A Cooperative Optimization Method for DDS-5G Networks Based on Fuzzy Logic

Abstract

This invention belongs to the field of industrial network performance optimization technology, specifically providing a DDS-5G network collaborative optimization method based on fuzzy logic. By setting static mapping rules, the QoS Policy parameters of the DDS are mapped to the QoS flow identifiers of the QoS flows in the 5G network, thereby achieving initial DDS-5G collaboration. Simultaneously, by establishing a fuzzy logic inference system, using the QoS Policy parameters and real-time network state parameters as input variables and 5QI as the output variable, and by optimizing and adjusting the QoS Policy parameters of the DDS through DQN and updating the fuzzy logic parameters of the fuzzy logic inference system, further DDS-5G network collaborative optimization is achieved. Through a cross-layer adaptation mechanism, efficient collaboration between the application layer and the 5G network is realized, making it suitable for scenarios with high requirements for low latency and high reliability, such as industrial automation and intelligent manufacturing.

Description

Technical Field

[0001] This invention belongs to the field of network performance optimization technology, specifically relating to a DDS-5G network collaborative optimization method based on fuzzy logic. Background Technology

[0002] In the communication control technology of networked intelligent equipment, ROS2 (Robot Operating System 2) achieves higher communication service quality and network environment adaptability through its distributed communication architecture of Data Distribution Service (DDS). DDS provides various Quality of Service (QoS) policy settings to control data transmission behavior and requirements. Developers can flexibly configure different QoS policies to ensure communication service quality according to the communication needs of different application scenarios. Meanwhile, in 5G networks, the Service Data Adaptation Protocol (SDAP) adds a QoS Flow ID (QFI) to the traversing data packets, responsible for mapping between QoS Flow and Data Radio Bearer (DRB), thereby achieving QoS for different service flows.

[0003] In existing intelligent equipment communication technologies, because the DDS quality of service strategy and the 5G network service data adaptation protocol are located at different network layers, they can independently implement quality of service control and cannot achieve effective synergy, resulting in low efficiency in optimizing network performance. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to overcome the defect in the prior art that the DDS and 5G networks of networked intelligent equipment are in different network layers and therefore cannot achieve collaborative optimization, so as to provide a DDS-5G network collaborative optimization method based on fuzzy logic.

[0005] This invention provides a DDS-5G network collaborative optimization method based on fuzzy logic, comprising the following steps:

[0006] Step S1: Set static mapping rules to map the DDS QoS Policy parameters to the QoS flow identifiers of the 5G network QoS flows;

[0007] Step S2: Monitor the real-time network status, communication system resource status, and control system resource status; when the network status changes exceed the threshold, or the performance indicators of the communication system or control system fail to meet the standards, proceed to DQN optimization;

[0008] Step S3: Establish a fuzzy logic reasoning system, using QoS Policy parameters and real-time network status parameters as input variables and 5QI as the output variable;

[0009] Step S4: In DQN optimization, based on the real-time collected network status and performance indicators, DQN selects the optimal composite action based on the ε-greedy policy; executes the optimal composite action, adjusts the QoS Policy parameters of DDS and updates the fuzzy logic parameters of the fuzzy logic inference system, calculates the instant reward value, and updates the Q network based on the instant reward value.

[0010] Furthermore, the static mapping rule includes the following method steps:

[0011] Step S1.1: Determine the packet delay budget based on the LIFESPAN and DEADLINE of the QoS Policy;

[0012] Step S1.2: Determine the packet error rate based on the RELIABILIT of the QoS Policy;

[0013] Step S1.3: Set the resource type according to the DEADLINE setting, HISTORY setting, and message publishing rate change rate of the QoS Policy;

[0014] Step S1.4: Determine the QoS flow identifier based on resource type, packet delay budget, packet error rate, and resource type.

[0015] Furthermore, when monitoring the real-time network status, the network status includes packet loss rate, latency, and congestion level.

