Adaptive routing based low power wireless ad hoc network dynamic configuration method and system

Abstract

The application provides a low-power wireless self-organizing network dynamic configuration method and system based on adaptive routing, relates to the technical field of network protocols, and comprises the following steps: acquiring node working state parameters, calculating an energy level value to determine a data transmission path, and setting a multi-level data cache occupation threshold; calculating a network load distribution state, and differentially scheduling nodes; selecting a low-power forwarding channel based on node routing applicability; and monitoring a network state in real time and updating parameters. The application improves network adaptability, reduces communication power consumption, prolongs the network life cycle, and improves data transmission efficiency.

Description

TECHNICAL FIELD

[0001] The present application relates to network protocol technology, and in particular to a low-power wireless ad hoc network dynamic configuration method and system based on adaptive routing. BACKGROUND

[0002] Low-power wireless ad hoc network technology has been widely used in the Internet of Things, smart home, industrial control and other fields due to its self-organizing and self-configuring characteristics. Such a network is usually composed of a large number of nodes deployed in a decentralized manner, each node having limited energy supply and computing resources. In practical applications, nodes need to work cooperatively to build an efficient data transmission path to ensure network reliability and prolong the overall network life cycle.

[0003] The existing low-power wireless ad hoc network technology has some technical defects in dynamic configuration and adaptability routing. The existing routing algorithm often adopts a fixed routing strategy and fails to fully consider the real-time energy status of nodes, resulting in unbalanced energy consumption and premature depletion of power in some nodes, forming a "energy hole" in the network and affecting the overall network connectivity. The existing technology lacks an effective monitoring mechanism for network area load distribution, and when some nodes in certain areas are heavily loaded, it cannot timely adjust the data flow distribution strategy, causing local network congestion and increasing data transmission delay. Most current routing algorithms lack the ability to perceive the data buffer state of nodes and cannot dynamically adjust the network configuration according to the buffer occupancy rate, which easily causes buffer overflow and data loss in data burst transmission scenarios.

[0004] With the diversification and complexity of Internet of Things application scenarios, the reliability, adaptability and energy efficiency of low-power wireless ad hoc networks are constantly improving, and there is an urgent need for an adaptive routing configuration method that can dynamically adjust according to the real-time state of nodes and network load to achieve rational allocation of network resources and efficient use of energy. SUMMARY

[0005] The embodiments of the present application provide a low-power wireless ad hoc network dynamic configuration method and system based on adaptive routing, which can solve the problems in the prior art.

[0006] In a first aspect, the embodiments of the present application provide a low-power wireless ad hoc network dynamic configuration method based on adaptive routing, comprising:

[0007] In a second aspect, the embodiments of the present application provide a low-power wireless ad hoc network dynamic configuration system based on adaptive routing, comprising:

[0008] obtaining the working state parameters of each node in the wireless ad hoc network;

[0009] According to the working state parameter, an energy level value of the node is calculated, a data flow transmission path is determined based on the energy level value, and a multi-level data cache occupancy threshold is set; when it is detected that the node data cache occupancy rate reaches a threshold of different levels, the energy level value of the node is correspondingly raised, different degrees of network collaborative adjustment are triggered, and the data flow transmission path is recalculated based on the updated energy level value;

[0010] Based on the working state parameter, a network region load distribution state is calculated, when it is detected that the load distribution is uneven, the nodes on the data flow transmission path are differentially scheduled in combination with the energy level value of the current node, and the data processing order and forwarding proportion are dynamically adjusted;

[0011] According to the energy level value of the node and the region load distribution state, a routing applicability of the node is calculated, and a node group with a routing applicability meeting a preset condition and a lowest communication power consumption is selected to form a data forwarding channel;

[0012] The network running state is monitored in real time, when a monitoring parameter exceeds a preset range, the working state parameter of the node is updated, and the energy level value of the node is calculated.

[0013] In an optional implementation,

[0014] According to the working state parameter, an energy level value of the node is calculated, a data flow transmission path is determined based on the energy level value, and a multi-level data cache occupancy threshold is set; when it is detected that the node data cache occupancy rate reaches a threshold of different levels, the energy level value of the node is correspondingly raised, different degrees of network collaborative adjustment are triggered, and the data flow transmission path is recalculated based on the updated energy level value;

[0015] The working state parameter includes historical performance indicators of the node and network topology location information;

[0016] According to the data transmission success rate, performance fluctuation degree and data processing capacity of the historical performance indicators in different time windows, a historical stability score is calculated, and the historical stability score is obtained by weighted calculation by giving different weight coefficients to different time windows;

[0017] Based on the network topology location information, the degree centrality, betweenness centrality and closeness centrality of the node are calculated, and a topology importance coefficient is obtained by weighted fusion, and the weight of the weighted fusion is dynamically adjusted according to the network topology change rate;

[0018] According to the resource state information of the node, a basic energy value is calculated, the basic energy value is adaptively weighted fused with the historical stability score and the topology importance coefficient to obtain an energy level value, the weight coefficient of the adaptive weighted fusion is dynamically adjusted according to a network stability indicator, a data flow transmission path is determined based on the energy level value, and a warning threshold, a critical threshold and an emergency threshold of the data cache occupancy rate are set.

