Link anti-collision control method and system based on service flow energy characteristic assessment

Abstract

This application discloses a link anti-collision control method and system based on service flow energy characteristic assessment, relating to the field of wireless communication link resource control. The method includes: collecting transmission power, transmission duration, and transmission time information of each service flow in a shared wireless medium to construct an initial energy time series; performing multi-scale energy characteristic analysis on the series to extract energy mean, fluctuation parameters, burstiness, and energy spectrum characteristics to form a service flow energy fingerprint; calculating the correlation of energy waveforms and burstiness differences between service flows based on the energy fingerprint to obtain a conflict coupling coefficient characterizing the degree of competition and overlap; constructing a conflict risk matrix and determining the conflict risk value and risk level of each service flow; and adaptively adjusting the medium access backoff window and scheduling priority according to the risk level. This achieves dynamic anti-collision control at the data link layer, improving link resource utilization and transmission stability.

Description

Technical Field

[0001] This application relates to the field of wireless communication link resource control technology, and in particular to a link anti-collision control method and system based on service flow energy characteristic assessment. Background Technology

[0002] In wireless sensor networks, the Internet of Things (IoT), and other communication systems sharing physical media, multiple nodes typically need to transmit data on the same channel or adjacent communication resources. When multiple nodes initiate transmissions simultaneously within a similar time window, signal overlap and data packet collisions can easily occur, further leading to problems such as retransmissions, increased access latency, and increased node energy consumption. To coordinate node access to shared channels, the data link layer typically manages the timing of node transmissions and the channel occupancy process through media access control protocols.

[0003] Currently, common link-layer anti-collision methods mainly include contention-based carrier sense multiple access / collision avoidance mechanisms, improved mechanisms based on random backoff, allocation-based time division multiple access scheduling mechanisms, and some hybrid control mechanisms. These mechanisms typically reduce the probability of collisions by listening to channel conditions, setting random backoff windows, pre-dividing communication time slots, or adding control handshake processes. They can play a certain role in mitigating collisions under normal load and relatively stable network environments.

[0004] However, as wireless sensor networks and IoT applications increasingly move towards higher density, heterogeneity, and diversified services, network traffic exhibits significant differences in transmission frequency, duration, burstiness, packet length, and energy consumption. If uniform backoff parameters, fixed time slot allocation, or coarse-grained priority rules are still used to homogenize different traffic flows, it becomes difficult to accurately reflect the varying impacts of each traffic flow on the shared channel contention. Especially in scenarios with dense nodes, complex topologies, or rapidly fluctuating traffic loads, some highly bursty or energy-intensive traffic flows may experience concentrated contention within a short period, while regular traffic flows may suffer unnecessary waiting, collisions, or retransmissions due to a uniform control strategy.

[0005] Therefore, existing link-layer anti-collision mechanisms still suffer from problems such as coarse control granularity, insufficient perception of service differences, and poor matching between scheduling strategies and actual channel contention states when facing heterogeneous dynamic service flows. These problems affect the channel utilization efficiency, access latency, and node energy efficiency of shared-medium communication systems, and are particularly detrimental to battery-powered wireless sensor networks and IoT scenarios with a large number of nodes and continuously changing service states. Summary of the Invention

[0006] This application provides a link anti-collision control method, system, storage medium, computer program product, and electronic device based on service flow energy characteristic assessment, which at least solves the problem of insufficient matching degree between link anti-collision control and actual channel contention state in heterogeneous dynamic service flow scenarios in the prior art.

[0007] In a first aspect, embodiments of this application provide a link anti-collision control method based on service flow energy characteristic assessment. The method includes: collecting transmission power data, transmission duration data, and transmission time information of each service flow during transmission in a shared wireless medium; constructing an initial energy time series corresponding to each service flow based on the transmission power data, transmission duration data, and transmission time information; performing multi-scale energy characteristic analysis on the initial energy time series of each service flow, extracting multidimensional statistics including energy mean, energy fluctuation parameters, and energy burstiness, and obtaining energy spectrum features to characterize the energy distribution state at different time scales; constructing a service flow energy fingerprint corresponding to each service flow based on the multidimensional statistics and the energy spectrum features; calculating the energy waveform correlation and burstiness difference characteristics between different service flows based on the service flow energy fingerprints of each service flow; and determining the conflict coupling coefficient between different service flows based on the energy waveform correlation and burstiness difference characteristics. The coupling coefficient is used to characterize the degree of correlation between different service flows competing and overlapping in a shared wireless medium; wherein, the energy waveform correlation is used to characterize the degree of correlation between the energy time series change trends of different service flows within the same observation window, and the burst degree difference feature is used to characterize the degree of difference between the energy burst degrees of different service flows within the same observation window; a conflict risk matrix is ​​constructed based on the conflict coupling coefficients between different service flows, and each service flow is taken as the target service flow. Based on a set of target conflict coupling coefficients corresponding to the target service flow in the conflict risk matrix, the conflict risk value of the target service flow is determined, and the conflict risk value of each service flow is compared with a preset risk level threshold to determine the conflict risk level of each service flow; according to the conflict risk level of each service flow, the medium access backoff window and scheduling priority of the corresponding service flow are adaptively adjusted to perform dynamic anti-collision control for the data link layer according to the adjusted medium access backoff window and scheduling priority.

[0008] Secondly, embodiments of this application provide a link anti-collision control system based on service flow energy characteristic assessment. The system includes: an energy sequence construction unit, used to collect transmission power data, transmission duration data, and transmission time information of each service flow in the shared wireless medium during the transmission process, and construct an initial energy time sequence corresponding to each service flow based on the transmission power data, the transmission duration data, and the transmission time information; an energy fingerprint generation unit, used to perform multi-scale energy characteristic analysis on the initial energy time sequence of each service flow, extract multi-dimensional statistics including energy mean, energy fluctuation parameters, and energy burstiness, and obtain energy spectrum features to characterize the energy distribution state at different time scales, and construct a service flow energy fingerprint corresponding to each service flow based on the multi-dimensional statistics and the energy spectrum features; and a conflict coupling calculation unit, used to calculate the energy waveform correlation and burstiness difference characteristics between different service flows based on the service flow energy fingerprints of each service flow, and determine the conflict coupling coefficient between different service flows based on the energy waveform correlation and the burstiness difference characteristics. The conflict coupling coefficient is used to characterize the degree of correlation between different service flows competing and overlapping in a shared wireless medium; wherein, the energy waveform correlation is used to characterize the degree of correlation between the energy time series change trends of different service flows within the same observation window, and the burst degree difference feature is used to characterize the degree of difference between the energy burst degrees of different service flows within the same observation window; the risk level determination unit is used to construct a conflict risk matrix based on the conflict coupling coefficient between different service flows, and to determine the conflict risk value of each target service flow based on a set of target conflict coupling coefficients corresponding to the target service flow in the conflict risk matrix, so as to compare the conflict risk value of each service flow with a preset risk level threshold to determine the conflict risk level of each service flow; the backoff scheduling control unit is used to adaptively adjust the medium access backoff window and scheduling priority of the corresponding service flow according to the conflict risk level of each service flow, so as to execute dynamic anti-collision control for the data link layer according to the adjusted medium access backoff window and scheduling priority.

[0009] Thirdly, an electronic device is provided, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the steps of the link anti-collision control method based on traffic flow energy characteristic assessment according to any embodiment of the present application.

[0010] Fourthly, embodiments of this application provide a storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the link anti-collision control method based on service flow energy characteristic assessment of any embodiment of this application.

[0011] Fifthly, embodiments of this application provide a computer program product, including a computer program / instructions, which, when executed by a processor, implement the steps of the link anti-collision control method based on service flow energy characteristic assessment of any embodiment of this application.

[0012] The link anti-collision control method and system based on service flow energy characteristic assessment provided in this application can achieve at least the following technical effects:

[0013] By collecting transmission power, transmission duration, and transmission time information for each service flow and constructing an initial energy time series, this scheme can transform the transmission behavior of service flows in a shared wireless medium into energy occupancy information with temporal continuity and intensity characterization capabilities. Based on this, through multi-scale energy feature analysis, energy mean, energy fluctuation parameters, energy burst intensity, and energy spectrum characteristics are extracted. This allows for the construction of service flow energy fingerprints based on average occupancy levels, fluctuation stability, burst concentration, and energy distribution at different time scales. Thus, the behavioral differences of each service flow during link contention are objectively quantified, providing a finer-grained basis for subsequent conflict risk identification and differentiated access control.

[0014] Based on the energy fingerprinting of service flows, the correlation and burstiness differences of energy waveforms among different service flows are calculated, and the conflict coupling coefficient is determined accordingly. This allows for the characterization of the mutual influence between service flows from the perspectives of time synchronization, burst superposition relationships, and competition overlap trends. After constructing a conflict risk matrix based on the conflict coupling coefficient, the conflict risk value and conflict risk level are determined by a set of conflict coupling relationships corresponding to the target service flow. This enables link control decisions to no longer rely solely on the state of a single service flow, but to make a comprehensive judgment based on the competitive relationships between concurrent service flows. This improves the matching degree between the conflict risk level classification and the actual shared medium competition state, providing a quantitative basis for the adaptive adjustment of medium access backoff windows and scheduling priorities.

[0015] This technical solution incorporates the actual energy consumption characteristics of service flows and their competitive relationships into the data link layer anti-collision control process. It establishes a continuous processing logic from feature extraction to risk quantification and parameter adjustment, enabling backoff windows and scheduling priorities to be dynamically mapped according to the service flow status and its coupled risks. Consequently, link layer control shifts from a static and uniform access method to dynamic collaborative control oriented towards the differences and competitive correlations of service flows. This effectively reduces the probability of packet collisions and retransmissions caused by concentrated competition among service flows, improving channel utilization efficiency and node energy usage performance in shared-medium communication systems. Attached Figure Description

[0016] 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 some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 A flowchart of an example of a link anti-collision control method based on service flow energy characteristic assessment according to an embodiment of this application is shown;

[0018] Figure 2 This document illustrates an example of an operation flowchart for constructing service flow energy fingerprints corresponding to each service flow according to an embodiment of the present application.

[0019] Figure 3 A flowchart illustrating an example of determining the conflict coupling coefficient between different service flows in a method according to an embodiment of this application is shown.

[0020] Figure 4 A schematic diagram illustrating the system operation mechanism of an example of a link anti-collision control method based on service flow energy characteristic assessment according to an embodiment of this application is shown.

[0021] Figure 5 A schematic diagram of a comparative experimental simulation showing the normalized throughput and average access delay of different methods under varying network load is presented.

[0022] Figure 6 This diagram illustrates a comparative simulation of the cumulative distribution of average energy consumption per unit of effective data successfully transmitted at a network node using different methods.

[0023] Figure 7 A schematic diagram showing the comparative experimental simulation results of an example of parameter robustness and closed-loop dynamic response verification in an embodiment of this application is illustrated.

[0024] Figure 8A structural block diagram of an example of a link anti-collision control system based on traffic flow energy characteristic assessment according to an embodiment of this application is shown. Detailed Implementation

[0025] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0026] In contention-based access mechanisms, CSMA / CA (Carrier Sense Multiple Access with Collision Avoidance) and its accompanying BEB (Binary Exponential Backoff) mechanism are commonly used. These mechanisms typically reduce the probability of multiple nodes transmitting simultaneously through channel sensing, contention windows, and random backoff. However, some studies indicate that due to wireless propagation delays, even if a node initiates transmission after detecting channel idleness, signal overlap with other nodes may still occur within adjacent propagation windows. Furthermore, wireless nodes often struggle to synchronously complete effective collision detection during transmission, and post-collision recovery frequently relies on acknowledgment feedback, timeout assessment, and retransmission processes. When network load increases or multiple nodes choose similar backoff durations, transmission windows may converge again, increasing waiting, retransmission, and control overhead.