[0016] Furthermore, a fuzzy logic reasoning system is established, including the following steps:

[0017] Step S3.1: Convert the QoS Policy parameters and real-time network status parameters, which are used as input, into semantic descriptions through fuzzification.

[0018] Step S3.2: Perform fuzzy inference, complete fuzzy inference based on the fuzzy rule base, and calculate the activation degree of each rule. The minimum value of the rule activation degree is the confidence degree.

[0019] Step S3.3: Perform defuzzification and introduce forced constraint logic, calculate the weighted average of all activation rules, and output the 5QI value after constraint.

[0020] Furthermore, in the fuzzy logic reasoning system, the fuzzy rule base includes 54 rules, and the inputs to the rules include one of two types of reliability, one of three types of delay budget, one of three types of bandwidth utilization, and one of three types of packet loss rate.

[0021] Furthermore, in the DQN optimization, the state space of the DQN network includes network performance and control performance. The control performance includes packet loss performance, end-to-end latency, network jitter, and bandwidth utilization. The control performance also includes pose error and point cloud stitching residual.

[0022] Furthermore, in the DQN optimization, the action space of the DQN network includes QoS Policy adjustment and fuzzy logic parameter optimization; the QoS Policy adjustment includes RELIABILIT policy adjustment, HISTORY DEPTH size adjustment, DURABILITY policy adjustment, and DEADLINE size adjustment; the fuzzy logic parameter optimization includes fuzzy rule weight correction and fuzzy logic membership function parameter update.

[0023] Furthermore, in the DQN optimization, the reward function of the DQN network is expressed as:

[0024] ;

[0025] in, Indicates a delay reward. Indicates a reliability bonus. This represents the confidence reward. Indicates the maximum allowable delay. Indicates packet loss rate, This represents the maximum allowable packet loss rate, and fuzzy_confidence represents the mapping confidence level.

[0026] Beneficial Effects: This invention provides a DDS-5G network collaborative optimization method based on fuzzy logic. By setting static mapping rules, the QoS Policy parameters of DDS are mapped to the QoS flow identifiers of QoS flows in the 5G network, thereby achieving initial DDS-5G collaboration. Simultaneously, by establishing a fuzzy logic inference system, QoS Policy parameters and real-time network state parameters are used as input variables, and 5QI is used as the output variable. The DDS QoS Policy parameters are optimized and adjusted through DQN, and the fuzzy logic parameters of the fuzzy logic inference system are updated, further realizing DDS-5G network collaborative optimization. Through a cross-layer adaptation mechanism, efficient collaboration between the application layer and the 5G network is achieved, which is suitable for scenarios with high requirements for low latency and high reliability, such as industrial automation and intelligent manufacturing. Attached Figure Description

[0027] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0028] Figure 1 This is a schematic flowchart illustrating the overall method steps of the present invention;

[0029] Figure 2 This is a schematic diagram of the static mapping rule process of the present invention;

[0030] Figure 3 This is a schematic diagram of the fuzzy rule process of the present invention. Detailed Implementation

[0031] To make the above-mentioned objectives, features, and advantages of this application more apparent and understandable, the specific embodiments of this application are described in detail below with reference to the accompanying drawings. Many specific details are set forth in the following description to provide a thorough understanding of this application. However, this application can be implemented in many other ways different from those described herein, and those skilled in the art can make similar modifications without departing from the spirit of this application. Therefore, this application is not limited to the specific embodiments disclosed below.

[0032] In aircraft manufacturing, aircraft skin assembly is a critical task directly related to the assembly accuracy and aerodynamic performance of the entire aircraft. Traditional aircraft skin measurement methods typically require manual intervention, which is time-consuming, labor-intensive, and costly. To improve measurement efficiency and accuracy, automated measurement devices such as laser scanners and collaborative robotic arms have been gradually introduced. These devices need to coordinate in real time during the measurement process, exchanging image data, control commands, and status information at high frequency through communication networks, which places higher demands on the bandwidth, latency, and reliability of the communication system. Therefore, adopting automation technology to measure aircraft skin is of great significance. In the context of automated aircraft skin measurement, the high reliability and real-time performance of the communication system have become key supporting technologies for ensuring measurement accuracy and efficiency. This embodiment provides a DDS-5G network collaborative optimization method based on fuzzy logic for data communication of automated measurement devices under aircraft skin, supporting the stable and efficient operation of networked robots in the aircraft skin measurement process.