[0019] In an optional implementation,

[0020] When a node's data cache occupancy rate is detected to reach different threshold levels, the node's energy level is increased accordingly, triggering varying degrees of network collaborative adjustment. The steps for recalculating the data flow transmission path based on the updated energy level include:

[0021] Based on the historical average and standard deviation of the node's cache utilization rate, a dynamic adjustment mechanism for different levels of thresholds is established. The dynamic adjustment mechanism dynamically updates the threshold adjustment coefficient based on the difference between the cache processing success rate and the preset target processing rate.

[0022] The cache growth trend is predicted based on historical time-series data sequences and attention weight matrices, and the cache pressure coefficient is calculated based on the prediction results.

[0023] When the cache utilization rate reaches different threshold levels, the node's energy level value is updated based on the cache pressure coefficient.

[0024] Based on the available resource ratio of nodes and network topology location information, select a collaborative node group and calculate the target load allocation ratio of each node in the collaborative node group.

[0025] The cost of the data stream transmission path is calculated based on the updated energy level value and the target load allocation ratio. The data stream transmission path is selected according to the cost value. The optimal transmission path is determined by comprehensively scoring the reliability, latency and stability of the data stream transmission path.

[0026] A gradual path switching mechanism is established to adaptively adjust the data flow allocation ratio of the old and new transmission paths according to the time process. The data flow allocation ratio is related to the comprehensive score.

[0027] In one alternative implementation,

[0028] A gradual path switching mechanism is established to adaptively adjust the data flow allocation ratio between the old and new transmission paths according to the time progress. The steps related to the data flow allocation ratio and the comprehensive score include:

[0029] Determine the sliding time window for path switching, calculate the optimal switching duration based on the total amount of data to be migrated, the current load level, and the maximum load threshold, and update the sliding time window according to the optimal switching duration;

[0030] The resource utilization efficiency of the new and old transmission paths is calculated. The resource utilization efficiency is obtained by weighted fusion of various resource utilization rates. The comprehensive score is then weighted and corrected based on the resource utilization efficiency to obtain the corrected comprehensive score.

[0031] The switching process value is calculated based on the time progress within the sliding time window. The switching process value increases non-linearly with the time progress within the sliding time window. The basic data flow allocation ratio of the old and new transmission paths is calculated based on the switching process value and the corrected comprehensive score.

[0032] Extract the business priority of the data stream, and calculate the actual allocation ratio by weighting the business priority with the basic allocation ratio of the data stream.

[0033] A performance evaluation matrix is ​​constructed based on the latency, packet loss, and bandwidth utilization metrics of the old and new transmission paths during the switching process. The deviation values ​​between the metrics in the performance evaluation matrix and the preset target metrics are calculated. The calculation parameters of the switching process value and the weighting coefficient of the comprehensive score are adjusted according to the deviation values ​​to dynamically optimize the path switching process.

[0034] In one alternative implementation,

[0035] Based on the aforementioned working status parameters, the network area load distribution status is calculated. When uneven load distribution is detected, the nodes on the data flow transmission path are differentiated and scheduled according to their energy level values, dynamically adjusting the data processing order and forwarding ratio. The steps include:

[0036] A load feature vector is constructed based on the node's working status parameters, and the load feature vector is weighted based on the load weight coefficient to obtain the node's comprehensive load value.

[0037] Calculate the standard deviation and mean of the comprehensive load value within the network area. When the ratio of the standard deviation to the mean exceeds a first preset threshold, and the deviation of the comprehensive load value of any node within the area from the mean exceeds a second preset threshold, it is determined that the load distribution is uneven. In response to the uneven load distribution, obtain the current energy level value of each node on the data stream transmission path, and calculate the energy weight coefficient based on the current energy level value.

[0038] Based on the processing priority of the calculation node according to the comprehensive load value and the energy weight coefficient, the data processing sequence is sorted according to the processing priority; based on the processing priority and the forwarding ratio of the calculation node according to the comprehensive load value, data traffic is allocated according to the forwarding ratio.

[0039] A scheduling effect evaluation index is constructed, which includes the ratio of regional load standard deviation to mean, node processing priority distribution, and energy weight coefficient. When the deviation between the calculated result of the scheduling effect evaluation index and the preset target threshold exceeds a preset range, the load weight coefficient is dynamically updated.