[0027] In low-power, short-range wireless communication scenarios, some current technologies employ ZigBee MAC or IEEE 802.15.4-like mechanisms, combining CSMA / CA with GTS (Guaranteed Time Slot) to balance random contention access with deterministic transmission for specific services. These mechanisms offer advantages such as simple implementation and mature protocols under typical loads, but their backoff counts, backoff windows, and guaranteed time slot configurations are often highly predefined. When network load increases periodically or service access demands change rapidly, excessive guaranteed time slot configuration may reduce channel utilization, while contention access may result in numerous retransmissions due to untimely parameter adjustments. Therefore, a trade-off exists between reliability, real-time performance, and resource utilization efficiency.

[0028] In complex topology environments, the hidden node problem further weakens the control effect of contention-based mechanisms. Due to differences in node communication range, obstruction conditions, and link quality, some transmitting nodes may not be able to directly detect each other, yet simultaneously send data to the same receiving node, resulting in collisions at the receiver. To address this issue, current technologies often employ RTS / CTS (Request To Send / Clear To Send) handshake mechanisms for coordination to reduce the probability of hidden node collisions. However, in wireless sensor networks and the Internet of Things (IoT), service data packets are typically short, and nodes are often constrained by battery capacity. Additional handshake frames consume channel resources and increase communication latency and node power consumption, thus limiting their application in high-density or high-frequency short-packet scenarios.

[0029] For allocation-based scheduling mechanisms, TDMA (Time Division Multiple Access) and its derived collision-free energy-saving MAC schemes typically reduce collisions by pre-setting transmit and receive time slots and putting nodes into a sleep state during non-communication periods to reduce idle listening power consumption. However, such mechanisms rely on strict time synchronization and scheduling maintenance. When the network scales up, node states change, or topology relationships are adjusted, time slot rearrangement, synchronization maintenance, and scheduling information exchange will incur additional system overhead; if the scheduling table is not updated in a timely manner, some time slots may be idle while some communication needs are queued and congested.

[0030] In general, current technologies have improved the link-layer anti-collision process in terms of contention backoff, guaranteed time slots, control handshake, and time-division scheduling. However, their control rules still mainly rely on channel sensing results, preset backoff parameters, handshake status, or periodic scheduling arrangements. In communication scenarios with dense nodes, complex topologies, significant load fluctuations, and limited node energy, the above mechanisms still have certain shortcomings in terms of control overhead, dynamic adaptability, channel utilization, and low-power operation.

[0031] It should be understood that the above description of the relevant technologies is intended only to help the public better understand the inventive spirit and motivation of this application, and is not intended to limit this application. Furthermore, the technical solutions described in the above-mentioned relevant technologies are not prior art, and may also be undisclosed technical solutions, such as those under research or in the laboratory stage.

[0032] The technical solutions in this application, including the collection, storage, use, processing, transmission, provision, and disclosure of users' personal information, comply with relevant laws and regulations and do not violate public order and good morals.

[0033] Figure 1A flowchart illustrating an example of a link anti-collision control method based on service flow energy characteristic assessment according to an embodiment of this application is shown.

[0034] Regarding the execution subject of the method in the embodiments of this application, it can be any controller or processor with computing or processing capabilities, such as an Internet of Things gateway controller, which implements the method in the embodiments of this application by running programs or instructions stored in a storage medium.

[0035] In some examples, it may be integrated into an electronic device or terminal through software, hardware, or a combination of both, and the type of terminal or electronic device may be diverse.

[0036] like Figure 1 As shown, in step S110, the transmission power data, transmission duration data, and transmission time information of each service stream in the shared wireless medium are collected during the transmission process, and the initial energy time series corresponding to each service stream is constructed based on the transmission power data, transmission duration data, and transmission time information.

[0037] Here, the shared wireless medium can be a wireless sensor network, the Internet of Things (IoT), an industrial IoT, a smart park wireless access network, a warehouse logistics wireless monitoring network, or other communication environments where multiple nodes compete for the same wireless channel. The various service flows can be data streams generated by different nodes, or data streams formed by different service types on the same node. For example, in a smart park scenario, temperature and humidity sensors can periodically upload environmental status data, access control or video security nodes can upload alarm data when abnormal events occur, and lighting or air conditioning control nodes can receive or send control command data. In an industrial IoT scenario, vibration sensors, motor status acquisition nodes, valve control nodes, and edge controllers may simultaneously generate different types of data services. These service flows typically differ significantly in transmission frequency, duration, data packet size, and transmission power; therefore, relying solely on the number of data packets or node identifiers is insufficient to accurately reflect the actual occupancy of the shared wireless medium by each service flow.

[0038] In some implementations, basic transmission data of each service flow during transmission can be collected at the node-side wireless transceiver module, link-layer media access control module, or network-side control node. Transmission power data characterizes the power occupancy level of the wireless medium during the transmission phase of the service flow; transmission duration data characterizes the duration of continuous occupancy of the shared wireless medium for that transmission event; and transmission time information determines the location of the transmission event on a unified timeline. Taking environmental monitoring and security alarm services in a smart park as examples, environmental monitoring services may transmit periodically with lower power and shorter duration, while security alarm services may transmit intensively for a longer duration or with higher transmission power within a short period. By synchronously collecting transmission power, transmission duration, and transmission time information, the system can simultaneously grasp the energy intensity, time occupancy, and temporal distribution of the service flows.

[0039] After obtaining the above data, the system can organize the discrete transmission events of each service flow according to the transmission time information, and combine the transmission power data and transmission duration data of the corresponding transmission events to form an initial energy time series that reflects the changes in energy consumption of the service flow during continuous observation. This initial energy time series is used to characterize the energy consumption trajectory of the service flow at different time positions, rather than simply counting how many data packets a certain service flow sent within a certain period of time. Therefore, for high-frequency short packet services, low-frequency long packet services, periodic status reporting services, and sudden alarm services, the system can perform subsequent analysis using a unified energy time series expression method.

[0040] In step S120, multi-scale energy feature analysis is performed on the initial energy time series of each service flow, extracting multi-dimensional statistics including energy mean, energy fluctuation parameters and energy burst degree, and obtaining energy spectrum features to characterize the energy distribution state at different time scales. Based on the multi-dimensional statistics and energy spectrum features, the service flow energy fingerprint corresponding to each service flow is constructed.

[0041] It should be noted that the energy consumption behavior of different service flows in a shared wireless medium may exhibit different temporal patterns. For example, periodic monitoring services such as temperature and humidity, smoke concentration, and equipment temperature rise typically show relatively stable low-intensity energy consumption; services such as equipment anomaly alarms, image capture uploads, and industrial equipment vibration change reporting may generate significant energy peaks in a short period of time; control command flows, although potentially small in data volume, may be highly sensitive to access latency and transmission timing. If these services are uniformly processed based solely on average traffic or packet count, it is difficult to reflect their true channel contention characteristics. Therefore, this embodiment performs multi-scale energy characteristic analysis on the initial energy time series to characterize the energy behavior of service flows from different perspectives.

[0042] Among these, the average energy value can be used to describe the overall energy consumption level of a service flow within the observation window. For example, the average energy value of a node continuously uploading device operation data may reflect its long-term occupancy of the shared medium. The energy fluctuation parameter can be used to describe the dispersion of the service flow's energy value around the average level. For example, periodic heartbeat services typically have small fluctuations, while abnormal alarm services may exhibit large fluctuations before and after triggering. The energy burst rate can be used to describe whether the service flow has rapidly changing transmission characteristics between adjacent or similar time steps. For example, video alarms, device fault reporting, or batch status synchronization may generate high burst rates in a short period of time. Through the above multidimensional statistics, the system can no longer only identify "whether a service flow is sent," but also further identify "in what form of energy the service flow is sent."

[0043] In addition to time-domain statistics, energy spectrum features can be obtained to characterize the energy distribution at different time scales. These energy spectrum features can reflect whether the energy changes in the service flow are mainly concentrated on longer or shorter time scales, or whether they exhibit multi-scale variations simultaneously. For example, periodic status reporting services may show significant energy concentration on more stable time scales, while sudden alarm services may show strong energy concentration on shorter time scales. By introducing energy spectrum features, we can supplement the scale distribution information that is difficult to fully express with energy mean, energy fluctuation parameters, and energy burstiness, giving the service flow energy fingerprint a more complete characterization capability.

[0044] After obtaining multidimensional statistics and energy spectrum characteristics, the system combines these characteristics into energy fingerprints for each service flow. These energy fingerprints can be understood as identifiers of the energy behavior of service flows within a shared wireless medium, reflecting not only the average occupancy intensity of the service flow but also its volatility, burstiness, and multi-scale distribution. By constructing these energy fingerprints, subsequent conflict analysis can be based on the actual energy behavior of the service flows, rather than simply treating different types of sensor data, alarm data, and control data as homogeneous traffic, thereby improving the system's ability to identify heterogeneous service flows.

[0045] In step S130, the energy waveform correlation and burstiness difference characteristics between different service flows are calculated based on the service flow energy fingerprint of each service flow. The conflict coupling coefficient between different service flows is then determined based on these characteristics. The conflict coupling coefficient characterizes the degree of correlation between different service flows that compete and overlap in the shared wireless medium. Here, the energy waveform correlation characterizes the correlation between the energy time series change trends of different service flows within the same observation window, and the burstiness difference characteristic characterizes the degree of difference in energy burstiness between different service flows within the same observation window.

[0046] It should be noted that conflicts in shared wireless media are not solely determined by the energy consumption of a single service flow, but also by the overlapping transmission times and energy variations among multiple service flows. For example, in an industrial workshop, multiple vibration sensors may simultaneously trigger reporting during the same period of abnormal equipment vibration; in a smart park, multiple security nodes may send alarm information concurrently after the same abnormal event occurs; in warehousing and logistics scenarios, multiple positioning nodes or environmental nodes may simultaneously initiate status updates due to the same scheduling cycle. Even if these service flows originate from different nodes, they may form synchronous or near-synchronous energy peaks within similar time windows, thereby increasing the risk of competition and overlap in the shared media.

[0047] Therefore, this embodiment proposes to calculate the energy waveform correlation between different service flows based on service flow energy fingerprints. If the energy time series change trends of two service flows are relatively similar within the same observation window, such as both rising simultaneously, reaching peaks simultaneously, or having similar fluctuation rhythms in similar time periods, it can be considered that the two have a high degree of correlation in competing and overlapping in the shared wireless medium. Through this energy waveform correlation, the system can identify which service flows are more likely to "collide" with each other in the time dimension, thereby providing a basis for subsequent pairwise conflict relationship assessment.

[0048] Furthermore, this embodiment also calculates the burst intensity difference characteristics between different service flows. These characteristics describe the difference in the intensity of short-term energy changes between two service flows. For example, a low-power periodic reporting service and a burst alarm service may overlap in transmission within a certain time window, but their impact on the shared medium is not the same. When two service flows with strong burst characteristics compete within similar time windows, it may put more significant pressure on the link layer scheduling. By introducing burst intensity difference characteristics, the system can consider not only whether the two changes synchronously, but also the impact of the difference in burst intensity on the competition overlap relationship when determining the conflict association between service flows.

[0049] After obtaining the correlation of energy waveforms and the difference in burstiness, the system merges the two to determine the conflict coupling coefficient between different service flows. This conflict coupling coefficient is used to transform the synchronization relationship and burst difference relationship of energy changes between two service flows into a unified pairwise conflict correlation index, enabling the system to complete the transformation from the energy fingerprint of a single service flow to the conflict coupling relationship between service flows.

[0050] In step S140, a conflict risk matrix is ​​constructed based on the conflict coupling coefficients between different business flows. Each business flow is taken as the target business flow, and the conflict risk value of the target business flow is determined based on a set of target conflict coupling coefficients corresponding to the target business flow in the conflict risk matrix. The conflict risk value of each business flow is compared with a preset risk level threshold to determine the conflict risk level of each business flow.

[0051] In practical implementation, in densely populated shared wireless media scenarios, a particular service flow typically does not compete with only a single service flow, but may be affected by multiple concurrent service flows simultaneously. For example, in an industrial IoT production line, a device status reporting flow may simultaneously compete with vibration monitoring, temperature monitoring, and control feedback flows from adjacent devices; in a smart campus network, a security alarm flow may simultaneously share the same wireless access resource with access control status flows, environmental monitoring flows, and lighting control flows. Therefore, merely judging the local conflict relationship between two service flows may be insufficient to assess the overall competitive pressure of a particular service flow in the current network.