[0033] The network collaborative optimization method provided in this embodiment uses a fuzzy logic algorithm to solve the adaptation problem between DDS QoS policies and QFI in 5G. First, the DDS QoS policy is mapped to the corresponding 5G QFI according to predefined mapping rules. During operation, key service quality indicators such as latency and packet loss rate are collected in real time through a network condition monitoring module. A finite state machine is used to dynamically adjust QoS, divided into QoS Policy adaptive adjustment state and QFI adjustment state. This method achieves efficient collaboration between the application layer and the 5G network through a cross-layer adaptation mechanism without modifying the 5G core network, and is suitable for scenarios with high requirements for low latency and high reliability, such as industrial automation and intelligent manufacturing.

[0034] Reference Figures 1-3 As shown, this embodiment provides a DDS-5G network collaborative optimization method based on fuzzy logic, including the following steps:

[0035] Step S1: Set static mapping rules to map the DDS QoS Policy parameters to the QoS flow identifiers of the 5G network QoS flows;

[0036] Step S2: Monitor the real-time network status, communication system resource status, and control system resource status; when the network status changes exceed the threshold, or the performance indicators of the communication system or control system fail to meet the standards, proceed to DQN optimization;

[0037] Step S3: Establish a fuzzy logic reasoning system, using QoS Policy parameters and real-time network status parameters as input variables and 5QI as the output variable;

[0038] Step S4: In DQN optimization, based on the real-time collected network status and performance indicators, DQN selects the optimal composite action based on the ε-greedy policy; executes the optimal composite action, adjusts the QoS Policy parameters of DDS and updates the fuzzy logic parameters of the fuzzy logic inference system, calculates the instant reward value, and updates the Q network based on the instant reward value.

[0039] Specifically, the static mapping rules in step S1 include the following method steps:

[0040] Step S1.1: Determine the packet delay budget based on the LIFESPAN and DEADLINE of the QoS Policy; LIFESPAN represents the maximum time data can survive in the ROS2 system, and DEADLINE represents the maximum time interval required for data updates. In this embodiment, the packet delay budget is selected as the minimum value of LIFESPAN and DEADLINE to ensure data availability and validity. This is expressed as:

[0041] ;

[0042] Step S1.2: Determine the packet error rate based on the RELIABILIT of the QoS Policy; when the RELIABILIT of the QoS Policy is reliable, its communication behavior will ensure that every data packet is received, and the packet error rate is set to [value missing]. When the RELIABILIT of the QoS Policy is set to best_effort, its communication behavior is best-effort and does not guarantee that every packet will be received. In this case, the packet error rate is set to... 。 is represented as:

[0043] ;

[0044] The process also includes the following steps: determining bandwidth requirements based on the number of historical message cache entries, message size, and maximum message update cycle according to the QoS Policy; the bandwidth requirement is calculated by dividing the product of the number of historical message cache entries and the message size in bits by the maximum message update cycle. This is represented as:

[0045] ;

[0046] The above method abstracts specific QoS policy settings from the application layer into specific service quality indicators.

[0047] Step S1.3: Based on the QoS Policy's DEADLINE setting, HISTORY setting, and message delivery rate change rate, set the resource type to match the data stream to the appropriate resource type. Resource types in 5G systems can be divided into three categories: Non-GBR (Non-Guaranteed Bit Rate), DC-GBR (Due to Latency Critical Guaranteed Bit Rate), and GBR (Guaranteed Bit Rate). Check if there is a DEADLINE setting and if the message delivery rate change rate exceeds 10%, which can be roughly divided into non-guaranteed bit rate and guaranteed bit rate; further subdivided into delay-critical guaranteed bit rate and guaranteed bit rate based on the HISTORY setting.

