Dynamic intelligent system for purple light laser marking machine

Abstract

The invention relates to the technical field of intelligent marking, and discloses a dynamic intelligent system for a purple light laser marking machine, which comprises the following modules: an intelligent identification module used for identifying and analyzing a workpiece, a data processing and analyzing module, a marking parameter dynamic adjusting module used for dynamically adjusting relevant parameters of the purple light laser marking machine in real time, the data processing and analysis module is used for processing and analyzing the data based on a preset algorithm, and the motion control module is used for controlling the marking head to move according to setting. According to the method, the self-adaptive control algorithm is referred to by the established model, the marking parameters can be adjusted in a self-adaptive manner by constructing the reference model and comparing the actual output error, the workpiece and environment change can be effectively coped, the marking quality stability and consistency are improved, and the marking efficiency is improved. And meanwhile, the system characteristics do not need to be accurately predicted, the laser marking parameters do not need to be manually controlled in the process, and the adaptive capacity and robustness of the marking system are enhanced.

Description

Technical Field

[0001] The present invention relates to the field of intelligent marking technology, and in particular to a dynamic intelligent system for a purple laser marking machine. Background Art

[0002] The ultraviolet laser marking machine is a device that uses a 355nm ultraviolet laser to mark materials. Its working principle is photochemical ablation, relying on short-wavelength lasers to break the molecular chain and engrave the content. It has significant characteristics, with a small focused spot, which can achieve ultra-fine marking; a small heat-affected zone, which is a cold processing, can reduce the mechanical deformation of the material, and is suitable for heat-sensitive materials; it is applicable to a wide range of materials, including a variety of materials except copper; the beam quality is good, the marking is clear and stable; the overall performance of the machine is stable, the size is small, the power consumption is low, and the life is long. It has a wide range of applications, and is used in the electronics industry to mark chips and other information, the pharmaceutical and food industries to mark related products, and is also used in the cosmetics, gifts and other industries to meet the different marking needs of various industries.

[0003] Laser marking technology has been widely used in many fields due to its high efficiency, precision and long-lasting marking characteristics. With its unique short wavelength advantage, the Ziguang laser marking machine can achieve extremely fine marking and plays a key role in industries such as high-end electronic components, precision medical equipment, and fine jewelry.

[0004] In existing laser marking machine motion control technology, traditional fixed parameter control strategies, on the one hand, are difficult to adapt to workpieces of varying materials due to the complex interaction between the laser and the material, often resulting in unstable marking depth and clarity, seriously affecting product quality consistency. For example, when marking workpieces made of a mixture of metal and plastic, it is impossible to account for the differences in laser energy absorption between the two, resulting in poor marking results in some areas. Furthermore, environmental factors (such as temperature and humidity changes) can affect the laser output characteristics and the performance of the marking machine's mechanical components. Traditional control technologies are unable to automatically adjust these characteristics, requiring frequent manual calibration, which significantly reduces production efficiency. In light of this, we propose a dynamic intelligent system for ultraviolet laser marking machines. Summary of the Invention

[0005] In response to the shortcomings of the existing technology, the present invention provides a dynamic intelligent system for a purple laser marking machine, which solves the problem that fixed parameters in the existing technology are difficult to match different marking requirements, and that due to environmental influences, there is a large difference between the actual marking effect and the preset marking effect.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: A dynamic intelligent system for a purple laser marking machine, comprising the following modules:

[0007] Intelligent recognition module: used to identify and analyze the material, shape and pattern to be marked of the workpiece;

[0008] Data processing and analysis module: connected to the intelligent recognition module, receiving the recognition data transmitted by it, and processing and analyzing the data based on the preset algorithm to generate a marking parameter adjustment strategy;

[0009] Marking parameter dynamic adjustment module: according to the adjustment strategy generated by the data processing and analysis module, the laser power, frequency, pulse width and marking speed of the purple laser marking machine are dynamically adjusted in real time;

[0010] Motion control module: used to control the marking head of the purple laser marking machine to perform marking movement according to the preset path and adjusted parameters;

[0011] Storage module: used to store recognition data, preset algorithms, marking parameters and marking task related information;

[0012] Communication module: used to realize data transmission and interaction between the system and external devices or host computers;

[0013] Fault diagnosis and early warning module: used to monitor the operating status of each module of the system in real time and issue an early warning signal when an abnormality is detected.