[0052] Therefore, in this embodiment, the conflict coupling coefficients between different service flows are organized according to the service flow index to construct a conflict risk matrix; each non-corresponding element in the conflict risk matrix can represent a conflict coupling relationship between a pair of service flows. Through matrix representation, the system can observe the competitive relationship between each service flow and other service flows from an overall perspective, rather than viewing a single link or node in isolation, making the matrix a structured representation of the competitive relationship between service flows in the current shared wireless medium.

[0053] After obtaining the conflict risk matrix, the system takes each service flow as the target service flow and extracts a set of target conflict coupling coefficients corresponding to that target service flow from the conflict risk matrix. These target conflict coupling coefficients reflect the overall correlation between the target service flow and other concurrent service flows. The system determines the conflict risk value of the target service flow based on these target conflict coupling coefficients. In other words, the conflict risk value of a target service flow is not simply determined by its own energy consumption level, but rather comprehensively reflects its competitive and overlapping relationship with other active service flows. Thus, even if a service flow itself has low transmission power, if it frequently undergoes synchronous changes with multiple service flows within similar time windows, it may still be identified as having a high conflict risk.

[0054] Furthermore, the system compares the conflict risk value of each business flow with a preset risk level threshold to determine the conflict risk level of each business flow. For example, within the current observation window, if a business flow has a strong conflict coupling relationship with multiple other business flows, its conflict risk value may reach a high level and be determined as a high-risk level; if a business flow has a weak conflict coupling relationship with other business flows, it can be determined as a low-risk level. Thus, the global business flow competition relationship can be converted into a risk level that each business flow can use for scheduling control, providing clear input for subsequent differentiated backoff and priority adjustment.

[0055] In step S150, based on the conflict risk level of each service flow, the media access backoff window and scheduling priority of the corresponding service flow are adaptively adjusted so as to perform dynamic anti-collision control for the data link layer according to the adjusted media access backoff window and scheduling priority.

[0056] In some implementations, the system differentiates the media access behavior of each service flow at the data link layer based on its conflict risk level. For service flows with a high conflict risk level, its media access backoff window can be appropriately increased, or its scheduling priority within the current contention cycle can be reduced, causing its transmission timing to stagger with other service flows and reducing the likelihood of it competing for shared wireless media with other service flows in close time windows. For example, when multiple security alarm flows, device anomaly reporting flows, or batch status synchronization flows exhibit strong competitive correlation within the same time period, the system can implement more conservative access control for the higher-risk service flows to reduce the possibility of collisions and retransmissions caused by concentrated transmission.

[0057] For traffic flows with low collision risk, the system can maintain a small backoff window or grant them relatively stable access priority to reduce unnecessary waiting. For example, periodic environmental monitoring data, device heartbeat data, or low-frequency status reporting data, if their collision coupling with other traffic flows is weak, can obtain a more timely transmission opportunity without significantly increasing collision risk. Through this differentiated control, the system avoids the inefficiency caused by uniformly extending the backoff time for all traffic flows, and also prevents high-risk traffic flows from continuing to compete for shared channels in an disorderly manner during network congestion.

[0058] In this embodiment, the media access backoff window can be used to control the waiting range of a service flow before initiating transmission, and the scheduling priority can be used to control the access order or transmission opportunity allocation among multiple candidate service flows. By simultaneously adjusting the backoff window and scheduling priority, the system can coordinate service flows from both the perspectives of transmission timing and access order. In this way, the link layer control strategy no longer relies solely on fixed backoff parameters or uniform priority rules, but can be dynamically adjusted according to the conflict risk level of the service flow within the current observation window.

[0059] This embodiment establishes a complete processing link, encompassing service flow energy data acquisition, energy fingerprint construction, inter-service flow conflict coupling analysis, conflict risk level determination, and dynamic anti-collision control at the data link layer. This processing link transforms the energy behavior of service flows during actual transmission into a basis for link-layer media access control. This enables the system to more effectively distinguish between stable service flows, bursty service flows, and highly competitive related service flows in high-density heterogeneous service scenarios such as wireless sensor networks, industrial IoT, and smart parks. This helps reduce packet collisions and invalid retransmissions in shared wireless media, improves channel utilization efficiency and access latency, and reduces node energy consumption caused by conflict retransmissions.

[0060] Regarding the implementation details of constructing the initial energy time series of each service flow in step S110, in some examples of embodiments of this application, firstly, for each service flow, the transmission event of the corresponding service flow is mapped to a discrete time axis with a preset sampling period using the transmission time information, and for each transmission event within the same sampling period, energy conversion is performed based on the corresponding transmission power data and transmission duration data to obtain the energy increment corresponding to each transmission event, and the energy increments within the same sampling period are aggregated in the time dimension to obtain the discrete original energy sequence of the corresponding service flow.

[0061] In practice, media access in wireless sensor networks, industrial IoT, or other shared wireless media networks typically occurs as discrete transmission events, and the data packet transmission times of different service flows are not completely synchronized. To perform unified time-series processing on different service flows, the system can pre-set the sampling period. The system utilizes the transmission time information carried by each transmission event to map the transmission events of the corresponding service flow to the corresponding time steps on the discrete time axis. If multiple transmission events occur in the same service flow within the same sampling period, the energy increment is determined based on the transmission power data and transmission duration data corresponding to each transmission event. The multiple energy increments within the same sampling period are then aggregated to obtain the discrete original energy value of the service flow at that time step.

[0062] For example, the first Each business flow at time step The discrete original energy value can be determined by the following formula:

[0063] Equation (1)

[0064] In the formula, Indicates the first Each business flow at time step Discrete original energy values, Indicates at time step The first sampling period in the corresponding sampling period A set of sending events occurring in each service flow. Represents any send event in the set of send events. Indicates the first Sending events in each business flow The corresponding transmission power data, Indicates the first Sending events in each business flow The corresponding transmission duration data. Through the above processing, transmission events with different packet lengths, transmission durations, and power levels can be uniformly represented as energy occupancy values ​​at time steps, thereby obtaining the discrete raw energy sequence of the corresponding service flow.

[0065] Then, the average packet transmission period of the corresponding service flow is determined based on the transmission time information, and the reference window length is determined based on the correlation ratio between the average packet transmission period and the preset sampling period.

[0066] It should be noted that the data generation period and transmission frequency of different service flows may vary significantly. For example, environmental status monitoring services may report at shorter intervals, device heartbeat services may report at longer intervals, and alarm services may send data in a concentrated manner shortly after an event is triggered. If all service flows use the same length of smoothing window, it may lead to over-smoothing of local changes in high-frequency service flows, or insufficient suppression of random noise in low-frequency service flows. Therefore, the system can use transmission time information to calculate the time interval between adjacent transmission events of the same service flow and determine the average packet transmission period of that service flow accordingly. Based on this, combined with the preset sampling period Determine the baseline window length corresponding to this business flow.

[0067] For example, the reference window length can be determined by the following formula:

[0068] Equation (2)

[0069] In the formula, Indicates the first The baseline window length corresponding to each business flow Indicates the first Average packet transmission period for each service flow Indicates the preset sampling period. This indicates the preset scaling factor. This indicates an up-rounding operation. In this way, the reference window length can be matched with the transmission rhythm of the service flow itself, giving service flows with different transmission frequencies a relatively suitable initial smoothing scale.

[0070] Next, the energy fluctuation characteristics of the discrete original energy sequence within a preset observation window are extracted, and the reference window length is dynamically adjusted using these energy fluctuation characteristics to determine the current window length of the adaptive sliding window filter. Here, the current window length is a positive integer, and the value of the current window length is negatively correlated with the intensity of the energy fluctuation characteristics.

[0071] Here, the discrete raw energy sequence of the service flow may simultaneously contain random measurement noise and actual energy changes caused by burst transmissions. To avoid insufficient noise reduction during the steady phase or excessive smoothing during the burst phase using a fixed window, the system can calculate energy fluctuation characteristics within a preset observation window. The energy fluctuation characteristic can be a normalized first-order difference characteristic, a local variance characteristic, or other characteristics that can characterize the severity of local energy changes. The system dynamically adjusts the reference window length based on this energy fluctuation characteristic, using a longer window to enhance smoothing when the traffic flow is stable, and a shorter window to preserve the edges of energy changes when the traffic flow exhibits significant fluctuations.

[0072] For example, the current window length can be determined by the following formula:

[0073] Equation (3)

[0074] In the formula, Indicates the first Each business flow at time step The current window length, Indicates the first The baseline window length corresponding to each business flow Indicates the first Each business flow at time step The corresponding energy fluctuation characteristics This indicates the preset sensitivity coefficient. This indicates a floor function. By setting a maximum value constraint, you can ensure that the current window length is not less than [the maximum value]. As can be seen from this formula, when the energy fluctuation characteristics increase, the current window length decreases accordingly; when the energy fluctuation characteristics decrease, the current window length increases accordingly or returns to a level close to the baseline window length.

[0075] Furthermore, based on the current window length, an adaptive sliding window filter is used to perform local mean-smoothing and noise reduction processing on the discrete original energy sequence to filter out random measurement noise from the wireless channel and output the initial energy time series. Here, the local mean-smoothing and noise reduction processing determines the smoothed energy value for the corresponding time step based on the discrete original energy values ​​within the current time step and its preceding historical time interval, which is limited by the current window length.

[0076] For example, the initial energy time series can be expressed by the following equation:

[0077] Equation (4)

[0078] In the formula, Indicates the first Each business flow at time step The initial energy time series energy values, Indicates the first Each business flow at time step The energy values ​​of the discrete original energy sequence, Indicates the first Each business flow at time step The current window length, This indicates the summation index. The formula shows that when determining the smoothed energy value at the current time step, the system performs local averaging based on the current time step and several preceding historical time steps, without relying on data from future time steps. Therefore, it is suitable for online data link layer monitoring processes.

[0079] Through the embodiments of this application, discrete transmission events of various service flows can be converted into energy sequences with a unified time base, and smoothed using a sliding window that matches the transmission rhythm and local fluctuation state of the service flows. On the one hand, determining the energy increment based on transmission power data and transmission duration data can more accurately reflect the intensity of service flow occupancy on the shared radio medium than simply counting the number of packets; on the other hand, determining the current window length based on the average packet transmission period and energy fluctuation characteristics can enhance noise suppression during stable phases and reduce the weakening of burst features during fluctuating phases. Therefore, the resulting initial energy time series has good stability and time sequence identification capability, and can more accurately characterize the energy occupancy changes of service flows in the shared radio medium.

[0080] Figure 2 A flowchart illustrating an example of constructing service flow energy fingerprints corresponding to each service flow according to an embodiment of this application is shown.

[0081] like Figure 2As shown, in step S210, for a preset observation window, the sample statistical mean of the initial energy time series within the preset observation window is extracted as the energy mean, and the statistical standard deviation of the initial energy time series within the preset observation window is calculated as the energy fluctuation parameter characterizing the degree of energy dispersion within the window.

[0082] Here, the occupancy of the shared wireless medium by different service flows exhibits different steady-state characteristics on a macroscopic level. The system first extracts basic multidimensional statistics in the time domain to construct the fundamental dimensions of the service flow energy fingerprint. For example, the technical logic of the energy mean and energy fluctuation parameters can be expressed by the following formula:

[0083] Equation (5)

[0084] Equation (6)

[0085] In the formula, Indicates the first Average energy of each business flow Indicates the first Energy fluctuation parameters for each business flow This indicates the total number of time steps contained in the preset observation window. Indicates the first Each business flow at time step The initial energy time series energy values. From a communication physics perspective, the energy mean. This reflects the average channel occupancy load of the service flow within the current observation window (e.g., the average for video streams is much higher than that for periodic heartbeat streams); while the energy fluctuation parameter... This reflects the degree of dispersion of the service flow energy from its average level, and is used to quantify the stability of its transmission behavior.

