wireless communication network includes a management unit and multiple node units. The management unit is directly connected to the management computer and is responsible for coordinating the communication tasks of each node unit. The node units are connected to the node control units of each shuttleless loom and are responsible for issuing instructions and uploading status data. The network adopts a hierarchical protocol design. The physical layer supports dual-band communication (2.4 GHz and 5 GHz), the data link layer realizes time slot allocation and conflict avoidance, and the network layer provides routing and load balancing functions.

[0026] The main channel of the wireless communication network adopts the TSCH (Time-Slotted Channel Hopping) protocol, and the time slot length is 10 milliseconds. Within each time slot, the node unit transmits data according to the allocated time slot and channel frequency. The time slot allocation strategy is dynamically adjusted based on node priorities and communication requirements. Nodes with higher priorities (such as nodes sending emergency stop instructions) are allocated more time slot resources. The channel hopping sequence is generated by the management unit, and the hopping interval is 5 milliseconds, effectively reducing co-channel interference.

[0027] The backup channel adopts UWB (Ultra-WideBand) pulse communication technology, and the operating frequency band is from 3.1 GHz to 4.8 GHz. The backup channel is mainly used to transmit emergency instructions, such as broken yarn alarms or emergency stop signals. When the instruction queue length of the main channel exceeds a preset threshold the system automatically switches to the backup channel. The transmission delay of the UWB channel is less than 1 millisecond, ensuring the timely delivery of critical instructions.

[0028] The resource allocation mechanism of the wireless communication network is based on a game theory model. Each node unit is regarded as a game participant, and its action set , where is the transmission power, is the time slot selection. The utility function of participant is defined as: ; where is the instruction transmission delay, is the energy consumption, , is the weight coefficient. Through the iterative strategy update algorithm, the system calculates the Nash equilibrium and dynamically allocates communication resources.

[0029] In some embodiments, the load balancing mechanism of the wireless communication network includes the following steps: Channel utilization monitoring: Real-time calculation of the channel utilization rate , where is the communication rate of node , is the total channel bandwidth; Load adjustment: When > 0.8, trigger the game theory scheduling module to reallocate resources; Priority adjustment: Dynamically adjust the time slot allocation ratio according to the node task priority. Nodes with higher priorities are allocated more time slots.

[0030] The fault handling mechanism of the wireless communication network includes: Link quality monitoring: Real-time monitoring of the link quality index , where is the received power, and is the noise power; Link switching: When < 10, automatically switch to the backup channel; Transmission mechanism: For unacknowledged instructions, start exponential backoff retransmission, and set the upper limit of the retransmission times to 3 times, and the timeout threshold = 200 ms.

[0031] The extended technical content of the wireless communication network includes dynamically adjusting the channel bandwidth: Bandwidth allocation algorithm: Dynamically adjust the channel bandwidth allocation ratio according to the node communication requirements. For example, when the task priority of a certain node is higher, allocate more bandwidth resources; Adjustment rule: ; where is the allocated bandwidth of node , is the reference bandwidth, is the node priority, is the highest priority, is the adjustment coefficient.

[0032] The information geometry optimization module is responsible for optimizing the control parameters based on the statistical characteristics of the system state distribution. It realizes dynamic optimization under multi-objective constraints by constructing a statistical manifold, calculating the Fisher information matrix, and updating the parameters along the geodesic. This module works closely with the management computer, receives the loom state data in real time, and generates the optimal control strategy to ensure the efficient operation of the system in the wireless network environment. The implementation method of the information geometry optimization module is described in detail below in combination with the technical solutions.

[0033] The information geometry optimization module is deployed on the high-performance computing unit of the management computer and adopts a multi-threaded parallel computing framework. Its input is the loom state data from the node control unit, and the output is the optimized control parameter vector , where represents the th control parameter (such as motor speed, weft tension, etc.). The core algorithms of the module include statistical manifold construction, Fisher information matrix calculation, and parameter update.

[0034] The process of constructing the statistical manifold is as follows: Assume that the system state follows a parametric probability distribution , where is the random vector containing the loom state variables, is the control parameter vector; Fit the distribution parameters by the maximum likelihood estimation method to ensure It can accurately describe the statistical characteristics of the system state.

[0035] The calculation process of the Fisher information matrix is as follows: Define the Fisher information matrix The -th row and -th column element of is: ; where is the expectation operator, is the log-likelihood function; Approximate the expected value by the Monte Carlo sampling method, and the number of sampling points N = 1000.