[0014] Preferably, the intelligent recognition module includes an image acquisition unit and a material detection unit. The image acquisition unit uses a high-resolution camera to capture images of the workpiece surface and the pattern to be marked, and the material detection unit detects and identifies the workpiece material through spectral analysis technology.

[0015] Preferably, the preset algorithm in the data processing and analysis module includes a deep learning algorithm, which optimizes the marking parameter adjustment strategy by learning a large amount of workpieces of different materials, shapes and marking pattern data.

[0016] Preferably, the marking parameter dynamic adjustment module adopts a closed-loop control method to obtain feedback data during the marking process in real time and transmit the feedback data to the data processing and analysis module.

[0017] Preferably, the motion control module uses a high-precision motor and transmission device and has a motion feedback mechanism, which can realize the positioning and movement of the marking head in three-dimensional space and can compensate and adjust for movement deviations.

[0018] Preferably, the storage module adopts blockchain distributed storage technology to distribute the key marking data, algorithms and task records, etc., to ensure the security, non-tamperability and traceability of the data, and facilitate data management in multi-device collaborative operations or data audit scenarios.

[0019] Preferably, the fault diagnosis and early warning module uses big data analysis and machine learning models to not only monitor the operating status of each module of the system in real time, but also predict potential fault risks based on historical fault data and provide maintenance prompts in advance.

[0020] Preferably, the fault diagnosis and warning module will simultaneously detect deviations or errors that occur during the marking process, including discontinuous marking lines and missing characters. The system can automatically start the intelligent error correction program and automatically re-mark and repair the erroneous parts by replanning the marking path.

[0021] Preferably, the communication module includes a data transmission module and a human-computer interaction module. The data transmission module is used to transmit various data during the movement and control process of the laser marking machine. The human-computer interaction module is equipped with an augmented reality display function. The operator can intuitively preview the marking effect on the workpiece through the AR device, adjust the marking position and angle parameters in real time, and interact with the system through gesture recognition technology.

[0022] Preferably, the communication module supports 5G communication technology and edge computing architecture, which is used to achieve high-speed, low-latency data transmission between the system and external devices, while performing partial data processing on local edge devices to reduce data transmission pressure.

[0023] The present invention provides a dynamic intelligent system for a purple laser marking machine. It has the following beneficial effects:

[0024] 1. The present invention establishes a model reference adaptive control algorithm, which can adaptively adjust marking parameters by building a reference model and comparing the actual output error, effectively responding to changes in the workpiece and environment, and improving the stability and consistency of marking quality. At the same time, there is no need to accurately predict system characteristics and no need for manual control of laser marking parameters during the process, thereby enhancing the adaptability and robustness of the marking system.

[0025] 2. The present invention utilizes an image processing-based deviation detection algorithm and path planning algorithm within the fault diagnosis and early warning module to automatically identify problems such as discontinuous lines and missing characters during the marking process, providing an accurate basis for subsequent error correction. The path planning algorithm can then quickly plan the optimal re-marking and repair path based on the detected deviations. The two algorithms work together to significantly reduce manual intervention, labor costs, and human error. Furthermore, they can promptly detect and correct marking errors, improving product quality and consistency.

[0026] 3. The present invention closely combines the established PID control algorithm with feedback control and applies it to the motion control module. Through the real-time feedback of the actual position information of the marking head, the error between the target and the actual position is continuously calculated, and the control signal is quickly adjusted through the proportional, integral and differential links. This not only speeds up the response speed of the marking head so that it can quickly approach the target position, but also effectively eliminates steady-state errors, reduces overshoot, enhances system stability, and allows the marking head to be accurately positioned in three-dimensional space, ensuring that the marking operation is carried out stably, efficiently and with high precision, significantly improving the marking quality and production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 This is the module diagram of the dynamic intelligent system for this purple laser marking machine;

[0028] Figure 2 Schematic diagram of the adaptive control algorithm flow of the present invention;

[0029] Figure 3 It is a schematic diagram of the PID control algorithm and feedback control flow of the present invention. DETAILED DESCRIPTION

