Port mixed service traffic cooperative transmission method based on 5G-TSN logic network bridge

Abstract

The invention relates to the technical field of port communication, in particular to a port mixed service traffic cooperative transmission method based on a 5G-TSN logic network bridge. The method comprises the following steps: periodically acquiring context information of a wireless channel between a 5G base station and a port mobile terminal by using the 5G base station, and collecting historical transmission feedback data; constructing a channel state evaluation model based on ridge regression, and establishing a 5G air interface retransmission time model based on the channel state information; a cluster-based collaborative learning mechanism is introduced, and a channel state evaluation model is jointly trained and updated; the method comprises the following steps: dynamically adjusting an attenuation factor and a time sensitive weight of a parameter in a channel state evaluation model according to a deviation between an actual transmission time delay and a time delay budget in a 5G domain, compensating time delay jitter of a 5G air interface through a 5G-TSN mixed flow scheduling strategy by using a deterministic transmission capability of the TSN, and determining the time delay jitter of the 5G air interface according to a service type, the time delay budget in the 5G domain and the size of a remaining time window. The flow priority is dynamically set, queue mapping is determined by 5G and TSN transmission conditions, and iterative adjustment is fed back.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of port communication, and in particular to a port mixed service traffic cooperative transmission method based on a 5G-TSN logical bridge. BACKGROUND

[0002] With the acceleration of the construction of smart ports and automated wharfs, automated quay cranes (QC), automated guided vehicles (AGV), yard cranes (RTG), high-definition video monitoring, and multi-source sensors need to achieve high reliability, low latency, and deterministic data transmission in a unified network. The current widely used fifth-generation mobile communication (5G) network has significant advantages in providing large bandwidth, low latency, and massive connections, while the time-sensitive network (Time-Sensitive Networking, TSN) can guarantee the real-time performance and jitter controllability of end-to-end transmission through time-aware scheduling and deterministic forwarding. Based on the 3GPP R16 standard, the 5G-TSN logical bridge architecture provides a new solution for the deep integration of wireless and wired networks in ports. However, the complexity of the port communication environment and the diversity of business needs pose unprecedented challenges to 5G-TSN joint transmission: (1) The port communication conditions have a high degree of dynamicity, and are affected by harsh physical factors such as salt spray and dust, as well as the high mobility of terminal nodes. It is difficult to accurately obtain and frequently change the channel state information (CSI). This makes it difficult to accurately predict the error rate and retransmission number of the 5G link. Traditional scheduling algorithms that rely on real-time CSI perform poorly in terms of end-to-end delay and jitter control, making it difficult to meet the requirements of AGV cooperative scheduling, quay crane precise control, and other key port businesses that require millisecond-level latency and ultra-high reliability.

[0003] (2) The port business types are highly heterogeneous, and the demand for network resources varies significantly between different types of businesses. The existing scheduling mechanism is difficult to balance low-latency control flow and high-bandwidth, multi-throughput data flow in the same 5G-TSN integrated network, resulting in low resource utilization efficiency and insufficient stability of key business transmission.

[0004] To address the above problems, the 5G-TSN cooperative transmission mechanism is gradually being applied to industrial internet and smart port scenarios due to its ability to integrate the high bandwidth and low latency access capabilities of 5G with the deterministic forwarding characteristics of TSN. However, existing research mostly only optimizes for wireless or wired single domains, and there is still a lack of cross-domain joint prediction and scheduling algorithms that address the high dynamicity, strong interference, and multi-service heterogeneity requirements of ports. How to fully utilize the flexible access capabilities of 5G and the deterministic forwarding capabilities of TSN under a unified 5G-TSN logical bridge to achieve TSN scheduling to compensate for 5G air interface latency jitter and ensure end-to-end deterministic transmission of multi-source heterogeneous services has become a key technical problem that needs to be solved in current intelligent port communication systems. SUMMARY

[0005] To overcome the problems of unpredictable delay jitter, slow scheduling convergence and difficulty in comprehensive guarantee of multi-dimensional communication performance of the existing 5G-TSN cooperative transmission in the high dynamic, strong interference and multi-service concurrent environment of the port, the core idea of the method is to compensate the delay jitter of the 5G air interface by using TSN scheduling, and to realize end-to-end deterministic transmission of multi-source heterogeneous services through cross-domain intelligent prediction and cooperative optimization.