[0036] The specific steps of parameter update include: Gradient calculation: Calculate the gradient of the loss function , and the loss function is defined as the weighted sum of delay, energy consumption, and control error: ; where , , is the weight coefficient, is the delay, is the energy consumption, is the variance of the control error; Parameter update: Update the control parameters along the geodesic direction of the statistical manifold: ; where is the learning rate, is the inverse matrix of the Fisher information matrix, is the gradient, is the loss function; Feasible region projection: Constrain the updated parameters to the physical feasible range by the Lagrange multiplier method.

[0037] The extended technical content of the information geometric optimization module includes dynamically adjusting the learning rate: Learning rate adjustment rule: Dynamically adjust the learning rate according to the optimization process , with the initial value = 0.1, and the adjustment formula is: ; where is the number of iterations, is the total number of iterations; Adaptive adjustment: When the change rate of the loss function is less than the threshold δ = 0.01, increase the learning rate to accelerate convergence. is the loss function value at the -th iteration, is the loss function value at the -th iteration.

[0038] The hardware acceleration implementation of the information geometry optimization module includes: Parallel computing architecture: A multi-core CPU and GPU are used for collaborative computing. The GPU is responsible for the parallel computing of the Fisher information matrix, and the CPU is responsible for parameter update and feasible region projection; Matrix inversion optimization: The QR decomposition method is used to accelerate the inversion operation of the Fisher information matrix, and the computational complexity is reduced from to .

[0039] In some embodiments, the fault handling mechanism of the information geometry optimization module includes: Abnormal detection: Real-time monitoring of abnormal situations during the optimization process, such as gradient explosion or matrix singularity; Parameter reset: When an abnormality is detected, immediately reset the control parameter to the initial value and restart the optimization process; Logging: Record the input data, intermediate results, and output parameters of each optimization to a log file for subsequent analysis and debugging.

[0040] The game theory scheduling module is responsible for optimizing the communication resource allocation through a non-cooperative game model. Its design aims to solve the resource competition problem brought about by multi-node concurrent communication, ensuring the timely transmission of high-priority instructions and the optimal overall system performance. This module works closely with the management computer, receiving node status data in real time and generating an equilibrium strategy to dynamically adjust the allocation ratio of communication resources.

[0041] The game theory scheduling module is deployed in the high-performance computing unit of the management computer and adopts a distributed computing framework. Its input is the communication demand data from the node control unit, and the output is the resource allocation scheme under the equilibrium strategy. The core algorithms of the module include game model definition, utility function design, and Nash equilibrium solution.

[0042] In some embodiments, the definition process of the game model is as follows: User set: Each node control unit is regarded as a game participant, and the user set , where is the total number of nodes; Action set: The action set of user , where is the transmission power, is the time slot selection; Strategy space: The strategy space of user is the action set Probability distribution on 。

[0043] The design process of the utility function is as follows: Definition of utility function: The utility function of user ii is the weighted negative value of delay and energy consumption: ; Where is the instruction transmission delay, is the energy consumption, , is the weight coefficient; Delay calculation: The delay consists of queuing delay and transmission delay: ; Where is the queuing delay, is the transmission delay; Energy consumption calculation: The energy consumption is determined by the transmission power and the communication duration: ; Where is the energy consumed by the node control unit during communication, is the transmission power of the node control unit.

[0044] The solution process of the Nash equilibrium is as follows: Strategy update: An iterative strategy update algorithm is adopted, and in each iteration, the user selects a strategy that maximizes its own utility: ; Where ∈(0, 1) is the strategy update rate, is the Dirac distribution of the best response action, is the strategy distribution of user 1 at the -th iteration; Convergence condition: When the strategy change amount is met, the algorithm converges, and ϵ = 0.01 is the convergence threshold.

[0045] In some embodiments, the resource allocation mechanism of the game theory scheduling module includes the following steps: Priority adjustment: Dynamically adjust the time slot allocation ratio according to the node task priority, and allocate more time slots to nodes with higher priority; Load balancing: Real-time monitor the channel utilization rate , when > 0.8U, trigger resource reallocation, is the maximum value; Collision Avoidance: Collisions caused by concurrent communication of multiple nodes are avoided through time slot allocation and channel hopping.