[0048] The rate of change in message publishing speed is expressed as:

[0049] ;

[0050] in This indicates the message publishing speed within the current time window. This represents the message publishing speed of the previous time window; the rate of change is calculated after calculating the difference between the two.

[0051] The method for determining resource type is as follows:

[0052] ;

[0053] Step S1.4: Determine the QoS flow identifier based on resource type, packet delay budget, packet error rate, and resource type. For unguaranteed bit rate types, find the specific 5QI value based on the packet delay budget. For delay-critical guaranteed bit rate types, find the specific 5QI value based on their bandwidth requirements. For guaranteed bit rate types, find the specific 5QI value based on the packet delay budget and packet error rate.

[0054] The specific process is as follows: Based on the calculated bandwidth requirements and the Maximum Data Burst Volume (MDBV) and Data Rate Averaging Window (DRAW) features in 5QI, the following parameters are calculated: Compare them. When BW < Under the premise of selecting The minimum 5QI value ensures that the mapping result satisfies the constraints and has optimal fit. BW and All units are in bps. When static mapping cannot confirm the specific mapping item of QoS Policy, the system will issue a warning and handle it using the default 5QI=9. The UE will automatically complete the mapping to QFI according to the value of 5QI through QoS Rules.

[0055] The calculation method is expressed as follows:

[0056] .

[0057] In step S2 of this embodiment, when the network state change rate exceeds the threshold of 20%, the network state change rate is calculated as: the difference between the current network state parameter and the network state parameter of the previous time node divided by the network state parameter of the previous time node, or when the performance indicators of the communication system or control system are not up to standard, i.e. the current end-to-end latency is greater than the initial data packet latency budget / packet loss rate is greater than the initial packet loss budget / network jitter is severe / bandwidth utilization is too low / pose error and point cloud stitching error are too large, DQN optimization is entered;

[0058] As a preferred embodiment, when monitoring the real-time network status, the network status includes packet loss rate, latency, and congestion level.

[0059] In step S3 of this embodiment, the fuzzy logic reasoning system is established, including the following steps:

[0060] Step S3.1: The QoS Policy parameters and real-time network status parameters, which are used as input, are converted into semantic descriptions through fuzzification. Using both the QoS Policy and real-time network status parameters as parameters solves the adaptation problem under different network conditions. The fuzzification of the input parameters is specifically shown in Table 1.

[0061] Table 1. Input Parameter Fuzzification Table

[0062] ;

[0063] Step S3.2: Perform fuzzy inference, complete fuzzy inference based on the fuzzy rule base, and calculate the activation degree of each rule. The minimum value of the rule activation degree is the confidence degree. In the fuzzy logic inference system, the fuzzy rule base includes 54 rules (2 reliability * 3 delay budget * 3 bandwidth utilization * 3 packet loss rate). The input of the rule includes one of the two types of reliability, one of the three types of delay budget, one of the three types of bandwidth utilization, and one of the three types of packet loss rate.

[0064] In this embodiment, two rules are given as examples:

[0065] Table 2: Example of Rules

[0066] ;

[0067] Step S3.3: Defuzzify and introduce mandatory constraint logic, calculate the weighted average of all activation rules, and output the 5QI value after constraint. The UE will map the 5QI to a specific QFI according to the QoS Rules built into the 5G system, thereby affecting the scheduling of QoS flow and realizing a complete mapping link from DDS QoS Policy -> 5QI -> QFI, thus ensuring the scheduling priority and transmission performance of data flow on the core network and air interface sides.