[0040] In one alternative implementation,

[0041] The process of constructing a scheduling performance evaluation index that includes the ratio of regional load standard deviation to mean, node processing priority distribution, and energy weight coefficient, and dynamically updating the load weight coefficient when the calculated result of the scheduling performance evaluation index deviates from a preset target threshold by more than a preset range, includes the following steps:

[0042] Construct a multi-scale time window, calculate the ratio of regional load standard deviation to mean in each time window, and weight the load balance assessment value by assigning different weight coefficients based on the time scale.

[0043] A priority distribution rationality score is calculated based on the dispersion of the processing priorities of the nodes. The priority distribution rationality score is obtained by calculating the entropy value of the processing priorities.

[0044] The network energy efficiency index is calculated based on the energy weighting coefficient. The network energy efficiency index reflects the data processing efficiency per unit energy consumption. The ratio of the network energy efficiency index to the historical best value is used as the energy utilization evaluation value.

[0045] The load balancing evaluation value, the priority distribution rationality score, and the energy utilization evaluation value are adaptively weighted and fused to obtain the scheduling effect evaluation index. The weight coefficients of the adaptive weighted fusion are dynamically adjusted according to the network performance requirements.

[0046] When the deviation between the scheduling effect evaluation index and the preset target threshold exceeds a preset range, the contribution of each dimension of the load feature vector is calculated based on the deviation, and the load weight coefficient is updated according to the contribution.

[0047] In one alternative implementation,

[0048] The steps for calculating the routing suitability of nodes based on their energy level and regional load distribution, and selecting nodes with the lowest communication power consumption that meet preset conditions to form a data forwarding channel, include:

[0049] A candidate route set is constructed based on the communication range of the nodes and the network topology. The hop count between each node in the candidate route set and the target node is calculated. The link quality from each node to the target node is determined based on the hop count. The link quality is obtained by weighting the signal strength, bit error rate and packet loss rate.

[0050] The remaining lifetime of a node is predicted based on its energy level value. The remaining lifetime prediction value decreases exponentially as the energy level value decreases. The maximum load carrying capacity of the node is then set based on the remaining lifetime prediction value.

[0051] The load balancing index in the regional load distribution status is combined with the maximum load carrying capacity of the node to calculate the load capacity score of the node.

[0052] Based on the link quality, the remaining lifetime prediction value, and the load capacity score, the routing suitability of the node is calculated, and the communication power consumption per unit data transmission is calculated for nodes whose routing suitability meets the preset conditions.

[0053] The node sequence with the highest routing suitability and lowest communication power consumption is selected to construct a data forwarding channel. The reliability index of the data forwarding channel is calculated. When the reliability index is lower than a preset reliability threshold, the node sequence with the second-best routing suitability is selected to construct a backup forwarding channel.

[0054] The first unit is used to obtain the working status parameters of each node in the wireless ad hoc network;

[0055] The second unit is used to calculate the energy level value of the node according to the working status parameters, determine the data flow transmission path based on the energy level value, and set multi-level data cache occupancy thresholds; when the node data cache occupancy rate is detected to reach different level thresholds, the energy level value of the node is increased accordingly, triggering different degrees of network collaborative adjustment, and the data flow transmission path is recalculated based on the updated energy level value.

[0056] The third unit is used to calculate the network area load distribution status based on the working status parameters. When uneven load distribution is detected, it performs differentiated scheduling of nodes on the data flow transmission path in combination with the energy level value of the current node, and dynamically adjusts the data processing order and forwarding ratio.

[0057] The fourth unit is used to calculate the routing suitability of nodes based on their energy level and regional load distribution, and select nodes with routing suitability that meet preset conditions and have the lowest communication power consumption to form a data forwarding channel.

[0058] The fifth unit is used to monitor the network operation status in real time. When the monitored parameters exceed the preset range, it updates the working status parameters of the nodes and returns the energy level value of the computing nodes.

[0059] A third aspect of the present invention provides an electronic device, comprising:

[0060] processor;

[0061] Memory used to store processor-executable instructions;

[0062] The processor is configured to invoke instructions stored in the memory to execute the aforementioned method.

[0063] A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the aforementioned method.

[0064] This invention obtains the working status parameters of each node, calculates the energy level value, and determines the data transmission path based on this. When the node cache occupancy rate reaches different thresholds, it triggers network collaborative adjustment, realizes adaptive optimization of network topology, and improves the load balancing and stability of the system.