[0046] The extended technical content of the game theory scheduling module includes dynamically adjusting the weight coefficients: Weight Adjustment Rule: Dynamically adjust the weight coefficients α and β in the utility function according to the system state. For example, increase β in an energy consumption sensitive scenario to optimize energy consumption; Adjustment Formula: ; where is the reference weight, is the task deadline, is the adjustment coefficient.

[0047] In some embodiments, the hardware acceleration implementation of the game theory scheduling module includes: Parallel Computing Architecture: Adopt the collaborative computing of multi-core CPU and FPGA. The FPGA is responsible for the parallel computing of the utility function, and the CPU is responsible for policy update and equilibrium solution; Matrix Operation Optimization: Adopt sparse matrix storage and calculation technology to reduce the calculation complexity.

[0048] The fault handling mechanism of the game theory scheduling module includes: Abnormality Detection: Real-time monitor the abnormal situations during the game process, such as policy oscillation or abnormal utility function values; Policy Reset: When an abnormality is detected, immediately reset the policy distribution to the initial value and restart the game solution; Log Recording: Record the input data, intermediate results and output policies of each game into the log file for subsequent analysis and debugging.

[0049] The hardware acceleration and execution module is responsible for the efficient calculation and execution of control instructions. Its design aims to improve the calculation efficiency through hardware acceleration technology to ensure the real-time and accuracy of control instructions. This module closely cooperates with the node control unit, receives the optimization instructions from the management computer and executes them quickly, and at the same time feeds back the execution results to the management computer. The implementation method of the hardware acceleration and execution module is described in detail below in combination with the technical solutions.

[0050] The hardware acceleration and execution module is deployed on the embedded platform of the node control unit, and adopts the collaborative computing architecture of multi-core processor and FPGA. Its input is the optimization instructions from the management computer, and the output is the execution results and status feedback data. The core functions of the module include statistical characteristic matrix calculation, game theory scheduling engine and dynamic voltage and frequency regulation.

[0051] In some embodiments, the specific implementation of the statistical characteristic matrix calculation unit includes the following steps: Data preprocessing: Receive the loom status data from the node control unit , and perform normalization processing; Matrix calculation: Calculate the statistical property matrix of the system state distribution , and its element in the th row and th column is: ; where is the expectation operator, is the log-likelihood function; Result output: Send the calculated statistical property matrix to the information geometry optimization module.

[0052] The specific implementation of the game theory scheduling engine includes the following steps: Utility function calculation: Calculate the utility function according to the node communication requirements , where is the delay, is the energy consumption; Strategy update: Adopt the iterative strategy update algorithm, and in each iteration, the user selects the strategy that maximizes its own utility: ; where ∈(0, 1) is the strategy update rate, is the Dirac distribution of the best response action, is the strategy distribution of user 1 at the th iteration; Result output: Send the calculated equilibrium strategy to the game theory scheduling module.

[0053] In some embodiments, the specific implementation of the dynamic voltage and frequency regulation unit includes the following steps: Load monitoring: Monitor the system load in real time , which is defined as the weighted sum of the CPU utilization rate and the memory utilization rate: ; where is the CPU utilization rate, is the memory utilization rate, is the weight coefficient; Frequency regulation: Dynamically adjust the computing frequency according to the system load : ; where is the reference frequency, is the load change rate.

[0054] The extended technical content of the hardware acceleration and execution module includes dynamically adjusting computing resources: Resource allocation algorithm: Dynamically adjust the computing resource allocation ratio according to task priorities. For example, when a certain task has a higher priority, allocate more computing resources; Adjustment rules: ; where is the allocated resource for task .

[0055] In some embodiments, the fault handling mechanism of the hardware acceleration and execution module includes: Abnormality detection: Real-time monitor the abnormal conditions during the computing process, such as computing overflow or memory leak; Resource reset: When an abnormality is detected, immediately reset the computing resource allocation to the initial value and restart the computing task; Log recording: Record the input data, intermediate results, and output data of each computation into a log file for subsequent analysis and debugging.

[0056] The fault detection and self-healing module is responsible for real-time monitoring of the system status and automatically handling faults. Its design aims to ensure the stable operation of the system in a complex industrial environment through an efficient fault detection and self-healing mechanism. This module closely cooperates with the node control unit, real-time collects sensor data, analyzes fault characteristics, and quickly initiates the self-healing process to reduce downtime. The implementation method of the fault detection and self-healing module is described in detail below in combination with the technical solutions.