[0030] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the specification of the present invention. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0031] Example:

[0032] Please see the attached Figure 1 - Attachment Figure 3 The embodiment of the present invention provides a dynamic intelligent system for a purple laser marking machine, including the following modules:

[0033] Intelligent recognition module: used to identify and analyze the material, shape and pattern to be marked of the workpiece;

[0034] Data processing and analysis module: connected to the intelligent recognition module, receiving the recognition data transmitted by it, and processing and analyzing the data based on the preset algorithm to generate a marking parameter adjustment strategy;

[0035] Marking parameter dynamic adjustment module: according to the adjustment strategy generated by the data processing and analysis module, the parameters of the purple laser marking machine including laser power, frequency, pulse width and marking speed are adjusted dynamically in real time;

[0036] Motion control module: used to control the marking head of the purple laser marking machine to perform marking movement according to the preset path and adjusted parameters;

[0037] Storage module: used to store recognition data, preset algorithms, marking parameters and marking task related information;

[0038] Communication module: used to realize data transmission and interaction between the system and external devices or host computers;

[0039] Fault diagnosis and early warning module: used to monitor the operating status of each module of the system in real time and issue an early warning signal when an abnormality is detected.

[0040] The intelligent recognition module includes an image acquisition unit and a material detection unit. The image acquisition unit uses a high-resolution camera to capture images of the workpiece surface and the pattern to be marked, and the material detection unit detects and identifies the workpiece material through spectral analysis technology.

[0041] The preset algorithm in the data processing and analysis module includes a deep learning algorithm, which optimizes the marking parameter adjustment strategy by learning a large amount of workpiece data of different materials, shapes and marking patterns.

[0042] The marking parameter dynamic adjustment module adopts a closed-loop control method to obtain feedback data during the marking process in real time and transmit the feedback data to the data processing and analysis module. The following algorithm is established for the dynamic adjustment of parameters during the marking process:

[0043] Establishment of system model

[0044] In the scenario of dynamic adjustment of laser marking parameters, the characteristics of the marking system will change due to changes in the workpiece material, shape, and environmental factors. The core idea of ​​model reference adaptive control is to build a reference model that represents the ideal performance expected to be achieved by the marking system. The actual marking system (controlled object) will be compared with the reference model, and the controller parameters will be adjusted in real time based on the error between the two outputs, so that the output of the actual system can track the output of the reference model as much as possible to adapt to different working conditions;

[0045] Assuming that the reference model is a linear time-invariant system, its state space expression can be expressed as:

[0046]

[0047] in:

[0048] x m (t) is the state vector of the reference model, with dimension n × 1;

[0049] A m is the n×n system matrix, which determines the dynamic characteristics of the reference model;

[0050] B mis the n×r input matrix, where r is the dimension of the input signal;

[0051] r(t) is the reference input signal, such as the parameter setting value corresponding to the desired marking effect;

[0052] C m is the p×n output matrix, where p is the dimension of the output signal;

[0053] y m (t) is the output of the reference model, that is, the ideal marking effect response. In a discrete time system, the reference model can be expressed as:

[0054]

[0055] The actual laser marking system can be expressed as:

[0056]

[0057] in:

[0058] x(t) is the state vector of the actual system.

[0059] A(t), B(t) and C(t) are the system matrix, input matrix and output matrix respectively. Since the system characteristics change with time, these matrices are time-varying.

[0060] u(t) is the control input, that is, the marking parameters that need to be adjusted (such as laser power, frequency, etc.).

[0061] y(t) is the output of the actual system, that is, the actual marking effect feedback.

[0062] In discrete-time systems, the actual system is represented as:

[0063]

[0064] Error Definition and Analysis

[0065] Define the error vector e(t)=y m (t)-y(t), which is the difference between the reference model output and the actual system output. The goal of model reference adaptive control is to minimize the error e(t) by adjusting the control input u(t) and eventually approach zero.

[0066] Design of adaptive law

[0067] The adaptive law is used to adjust the controller parameters according to the error e(t). Common design methods are based on the Lyapunov stability theory to ensure the stability of the system.