[0006] The technical means adopted by the present application are as follows: The method comprises the following steps: periodically collecting context information of a wireless channel between a 5G base station and a port mobile terminal, and collecting historical transmission feedback data; constructing a channel state evaluation model based on ridge regression, estimating the current channel state, and establishing a 5G air interface retransmission frequency model based on the channel state information, and comprehensively calculating the delay budget in the 5G domain; introducing a cooperative learning mechanism based on clustering, dynamically dividing the mobile terminals into clusters according to the context similarity, sharing the channel state information and historical transmission feedback data within the cluster based on the transfer learning theory, and updating the channel state evaluation model through joint training; dynamically adjusting the decay factor and time-sensitive weight of the parameters in the channel state evaluation model according to the deviation of the actual transmission delay and the delay budget in the 5G domain, so as to enhance the adaptive ability of the model to the high dynamic and poor communication environment of the port; through the 5G-TSN mixed traffic scheduling strategy, the deterministic transmission capability of TSN is used to compensate the delay jitter of the 5G air interface, the traffic priority is dynamically set according to the remaining time window size according to the type of service and the delay budget in the 5G domain, the queue mapping is determined by the 5G and TSN transmission conditions, and feedback iteration adjustment is performed.

[0007] Further, the context information includes signal strength, terminal speed, motion direction, system capacity and historical record, which constitutes a six-dimensional feature vector , and the transmission feedback of the communication nodes in the past communication rounds is collected synchronously to construct a wireless channel state estimation basic data set.

[0008] Further, the channel state evaluation model based on ridge regression is modeled according to the feedback of the historical context to obtain the regression parameters :

[0009] wherein, is a regularization parameter, represents a historical context matrix, indicates historical channel state feedback, is a unit matrix.

[0010] estimate the current transmission accuracy:

[0011] wherein, indicates the transmission accuracy estimate between the 5G base station and the terminal at time t.

[0012] Further, the 5G air interface retransmission number model starts retransmission when the receiving end fails to confirm, and defines the delay of each retransmission as , the maximum retransmission number of the service flow is set semi-statically, and when the upper limit is reached and correct reception is still not achieved, the data packet will be discarded:

[0013] wherein, , , respectively indicate the maximum transmission number under different preset channel conditions, is a channel state mapping function, indicates the upper threshold of the predicted channel state, indicates the lower threshold of the predicted channel state.

[0014] The calculation method of the 5G air interface transmission delay budget is expressed as:

[0015] wherein, indicates the data packet size, indicates the air interface transmission rate, indicates the time length in microslots; The 5G domain delay budget is expressed as:

[0016] wherein, is the air interface transmission independent delay.

[0017] Further, the cluster set is formed by context similarity, specifically, in each round of time , the terminal node compares the context feature quantization difference with other nodes , if the difference does not exceed the preset threshold , it is divided into the same cluster set:

[0018] wherein, denotes a cluster set partitioned according to channel context feature vectors, denotes a set of all terminal nodes that can participate in collaborative learning.

[0019] Further, the cluster set shares channel state information and historical transmission feedback data based on the transfer learning theory, which means that the terminal nodes in the same cluster set share context information and feedback information for estimating the channel state, and the ridge regression process parameters are jointly trained in the transfer learning manner. When estimating the channel state of the members in the same cluster set , the process parameters used can be expressed as:

[0020] wherein, denotes a channel state context feature vector and its corresponding return value, denotes a cluster set collaborative channel state estimation parameter vector.

[0021] Further, the attenuation factor is dynamically adjusted according to the error between the channel state estimation value and the actual feedback:

[0022] wherein, , is a set range boundary, denotes a loss function, denotes the transmission accuracy between the base station and the terminal at time .

[0023] Further, the 5G-TSN hybrid traffic scheduling strategy divides the TSN forwarding delay interval by the number of port queues:

[0024] wherein, denotes a delay interval span, denotes the maximum forwarding delay of TSN, denotes the minimum forwarding delay of TSN, is the number of port queues; The forwarding deadline of the first priority queue is expressed as:

[0025] The TSN delay of time slot t is expressed as: ​

[0026] wherein, denotes the number of TSN switches passed, denotes the number of high-priority data streams scheduled at denotes the amount of data that can be transmitted by one resource block, denotes the forwarding deadline corresponding to each queue, denotes the minimum data rate supported by the TSN system, denotes the forwarding period of the TSN system.