[0086] In step S220, for the energy changes between adjacent time steps caused by burst transmission of service flows, the absolute energy change between adjacent time steps is extracted, and the relative change amplitude of a single step is determined by combining it with the energy baseline value of the previous time step. Then, the energy burst degree is obtained by averaging the relative change amplitudes of the single steps within a preset observation window. Here, when determining the relative change amplitude of a single step, a preset positive real number anti-zero factor is introduced to maintain the stability of numerical calculation when the energy value of the previous time step is too small or empty.

[0087] In practical implementation, in industrial control or abnormal alarm scenarios, business flows often burst within a very short time, generating high-intensity transient impacts. To accurately capture such behavior, the system calculates the energy burst rate. For example, the energy burst rate can be expressed by the following formula:

[0088] Equation (7)

[0089] In the formula, Indicates the first Energy burst rate of each business flow This represents the energy value of the initial energy time series at the previous time step. This is a pre-defined positive real-valued zero-prevention factor used to avoid a denominator of zero. The logic of this formula is that the numerator measures the absolute energy jump between adjacent time steps, while the denominator converts it into a relative rate of change. This is because wireless nodes may be in a dormant state during non-transmission periods. Approaching zero, introducing This not only prevents program crashes due to division by zero overflow, but also ensures the robustness of burst assessment. By employing an arithmetic mean based on the relative rate of change, it can sensitively reflect hidden "high-pulse" traffic in the network.

[0090] In step S230, continuous wavelet transform is used to decompose the initial energy time series into multiple scales to obtain wavelet coefficients within a preset scale set and a preset observation window. Based on the relative proportion of the integrated energy of the wavelet coefficients at a specific scale in the total integrated energy corresponding to the preset scale set, the normalized energy spectral density is obtained as the energy spectral feature.

[0091] In practice, the energy evolution of business flows is often non-stationary, and simple time-domain statistics cannot distinguish its inherent frequency-domain patterns. The system employs Continuous Wavelet Transform (CWT) as a time-frequency analysis tool, utilizing different scaling scales. Translational parameters The wavelet basis functions are used to perform inner product operations on the sequence. Large-scale operations can capture the low-frequency periodic macroscopic envelope of traffic flows, while small-scale operations are extremely adept at capturing local details and internal oscillatory energy components caused by instantaneous collisions or continuous transmission of tiny data packets. For example, the normalized energy spectral density can be expressed by the following formula:

[0092] Equation (8)

[0093] In the formula, Indicates the first A business flow at a specific scale Normalized energy spectral density under the following conditions Represents a predefined scale set Any typical scale in, Indicated in scale and time displacement The wavelet coefficients below, Indicates the preset observation window. The denominator represents the set of preset scales, and the summation of the wavelet coefficient energies at each scale within the preset scale set. Through the above normalization process, the influence of the absolute power of the signal is eliminated, enabling the system to characterize the time-frequency characteristics of the service flow purely based on the "distribution ratio structure" of energy at each scale.

[0094] In step S240, the energy mean, energy fluctuation parameters and energy burst degree are combined into a multidimensional statistic, and the multidimensional statistic is jointly vectorized and concatenated with the energy spectrum features under multiple typical scales to form a multidimensional comprehensive feature vector, which is then used as the energy fingerprint of each business flow.

[0095] In some implementations, to facilitate structured pairwise conflict correlation calculations in subsequent stages, the system needs to align and encapsulate the heterogeneous features extracted in the preceding steps. The system selects representative features from a preset scale set. At typical scales, the corresponding normalized energy spectral density sequences are extracted and concatenated with the previously obtained time-domain scalar features along with mathematical dimensions. For example, this joint concatenation process can be abstracted as a feature vector... Construction:

[0096] Equation (9)

[0097] In the formula, That is, the first The business flow energy fingerprint constructed from each business flow This represents the vector transpose operation; to For the pre-selected typical scales, this vectorized high-dimensional structure is equivalent to generating a series of business flow fingerprints with a unified structure for each complex business flow, so that any type of data flow can be quantified and cross-compared under the same mathematical framework.

[0098] Through the embodiments of this application, an extremely sophisticated non-stationary signal sensing mechanism is established above the data link layer, deeply integrating multi-dimensional time-domain statistics (steady-state load, discrete fluctuations, transient bursts) with time-frequency dual-domain analysis (multi-scale frequency band distribution ratios). The resulting high-dimensional energy fingerprint vector can not only accurately extract and quantify the hidden high-frequency oscillation energy components and low-frequency periodic behaviors in the service flow, but also ensures the extreme robustness of the evaluation model in dealing with extremely dynamic conditions and even channel fading through normalization processing and the introduction of zero-factor prevention.

[0099] Figure 3 A flowchart illustrating an example of determining the conflict coupling coefficient between different service flows in a method according to an embodiment of this application is shown.

[0100] like Figure 3 As shown, in step S310, for any two different first and second service flows in the shared wireless medium, the initial energy time series and the energy mean in the service flow energy fingerprints corresponding to the first and second service flows are extracted. Based on the centralized deviation distribution of the initial energy time series relative to the corresponding energy mean, the original statistical correlation coefficient between the two service flows is determined. Here, the original statistical correlation coefficient is used to characterize the degree of synchronization of the energy change trends of the two service flows, and a zero-prevention factor is introduced in the normalization process of the original statistical correlation coefficient to maintain the stability of the correlation calculation.

[0101] In some implementations, any two active service flows in the shared wireless medium can be considered as an evaluation pair. For example, the first service flow can be the second... The first business flow, the second business flow can be the first Each service flow is evaluated. For this pair, the system acquires the initial energy time series of both within the same observation window and obtains their corresponding energy mean values. To avoid interference from differences in the baseline energy levels of different service flows in the correlation judgment, the system first centers the initial energy time series of each service flow, i.e., using the corresponding energy mean as a benchmark, and extracts the deviation of the energy value at each time step relative to this benchmark. In this way, the original statistical correlation coefficient mainly reflects the degree of synchronization between the energy change trends of the two service flows, rather than simply reflecting their absolute energy magnitude.

[0102] For example, the original statistical correlation coefficient can be calculated using the following formula:

[0103] Equation (10)

[0104] In the formula, Indicates the first The first business flow and the first The original statistical correlation coefficients between individual business flows and They represent the first The first business flow and the first Each business flow at time step The energy value, and They represent the first The first business flow and the first Average energy of each business flow This is a preset positive real number anti-zero factor used to avoid the denominator being zero.

[0105] In equation (10), the numerator is used to characterize whether the centralized energy deviation of the two traffic flows has a consistent trend within the same observation window; the denominator is used to normalize based on the energy dispersion of each flow, making traffic flows of different energy levels comparable. By introducing a zero-prevention factor, when a traffic flow is basically stable within the observation window and has a low or zero dispersion, the normalization process can be avoided to prevent numerical instability, thereby improving the stability of the original statistical correlation coefficient calculation process.

[0106] In step S320, the original statistical correlation coefficient is subjected to non-negative truncation to filter out negative correlation components, and the retained non-negative correlation components are used as energy waveform correlation.

[0107] In practice, the original statistical correlation coefficient may be positive, zero, or negative. Positive correlation typically indicates that two traffic flows have similar energy change trends within the observation window, such as rising simultaneously or reaching energy peaks at the same time within similar time periods. Negative correlation indicates that the two traffic flows have opposite trends in energy change; for example, when one traffic flow's energy increases, the other's energy decreases or remains in a relatively low-occupancy state. In the context of conflict prevention and control, it is more important to focus on the competitive overlap resulting from positive synchronous changes, while negative correlation changes should generally not be interpreted as a positive conflict risk.

[0108] Therefore, the system performs non-negative truncation on the original statistical correlation coefficients. For example, the energy waveform correlation can be expressed as follows:

[0109] Equation (11)

[0110] In the formula, Indicates the first The first business flow and the first The energy waveform correlation between the two service flows. In equation (11), when the original statistical correlation coefficient is less than zero, the energy waveform correlation is set to zero; when the original statistical correlation coefficient is greater than or equal to zero, its non-negative part is retained as the energy waveform correlation. Through the above processing, the energy waveform correlation can reflect only the positive synchronous change relationship between the two service flows, avoiding unnecessary interference from negative correlation components to the evaluation of conflict coupling.

[0111] In step S330, the energy burst rate in the energy fingerprint of the first service flow and the second service flow is extracted, and the absolute difference between the energy burst rates of the first service flow and the second service flow is obtained, which is used as the burst rate difference feature.

[0112] In practice, energy waveform correlation mainly describes the degree of synchronization between two service flows in terms of their temporal change trends. However, even if different service flows have similar change trends, their short-term energy change intensities may differ. For example, one service flow may exhibit stable periodic reporting, while another service flow may experience short-term high-intensity transmission within a similar time period; or both service flows may exhibit burst behavior, but the burst intensities may differ. To further characterize this difference in short-term energy change intensity, the system extracts the energy burst degree of the two service flows and determines the burst degree difference characteristics based on the absolute difference between their burst degrees.

[0113] For example, the burstiness difference feature can be expressed by the following formula:

[0114] Equation (12)

[0115] In the formula, Indicates the first The first business flow and the first The characteristics of burstiness differences between individual business flows and They represent the first The first business flow and the first The burst intensity of each service flow. By using the absolute difference magnitude, the burst intensity difference feature can be made independent of the order of service flows, and can only be used to characterize the difference in burst intensity between two service flows.

[0116] In step S340, a burst degree difference weighting function based on the natural index is introduced, with the burst degree difference feature as the independent variable, and the difference amplification weight is determined by combining the preset burst degree sensitive adjustment factor; the energy waveform correlation and the difference amplification weight are nonlinearly fused to calculate the conflict coupling coefficient, thereby using the burst degree sensitive adjustment factor to adjust the degree of influence of the burst degree difference feature on the conflict coupling coefficient.

[0117] In practical implementation, the correlation of energy waveforms can serve as the basic evaluation metric for the positive synchronous change of two service flows, while the burst degree difference characteristic can be used to adjust the sensitivity of this basic evaluation metric to heterogeneous burst service flows. To enable the burst degree difference to exert a nonlinear adjustment effect on the conflict coupling evaluation, the system introduces a burst degree difference weighting function based on the natural index, and controls the amplification intensity of this weighting function through a preset burst degree sensitivity adjustment factor.

[0118] For example, the conflict coupling coefficient can be calculated using the following formula:

[0119] Equation (13)

[0120] In the formula, Indicates the first The first business flow and the first The conflict coupling coefficient between individual business flows Indicates the first The first business flow and the first Energy waveform correlation between individual service flows Indicates the first The first business flow and the first The characteristics of burstiness differences between individual business flows This is a preset burst sensitivity adjustment factor used to adjust the influence weight of burst difference characteristics on the conflict coupling coefficient.

[0121] In equation (13), the conflict coupling coefficient is based on the nonnegative energy waveform correlation and uses the burst intensity difference feature to form a nonlinear weighting. When the energy waveform correlation of two service flows is low, even if the burst intensity difference is large, their conflict coupling coefficient will be limited by the basic correlation. When the energy waveform correlation of two service flows is high, the burst intensity difference feature will further adjust their conflict coupling coefficient, so that the system has a higher degree of discrimination for service flow combinations with synchronous change relationship and obvious burst intensity difference. By adjusting the burst intensity sensitive adjustment factor, the influence intensity of the burst intensity difference feature on the conflict coupling coefficient can be changed to adapt to the anti-conflict control requirements under different network loads or service types.

[0122] Through the embodiments of this application, a centralized deviation distribution can be used to describe the degree of synchronization of energy changes between two service flows within the same observation window, and non-negative truncation processing can be used to focus the correlation of energy waveforms on a positive synchronous change relationship. Simultaneously, by introducing burst degree difference characteristics and a nonlinear weighting function, service flow combinations with different short-term energy change intensities can be further distinguished. Therefore, the resulting conflict coupling coefficient can stably characterize the degree of competition and overlap between two service flows, and provides a quantitative basis for evaluating the conflict correlation of paired service flows in a shared wireless medium.