[0068] Direct Adaptive Control

[0069] Assume that the controller structure is u(t) = K(t)r(t), where K(t) is the adaptive gain matrix that needs to be adjusted in real time according to the error;

[0070] Define a Lyapunov function in K * is an ideal gain matrix. Choose a suitable Lyapunov function so that its derivative Negative definite or semi-negative definite;

[0071] For example, for a simple single-input single-output system, choose Where γ>0 is the step size of adaptive gain adjustment;

[0072] right Derivative:

[0073]

[0074] By substituting the state equation and error equation of the system, after a series of derivations and simplifications, the adaptive law is obtained:

[0075]

[0076] In discrete-time systems, the adaptive law can be expressed as:

[0077] ΔK[k]=γe[k]r[k]

[0078] Where ΔK[k]=K[k+1]-K[k]

[0079] Indirect adaptive control

[0080] Indirect adaptive control first estimates the system parameters (such as A(t), B(t)) online, and then designs the controller based on the estimated parameters. For example, the least squares method is used to estimate the parameters. The update formula is:

[0081]

[0082] Where Γ is the gain matrix and φ(t) is the regression vector.

[0083] Application in laser marking parameter adjustment

[0084] Parameter initialization: Before starting marking, set the reference model parameter A m 、B m 、C m , and the initial parameter K(0) of the adaptive controller;

[0085] Real-time feedback and adjustment: During the marking process, the actual system output y(t) and the reference model output y m (t) are obtained in real time, and the error e(t) is calculated. According to the adaptive law, the parameters K(t) of the controller are adjusted, thereby changing the control input u(t), i.e. adjusting the marking parameters such as laser power, frequency, pulse width and marking speed;

[0086] Continuous optimization: As the marking process proceeds, the above feedback and adjustment process is repeatedly performed, so that the actual marking effect gradually approaches the ideal effect represented by the reference model, realizing adaptive adjustment for different workpieces and working conditions.

[0087] The motion control module adopts high-precision motors and transmission devices, and has a motion feedback mechanism, which can realize the positioning and motion of the marking head in three-dimensional space, and can compensate and adjust for movement deviation. For motor control and transmission device control during the motion process, we establish the following algorithm:

[0088] PID (Proportional-Integral-Derivative) control algorithm is used to realize the precise position control of the marking head in three-dimensional space. The core idea is to generate corresponding control signals by calculating the error between the target position and the actual position of the marking head through the proportional (P), integral (I) and derivative (D) three links, to drive the high-precision motor and transmission device, so that the marking head moves towards the target position and finally stabilizes at the target position;

[0089] Formula

[0090]

[0091] Where e(t) = r(t) - y(t);

[0092] r(t) is the target position;

[0093] y(t) is the actual position;

[0094] K p is the proportional coefficient, which determines the response strength of the controller to the current error;

[0095] K i is the integral coefficient, used to handle the steady-state error of the system;

[0096] K d is the derivative coefficient, mainly used to predict the trend of error change and adjust in advance;

[0097] Role of each link

[0098] Proportional link u P (t) = K p e(t), to speed up the response;

[0099] Points Eliminate steady-state errors;

[0100] Differentiation link Predict error trends, reduce overshoot, and be sensitive to noise.

[0101] Derivation of discrete-time formula

[0102] The sampling period is T, k represents the discrete moment, and the integral term is approximately The differential term is approximately The discrete formula is:

[0103]

[0104] Parameter tuning method

[0105] Trial and error method: adjust K first p To a slight overshoot, add integral link to adjust K i Eliminate the steady-state error and add a differential link to adjust K d Reduce overshoot;

[0106] Application process in motion control module

[0107] Initialization: Set the target position r[k] and initialize the parameter K of the PID controller p , K i and K d , and the initial value of the integral term

[0108]

[0109] Sampling: At each sampling time k, the actual position y[k] of the marking head is obtained through the motion feedback mechanism;

[0110] Error calculation: calculate the error at the current moment e[k] = r[k] - y[k];

[0111] PID calculation: Calculate the control output u[k] according to the discrete-time PID controller formula;

[0112] Control execution: convert the control output u[k] into the control signal of the motor, drive the motor and transmission device, and move the marking head;

[0113] Loop iteration: Repeat the above sampling, error calculation, PID calculation and control execution steps until the marking head reaches the target position and the error e[k] is less than the set threshold.