[0027] According to the total delay requirement of the arriving data and the 5G latency budget , the optimal priority division is determined:

[0028] wherein, denotes the end-to-end delay requirement of the arriving data, denotes the 5G latency budget.

[0029] Compared with the prior art, the present application has the following advantages: The 5G-TSN logical bridge-based port mixed service traffic cooperative transmission method provided by the present application has improved communication environment adaptability, model convergence speed, multi-terminal cooperative capability and multi-index optimization performance, and has wide practical value and popularization potential, and is particularly suitable for complex, dynamic and heterogeneous port communication network environments.

[0030] The present application effectively solves the problems of large 5G air interface delay jitter and difficulty in meeting the needs of multiple heterogeneous services in the port environment, and significantly improves the communication reliability and real-time performance of automated equipment under complex working conditions. DETAILED DESCRIPTION

[0031] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or prior art description will be briefly introduced. Obviously, the drawings in the following description are some embodiments of the present application, and those skilled in the art can obtain other drawings according to these drawings without creative labor.

[0032] Figure 1 The structure block diagram of the 5G-TSN logical bridge-based port mixed service traffic cooperative transmission method in the present application.

[0033] Figure 2 The port 5G-TSN logical bridge architecture in the present application.

[0034] ​Figure 3 is a process diagram based on cluster set cooperative learning mechanism in the present application.

[0035] Figure 4 is a diagram of time sensitive coefficient in the present application.

[0036] Figure 5 is a diagram of service traffic priority dynamic setting strategy.

[0037] Figure 6 is a simulation image of high throughput data accumulation delay in the embodiment.

[0038] Figure 7 is a simulation image of high throughput data accumulation network throughput in the embodiment.

[0039] Figure 8 is a simulation image of delay sensitive data accumulation delay in the embodiment.

[0040] Figure 9 is a simulation image of delay sensitive data accumulation throughput in the embodiment. DETAILED DESCRIPTION

[0041] It should be noted that the embodiments and features of the embodiments in the present application can be combined with each other without conflict. The present application will be described in detail below with reference to the accompanying drawings and in conjunction with the embodiments.

[0042] In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions of the embodiments of the present application will be described clearly and completely below in conjunction with the drawings of the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. The description of the at least one exemplary embodiment is actually only illustrative, but not as any limitation on the present application and its application or use. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0043] It should be noted that the terms used herein are only for describing specific embodiments, and are not intended to limit the exemplary embodiments according to the present application. As used herein, the singular form is intended to include the plural form unless the context clearly indicates otherwise, and it should also be understood that when the terms "comprise" and / or "include" are used in the specification, there is a feature, step, operation, device, component and / or combination thereof.

[0044] The relative arrangement of parts and steps, numerical expressions, and numerical values set forth in these embodiments do not limit the scope of the present application unless otherwise specified. Meanwhile, it should be clear that the sizes of the various parts shown in the drawings are not drawn in actual proportion for the convenience of description. The techniques, methods, and devices known to those of ordinary skill in the relevant art can not be discussed in detail, but should be considered as part of the specification under appropriate circumstances. In all examples shown and discussed here, any specific value should be interpreted as merely exemplary, not as a limitation. Therefore, other examples of exemplary embodiments can have different values. It should be noted that similar reference numbers and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further discussed in subsequent drawings.

[0045] As shown in Figure 1 , the present application provides a port mixed service traffic cooperative transmission method based on 5G-TSN logical bridge, comprising: periodically collecting context information of wireless channel between 5G base station and port mobile terminal, and collecting historical transmission feedback data; in specific implementation, as a preferred embodiment of the present application, the context information includes signal strength, terminal speed, motion direction, system capacity and historical record, constituting a six-dimensional feature vector , and synchronously collecting transmission feedback of communication nodes in past communication rounds to build wireless channel state estimation basic data set.

[0046] The 5G base station will perceive each mobile terminal within the communication range in each time slot , at the same time, the 5G base station records historical interaction feedback as the environmental reward input of strategy optimization.

[0047] A channel state evaluation model based on ridge regression is constructed to estimate the current channel state, and a 5G air interface retransmission number model is established based on channel state information, and the time delay budget in 5G domain is calculated comprehensively; in specific implementation, as a preferred embodiment of the present application, the channel state evaluation model based on ridge regression is modeled to obtain regression parameters according to the feedback of historical context :

[0048] Among them, is the regularization parameter, represents the historical context matrix, represents the historical channel state feedback, is the unit matrix.