[0123] Regarding the implementation details of determining the conflict risk level of each service flow in step S140, in some examples of embodiments of this application, the conflict coupling coefficients between pairs of active service flows in the current shared radio medium are arranged according to the service flow index to construct a conflict risk matrix. Here, the main diagonal elements of the conflict risk matrix are zero, and the off-diagonal elements are used to characterize the conflict coupling coefficients between corresponding two service flows.

[0124] It should be noted that in a multi-node concurrent shared wireless medium network, any service flow may simultaneously compete with multiple other service flows. For example, in an industrial wireless sensor network, a device status reporting flow may compete for the same wireless medium with vibration monitoring flow, temperature monitoring flow, and control feedback flow from adjacent devices; in a smart campus wireless network, a security alarm flow may compete for medium access with access control status flow, environmental monitoring flow, and lighting control flow within a similar time window. Therefore, relying solely on the conflict coupling relationship between a single service flow and other single service flows is insufficient to fully reflect the overall competitive status of that service flow in the current network. To address this, the system matrix-arranges the conflict coupling coefficients between pairs of currently active service flows according to their service flow indices to form a conflict risk matrix.

[0125] For example, the constructed conflict risk matrix can be expressed by the following formula:

[0126] Equation (14)

[0127] In the formula, Represents the conflict risk matrix. This indicates the total number of active service flows in the current shared wireless medium. Indicates the first The first business flow and the first The matrix represents the conflict coupling coefficients between various service flows. The rows and columns of this matrix correspond to the active service flows in the current shared wireless medium. Since a single service flow does not compete with itself for medium access, the diagonal elements of the matrix are set to zero; the off-diagonal elements represent the conflict coupling relationships between two different service flows. This matrix representation organizes the dispersed pairwise conflict coupling relationships into a unified data structure, enabling the system to describe the overall competitive relationship between service flows in the current network.

[0128] Then, for the conflict risk matrix, each business flow is taken as the target business flow, and a set of matrix elements in the conflict risk matrix corresponding to the target business flow and excluding the main diagonal elements are extracted as the corresponding set of target conflict coupling coefficients. The set of target conflict coupling coefficients is then averaged and aggregated to obtain the conflict risk value that characterizes the overall conflict correlation strength corresponding to the target business flow.

[0129] In specific implementation, the first When a business flow is used as the target business flow, the system can extract the first business flow from the conflict risk matrix. Each element outside the main diagonal of the row is considered a set of target conflict coupling coefficients between the target business flow and other active business flows. In other implementations, equivalent extraction can be performed based on the column vector elements corresponding to the target business flow in the matrix. If the conflict coupling coefficients are calculated symmetrically, the corresponding row and column vectors can have the same or equivalent risk representation meaning. The system performs mean aggregation based on the above set of target conflict coupling coefficients to obtain the conflict risk value of the target business flow. This conflict risk value reflects the overall competitive correlation level of the target business flow relative to other active business flows, rather than just reflecting the local impact of a single business flow.

[0130] For example, the conflict risk value can be determined by the following formula:

[0131] Equation (15)

[0132] In the formula, Indicates as the first The conflict risk value of the target business flow for each business flow. This indicates the total number of active service flows in the current shared wireless medium. Indicates the first The first business flow and the first The target conflict coupling coefficient between service flows. The above formula indicates that when there are at least two active service flows in the current shared radio medium, the system performs average aggregation on the target conflict coupling coefficient between the target service flow and other active service flows; when the total number of currently active service flows is less than two, it indicates that there are no other service flows that can form a concurrent competition relationship with the target service flow, and at this time, the conflict risk value of the target service flow is assigned to zero.

[0133] By employing the aforementioned mean-based aggregation process, the direct impact of changes in the number of active service flows on the absolute magnitude of the risk value can be reduced, allowing the conflict risk value to more accurately represent the average competitive correlation strength between the target service flow and other service flows. Thus, in both lightly loaded and heavily loaded networks, the system can compare different service flows based on a relatively consistent risk quantification method. Simultaneously, through... By assigning zeros to boundary cases, meaningless calculations can be avoided when there are no concurrent competing objects, and the risk value can be made to conform to the actual semantics of the shared medium competition relationship.

[0134] Next, the conflict risk value of the target business flow is compared with a preset risk level threshold, which is used to classify the conflict risk level of the business flow. On the one hand, when the conflict risk value of the target business flow is greater than or equal to the preset risk level threshold, the conflict risk level of the target business flow is marked as high risk. On the other hand, when the conflict risk value of the target business flow is less than the preset risk level threshold, the conflict risk level of the target business flow is marked as low risk.

[0135] In some implementations, the preset risk level threshold can be pre-set based on the channel capacity of the shared wireless medium, the expected collision rate, access latency requirements, service type, or network operation experience, or it can be adjusted by the system according to the operating status. This threshold is used to convert continuous collision risk values ​​into discrete risk levels that can be used by medium access control. When the collision risk value of a target service flow reaches or exceeds the threshold, it indicates that the overall collision correlation strength between the target service flow and other active service flows is high, and it can be marked as a high-risk level. When the collision risk value of a target service flow is lower than the threshold, it indicates that the overall collision correlation strength between the target service flow and other active service flows is low, and it can be marked as a low-risk level.

[0136] Through the embodiments of this application, the system can convert the pairwise conflict coupling relationship between different service flows into a conflict risk value for a single target service flow, and further convert the conflict risk value into a conflict risk level that is easy to use for link layer control. Thus, it retains the overall information of the competitive relationship between service flows, and reduces the calculation bias caused by changes in network scale through average aggregation, making the risk level determination of each service flow have good stability and comparability.

[0137] Regarding the details of the adaptive adjustment of the media access backoff window and scheduling priority of the service flow in the method of this application embodiment, in some examples of this application embodiment, firstly, for the target service flow determined to be of a high-risk level, a nonlinear backoff mapping model based on a sigmoid activation function is invoked. The deviation between the conflict risk value of the target service flow and the preset risk level threshold is used as the driving independent variable, and a smoothing parameter controlling the backoff growth slope is combined to determine the nonlinear backoff mapping ratio. Then, based on the backoff boundary interval formed by the minimum and maximum backoff time slots allowed by the protocol, a numerical projection is performed within the backoff boundary interval using the nonlinear backoff mapping ratio, and the projection result is rounded off by time slot alignment to determine the extended media access backoff window.

[0138] In practical implementation, during the media contention access process at the data link layer, the backoff window is used to control the waiting time range of a service flow before initiating transmission. If the target service flow has been determined to be high-risk, it indicates a strong competitive overlap with other active service flows. In this case, if a fixed backoff window or uniform backoff rule is still applied, it may be difficult to effectively stagger the high-risk service flow from other service flows in time. Therefore, this embodiment continuously and restrictively adjusts the backoff window based on the degree to which the conflict risk value of the target service flow exceeds the preset risk level threshold, so that the higher the risk of the service flow, the longer the backoff window is obtained, while ensuring that the backoff window is always kept within the range allowed by the protocol.

[0139] For example, the extended media access backoff window can be determined based on the number of backoff time slots using the following formula:

[0140] Equation (16)

[0141] In the formula, This indicates the number of backoff time slots corresponding to the adjusted media access backoff window. and These represent the minimum and maximum number of backoff time slots allowed by the protocol, respectively. This indicates a preset risk level threshold. To control the smoothing parameter of the retreat growth slope, This indicates the conflict risk value of the target business flow.

[0142] In equation (16), the system uses the deviation between the conflict risk value of the target service flow and the preset risk level threshold as input, obtains the backoff mapping ratio through the S-shaped activation function, and maps this ratio to the backoff boundary interval formed by the minimum backoff time slots and the maximum backoff time slots. When the conflict risk value of the target service flow exceeds the preset risk level threshold by a small margin, the number of backoff time slots increases in a relatively gradual manner; when the margin increases, the number of backoff time slots increases towards the maximum number of backoff time slots. Since the S-shaped activation function has saturation characteristics, the backoff window will not increase indefinitely, which helps to avoid excessively prolonged backoff waiting time. By rounding up, the backoff results obtained by continuous calculation can be matched with the backoff timing method of the data link layer, which uses discrete time slots as units.

[0143] Subsequently, for target service flows determined to be high-risk, a probabilistic priority downgrade mechanism is triggered. Based on the deviation between the target service flow's conflict risk value and a preset priority downgrade threshold, along with a preset suppression strength coefficient, a dynamic downgrade probability is calculated to delay transmission requests or reduce contention priority. Within the data link layer's media contention access cycle, the target service flow is controlled to relinquish its current channel contention opportunity according to the dynamic downgrade probability, thereby implementing a scheduling priority adjustment strategy. Here, the priority downgrade threshold is greater than or equal to the preset risk level threshold.

[0144] In some implementations, extending the backoff window is mainly used to adjust the transmission waiting time of the target service flow. However, when multiple high-risk service flows are simultaneously in a competitive state, relying solely on backoff window adjustment may still result in multiple service flows competing again at similar times. Therefore, this embodiment further supplements the adjustment of the scheduling priority of high-risk service flows through a probabilistic priority downgrading method. The priority downgrading threshold can be set to be greater than or equal to a preset risk level threshold, so that this adjustment mainly affects service flows with higher risk levels. The closer the conflict risk value of the target service flow is to or exceeds the priority downgrading threshold, the higher the probability that it will be delayed in transmission or have its competition priority reduced.

[0145] For example, the dynamic degradation probability can be expressed by the following formula:

[0146] Equation (17)

[0147] In the formula, This indicates the probability of dynamic downgrade. Degrade to the preset priority threshold and satisfy , This is the suppression intensity coefficient.

[0148] In equation (17), the system uses the deviation between the conflict risk value of the target service flow and the priority degradation threshold as input, and calculates the dynamic degradation probability between zero and one using a sigmoid activation function. The suppression strength coefficient is used to adjust the sensitivity of the dynamic degradation probability to changes in risk deviation. During the media contention access cycle, the system can determine whether the target service flow should postpone its current transmission request, reduce its current contention priority, or relinquish its current channel contention opportunity based on this dynamic degradation probability. Through the above probabilistic processing, a completely consistent deterministic delay strategy can be avoided for all high-risk service flows, thereby reducing the possibility of multiple high-risk service flows competing together again in subsequent contention cycles.

[0149] Furthermore, for target service flows determined to be of low risk level, their corresponding media access backoff window is set to the minimum backoff time slot, and they are given an access priority no lower than the basic transmission priority.

[0150] Here, "low-risk level" indicates that the overall conflict correlation between the target service flow and other active service flows is low within the current observation window. For this type of service flow, configuring a longer backoff window or a lower scheduling priority may result in unnecessary waiting and idle channel resources. Therefore, this embodiment sets the medium access backoff window for low-risk target service flows to the minimum number of backoff time slots allowed by the protocol, and assigns an access priority no lower than the basic transmission priority. This ensures that low-risk service flows receive timely transmission opportunities without significantly increasing the risk of contention overlap.

[0151] Through the embodiments of this application, the conflict risk level of service flows can be converted into backoff windows and scheduling priority control parameters executable at the data link layer. For high-risk service flows, the system extends their medium access waiting range through a restricted nonlinear backoff mapping and adjusts their contention priority through dynamic degradation probability; for low-risk service flows, the system maintains a smaller backoff window and a stable access priority. Thus, differentiated access control can be implemented for service flows of different risk levels within the same shared radio medium, alleviating concentrated contention for high-risk service flows while reducing unnecessary waiting for low-risk service flows.

[0152] Regarding the implementation details of dynamic anti-collision control for the data link layer in step S150, in some examples of embodiments of this application, firstly, during the process of executing data link layer media access control according to the adjusted media access backoff window and scheduling priority, the actual packet collision rate and the average energy consumption of nodes successfully transmitting a unit of valid data in the current network data link layer are collected within a preset dynamic monitoring period to obtain the current observed collision rate and the current observed energy consumption, respectively.