[0065] By calculating and monitoring the load distribution status of network areas in real time, this invention can perform differentiated scheduling based on node energy level values ​​when uneven load distribution is detected. This dynamically adjusts the data processing order and forwarding ratio, effectively avoiding network congestion and excessive node energy consumption, and extending the life cycle of the entire network.

[0066] Based on the node-based routing suitability calculation and selection mechanism, this invention realizes the construction of a data forwarding channel with optimal communication power consumption. Combined with real-time monitoring and parameter update mechanisms, the system can adjust the routing strategy in a timely manner according to changes in the network environment, significantly reducing the overall energy consumption of the wireless ad hoc network and improving the network's adaptability and reliability. It is particularly suitable for resource-constrained IoT application scenarios. Attached Figure Description

[0067] Figure 1 This is a flowchart illustrating the dynamic configuration method for low-power wireless ad hoc networks based on adaptive routing, according to an embodiment of the present invention.

[0068] Figure 2 A flowchart for dynamically adjusting the energy level of network nodes and optimizing data flow transmission paths based on cache pressure. Detailed Implementation

[0069] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0070] The technical solution of the present invention will be described in detail below with reference to specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0071] Figure 1 This is a flowchart illustrating the dynamic configuration method for low-power wireless ad hoc networks based on adaptive routing, as described in an embodiment of the present invention. Figure 1 As shown, the method includes:

[0072] Obtain the working status parameters of each node in the wireless ad hoc network;

[0073] The energy level value of the node is calculated based on the working status parameters. The data flow transmission path is determined based on the energy level value, and multi-level data cache occupancy thresholds are set. When the data cache occupancy rate of a node is detected to reach different level thresholds, the energy level value of the node is increased accordingly, triggering different degrees of network collaborative adjustment. The data flow transmission path is recalculated based on the updated energy level value.

[0074] The network area load distribution status is calculated based on the working status parameters. When uneven load distribution is detected, the nodes on the data flow transmission path are differentiated and scheduled in combination with the energy level value of the current node, and the data processing order and forwarding ratio are dynamically adjusted.

[0075] The routing suitability of a node is calculated based on its energy level and regional load distribution. Nodes with the lowest communication power consumption that meet the preset conditions are selected to form a data forwarding channel.

[0076] The system monitors the network's operational status in real time. When the monitored parameters exceed the preset range, it updates the node's operational status parameters and returns the energy level value of the computing node.

[0077] In one optional implementation, the steps of calculating the node's energy level value based on the working status parameters, determining the data stream transmission path based on the energy level value, and setting multi-level data buffer occupancy thresholds include:

[0078] The working status parameters include the node's historical performance indicators and network topology location information;

[0079] Based on the data transmission success rate, performance fluctuation and data processing capability of the historical performance indicators in different time windows, a historical stability score is calculated. The historical stability score is obtained by weighting different time windows by assigning different weight coefficients to them.

[0080] Based on the network topology location information, the degree centrality, betweenness centrality, and proximity centrality of nodes are calculated, and a weighted fusion is performed to obtain the topological importance coefficient. The weights of the weighted fusion are dynamically adjusted according to the network topology change rate.

[0081] The basic energy value is calculated based on the resource status information of the node. The basic energy value is then adaptively weighted and fused with the historical stability score and the topology importance coefficient to obtain the energy level value. The weight coefficient of the adaptive weight fusion is dynamically adjusted according to the network stability index. The data flow transmission path is determined based on the energy level value, and warning thresholds, critical thresholds, and emergency thresholds for data cache occupancy are set.

[0082] In real-world network environments, the data stream transmission path determination method based on node energy level values ​​optimizes data transmission efficiency by evaluating node status from multiple dimensions. This embodiment details the technical implementation process of calculating energy level values ​​based on node operating status parameters, determining data stream transmission paths accordingly, and setting multi-level buffer thresholds.

[0083] For example, the working status parameters of a node include two main categories of data: historical performance metrics and network topology location information. Historical performance metrics cover the long-term performance characteristics of the node in the network, while network topology location information reflects the node's position and importance in the overall network structure.

[0084] When calculating the historical stability score, the data transmission success rate, performance volatility, and data processing capability of the nodes are collected within different time windows. Three time scales are set: short-term window (e.g., the last 24 hours), medium-term window (e.g., the last 7 days), and long-term window (e.g., the last 30 days). For each time window, a data transmission success rate is calculated, such as 98.5% for the short-term window, 97.2% for the medium-term window, and 96.8% for the long-term window. Performance volatility is obtained by calculating the standard deviation, such as 1.2 for the short-term window, 1.8 for the medium-term window, and 2.3 for the long-term window. Data processing capability is expressed as the number of packets processed per second, such as 850 packets / second for the short-term window, 820 packets / second for the medium-term window, and 800 packets / second for the long-term window. Different weighting coefficients are assigned to each window, for example, 0.5 for the short-term window, 0.3 for the medium-term window, and 0.2 for the long-term window. The overall historical stability score is obtained by multiplying the index values ​​of the three windows by their respective weights and summing the results. For example, the overall score for data transmission success rate is 98.5%×0.5+97.2%×0.3+96.8%×0.2=97.83%. Similarly, the overall score for performance fluctuation and data processing capability is calculated, and the three are finally combined according to a preset weight (e.g., 4:3:3) to form a historical stability score.