[0057] In this embodiment, the fault detection and self-healing module is integrated into the embedded platform of the node control unit, and includes a broken yarn detection unit, a time delay compensation unit, and a self-healing control unit. Its input is real-time data from the sensor interface (such as optoelectronic sensor signals, communication time delay metrics), and the output is a fault alarm signal and a self-healing control instruction. The module realizes the rapid identification and recovery of faults through a multi-level threshold judgment and prediction algorithm.

[0058] In some embodiments, the implementation of the broken yarn detection unit includes the following steps: Signal acquisition: Collect the yarn tension signal through an optoelectronic sensor array , with a sampling frequency of 1 MHz; Abnormality determination: Calculate the moving window mean and standard deviation of the signal amplitude. When = 5 consecutive sampling points satisfy , it is determined as a broken yarn event; Standby yarn switching: Trigger the solenoid valve to control the standby yarn to cut in, and the switching time <50 .

[0059] In some embodiments, the implementation of the time delay compensation unit includes the following steps: Time delay modeling: Establish a stochastic process model of the communication time delay , where is the fixed time delay, obeys the normal distribution random perturbation; Predictive observer design: Design a predictive observer based on the delay differential equation: ; where , , is the system state matrix, is the observer gain matrix, which is calculated by solving the Riccati equation ; Compensation control: Generate a compensation instruction according to the predicted state , where is the feedback gain matrix.

[0060] In some embodiments, the implementation of the self-healing control unit includes the following steps: Fault classification: Classify faults into three levels: Level 1 fault (urgent): Yarn breakage, mechanical overload; Level 2 fault (important): Communication time delay exceeds the limit > 200 ms; Level 3 fault (general): Sensor data drift; Self-healing strategy: Level 1 fault: Immediately stop the machine and start the standby control loop; Level 2 fault: Reduce the node communication rate and retransmit the instruction; Level 3 fault: Switch to the redundant sensor and report to the management computer.

[0061] The extended technical content of the fault detection and self-healing module includes dynamically adjusting the detection threshold: Adaptive threshold algorithm: Dynamically adjust the yarn breakage detection threshold according to historical fault data : Threshold constraint: Set the upper and lower limits = 0.1, = 0.5, to prevent over-sensitivity or insensitivity.

[0062] The hardware implementation of the fault detection and self-healing module includes: High-speed comparator: Use FPGA to implement real-time signal amplitude comparison, with a delay < 1 μs; Predictive Observer Acceleration: Matrix operation acceleration is achieved through a coprocessor (such as an ARM Cortex-M7), with a computing frequency = 100 kHz; Redundant Control Loop: A dual-MCU architecture is deployed, automatically switching to the standby controller in case of main controller failure.

[0063] The logging and debugging functions of the fault detection and self-healing module include: Data Logging: Store sensor data, fault events, and self-healing operations for the most recent 24 hours; Remote Diagnosis: Upload fault logs to the management computer via a wireless communication network, supporting remote parameter adjustment; Self-Test Mode: Periodically perform sensor calibration and communication link tests to ensure the reliability of the module.

[0064] Through the above implementation, the fault detection and self-healing module realizes a closed-loop control from fault detection, classification to the execution of self-healing strategies, significantly enhancing the robustness and fault tolerance of the system.

[0065] Working Principle: Connect the management computer and multiple node control units through a wireless communication network to form an efficient and collaborative control system. The management computer is responsible for global optimization and task scheduling, generating optimal control strategies based on the information geometry optimization module and the game theory scheduling module, and sending them to each node control unit via the wireless network. The node control unit executes the instructions and collects loom status data, feeding it back to the management computer for dynamic strategy adjustment. The hardware acceleration and execution module ensure the real-time nature of computing and control, while the fault detection and self-healing module monitors the system status in real time, automatically handling faults to ensure the stable operation of the system.