[0153] In practice, the dynamic monitoring period can be a sliding time window covering several media contention access periods. Within this dynamic monitoring period, the data link layer or node-side communication module can collect operational data such as the number of data packets sent, the number of retransmissions, the status of acknowledgment frame reception, media eavesdropping energy consumption, data transmission energy consumption, and acknowledgment frame reception energy consumption. The current observed collision rate can be used to characterize the retransmission level caused by collisions or suspected collisions during the current media access process, and the current observed energy consumption can be used to characterize the average energy consumed by the node to successfully transmit a unit of valid data.

[0154] For example, the current observation collision rate and the current observation energy consumption can be determined by the following formula:

[0155] Equation (18)

[0156] Equation (19)

[0157] In the formula, Indicates the current observed collision rate. This indicates the total number of data packet transmissions initiated by the data link layer within the dynamic monitoring period. This indicates the number of retransmissions that occur within the dynamic monitoring period due to the lack of an acknowledgment frame or the triggering of a retransmission condition. This represents a positive real number zero-prevention factor used to avoid a denominator of zero; This indicates the current energy consumption for observation. This represents the total energy used by a node for media listening, data transmission, and acknowledgment frame reception during the dynamic monitoring period. This represents the amount of payload data successfully delivered within the dynamic monitoring period. This represents a positive real number used to prevent the denominator from being zero. Through this quantification method, the actual link operating status within the dynamic monitoring period can be converted into observational indicators that can be used for parameter updates.

[0158] Then, using a closed-loop feedback self-learning mechanism, the collision residual between the current observed collision rate and the preset expected target collision rate, as well as the energy consumption residual between the current observed energy consumption and the preset expected target energy consumption, are calculated.

[0159] Here, the expected target collision rate can be preset based on service reliability requirements, allowable retransmission levels, or link service quality requirements; the expected target energy consumption can be preset based on node battery capacity, unit effective data transmission energy consumption requirements, or network energy-saving operation targets. The system compares the current observed collision rate with the expected target collision rate to obtain the collision residual; and compares the current observed energy consumption with the expected target energy consumption to obtain the energy consumption residual. The collision residual is used to characterize the direction and degree of deviation of the current network collision level from the target collision level, and the energy consumption residual is used to characterize the direction and degree of deviation of the current unit effective data transmission energy consumption from the target energy consumption level.

[0160] Subsequently, based on the adaptive update rule, feedback learning rates corresponding to specific parameters are introduced. Collision residuals are used to perform reverse iterative compensation updates on the preset risk level thresholds used to classify risk levels, and energy consumption residuals are used to perform forward iterative compensation updates on the smoothing parameters controlling the backoff growth slope. Here, when the current observed collision rate is higher than the expected target collision rate, the preset risk level threshold is reduced to improve the sensitivity of high-risk level determination in the next dynamic monitoring cycle; when the current observed collision rate is lower than the expected target collision rate, the preset risk level threshold is increased to reduce the possibility of excessive suppression of low-conflict business flows; when the current observed energy consumption is higher than the expected target energy consumption, the smoothing parameter is increased to enhance the response strength of the nonlinear backoff mapping model to changes in conflict risk; when the current observed energy consumption is lower than the expected target energy consumption, the smoothing parameter is decreased to reduce the adjustment sensitivity of the nonlinear backoff mapping model.

[0161] Here, the preset risk level threshold is used to control the triggering conditions for a business flow to be judged as high-risk. If the current observed collision rate is higher than the expected target collision rate, it indicates that the current high-risk business flow identification may not be sensitive enough. The system lowers the preset risk level threshold to include more business flows with competing and overlapping relationships in the high-risk level. If the current observed collision rate is lower than the expected target collision rate, the preset risk level threshold can be appropriately increased to reduce overly conservative control. The smoothing parameter is used to adjust the sensitivity of the backoff window in the nonlinear backoff mapping model to changes in conflict risk. If the current observed energy consumption is higher than the expected target energy consumption, the smoothing parameter can be increased to make the backoff mapping response to high-risk business flows more concentrated; if the current observed energy consumption is lower than the expected target energy consumption, the smoothing parameter can be decreased to make the backoff mapping change more gradually.

[0162] It should be noted that during the reverse iterative compensation update and forward iterative compensation update processes, limit value truncation protection is applied to the updated preset risk level threshold and smoothing parameter based on pre-set upper and lower limit ranges for the threshold and smoothing parameter, respectively, to achieve bounded iteration to prevent out-of-bounds errors. For example, based on the adaptive update rule, the preset risk level threshold used to classify risk levels is updated iteratively using collision residuals, and the smoothing parameter controlling the backoff growth slope is updated iteratively using energy consumption residuals.

[0163] Equation (20)

[0164] Equation (21)

[0165] In the formula, and These represent the preset risk level thresholds before and after the update, respectively. and These represent the lower and upper limits of the preset risk level threshold, respectively; and These represent the smoothing parameters before and after the update, respectively. and These represent the lower and upper limits of the smoothing parameter, respectively. and These represent the current observed collision rate and the expected target collision rate, respectively. Characterizes the collision residual; and These represent the current observed energy consumption and the expected target energy consumption, respectively. Characterizes the residual energy consumption; The learning rate corresponding to the collision residual. The learning rate, which matches the dimensions of the energy consumption residual, is used to convert the energy consumption residual into an update amount for the smoothing parameters.

[0166] In the above formula, the preset risk level threshold is updated in the opposite direction to the collision residual. When the current observed collision rate is higher than the expected target collision rate, the collision residual is positive, and the updated preset risk level threshold is relatively lower; when the current observed collision rate is lower than the expected target collision rate, the collision residual is negative, and the updated preset risk level threshold is relatively higher. The smoothing parameter is updated in the same direction as the energy consumption residual. When the current observed energy consumption is higher than the expected target energy consumption, the updated smoothing parameter is relatively larger; when the current observed energy consumption is lower than the expected target energy consumption, the updated smoothing parameter is relatively smaller. The learning rate can control the magnitude of each parameter update, avoiding excessive parameter changes due to short-term fluctuations.

[0167] In addition, through and The combination of these parameters can limit the preset risk level threshold and smoothing parameters to their respective preset upper and lower limits. This ensures that even if collision residuals or energy consumption residuals fluctuate significantly during the dynamic monitoring period, the updated parameters will not exceed the pre-allowed range, thus guaranteeing that the parameter update results meet the executable requirements of the data link layer scheduling control.

[0168] Furthermore, the preset risk level threshold and smoothing parameters, after being updated by the limit value truncation protection, are fed back and applied to the conflict risk level determination step and the nonlinear backoff mapping model in the next dynamic monitoring cycle, so that the anti-conflict scheduling strategy of the data link layer can be adaptively corrected according to the feedback of the current network operation.

[0169] In practice, the updated preset risk level threshold is used to determine the risk level of service flow conflicts in the next dynamic monitoring cycle, and the updated smoothing parameter is used for the nonlinear mapping of the media access backoff window in the next dynamic monitoring cycle. Therefore, the collision rate and unit effective data energy consumption observed in the current dynamic monitoring cycle can influence the sensitivity of risk level determination and the intensity of backoff window adjustment in the next dynamic monitoring cycle. In this way, the system can periodically correct the anti-collision control parameters according to the link operating status, ensuring that the control parameters match the current load and energy consumption status of the shared wireless medium.

[0170] Through the embodiments of this application, during the execution of dynamic anti-collision control at the data link layer, the current observed collision rate and current observed energy consumption can be introduced as operational feedback, and the preset risk level threshold and smoothing parameters can be updated in a bounded manner based on the collision residual and energy consumption residual. This allows the judgment of service flow risk level and the adjustment of backoff window to change with the network operating status, while upper and lower limit truncation prevents parameter update results from exceeding the executable range, thereby improving the stability and adaptability of dynamic anti-collision control under load fluctuation scenarios.

[0171] Figure 4 The diagram illustrates an example of a link anti-collision control method based on service flow energy characteristic assessment according to an embodiment of this application. The system operation mechanism is divided into four logical blocks: input area, core model area, scheduling control area, and output result area, forming a complete closed-loop processing link from bottom-level data perception and risk quantification assessment to top-level adaptive scheduling.

[0172] like Figure 4 As shown, firstly, in the input area, the system performs service flow energy acquisition and sliding window smoothing preprocessing, transforming the underlying discrete physical transmission events into a smooth and stable initial energy time series. Subsequently, the data flows into the core model area, where the system sequentially performs energy fingerprint construction (extracting features such as energy burstiness and multi-scale spectrum) and conflict risk matrix assessment (calculating the conflict coupling coefficient between concurrent service flows). After deep feature fusion and risk quantification in the core model area, the system finally outputs the high and low conflict risk level determination results for each service flow.

[0173] Next, the scheduling control area receives the high and low risk levels output above and executes adaptive backoff window control and energy-aware priority scheduling at the data link layer accordingly. This achieves fine-grained suppression of high-risk traffic flows and smooth passage of low-risk traffic flows. Simultaneously, based on the "performance feedback and update" module, the actual collision rate and energy consumption of the current network are collected in real time. This not only directly applies dynamic control guidance to the scheduling control area but also feeds back errors to the core model area through parameter closed-loop adjustment of the links. This drives the underlying anti-collision scheduling parameters to undergo bounded iteration and self-learning correction as network load fluctuates.

[0174] This significantly reduces the probability of data packet collisions in shared wireless media, effectively improves the overall data throughput of the network in high-density heterogeneous concurrent scenarios, and greatly saves the total energy consumption of nodes due to invalid contention and retransmission, thereby achieving a comprehensive improvement in wireless link communication performance at the global level.

[0175] To objectively verify the effectiveness and system boundaries of the link anti-collision control method based on service flow energy characteristic assessment proposed in this application, a high-fidelity simulation platform was built using a discrete event network simulator (such as NS-3). The experiment specifically simulated 100 static wireless sensor nodes deployed in a 100m × 100m area. The physical layer adopted the IEEE 802.15.4 standard, the operating frequency band was set to 2.4GHz, and the maximum data rate was set to 250kbps.

[0176] To accurately reflect the complex concurrent scenarios in shared wireless media networks, the service flow model in the experiment was set as a hybrid heterogeneous model. This model specifically includes low-energy periodic service flows exhibiting a Poisson distribution and high-energy bursty service flows exhibiting a Pareto distribution, thereby comprehensively covering the traffic characteristics of different energy forms, such as stable occupancy and drastic fluctuations.

[0177] The experiment mainly compared the performance of three medium access control (MAC) mechanisms: the first is the CSMA / CA (Carrier Sense Multiple Access with Collision Avoidance) mechanism, which is a traditional requestless send / clear send (RTS / CTS) handshake mechanism, serving as the baseline; the second is the improved informed backoff scheme protocol (IBSP); and the third is the dynamic anti-collision control scheme based on energy characteristics proposed in this application.

[0178] Figure 5 This diagram illustrates a comparative simulation of normalized throughput and average access delay under varying network load conditions, illustrating an example of different methods. This comparative experiment evaluated the system's channel utilization and latency performance under different network load conditions. During the simulation, the network load was adjusted by gradually increasing the total data packet generation rate, and a dual Y-axis line graph was used to visually represent the dynamic trends of normalized throughput (left Y-axis) and average access delay (right Y-axis).

[0179] like Figure 5The performance under high load conditions is shown. When the network load exceeds the 40% threshold, the traditional CSMA / CA mechanism and the improved Informed Backoff Protocol (IBSP), used as benchmarks, rapidly reach throughput saturation due to inherent collision "vulnerability times" and hidden nodes, accompanied by severe congestion and collapse, while the average access latency spikes exponentially. In contrast, the throughput curve of the method in this application only enters the saturation plateau when the load reaches 70%, and there is no sharp drop in throughput. This fully demonstrates that by introducing a collision risk matrix and a nonlinear backoff mapping model based on collision risk values ​​and smoothing parameters, this application can effectively stagger the transmission windows of high-energy-consuming burst traffic flows on the time axis, fundamentally avoiding retransmission storms in heavily loaded networks.