[0068] In step S4 of this embodiment, the DQN algorithm modifies the rule weights in the fuzzy rules based on network feedback and learns and adjusts the QoS Policy to reduce unnecessary network overhead and improve network adaptability. The DQN algorithm combines deep learning and reinforcement learning, using a neural network to approximate the Q-value function. ,in Indicates an action, that is, a state. Under these conditions, the agent can choose any possible action. The DQN agent continuously interacts with the environment to obtain the current environmental state. In the current Next, according to the agent's policy Select Action After this action is performed, the environment will transition from the current state to the next state with a certain probability. Simultaneously, the agent evaluates the quality of its actions based on rewards received from environmental feedback. A memory bank stores historical interaction data. ,in This represents the immediate reward obtained at time step t. When there is enough data in the memory bank, data that disrupts correlation is randomly extracted for training, thus updating the algorithm.

[0069] The method by which an agent selects actions is represented as follows:

[0070] ;

[0071] The state space is represented as:

[0072] ;

[0073] ;

[0074] Where S represents the state space. Indicates packet loss performance, Indicates end-to-end delay. Indicates network jitter. Indicates bandwidth utilization. Indicates the actual bandwidth used. Indicates the total bandwidth. Indicates pose error. The point cloud stitching residual is represented in the DQN optimization. The state space of the DQN network includes network performance and control performance. The control performance includes packet loss performance, end-to-end latency, network jitter and bandwidth utilization. The control performance includes pose error and point cloud stitching residual.

[0075] Pose error is the error between the robot's actual pose and the expected target pose. It is sensitive to communication network latency and can reflect the robot's real-time control performance. Point cloud stitching residual is the geometric consistency error of point cloud data collected by multiple sensors after stitching and fusion. It is sensitive to communication packet loss.

[0076] In this embodiment, the action space of the DQN network in the DQN optimization includes QoS Policy adjustment and fuzzy logic parameter optimization; the QoS Policy adjustment includes RELIABILIT policy adjustment, HISTORY DEPTH size adjustment, DURABILITY policy adjustment, and DEADLINE size adjustment; the fuzzy logic parameter optimization includes fuzzy rule weight correction and fuzzy logic membership function parameter update.

[0077] The action space is represented as:

[0078] ;

[0079] In the DQN optimization described above, the reward function of the DQN network is expressed as:

[0080] ;

[0081] in, Indicates a delay reward. Indicates a reliability bonus. This represents the confidence reward. This indicates the maximum allowable latency; the closer the latency is to the maximum allowable latency, the better. The delay reward will decay exponentially; Indicates packet loss rate, This represents the maximum allowable packet loss rate. When the packet loss rate exceeds the maximum allowable packet loss rate, the reliability reward will be converted into a penalty. fuzzy_confidence represents the mapping confidence. The higher the matching degree, the higher the reward.

[0082] This embodiment provides a DDS-5G network collaborative optimization method based on fuzzy logic. By setting static mapping rules, the QoS Policy parameters of DDS are mapped to the QoS flow identifiers of 5G network QoS flows, thereby achieving initial DDS-5G collaboration. Simultaneously, by establishing a fuzzy logic inference system, QoS Policy parameters and real-time network status parameters are used as input variables, and 5QI is used as the output variable. The QoS Policy parameters of DDS are optimized and adjusted through DQN, and the fuzzy logic parameters of the fuzzy logic inference system are updated, further realizing DDS-5G network collaborative optimization. Through a cross-layer adaptation mechanism, efficient collaboration between the application layer and the 5G network is achieved, which is suitable for scenarios with high requirements for low latency and high reliability, such as industrial automation and intelligent manufacturing.