[0114] The storage module adopts blockchain distributed storage technology to distribute the key marking data, algorithms and task records, ensuring the security, non-tamperability and traceability of the data, and facilitating data management in multi-device collaborative operations or data audit scenarios.

[0115] The fault diagnosis and early warning module uses big data analysis and machine learning models to not only monitor the operating status of each module in the system in real time, but also predict potential fault risks based on historical fault data and provide maintenance prompts in advance.

[0116] The fault diagnosis and warning module will simultaneously detect deviations or errors that occur during the marking process, including discontinuous marking lines and missing characters. The system can automatically start the intelligent error correction program and automatically re-mark and repair the erroneous parts by replanning the marking path. The following algorithm is established in the process:

[0117] Deviation Detection Algorithm Based on Image Processing

[0118] Principle: During the marking process, a high-resolution camera is used to capture images of the marking area in real time, which are then compared and analyzed with pre-set standard marking images. Image processing technologies such as edge detection and feature matching are used to detect deviations such as discontinuous marking lines and missing characters.

[0119] Algorithm formula:

[0120] Filtering: Filter the collected image I(x,y) to obtain the smoothed image G(x,y). The filtering formula is:

[0121]

[0122] Where σ is the standard deviation of the distribution, (x0, y0) is the pixel coordinates* represents the convolution operation;

[0123] Calculate the gradient magnitude and direction: Calculate the gradient magnitude M(x,y) and direction θ(x,y) of the smoothed image:

[0124]

[0125] Among them G x and G y are the gradients of the image in the x and y directions respectively;

[0126] Non-maximum suppression: Perform non-maximum suppression on the gradient amplitude, retain the maximum value of the edge, and remove non-edge noise points;

[0127] Dual threshold detection and edge connection: Set a high threshold T h and low threshold T l , the gradient amplitude is greater than T jThe point is determined as a strong edge point, which is greater than T l And the points connected to the strong edge points are determined as weak edge points, thus detecting the edges in the image. By comparing the standard image with the detected edges, it is determined whether there is any deviation in the marking;

[0128] Path planning algorithm, used to replan the marking path

[0129] Principle: The marking area is abstracted into a graph structure. The nodes in the graph represent the possible locations of the marking head, the edges represent the connection relationship between nodes, and the edge weight represents the cost (such as time, distance, etc.) required for the marking head to move from one node to another. The algorithm starts from the starting node (i.e., the current location of the marking head) and gradually searches for the shortest path to the target node (i.e., the location that needs to be re-marked or repaired).

[0130] Algorithm formula:

[0131] Initialization: Let G = (V, E), where V is the set of nodes and E is the set of edges. Let the starting node be s, and the distance array d[v] represents the shortest distance from the starting node s to the node v. Initially, d[s] = 0. For other nodes v ≠ s, d[v] = ∞. Let S be a set to record the nodes for which the shortest path has been determined. Initially, S = {s}.

[0132] Iteration: In each iteration, select a node u from VS so that d[u] is minimized. Add node u to set S. For a node v adjacent to node u, if d[u]+w(u,v) <d[v](其中w(u,v)是边(u,v)的权重),则更新d[v]=d[u]+w(u,v)。重复迭代过程,直到所有节点都被加入集合S或者找到目标节点。最终得到的d数组中,目标节点对应的d值即为从起始节点到目标节点的最短路径长度,通过回溯可以得到具体的路径;

[0133] Therefore, a deviation detection algorithm and a path planning algorithm based on image processing are jointly utilized. The former uses image processing technology to accurately compare the real-time marking image with the standard image, sensitively detecting subtle deviations such as discontinuous marking lines and missing characters, providing an accurate basis for subsequent error correction. The latter constructs the marking area into a graph structure, efficiently finding the optimal path from the current position to the area requiring repair. The two algorithms work together to provide automatic error correction capabilities for the marking process, significantly improving marking quality, reducing scrap rates, and reducing manual intervention, thereby improving production efficiency and effectively ensuring the precise and efficient execution of marking work.