[0085] When calculating the topological importance coefficient of a node based on network topology location information, a complete network topology graph must first be obtained. Degree centrality represents the number of other nodes directly connected to a node. For example, if a node is connected to 8 other nodes, its original degree centrality value is 8. Betweenness centrality reflects the frequency with which a node lies on the shortest path between other node pairs in the network, and is calculated to a standardized value of 0.35. Proximity centrality measures the reciprocal of the average distance from a node to all other nodes in the network, and is standardized to 0.62. Considering the dynamic changes in network topology, the weights of the three centrality indicators are dynamically adjusted according to the rate of change of the network topology. When the rate of change of the network topology is high (e.g., the rate of change per hour exceeds 5%), the weight of degree centrality is increased to 0.5, the weight of betweenness centrality is 0.3, and the weight of proximity centrality is 0.2; when the network is relatively stable (e.g., the rate of change per hour is less than 1%), the weight of degree centrality is adjusted to 0.2, the weight of betweenness centrality is 0.5, and the weight of proximity centrality is 0.3. Through this dynamic weight fusion, the topological importance coefficient of the node is calculated.

[0086] The basic energy value of a node is calculated using a resource status weighted subtraction method, i.e., basic energy value = 100 - [(CPU utilization × 0.35) + (memory utilization × 0.3) + ((1 - remaining storage percentage) × 0.15) + (current load × 0.2)], where 0.35, 0.3, and 0.15 are set weights. For example, with a CPU utilization of 35%, a memory utilization of 42%, remaining storage of 68GB (total capacity of 100GB), and a current load of 30%, the calculation process is 100 - [12.25 + 12.6 + 4.8 + 6] = 78.35, which is rounded to 78.5.

[0087] The basic energy value, historical stability score, and topology importance coefficient are adaptively weighted and fused, with network stability determining the weight allocation: when the network is highly stable (e.g., jitter rate less than 0.5%), the basic energy value has a weight of 0.4, the historical stability score has a weight of 0.25, and the topology importance coefficient has a weight of 0.35; when the network fluctuates significantly (e.g., jitter rate greater than 2%), the basic energy value weight is adjusted to 0.3, the historical stability score weight increases to 0.45, and the topology importance coefficient weight remains at 0.25. For example, under stable network conditions, the final energy level value of a node is calculated as 78.5 × 0.4 + 82.3 × 0.25 + 75.6 × 0.35 = 78.61.

[0088] The data stream transmission path is determined based on the calculated energy level values, which are divided into five levels: 90-100 (Level A), 80-89 (Level B), 70-79 (Level C), 60-69 (Level D), and below 60 (Level E). The data stream preferentially selects nodes with higher energy levels as transmission paths, while also considering path hop count limitations. For example, for critical data streams, paths consisting of Level A nodes are preferred; if this is not possible, the shortest path involving Level B nodes is selected as a fallback.

[0089] Different data cache occupancy thresholds are set for nodes of different energy levels. For Level A nodes, the warning threshold is 75%, the critical threshold is 85%, and the emergency threshold is 95%; for Level B nodes, the corresponding thresholds are 70%, 80%, and 90%; for Level C nodes, they are 65%, 75%, and 85%; for Level D nodes, they are 60%, 70%, and 80%; and for Level E nodes, they are 55%, 65%, and 75%. When a node's cache occupancy rate reaches the warning threshold, an alert is issued and backup paths are prepared; when the critical threshold is reached, partial traffic migration is initiated; and when the emergency threshold is reached, a full traffic transfer mechanism is triggered to ensure the continuity and reliability of data transmission.

[0090] This invention constructs a more comprehensive node energy evaluation system through multi-dimensional historical performance index analysis and network topology location assessment. The adaptive weight fusion mechanism dynamically adjusts the weights of various factors based on network stability, making the energy level calculation more consistent with the actual network state. This multi-dimensional adaptive evaluation method significantly improves the accuracy and robustness of network routing decisions, reduces the negative impacts of network fluctuations and topology changes, and provides a more reliable basis for data flow transmission path selection.