[0083] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0084] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A method for collaborative optimization of a DDS-5G network based on fuzzy logic, characterized in that, The method comprises the following steps: Step S1: setting a static mapping rule to map the QoS Policy parameter of the DDS to the QoS flow identifier of the QoS flow of the 5G network; Step S1.1: determining the packet delay budget according to the LIFESPAN and DEADLINE of the QoS Policy; the packet delay budget selects the minimum value of the LIFESPAN and the DEADLINE; Step S1.2: determining the packet error rate according to the RELIABILIT of the QoS Policy; When the RELIABILIT of the QoS Policy is reliable, the packet error rate is set to When the RELIABILIT of the QoS Policy is best_effort, the packet error rate is set to ; Step S1.3: Set the resource type according to the DEADLINE setting, the HISTORY setting and the message publishing speed change rate of the QoS Policy; when the message publishing speed change rate is greater than and there is a DEADLINE setting, set the resource type as non-guaranteed bit rate; when the message publishing speed change rate is less than and the HISTORY setting is , set the resource type as delay-critical guaranteed bit rate; when the message publishing speed change rate is less than and the HISTORY setting is , set the resource type as guaranteed bit rate; Step S1.4: determining the QoS flow identifier according to the resource type, the packet delay budget, the packet error rate and the resource type; the non-guaranteed bit rate type looks up the 5QI value according to the packet delay budget, the delay-critical guaranteed bit rate type looks up the 5QI value according to the bandwidth demand, and the guaranteed bit rate type looks up the 5QI value according to the packet delay budget and the packet error rate; Step S2: monitoring the real-time network state and the communication system resource state and the control system resource state; when the network state changes by more than a threshold value or the communication system or control system performance indicators do not meet the standards, entering the DQN optimization; Step S3: establishing a fuzzy logic reasoning system, taking the QoS Policy parameter and the real-time network state parameter as input variables, and taking the 5QI as an output variable; Step S4: in the DQN optimization, based on the real-time collected network state and performance indicators, the DQN selects the optimal composite action based on the ε-greedy strategy; executing the optimal composite action, adjusting the QoS Policy parameter of the DDS and updating the fuzzy logic parameter of the fuzzy logic reasoning system, calculating the immediate reward value, and updating the Q network based on the immediate reward value.

2. The method according to claim 1, wherein, When monitoring the real-time network state, the network state includes the packet loss rate, the delay, and the congestion degree.

3. The method of claim 1, wherein the method is based on fuzzy logic. The fuzzy logic reasoning system comprises the following steps: Step S3.1: converting the input QoS Policy parameter and the real-time network state parameter into semantic description through fuzzification processing; Step S3.2: performing fuzzy reasoning, completing fuzzy reasoning according to the fuzzy rule base, and calculating the activation degree of each rule, and the minimum value in the rule activation degree is the confidence degree; Step S3.3: performing defuzzification and introducing a forced constraint logic, calculating the weighted average value of all activated rules, and outputting the 5QI value after constraint.

4. The method of claim 1, wherein the method is based on fuzzy logic. In the fuzzy logic reasoning system, the fuzzy rule base comprises 54 rules, and the inputs of the rules include one of two types of reliability, one of three types of delay budget, one of three types of bandwidth utilization, and one of three types of packet loss rate.

5. The method of claim 1, wherein the method is based on fuzzy logic. In the DQN optimization, the state space of the DQN network includes network performance and control performance, the control performance includes packet loss performance, end-to-end delay, network jitter and bandwidth utilization, and the control performance includes pose error and point cloud stitching residual error.

6. The method of claim 1, wherein the method is based on fuzzy logic. In the DQN optimization, an action space of the DQN network comprises QoS Policy adjustment and fuzzy logic parameter optimization; the QoS Policy adjustment comprises RELIABILIT policy adjustment, HISTORY DEPTH size adjustment, DURABILITY policy adjustment, and DEADLINE size adjustment; and the fuzzy logic parameter optimization comprises fuzzy rule weight correction and fuzzy logic membership function parameter update.

7. The method according to claim 1, wherein, In the DQN optimization, a reward function of the DQN network is represented as: ； wherein, represents a latency reward, represents a reliability reward, represents a mapping confidence reward; represents a maximum allowed latency, represents a packet loss rate, represents a maximum allowed packet loss rate, fuzzy_confidence represents a mapping confidence.

Images