[0049] Estimate the current transmission accuracy:

[0050] wherein, represents the transmission accuracy estimation value between the 5G base station and the terminal at the moment.

[0051] To ensure transmission reliability, the system adopts a hybrid automatic repeat request mechanism. In a specific implementation, as a preferred embodiment of the present application, the 5G air interface retransmission number model starts retransmission when the receiving end fails to confirm, and defines the delay of each retransmission as For ultra-reliable low-latency communication, 5G usually uses SPS (Semi-Persistent Scheduling) mode, so that the first transmission and subsequent retransmission use the same resource block and MCS level, to reduce control overhead and meet the low-latency requirement. According to the 3GPP protocol, the maximum number of retransmissions of a service flow is set semi-statically, and when the upper limit is reached and the data packet cannot be correctly received, the data packet will be discarded:

[0052] wherein, , , respectively represent the maximum number of transmissions under different preset channel conditions, is a channel state mapping function, represents the upper threshold of the predicted channel state, represents the lower threshold of the predicted channel state.

[0053] The number of maximum retransmissions is determined according to the wireless resource channel quality information. When is lower than the minimum threshold, the wireless channel state is poor at this time, so is increased to increase the 5G system transmission delay budget; when the predicted channel quality is good, the probability of retransmission is low, and the value of is reduced. The calculation method of the 5G air interface transmission delay budget is represented as:

[0054] wherein, represents the data packet size, represents the air interface transmission rate, represents the time length in microslots; The 5G domain delay budget is represented as:

[0055] wherein, is the delay unrelated to air interface transmission.

[0056] To efficiently capture the changes of channel state in dynamic environments, this paper introduces cluster sets for channel state collaborative learning in port communication networks, as shown in Figure 3 By dividing cluster sets, 5G base stations accelerate the learning process by sharing the same terminal node channel state data within a cluster set, improving the convergence speed of each node within the cluster set. The introduction of a cluster-based collaborative learning mechanism dynamically divides mobile terminals into cluster sets based on context similarity. Within the cluster set, channel state information and historical transmission feedback data are shared based on the transfer learning theory to jointly train and update the channel state evaluation model. In specific implementations, as a preferred embodiment of the present invention, the cluster set is formed by context similarity. Specifically, at each round of time , the terminal node compares the context feature quantization difference with other nodes . If the difference does not exceed the preset threshold , it is divided into the same cluster set:

[0057] wherein, represents the cluster set divided according to the channel context feature vector, represents the set of all terminal nodes that can participate in collaborative learning.

[0058] In specific implementations, as a preferred embodiment of the present invention, the cluster set shares channel state information and historical transmission feedback data based on the transfer learning theory. This means that the estimation of channel state for terminal nodes within the same cluster set shares context information and feedback information. The process parameters used in the ridge regression process are jointly trained using transfer learning. The process parameters used in the channel state estimation of members within the same cluster set and can be represented as:

[0059] wherein, represents the channel state context feature vector and its corresponding feedback value, represents the collaborative channel state estimation parameter vector of the cluster set .

[0060] As shown in Figure 3 , the collaborative learning mechanism significantly improves modeling accuracy and strategy stability, reducing the exploration cost in the learning process and is suitable for deployment in multi-terminal collaborative scenarios in port environments.

[0061] ​To enhance the adaptability of the system to the dynamic communication environment, the application constructs a parameter updating mechanism to improve the timeliness of data. The mechanism dynamically adjusts the attenuation factor according to the deviation of the 5G air interface performance estimation to reduce the error caused by outdated data. According to the deviation of the actual transmission delay and the delay budget in the 5G domain, the attenuation factor and the time-sensitive weight of the parameters in the channel state evaluation model are dynamically adjusted, as shown in Figure 4

[0062] An online parameter updating mechanism with a dynamic attenuation factor is introduced in the 5G air interface transmission performance estimation. The mechanism dynamically adjusts the attenuation factor according to the difference between the actual value and the estimated value, so as to more accurately perceive the changes of the communication conditions. When the channel state fluctuates, the 5G air interface transmission performance changes accordingly. In order to capture the influence of the changes of the communication conditions over time on the transmission performance, the following loss function is defined to measure the deviation between the actual transmission accuracy and the estimated value:

[0063] The loss function reflects the effectiveness of the process parameters in the current communication environment. When the loss is large, it means that the process parameters have become invalid and the model parameters need to be updated in time. At this time, the attenuation factor should be reduced to reduce the weight of outdated data, so that the model can better respond to the current communication environment: In specific implementation, as a preferred embodiment of the application, the attenuation factor is dynamically adjusted according to the error between the channel state estimation value and the actual feedback:

[0064] Wherein, , is the set range boundary, represents the loss function, represents the transmission accuracy between the base station and the terminal at time .