[0134] The communication module includes a data transmission module and a human-computer interaction module. The data transmission module is used to transmit various data during the movement and control process of the laser marking machine. The human-computer interaction module is equipped with an augmented reality display function. The operator can intuitively preview the marking effect on the workpiece through the AR device, adjust the marking position and angle parameters in real time, and interact with the system through gesture recognition technology.

[0135] The communication module supports 5G communication technology and edge computing architecture, and is used to achieve high-speed, low-latency data transmission between the system and external devices, while performing partial data processing on local edge devices to reduce data transmission pressure.

[0136] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A dynamic intelligent system for a purple laser marking machine, characterized in that: Includes the following modules: Intelligent recognition module: used to identify and analyze the material, shape and pattern to be marked of the workpiece; Data processing and analysis module: connected to the intelligent recognition module, receiving the recognition data transmitted by it, and processing and analyzing the data based on the preset algorithm to generate a marking parameter adjustment strategy; Marking parameter dynamic adjustment module: according to the adjustment strategy generated by the data processing and analysis module, the parameters of the purple laser marking machine including laser power, frequency, pulse width and marking speed are adjusted dynamically in real time; Motion control module: used to control the marking head of the purple laser marking machine to perform marking movement according to the preset path and adjusted parameters; Storage module: used to store recognition data, preset algorithms, marking parameters and marking task related information; Communication module: used to realize data transmission and interaction between the system and external devices or host computers; Fault diagnosis and early warning module: used to monitor the operating status of each module of the system in real time and issue an early warning signal when an abnormality is detected.

2. A dynamic intelligent system for a purple laser marking machine according to claim 1, characterized in that: The intelligent recognition module includes an image acquisition unit and a material detection unit. The image acquisition unit uses a high-resolution camera to capture images of the workpiece surface and the pattern to be marked, and the material detection unit detects and identifies the workpiece material through spectral analysis technology.

3. The dynamic intelligent system for a purple laser marking machine according to claim 1, characterized in that: The preset algorithm in the data processing and analysis module includes a deep learning algorithm, which optimizes the marking parameter adjustment strategy by learning a large amount of workpiece data of different materials, shapes and marking patterns.

4. The dynamic intelligent system for a purple laser marking machine according to claim 1, characterized in that: The marking parameter dynamic adjustment module adopts a closed-loop control method to obtain feedback data during the marking process in real time and transmits the feedback data to the data processing and analysis module.

5. The dynamic intelligent system for a purple laser marking machine according to claim 1, characterized in that: The motion control module uses a high-precision motor and transmission device and has a motion feedback mechanism, which can realize the positioning and movement of the marking head in three-dimensional space and can compensate and adjust for movement deviations.

6. The dynamic intelligent system for a purple laser marking machine according to claim 1, characterized in that: The storage module adopts blockchain distributed storage technology to distribute the key marking data, algorithms and task records, ensuring the security, non-tamperability and traceability of the data, and facilitating data management in multi-device collaborative operations or data audit scenarios.

7. The dynamic intelligent system for a purple laser marking machine according to claim 1, characterized in that: The fault diagnosis and early warning module uses big data analysis and machine learning models to not only monitor the operating status of each module in the system in real time, but also predict potential fault risks based on historical fault data and provide maintenance prompts in advance.

8. The dynamic intelligent system for a purple laser marking machine according to claim 7, characterized in that: The fault diagnosis and warning module will simultaneously detect deviations or errors that occur during the marking process, including discontinuous marking lines and missing characters. The system can automatically start the intelligent error correction program and automatically re-mark and repair the erroneous parts by replanning the marking path.

9. The dynamic intelligent system for a purple laser marking machine according to claim 1, characterized in that: The communication module includes a data transmission module and a human-computer interaction module. The data transmission module is used to transmit various data during the movement and control process of the laser marking machine. The human-computer interaction module is equipped with an augmented reality display function. The operator can intuitively preview the marking effect on the workpiece through the AR device, adjust the marking position and angle parameters in real time, and interact with the system through gesture recognition technology.

10. A dynamic intelligent system for a purple laser marking machine according to claim 9, characterized in that: The communication module supports 5G communication technology and edge computing architecture, and is used to achieve high-speed, low-latency data transmission between the system and external devices, while performing partial data processing on local edge devices to reduce data transmission pressure.

Images