[0091] In one optional implementation, when the node's data cache occupancy rate is detected to reach different level thresholds, the node's energy level value is increased accordingly, triggering different degrees of network collaborative adjustment. The step of recalculating the data flow transmission path based on the updated energy level value includes:

[0092] Based on the historical average and standard deviation of the node's cache utilization rate, a dynamic adjustment mechanism for different levels of thresholds is established. The dynamic adjustment mechanism dynamically updates the threshold adjustment coefficient based on the difference between the cache processing success rate and the preset target processing rate.

[0093] The cache growth trend is predicted based on historical time-series data sequences and attention weight matrices, and the cache pressure coefficient is calculated based on the prediction results.

[0094] When the cache utilization rate reaches different threshold levels, the node's energy level value is updated based on the cache pressure coefficient.

[0095] Based on the available resource ratio of nodes and network topology location information, select a collaborative node group and calculate the target load allocation ratio of each node in the collaborative node group.

[0096] The cost of the data stream transmission path is calculated based on the updated energy level value and the target load allocation ratio. The data stream transmission path is selected according to the cost value. The optimal transmission path is determined by comprehensively scoring the reliability, latency and stability of the data stream transmission path.

[0097] A gradual path switching mechanism is established to adaptively adjust the data flow allocation ratio of the old and new transmission paths according to the time process. The data flow allocation ratio is related to the comprehensive score.

[0098] Combination Figure 2 The flowchart illustrates the dynamic adjustment of network node energy levels and optimization of data flow transmission paths based on cache pressure: For example, historical cache occupancy data is collected for each node, and the historical mean and standard deviation are calculated. For instance, if a node's historical cache occupancy mean is 65% and the standard deviation is 8%, three threshold levels are set: Warning level (mean + 0.5 × standard deviation) is 69%, Critical level (mean + 1.5 × standard deviation) is 77%, and Emergency level (mean + 2.5 × standard deviation) is 85%. The data processing success rate of the nodes is monitored daily. If the current success rate is 98.2% and the preset target success rate is 99%, the difference is -0.8%. The threshold adjustment coefficient is dynamically updated accordingly. The adjustment coefficient is calculated as 1 + (difference × 10) = -7%. After application, the three threshold levels are adjusted to 64.17%, 71.61%, and 79.05%, respectively.

[0099] A time-series prediction method is used to analyze cache growth trends. Recent cache occupancy time-series data for nodes is collected, such as data sequences sampled every 10 minutes over the past 24 hours. An attention weight matrix is ​​used to assign different weights to data points at different time points, identifying data points with greater predictive value. For example, for a node's cache occupancy sequence, data points from the most recent 4 hours are given higher weights (e.g., 0.08), while earlier data points have lower weights (e.g., 0.02). Based on the weighted data sequences, the cache occupancy trend for the next 2 hours is predicted. When the prediction shows that the cache occupancy will grow at a rate of 5% per hour, and the current occupancy has reached 68%, a higher cache pressure coefficient is calculated, such as 0.78 (out of 1).

[0100] When a node's cache occupancy reaches a set threshold, the node's energy level is updated based on the cache pressure coefficient. For example, if a level C node (energy level 75) reaches the warning threshold of 65% cache occupancy and has a cache pressure coefficient of 0.78, the node's energy level is reduced to 68, i.e., downgraded from level C to level D. As the difference between the cache occupancy and the warning threshold widens, the reduction in energy level value increases non-linearly. Energy level value reduction = base reduction × (1 + α × (difference between occupancy and threshold)) 2 ), where α is a coefficient that adjusts the degree of nonlinearity. For example, if the base reduction is 5 and α is 0.4, when the threshold is exceeded by 5%, the reduction is 5×(1+0.4×5²)=5×(1+0.4×25)=5×11=55, the original energy level is 75, and after the update it is reduced to 75-13=62; when the threshold is exceeded by 10%, the reduction is 5×(1+0.4×10²)=5×(1+0.4×100)=5×41=205, and it is reduced to 75-20=55.

[0101] The available resource ratio is a comprehensive evaluation of indicators such as CPU idle rate, memory availability, and bandwidth margin. For example, nodes with a CPU idle rate greater than 50%, a memory availability greater than 40%, and a topological distance of no more than 2 hops from the target node are selected as candidate cooperative nodes. From the candidate nodes, nodes with higher topological importance coefficients are further selected to form a cooperative node group. For a cooperative node group of 5 nodes, the target load allocation ratio is calculated based on the available resource status of each node, such as node A allocating 25%, node B allocating 20%, node C allocating 20%, node D allocating 15%, and node E allocating 20%.