[0065] The regression matrix and vector are updated using the attenuation factor:

[0066]

[0067] Through the above steps, the model can continuously adapt to dynamic changes and improve the accuracy of transmission rate estimation.

[0068] ​By means of the 5G-TSN mixed traffic scheduling strategy, the deterministic transmission capability of TSN is utilized to compensate for the delay jitter of the 5G air interface, the traffic priority is dynamically set according to the remaining time window size, the queue mapping is determined by the 5G and TSN transmission conditions, and feedback iteration adjustment is performed.

[0069] wherein, denotes the span of the delay interval, denotes the maximum forwarding delay of TSN, denotes the minimum forwarding delay of TSN, is the number of port queues; the forwarding deadline of the first priority queue is represented as:

[0070] the TSN delay of the time slot t is represented as:

[0071] wherein, denotes the number of TSN switch hops passed, denotes the number of high-priority data streams scheduled at time, denotes the amount of data that can be transmitted by one resource block, denotes the forwarding deadline corresponding to each queue, denotes the minimum data rate supported by the TSN system, denotes the forwarding period of the TSN system.

[0072] According to the total delay requirement of the arriving data and the 5G delay budget , the optimal priority division is determined:

[0073] wherein, denotes the end-to-end delay requirement of the arriving data, denotes the 5G delay budget.

[0074] As shown in Figure 5 , within the TSN domain, the traffic flow priority is dynamically set according to the remaining time window size and the actual transmission delay deviation of 5G, the data flow with high delay deviation is injected into the high-priority queue, so as to compensate for the delay jitter of the 5G air interface by means of TSN scheduling, and realize end-to-end deterministic transmission.​

[0075] Embodiment This embodiment adopts a port heterogeneous communication network simulation environment based on the real topological structure of Qinhuangdao Port. The simulation takes autonomous surface ships as the test object, and considers the influence of its mobility on the dynamic change of the communication topology, thereby exacerbating the dynamic change of the channel state. The communication data is generated by MATLAB R2023b, covering the following influencing factors: path loss, environmental attenuation (salt spray, dust, typhoon), multipath fading, shadow effect, etc. Each base station adopts an M / M / 1 queuing model, and sets the queuing delay under different terminal densities to simulate the influence of network congestion on communication performance.

[0076] Firstly, this embodiment compares the performance of each algorithm under high throughput demand data. Figure 6 The simulation image of the cumulative delay of high throughput data in the embodiment is shown; Figure 7 The simulation image of the cumulative network throughput of high throughput data. This invention combines the cluster collaborative learning mechanism and the dynamic parameter updating rule, and performs best in all indicators. The cluster collaborative learning mechanism based on it enables the algorithm to quickly adapt to changes in communication conditions, while the dynamic parameter updating rule improves the accuracy of perception, thereby achieving higher cumulative throughput and lower cumulative delay in a dynamic environment, and the convergence speed is significantly faster than other methods. In contrast, DB-LinUCB only uses dynamic parameter updating but lacks a cluster collaborative learning mechanism, CB-LinUCB uses a cluster collaborative learning mechanism but lacks dynamic updating, and LinUCB lacks both, so it performs poorly in a dynamic topology. UCB and LinUCB rely only on the basic exploration-exploitation mechanism, and respond slowly to changes in channel quality, resulting in significantly poorer performance in dynamic communication conditions.