[0066] Although the embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill 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 present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A control system for multiple shuttleless looms based on a wireless network, characterized in that, Including: A management computer for global optimization and task scheduling; A wireless communication network, including a management unit and multiple node units, where the management unit is connected to the management computer through an Ethernet interface; Multiple node control units, where the node units are connected to the node control units through radio frequency modules, used to realize data transmission between the management computer and the node control units. Each node control unit is installed in a control box of a shuttleless loom and is connected to the control function unit of the shuttleless loom control box through a serial communication interface, used to execute control instructions and collect loom status data; An information geometry optimization module, connected to the management computer through a software interface, used to optimize control parameters based on the statistical characteristics of the system state distribution; A game theory scheduling module, connected to the management computer through a software interface, used to optimize communication resource allocation through a non-cooperative game model; A hardware acceleration and execution module, connected to the node control units through an SPI bus, used to achieve efficient computing and execution of control instructions; A fault detection and self-healing module, connected to the node control units through a GPIO interface, used to monitor the system state in real time and achieve fault self-healing; Among them, the management computer issues control instructions to the node control units through the wireless communication network, and the node control units distribute the control instructions to each control function unit of the shuttleless loom control box through the serial communication interface and feedback the status data to the management computer.

2. The multi - shuttleless loom control system based on a wireless network according to claim 1, characterized in that, The wireless communication network includes a main channel and a backup channel. The main channel adopts a time slot allocation protocol, and the time slot length is 10 milliseconds - 13 milliseconds. The backup channel adopts a pulse communication protocol. When the instruction queue length of the node control unit exceeds a preset threshold, it automatically switches to the backup channel to transmit emergency instructions.

3. The multi-shuttleless loom control system based on a wireless network according to claim 1, characterized in that, The information geometry optimization module updates the control parameters by calculating the statistical characteristic matrix of the system state distribution and along the statistical manifold, specifically including: Calculating the statistical characteristic matrix of the system state distribution; Updating the control parameters based on the inverse matrix of the statistical characteristic matrix and the gradient of the loss function; Projecting the updated control parameters onto the feasible region.

4. The multi-loom control system based on a wireless network according to claim 1, characterized in that, The game theory scheduling module calculates the equilibrium strategy by defining a non-cooperative game model, specifically including: Defining the user set as all node control units, the action set as transmission power and time slot selection, and the utility function as the weighted negative value of delay and energy consumption; Calculating the equilibrium strategy by iteratively updating the strategy distribution; Using the equilibrium strategy result for communication resource allocation.

5. The multi-loom control system based on a wireless network according to claim 1, characterized in that, The hardware acceleration and execution module includes: A statistical characteristic matrix calculation unit, used to calculate the statistical characteristic matrix of the system state distribution in parallel and send the calculation result to the information geometry optimization module; A game theory scheduling engine, used to quickly calculate the equilibrium strategy and send the calculation result to the game theory scheduling module; A dynamic voltage and frequency regulation unit, used to dynamically adjust the calculation frequency according to the system load and send the adjustment result to the node control units.

6. The multi-shuttleless loom control system based on a wireless network according to claim 1, characterized in that, The fault detection and self-healing module includes: A broken yarn detection unit, used to detect broken yarn events through a photoelectric sensor array and a high-speed analog-to-digital converter and send the detection result to the node control unit; The time-delay compensation unit is used to compensate for the communication time-delay through a predictive observer and send the compensation result to the node control unit; The self-healing control unit is used to automatically start a standby control strategy when a fault is detected and send the fault information to the management computer.

7. The multi-shuttleless loom control system based on a wireless network according to claim 1, characterized in that, The management computer issues control instructions to the node control unit through a wireless communication network. The control instructions include communication rate, computing frequency, and control parameters. The node control unit distributes the control instructions to each control function unit of the shuttleless loom control box through a serial communication interface and feeds back the execution result to the management computer.

8. The multi - shuttleless loom control system based on a wireless network according to claim 1, characterized in that, The node control unit includes a main control chip, a communication interface, and a sensor interface. The main control chip uses a multi-core processor. The communication interface supports a serial bus and a controller area network bus. The sensor interface is used to connect an optoelectronic sensor array.

9. The multi-shuttleless loom control system based on a wireless network according to claim 1, characterized in that, The information geometry optimization module and the game theory scheduling module are called by the management computer. The management computer dynamically adjusts the optimization objectives and constraints according to the status data of the node control unit. The optimization objectives include the weighted sum of time-delay, energy consumption, and control error.

10. The multi-shuttleless loom control system based on a wireless network according to claim 1, characterized in that, The fault detection and self-healing module collects sensor data in real time through the node control unit. When a broken yarn or a communication time-delay limit is detected, it automatically triggers a self-healing process and reports the fault information to the management computer. The management computer recalculates the optimization strategy according to the fault information and issues it.

Images