[0102] The cost of a data stream transmission path considers factors such as the energy level, link quality, and load level of each node on the path. For example, a path with three nodes has energy level values ​​of 85, 78, and 72, link quality scores of 92 and 88, and load levels of 30%, 45%, and 40%. The path cost is calculated by weighting the harmonic mean of the energy level values, the arithmetic mean of the link quality scores, and the arithmetic mean of the load levels of each node on the path with weights of 0.4, 0.3, and 0.3 respectively, and then subtracting the result from 100 to obtain a final value of 65 (the maximum score is 100, with lower scores indicating higher costs). The cost of multiple possible paths is compared, and the path with the lowest cost is selected as the candidate path. Further evaluation is then performed: reliability is calculated using the link's historical data packet transmission success rate and bit error rate; latency is assessed based on the ratio of end-to-end transmission latency to a preset baseline latency; and stability is comprehensively measured by the historical fluctuation amplitude and frequency of link performance parameters. If the reliability score is 85, the latency score is 90, and the stability score is 82, the overall score calculated using a 4:3:3 weighting is 85.4. By comparing the overall scores of different paths, the path with the highest overall score is determined as the optimal transmission path.

[0103] Establish a gradual path switching mechanism to ensure a smooth data migration. Define a switching time window, such as 30 minutes. Within this window, dynamically adjust the data flow allocation ratio between the old and new paths based on the progress of the switch. In the initial stage of the switch (e.g., the first 5 minutes), the old path still handles 90% of the data flow, while the new path handles only 10%. In the middle stage (e.g., after 15 minutes), the old and new paths each handle 50%. In the later stage (e.g., after 25 minutes), the new path handles more than 90% of the data flow. This data flow allocation ratio is related to the path's overall score; the new path with a higher overall score will achieve a faster switching rate. For example, when the new path's overall score is 85.4 points, while the old path's is 72.6 points, the switching process will be accelerated, potentially completing 90% of the data migration within 20 minutes.

[0104] This invention implements an intelligent network collaborative adjustment mechanism based on cache utilization, significantly improving the network's ability to cope with cache pressure. The dynamic threshold adjustment mechanism can personalize threshold settings according to node characteristics; time-series prediction supports early detection of cache pressure; the nonlinear energy level adjustment strategy ensures that the response matches the level of cache pressure; the collaborative node selection and load distribution mechanism achieves efficient utilization of network resources; and the progressive path switching mechanism ensures smooth data flow migration. This multi-level collaborative control significantly improves the stability and reliability of the network when facing cache pressure and reduces the risk of data loss.

[0105] In one optional implementation, a gradual path switching mechanism is established to adaptively adjust the data flow allocation ratio between the old and new transmission paths according to the time progress. The steps related to the data flow allocation ratio and the comprehensive score include:

[0106] Determine the sliding time window for path switching, calculate the optimal switching duration based on the total amount of data to be migrated, the current load level, and the maximum load threshold, and update the sliding time window according to the optimal switching duration;

[0107] The resource utilization efficiency of the new and old transmission paths is calculated. The resource utilization efficiency is obtained by weighted fusion of various resource utilization rates. The comprehensive score is then weighted and corrected based on the resource utilization efficiency to obtain the corrected comprehensive score.

[0108] The switching process value is calculated based on the time progress within the sliding time window. The switching process value increases non-linearly with the time progress within the sliding time window. The basic data flow allocation ratio of the old and new transmission paths is calculated based on the switching process value and the corrected comprehensive score.

[0109] Extract the business priority of the data stream, and calculate the actual allocation ratio by weighting the business priority with the basic allocation ratio of the data stream.

[0110] A performance evaluation matrix is ​​constructed based on the latency, packet loss, and bandwidth utilization metrics of the old and new transmission paths during the switching process. The deviation values ​​between the metrics in the performance evaluation matrix and the preset target metrics are calculated. The calculation parameters of the switching process value and the weighting coefficient of the comprehensive score are adjusted according to the deviation values ​​to dynamically optimize the path switching process.

[0111] For example, an initial sliding time window is determined based on the current network status data, typically set to a base window value, such as 30 minutes. Then, the total amount of data to be migrated is obtained, for example, 500MB. The current network load level is also detected, such as an average load of 65% for major nodes and a preset maximum load threshold of 85%. Based on these parameters, the optimal handover duration is calculated as follows: the total amount of data to be migrated divided by the available bandwidth (maximum load threshold minus current load level multiplied by link bandwidth), e.g., (500MB) / ((85%-65%)×10MB / s)=25 minutes. The calculated optimal handover duration is compared with the initial sliding time window. If the optimal handover duration is less than the initial window value, the initial window setting is maintained; if it is greater than the initial window value, the sliding time window is extended to the optimal handover duration to ensure a smooth data migration process.