[0077] Then, this embodiment compares the invention with DB-UCB, CB-UCB, UCB, and a random selection strategy in terms of processing delay-sensitive data, with the main performance indicators being system cumulative throughput and cumulative delay. Figure 8 It is shown that, thanks to the combination of the cluster collaborative learning mechanism and the dynamic parameter updating rule, the invention can quickly and accurately adapt to changes in channel state, thereby achieving the lowest cumulative delay. In contrast, DB-UCB lacks a cluster collaborative learning mechanism, and UCB lacks both dynamic updating and a cluster collaborative learning mechanism, and converges slowly when evaluating channel state. Although CB-UCB uses a cluster collaborative learning mechanism, it performs better than UCB in terms of performance, but due to the lack of dynamic updating, its performance is still inferior to the invention. Figure 9It can be seen that the application also performs best in maximizing system throughput, can effectively adapt to channel state changes, and fully utilizes communication resources. Although DB-UCB improves performance to some extent through dynamic parameter updating, due to the lack of cluster collaborative learning mechanism, the overall effect is still not as good as the application. The application realizes the lowest cumulative delay by quickly learning the dynamic changes of the channel state, while UCB and CB-UCB have higher delay due to the slower response speed to the real-time demand of the network. Although DB-UCB is better than UCB in terms of delay, it still lags behind the application due to the slower adaptation speed to the dynamic communication environment.

[0078] In the above embodiments, the application provides a port mixed traffic collaborative transmission method based on a 5G-TSN logical bridge, which is used to cope with the cross-domain collaborative transmission challenges brought by poor wireless environment and concurrent multi-source heterogeneous services in an intelligent port. In the port communication network, environmental factors such as salt spray, dust, and multipath reflection, as well as the high-speed movement of automated equipment, make the 5G air interface channel state information dynamically change and difficult to predict, and cause the random fluctuations of the delay and jitter of service data; at the same time, the different types of data such as control instructions, high-definition video streams, and sensor monitoring have significant differences in the demand for low latency, high bandwidth, and high reliability of the network, further increasing the difficulty of cross-domain joint scheduling. The existing 5G-TSN collaborative transmission scheme often lacks context awareness and comprehensive optimization of multi-dimensional performance when facing the above complex and rapidly changing scenarios, resulting in insufficient end-to-end deterministic guarantee, affecting the continuity and safety of automated port production. The application can predict the number of retransmissions triggered by 5G air interface errors by periodically collecting the context information and historical transmission feedback of 5G base stations and terminals, and constructing a channel state evaluation model based on ridge regression, dynamically calculating the complete 5G intra-domain latency budget and real-time updating of key parameters, and realizing accurate estimation and rapid compensation of delay and jitter. Further, the introduction of a cluster-based collaborative learning mechanism enables terminal nodes with similar contexts to form clusters to share channel experience data, and uses transfer learning to improve the convergence speed and strategy generalization ability of channel state estimation, thereby enhancing the adaptability of the entire network in a high dynamic environment. On this basis, the application dynamically sets the TSN traffic priority according to the 5G real-time transmission delay and budget deviation, and accurately gates and queues key business streams in the remaining time window, achieving the core goal of compensating for 5G wireless communication delay jitter using TSN scheduling. Through cross-domain joint scheduling and priority adaptive control, the application can simultaneously meet the millisecond-level low-latency demand of port control instructions, the high-bandwidth demand of high-definition video, and the high-throughput demand of sensor data. Simulation and test results show that, in the port scenario of high dynamic topology and complex concurrent services, compared with the traditional independent scheduling or single-domain optimization scheme, the application significantly reduces the end-to-end average delay and jitter, improves the arrival rate and communication reliability of key services, and realizes the deterministic transmission and resource efficient utilization of multi-source heterogeneous services in the 5G-TSN fusion network.

[0079] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the application, and not to limit them; although the application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or make equivalent substitutions for part or all of the technical features; and these modifications or substitutions do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the application.

Claims

1. A port mixed service traffic collaborative transmission method based on a 5G-TSN logical bridge, characterized in that, Comprise: Periodically collect the context information of the wireless channel between the 5G base station and the port mobile terminal, and collect the historical transmission feedback data; Construct a channel state evaluation model based on ridge regression, estimate the current channel state, and establish a 5G air interface retransmission number model based on the channel state information, and comprehensively calculate the 5G domain delay budget; Introduce a cluster-based collaborative learning mechanism, dynamically divide the mobile terminal into clusters according to the context similarity, share channel state information and historical transmission feedback data within the cluster based on the transfer learning theory, and jointly train and update the channel state evaluation model; According to the deviation of the actual transmission delay and the 5G domain delay budget, dynamically adjust the decay factor and time-sensitive weight of the parameters in the channel state evaluation model to enhance the adaptive ability of the model to the high dynamic and bad communication environment of the port; Through the 5G-TSN hybrid traffic scheduling strategy, use the deterministic transmission capability of TSN to compensate for the delay jitter of 5G air interface, dynamically set the traffic priority according to the remaining time window size according to the business type and the 5G domain delay budget, and the queue mapping is determined by the 5G and TSN transmission, and the feedback is iteratively adjusted.