[0180] Furthermore, analyzing the performance trade-offs under low load conditions, in sparse scenarios with extremely low network load (less than 15%), the average access latency of the proposed method (approximately 12ms) is slightly higher than that of the Baseline mechanism (approximately 8ms), representing a negligible cost under low load. This is because the proposed method introduces sliding window smoothing preprocessing and multi-dimensional feature extraction of service flow energy fingerprints at the data link layer, resulting in certain baseband computation and processing overhead. However, as the network load increases, the retransmission latency of traditional methods rises uncontrollably and rapidly, while the latency growth of the proposed method remains relatively gradual. In summary, the negligible computational cost of the proposed method under extremely low load successfully yields significant overall benefits in anti-collision performance and throughput during medium-to-high load phases.

[0181] Figure 6 This diagram illustrates a comparative simulation of the cumulative distribution of average energy consumption per unit of effective data successfully transmitted by different methods at network nodes. The purpose of this comparative experiment is to evaluate the system's performance in eliminating long-tail energy consumption and the cumulative distribution function (CDF) of overall network node energy efficiency. In shared wireless media, nodes in complex interference topologies often face severe "long-tail" energy consumption problems, where continuous collisions and endless retransmissions lead to a surge in communication energy consumption and rapid battery depletion for individual nodes, easily creating network holes in the local area. Therefore, this experiment statistically analyzed the average energy consumption required for successful transmission of each byte of data by all simulated nodes and plotted the corresponding cumulative distribution function curves to visually compare the differences in energy consumption fairness and overall performance of different control methods.

[0182] like Figure 6The experimental results show that the cumulative distribution function curve of the traditional CSMA / CA mechanism (Baseline) exhibits an extremely long "right-tail" shape. This indicates that under the traditional disorderly competition mechanism, approximately 20% of nodes (usually hidden nodes or victim nodes in heavily interfered areas) experience significantly deteriorated energy consumption per unit of data transmission, far exceeding the average level of conventional nodes. In contrast, the cumulative distribution function curve of the proposed method not only shifts significantly to the left, resulting in a significant reduction in the average communication energy consumption of the entire network, but also achieves rapid convergence at the 95th percentile.

[0183] It should be noted that this application introduces a dynamic anti-collision control strategy based on the energy characteristics of service flows at the data link layer. By conducting a collision risk matrix assessment of global concurrent service flows, this application can accurately identify and grant low-risk service flows smooth priority access, while applying restricted nonlinear backoff mapping and probabilistic degradation scheduling to high-energy-consuming, bursty, and highly interfering service flows. This refined collaborative control effectively suppresses the blind preemption of high-energy-consuming flows and weakens the long-tail effect of energy consumption caused by retransmission storms. This not only optimizes the overall energy utilization of the link but also effectively balances and extends the lifespan of all network nodes from a global perspective.

[0184] Figure 7 This diagram illustrates a comparative simulation of an example of parameter robustness and closed-loop dynamic response verification in an embodiment of this application. This set of comparative experiments comprehensively verifies the robustness of the core control logic of the data link layer in this application from two dimensions: static parameter sensitivity and dynamic time-axis response.

[0185] like Figure 7 Part (a) refers to the preset risk level threshold (i.e., the retreat threshold in the figure). ) and the suddenness sensitivity modulator (i.e., the suddenness sensitivity factor in the figure). A two-dimensional mesh scan was performed, and contour plots were drawn to show the improvement in network collision rate. The contour distribution indicates that when... and When the system is positioned within the optimal intermediate joint interval shown in the diagram, it can most accurately isolate service flows with high collision potential. Conversely, if... When the energy density approaches zero (i.e., the control logic ignores the transient energy burst differences between concurrent service flows), the system degenerates into a single feature evaluation mode that relies solely on the correlation of energy waveforms. This fails to effectively distinguish high-risk service flow combinations with similar trends but vastly different burst intensities, resulting in a significant reduction in the improvement of the collision rate. Therefore, this static dimension indirectly confirms the necessity of extracting multi-scale service flow energy fingerprints and introducing burst difference features into the conflict coupling coefficient, as proposed in this application.

[0186] like Figure 7Part (b) further demonstrates the closed-loop dynamic response process of the system when faced with drastic changes in external traffic. At 50 seconds into the simulation, a sudden surge of high-load traffic is injected into the network. This instantaneous and drastic load change immediately triggers the closed-loop performance feedback and update mechanism of the data link layer. As shown in the figure, within an extremely short transient oscillation period (approximately 2.5 seconds), the system, based on the feedback learning rate parameter... and The system rapidly updates the preset risk level threshold and smoothing parameters in a bounded iterative manner, thereby increasing the average backoff window time slots for high-risk service flows. This dynamic adjustment mitigates local collisions and retransmissions caused by sudden traffic surges, allowing network throughput to stabilize after brief fluctuations and gradually recover to higher operating efficiency. Based on dynamic test results, it fully demonstrates that the energy-characteristic-driven anti-collision scheduling strategy proposed in this application possesses strong anti-interference and self-healing capabilities in complex and variable real-world engineering deployments subject to tidal impacts.

[0187] In summary, this application addresses the problem of traditional media access control technologies neglecting the heterogeneous energy characteristics of service flows by proposing a dynamic anti-collision control method for links based on service flow energy characteristic assessment. This scheme overcomes the limitations of "homogeneous blind backoff" in traditional protocols by introducing multi-scale service flow energy fingerprints at the data link layer for the first time to model underlying behavior, and constructs a coupled assessment model for collision risk by combining waveform correlation and burstiness differences. Based on this, the system implements refined adaptive suppression of high-risk service flows by invoking a nonlinear backoff mapping model based on a sigmoid activation function and a probabilistic priority degradation mechanism, while providing priority access guarantees for low-risk service flows. Furthermore, by introducing a closed-loop feedback self-learning mechanism based on both observed collision rate and energy consumption indicators, bounded iterative updates of core control parameters are achieved, enabling the system to highly adapt to complex and ever-changing network loads and topology evolution.

[0188] High-fidelity network simulation experiments objectively and comprehensively verified the effectiveness and system boundaries of the proposed solution. Experimental results show that the proposed solution not only effectively delays network congestion and collapse under high-load concurrent scenarios and significantly eliminates the "long tail of energy consumption" caused by hidden nodes, but also demonstrates extremely fast closed-loop convergence capability when facing sudden traffic surges; its minimal computational overhead under low load successfully achieves a significant improvement in global network throughput and efficiency. Thus, this application takes into account both underlying physical state awareness and link-layer collaborative scheduling.

[0189] It should be understood that the core technical architecture proposed in this application is not only applicable to wireless sensor networks and IoT environments, but can also be smoothly extended to shared medium communication systems with higher requirements for determinism and reliability, such as industrial IoT and vehicle networking, in the future. Furthermore, its application potential can be further released through dedicated baseband chip solidification and protocol stack standardization.

[0190] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of combined actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Secondly, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application. In the above embodiments, the descriptions of each embodiment have their own emphasis; for parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0191] Figure 8 A structural block diagram of an example of a link anti-collision control system based on traffic flow energy characteristic assessment according to an embodiment of this application is shown.

[0192] like Figure 8 As shown, the link anti-collision control system 800 based on service flow energy characteristic assessment includes an energy sequence construction unit 810, an energy fingerprint generation unit 820, a conflict coupling calculation unit 830, a risk level determination unit 840, and a backoff scheduling control unit 850.

[0193] The energy sequence construction unit 810 is used to collect the transmission power data, transmission duration data and transmission time information of each service stream in the shared wireless medium during the transmission process, and to construct the initial energy time sequence corresponding to each service stream based on the transmission power data, transmission duration data and transmission time information.

[0194] The energy fingerprint generation unit 820 is used to perform multi-scale energy feature analysis on the initial energy time series of each service flow, extract multi-dimensional statistics including energy mean, energy fluctuation parameters and energy burst degree, and obtain energy spectrum features to characterize the energy distribution state at different time scales. Based on the multi-dimensional statistics and the energy spectrum features, the energy fingerprint of each service flow is constructed.

[0195] The conflict coupling calculation unit 830 is used to calculate the energy waveform correlation and burst degree difference characteristics between different service flows based on the service flow energy fingerprint of each service flow, and to determine the conflict coupling coefficient between different service flows according to the energy waveform correlation and the burst degree difference characteristics. The conflict coupling coefficient is used to characterize the degree of correlation of different service flows competing and overlapping in the shared wireless medium. The energy waveform correlation is used to characterize the degree of correlation between the energy time series change trends of different service flows within the same observation window, and the burst degree difference characteristics are used to characterize the degree of difference between the energy burst degrees of different service flows within the same observation window.

[0196] The risk level determination unit 840 is used to construct a conflict risk matrix based on the conflict coupling coefficients between different business flows, and to determine the conflict risk value of the target business flow based on a set of target conflict coupling coefficients corresponding to the target business flow in the conflict risk matrix, and to compare the conflict risk value of each business flow with a preset risk level threshold to determine the conflict risk level of each business flow.

[0197] The backoff scheduling control unit 850 is used to adaptively adjust the media access backoff window and scheduling priority of the corresponding service flow according to the conflict risk level of each service flow, so as to perform dynamic anti-collision control for the data link layer according to the adjusted media access backoff window and scheduling priority.

[0198] In some embodiments, this application provides a non-volatile computer-readable storage medium storing one or more programs including execution instructions. The execution instructions can be read and executed by electronic devices (including but not limited to computers, servers, or network devices) to perform the steps of any of the link anti-collision control methods based on service flow energy characteristic assessment described above.

[0199] In some embodiments, this application also provides a computer program product, the computer program product including a computer program stored on a non-volatile computer-readable storage medium, the computer program including program instructions, which, when executed by a computer, cause the computer to perform any of the steps of the link anti-collision control method based on service flow energy characteristic assessment described above.

[0200] In some embodiments, this application also provides an electronic device, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the steps of a link anti-collision control method based on traffic flow energy characteristic assessment.

[0201] The above-described product can perform the methods provided in the embodiments of this application, and has the corresponding functional modules and beneficial effects for performing the methods. Technical details not described in detail in this embodiment can be found in the methods provided in the embodiments of this application.

[0202] The electronic devices in this application can exist in various forms, including but not limited to: mobile communication devices, ultra-mobile personal computer devices, portable entertainment devices, or other airborne electronic devices with data interaction functions.

[0203] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0204] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented using software plus a general-purpose hardware platform, or of course, using hardware. Based on this understanding, the above technical solutions, in essence or the parts that contribute to the related technology, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0205] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A link anti-collision control method based on service flow energy characteristic assessment, characterized in that, The method includes: The system collects transmission power data, transmission duration data, and transmission time information of each service stream in the shared wireless medium during the transmission process, and constructs the initial energy time series corresponding to each service stream based on the transmission power data, the transmission duration data, and the transmission time information. Multi-scale energy feature analysis is performed on the initial energy time series of each service flow to extract multi-dimensional statistics including energy mean, energy fluctuation parameters and energy burst degree, and energy spectrum features are obtained to characterize the energy distribution state at different time scales. Based on the multi-dimensional statistics and the energy spectrum features, the service flow energy fingerprint corresponding to each service flow is constructed. Based on the energy fingerprint of each service flow, the energy waveform correlation and burst degree difference characteristics between different service flows are calculated, and the conflict coupling coefficient between different service flows is determined according to the energy waveform correlation and the burst degree difference characteristics. The conflict coupling coefficient is used to characterize the degree of correlation of competition and overlap between different service flows in the shared wireless medium. The energy waveform correlation is used to characterize the correlation between the energy time series change trends of different service flows within the same observation window, and the burst degree difference characteristics are used to characterize the degree of difference between the energy burst degrees of different service flows within the same observation window. A conflict risk matrix is ​​constructed based on the conflict coupling coefficients between different business flows. Each business flow is taken as the target business flow. The conflict risk value of the target business flow is determined based on a set of target conflict coupling coefficients corresponding to the target business flow in the conflict risk matrix. The conflict risk value of each business flow is compared with a preset risk level threshold to determine the conflict risk level of each business flow. Based on the conflict risk level of each service flow, the media access backoff window and scheduling priority of the corresponding service flow are adaptively adjusted to perform dynamic anti-conflict control for the data link layer according to the adjusted media access backoff window and scheduling priority. The step of calculating the energy waveform correlation and burstiness difference characteristics between different service flows based on the energy fingerprint of each service flow, and determining the conflict coupling coefficient between different service flows based on the energy waveform correlation and burstiness difference characteristics, includes: For any two different first and second service flows in a shared wireless medium, the initial energy time series corresponding to the first and second service flows and the mean energy value in the service flow energy fingerprint are extracted. Based on the centralized deviation distribution of the initial energy time series relative to the corresponding mean energy value, the original statistical correlation coefficient between the two service flows is determined. The original statistical correlation coefficient is used to characterize the degree of synchronization of the energy change trends of the two service flows. A zero-prevention factor is introduced in the normalization process of the original statistical correlation coefficient to maintain the stability of the correlation calculation. The original statistical correlation coefficients are subjected to non-negative truncation to filter out negative correlation components, and the retained non-negative correlation components are used as energy waveform correlation. Extract the energy burstiness from the energy fingerprints of the first and second service flows, obtain the absolute difference between the energy burstiness of the first and second service flows, and use it as the burstiness difference feature; By introducing a burst degree difference weighting function based on the natural index, the burst degree difference feature is used as the independent variable, and a preset burst degree sensitivity adjustment factor is used to determine the difference amplification weight; the energy waveform correlation and the difference amplification weight are nonlinearly fused to calculate the conflict coupling coefficient, thereby using the burst degree sensitivity adjustment factor to adjust the degree of influence of the burst degree difference feature on the conflict coupling coefficient.