2. The port mixed traffic collaborative transmission method based on the 5G-TSN logical bridge according to claim 1, characterized in that, The context information includes signal strength, terminal speed, motion direction, system capacity and historical record, constituting a six-dimensional feature vector And synchronously collect transmission feedback of communication nodes in past communication rounds To build a wireless channel state estimation basic data set.

3. The port mixed traffic collaborative transmission method based on the 5G-TSN logical bridge of claim 1, characterized in that, The channel state evaluation model based on ridge regression, according to the feedback of historical context Modeling to obtain regression parameters : wherein is a regularization parameter, denotes a historical context matrix, denotes a historical channel state feedback, is an identity matrix; Estimate the current transmission accuracy: wherein indicates The transmission accuracy estimation value between the 5G base station and the terminal at time point 5.

4. The port mixed traffic collaborative transmission method based on the 5G-TSN logical bridge of claim 1, characterized in that, The 5G air interface retransmission number model starts retransmission when the receiving end fails to confirm, and defines the delay of each retransmission as The maximum retransmission number of the service flow Semi-static setting is adopted, and when the upper limit is reached and correct reception is still not achieved, the data packet will be discarded: wherein, , , respectively represent the maximum number of transmissions under different preset channel conditions, is a channel state mapping function, represents an upper threshold of the predicted channel state, represents a lower threshold of the predicted channel state; A method for calculating a 5G air interface transmission latency budget is expressed as: wherein, represents a data packet size, represents an air interface transmission rate, represents a time length of a micro time slot; The 5G domain delay budget is expressed as: wherein, is a delay independent of air interface transmission.

5. The port mixed traffic collaborative transmission method based on 5G-TSN logical bridge according to claim 1, characterized in that, The cluster set is formed by context similarity, specifically, in each round of time , terminal nodes Compare the context feature with other nodes If the difference does not exceed the preset threshold , it is divided into the same cluster set: wherein, denotes a cluster set partitioned according to the channel context feature vector, denotes a set of all terminal nodes that can participate in the collaborative learning.

6. The port mixed traffic collaborative transmission method based on the 5G-TSN logical bridge of claim 1, characterized in that, The cluster set shares channel state information and historical transmission feedback data based on the migration learning theory, which means that the estimation of the channel state of the terminal nodes in the same cluster set shares context information and feedback information, and the ridge regression process parameters are jointly trained in the migration learning manner. When the channel state estimation is performed on the members in the same cluster set , the process parameters used may be represented as:​ wherein, denotes a channel state context feature vector and its corresponding return value, denotes a cluster set of cooperative channel state estimation parameter vectors.

7. The port mixed traffic collaborative transmission method based on the 5G-TSN logical bridge of claim 1, characterized in that, The decay factor is dynamically adjusted according to the error between the channel state estimation value and the actual feedback: wherein, , to set a range boundary, indicates that a loss function is determined by a deviation of an actual value from an estimated value, indicates a time point a transmission accuracy between a base station and a terminal .

8. The port mixed traffic collaborative transmission method based on the 5G-TSN logical bridge of claim 1, characterized in that, The 5G-TSN hybrid traffic scheduling strategy divides the TSN forwarding delay interval by equal parts according to the number of port queues: wherein, denotes a latency interval span, denotes a maximum forwarding delay for TSN, denotes a minimum forwarding delay for TSN, is the number of port queues; The first The forwarding deadline of the priority queue is represented as: TSN latency of a time slot t is represented as: wherein, denotes the number of TSN switches passed, denotes the total number of high-priority sensors scheduled at denotes the total number of high-priority sensors scheduled at denotes the amount of data that can be transmitted by one resource block, denotes the forwarding deadline corresponding to each queue, denotes the minimum data rate supported by the TSN system, denotes the forwarding period of the TSN system; According to the total latency requirement of the arrival data and the 5G latency budget , determine the optimal priority division: wherein, denotes an end-to-end latency requirement to reach the data, denotes a 5G latency budget.

Images