2. The method according to claim 1, characterized in that, The step of constructing the initial energy time series corresponding to each service flow based on the transmission power data, the transmission duration data, and the transmission time information includes: For each service flow, the transmission time information is used to map the transmission events of the corresponding service flow onto a discrete time axis with a preset sampling period. For each transmission event within the same sampling period, energy conversion is performed based on the corresponding transmission power data and transmission duration data to obtain the energy increment corresponding to each transmission event. The energy increments within the same sampling period are then aggregated in the time dimension to obtain the discrete original energy sequence of the corresponding service flow. The average packet transmission period of the corresponding service flow is determined based on the transmission time information, and the reference window length is determined based on the correlation ratio between the average packet transmission period and the preset sampling period. The energy fluctuation characteristics of the discrete original energy sequence within a preset observation window are extracted, and the length of the reference window is dynamically adjusted using the energy fluctuation characteristics to determine the current window length of the adaptive sliding window filter; wherein, the current window length is a positive integer, and the value of the current window length is negatively correlated with the intensity of the energy fluctuation characteristics; Based on the current window length, the adaptive sliding window filter is used to perform local mean-smoothing and noise reduction processing on the discrete original energy sequence to filter out random measurement noise in the wireless channel and output the initial energy time series; wherein, the local mean-smoothing and noise reduction processing determines the smoothed energy value of the corresponding time step based on the discrete original energy values ​​in the current time step and its preceding historical time interval, and the preceding historical time interval is limited by the current window length.

3. The method according to claim 2, characterized in that, The initial energy time series of each service flow is subjected to multi-scale energy feature analysis, extracting multidimensional statistics including energy mean, energy fluctuation parameters, and energy burst degree, and obtaining energy spectrum features to characterize the energy distribution state at different time scales. Based on the multidimensional statistics and the energy spectrum features, a service flow energy fingerprint corresponding to each service flow is constructed, including: For the preset observation window, the sample statistical mean of the initial energy time series within the preset observation window is extracted as the energy mean, and the statistical standard deviation of the initial energy time series within the preset observation window is calculated as the energy fluctuation parameter characterizing the degree of energy dispersion within the window; To address the energy changes between adjacent time steps caused by burst transmission of service flows, the absolute energy change between adjacent time steps is extracted, and the relative change amplitude of a single step is determined by combining it with the energy baseline value of the previous time step. Then, the energy burst degree is obtained by averaging the relative change amplitude of a single step within the preset observation window. In determining the relative change amplitude of a single step, a preset positive real number anti-zero factor is introduced to maintain numerical calculation stability when the energy value of the previous time step is too small or empty. The initial energy time series is decomposed into multiple scales using continuous wavelet transform to obtain wavelet coefficients within a preset scale set and the preset observation window. Based on the relative proportion of the integrated energy of the wavelet coefficients at a specific scale in the total integrated energy corresponding to the preset scale set, the normalized energy spectral density is obtained as the energy spectral feature. The energy mean, the energy fluctuation parameter, and the energy burst degree are combined into a multidimensional statistic. The multidimensional statistic is then jointly vectorized and concatenated with the energy spectrum features at multiple typical scales to form a multidimensional comprehensive feature vector. This multidimensional comprehensive feature vector is then used as the energy fingerprint of each service flow.

4. The method according to claim 3, characterized in that, The process involves constructing a conflict risk matrix based on the conflict coupling coefficients between different service flows, and taking each service flow as a target service flow. Based on a set of target conflict coupling coefficients corresponding to the target service flow in the conflict risk matrix, the conflict risk value of each service flow is determined. The conflict risk value of each service flow is then compared with a preset risk level threshold to determine the conflict risk level of each service flow. The conflict coupling coefficients between each pair of active service flows in the current shared wireless medium are arranged according to the service flow index to construct a conflict risk matrix; wherein, the main diagonal elements of the conflict risk matrix are zero, and the off-diagonal elements are used to characterize the conflict coupling coefficients between the corresponding two service flows. For the conflict risk matrix, each business flow is taken as the target business flow. A set of matrix elements in the conflict risk matrix that correspond to the target business flow and are excluding the main diagonal elements are extracted as a corresponding set of target conflict coupling coefficients. The set of target conflict coupling coefficients are then averaged and aggregated to obtain the conflict risk value that characterizes the overall conflict association strength corresponding to the target business flow. During the average aggregation calculation process, the number of valid aggregation items is determined based on the total number of active service flows in the current shared wireless medium; when the total number of service flows is less than two, the conflict risk value of the target service flow is configured to zero. The conflict risk value of the target business flow is compared with a preset risk level threshold, which is used to classify the conflict risk level of the business flow. When the conflict risk value of the target service flow is greater than or equal to the preset risk level threshold, the conflict risk level of the target service flow is marked as high risk level; When the conflict risk value of the target service flow is less than the preset risk level threshold, the conflict risk level of the target service flow is marked as low risk level.

5. The method according to claim 4, characterized in that, The adaptive adjustment of the media access backoff window and scheduling priority of the corresponding service flow based on the conflict risk level of each service flow includes: For the target business flow that is determined to be of the high-risk level, a nonlinear backoff mapping model based on the S-shaped activation function is invoked. The excess deviation between the conflict risk value of the target business flow and the preset risk level threshold is used as the driving independent variable, and the nonlinear backoff mapping ratio is determined by combining the smoothing parameter that controls the backoff growth slope. Based on the backoff limit range formed by the minimum and maximum backoff time slots allowed by the protocol, the nonlinear backoff mapping ratio is used to project values ​​within the backoff limit range, and the projection results are rounded by time slot alignment to determine the extended media access backoff window. For target service flows determined to be at the high-risk level, a probabilistic priority downgrade mechanism is triggered. The deviation between the conflict risk value of the target service flow and a preset priority downgrade threshold, along with a preset suppression strength coefficient, is used as the evaluation basis to calculate a dynamic downgrade probability for delaying the transmission request or reducing the contention priority. Within the media contention access period of the data link layer, the target service flow is controlled to relinquish its current channel contention opportunity according to the dynamic downgrade probability, thereby implementing a scheduling priority adjustment strategy. The priority downgrade threshold is greater than or equal to the preset risk level threshold. For target service flows determined to be of the low-risk level, their corresponding media access backoff window is set to the minimum backoff time slot number, and an access priority of not less than the basic transmission priority is assigned.

6. The method according to claim 5, characterized in that, The step of performing dynamic anti-collision control for the data link layer according to the adjusted media access backoff window and the scheduling priority includes: During the execution of data link layer media access control according to the adjusted media access backoff window and scheduling priority, the actual packet collision rate and the average energy consumption of nodes successfully transmitting a unit of valid data in the current network data link layer are collected within a preset dynamic monitoring period to obtain the current observed collision rate and current observed energy consumption, respectively. Using a closed-loop feedback self-learning mechanism, the collision residual between the current observed collision rate and the preset expected target collision rate, as well as the energy consumption residual between the current observed energy consumption and the preset expected target energy consumption, are calculated. Based on adaptive update rules, feedback learning rates corresponding to specific parameters are introduced. The collision residual is used to perform reverse iterative compensation update on the preset risk level threshold used to classify risk levels, and the energy consumption residual is used to perform forward iterative compensation update on the smoothing parameter controlling the backoff growth slope. Specifically, when the current observed collision rate is higher than the expected target collision rate, the preset risk level threshold is reduced to improve the sensitivity of high-risk level determination in the next dynamic monitoring cycle; when the current observed collision rate is lower than the expected target collision rate, the preset risk level threshold is increased to reduce the possibility of excessive suppression of low-conflict service flows; when the current observed energy consumption is higher than the expected target energy consumption, the smoothing parameter is increased to enhance the response strength of the nonlinear backoff mapping model to changes in conflict risk; when the current observed energy consumption is lower than the expected target energy consumption, the smoothing parameter is decreased to reduce the adjustment sensitivity of the nonlinear backoff mapping model. During the reverse iterative compensation update and the forward iterative compensation update, limit value truncation protection is performed on the updated preset risk level threshold and the smoothing parameter according to the preset threshold upper and lower limit range and the smoothing parameter upper and lower limit range, respectively. The preset risk level threshold and the smoothing parameter, after being updated by the limit value truncation protection, are fed back and applied to the conflict risk level determination step and the nonlinear backoff mapping model in the next dynamic monitoring cycle, so that the anti-conflict scheduling strategy of the data link layer can be adaptively corrected according to the feedback of the current network operation.

7. A link anti-collision control system based on service flow energy characteristic assessment, characterized in that, The system is used to implement the method as described in any one of claims 1-6; the system comprises: An energy sequence construction unit is used to collect transmission power data, transmission duration data, and transmission time information of each service stream in the shared wireless medium during the transmission process, and to construct an initial energy time sequence corresponding to each service stream based on the transmission power data, the transmission duration data, and the transmission time information. The energy fingerprint generation unit is used to perform multi-scale energy feature analysis on the initial energy time series of each service flow, extract multi-dimensional statistics including energy mean, energy fluctuation parameters and energy burst degree, and obtain energy spectrum features to characterize the energy distribution state at different time scales. Based on the multi-dimensional statistics and the energy spectrum features, the energy fingerprint of each service flow is constructed. The conflict coupling calculation unit is used to calculate the energy waveform correlation and burst degree difference characteristics between different service flows based on the service flow energy fingerprint of each service flow, and to determine the conflict coupling coefficient between different service flows according to the energy waveform correlation and the burst degree difference characteristics. The conflict coupling coefficient is used to characterize the degree of correlation of different service flows competing and overlapping in a shared wireless medium. The energy waveform correlation is used to characterize the correlation between the energy time series change trends of different service flows within the same observation window, and the burst degree difference characteristics are used to characterize the degree of difference between the energy burst degrees of different service flows within the same observation window. The risk level determination unit is used to construct a conflict risk matrix based on the conflict coupling coefficients between different business flows, and to determine the conflict risk value of the target business flow based on a set of target conflict coupling coefficients corresponding to the target business flow in the conflict risk matrix, and to compare the conflict risk value of each business flow with a preset risk level threshold to determine the conflict risk level of each business flow. The backoff scheduling control unit is used to adaptively adjust the media access backoff window and scheduling priority of each service flow according to the conflict risk level of each service flow, so as to perform dynamic anti-collision control for the data link layer according to the adjusted media access backoff window and scheduling priority.

8. An electronic device, comprising: At least one processor; as well as A memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the steps of the method as described in any one of claims 1-6.

9. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1-6.

Images