[0112] Collect resource metrics such as CPU utilization, memory utilization, and link bandwidth utilization for each node on both the old and new paths. For example, on the old path, the CPU utilization of three nodes is 75%, 68%, and 70%, memory utilization is 65%, 60%, and 63%, and link bandwidth utilization is 78% and 72%. On the new path, the CPU utilization of three nodes is 45%, 50%, and 48%, memory utilization is 40%, 45%, and 42%, and link bandwidth utilization is 50% and 55%. Assign weights to each resource metric, such as a weight of 0.4 for CPU utilization, 0.3 for memory utilization, and 0.3 for bandwidth utilization. Calculate the resource utilization rate using a weighted average. The resource utilization rate of the old route is (75%×0.4+65%×0.3+78%×0.3)×0.5+(68%×0.4+60%×0.3+72%×0.3)×0.3+(70%×0.4+63%×0.3+72%×0.3)×0.2=70.1%; the resource utilization rate of the new route is 47.7%. Defining resource utilization efficiency as (100%-resource utilization rate), the resource utilization efficiency of the old route is 29.9%, and that of the new route is 52.3%. The resource utilization efficiency is then combined with the overall score to adjust the overall score. Assuming the overall score of the old route is 75 points and that of the new route is 85 points, the adjustment formula is the original overall score multiplied by (1+resource utilization efficiency×adjustment coefficient), with the adjustment coefficient set to 0.3. After the correction, the overall score of the old path is 75×(1+29.9%×0.3)=81.7 points, and the score of the new path is 85×(1+52.3%×0.3)=98.3 points.

[0113] The handover progress value describes the completion rate of the path handover within the time window, with an initial value of 0 and a value of 1 after a complete handover. The handover progress value increases non-linearly over time, exhibiting an S-shaped curve. The increase is slow in the initial stage of the handover, rapid in the middle stage, and slows down in the later stage. For example, in a 30-minute sliding time window, the handover progress value is 0.1 at the 5th minute, 0.5 at the 15th minute, and 0.9 at the 25th minute. The basic data flow allocation ratio between the old and new paths is calculated based on the handover progress value and the corrected comprehensive score. The data flow allocation ratio for the old path is (1 - handover progress value) × corrected comprehensive score of the old path / total score, and the allocation ratio for the new path is the handover progress value × corrected comprehensive score of the new path / total score. The total score is the sum of the corrected comprehensive scores of both paths. For example, at the 15th minute, the switching process value is 0.5, the old path data flow allocation ratio is (1-0.5)×81.7 / (81.7+98.3)=0.5×81.7 / 180=0.227, or 22.7%; the new path allocation ratio is 0.5×98.3 / 180=0.273, or 27.3%.

[0114] Business priorities are typically categorized into multiple levels, such as urgent (priority 1), high (priority 2), medium (priority 3), and low (priority 4). Different switching strategies are employed for data streams with different priorities. For urgent business data with priority 1, the tendency is to migrate it to a new path with better performance more quickly; while for lower priority data, it can remain on the old path for a longer period. Business priorities are converted into priority weights, such as a weight of 1.5 for priority 1, 1.2 for priority 2, 1.0 for priority 3, and 0.8 for priority 4. The business priority weights are combined with the basic allocation ratio of the data stream to calculate the actual allocation ratio. For example, for a business with priority 2, at the 15th minute, the actual allocation ratio of the old path is the basic ratio × (2 - priority weight) = 22.7% × (2 - 1.2) = 18.16%; the actual allocation ratio of the new path is the basic ratio × priority weight = 27.3% × 1.2 = 32.76%. After normalizing the two ratios, the old path is 18.16% / (18.16%+32.76%)=35.7%, and the new path is 64.3%.

[0115] During path switching, performance metrics of the old and new paths are continuously monitored, and a performance evaluation matrix is ​​constructed. This matrix includes latency, packet loss, and bandwidth utilization metrics at different time points. For example, 10 minutes after the switch begins, the old path has an average latency of 85ms, a packet loss rate of 1.2%, and a bandwidth utilization rate of 75%; the new path has an average latency of 65ms, a packet loss rate of 0.5%, and a bandwidth utilization rate of 60%. These actual metrics are compared with preset target metrics, and the deviation values ​​are calculated. Assuming a target latency of 75ms, a target packet loss rate of 0.8%, and a target bandwidth utilization of 70%, the latency deviation of the old path is (85-75) / 75 = 13.3%, the packet loss rate deviation is (1.2-0.8) / 0.8 = 50%, and the bandwidth utilization deviation is (75-70) / 70 = 7.1%. The latency deviation of the new path is (65-75) / 75 = -13.3%, the packet loss rate deviation is (0.5-0.8) / 0.8 = -37.5%, and the bandwidth utilization deviation is (60-70) / 70 = -14.3%. The calculation parameters for the handover process value and the weighting coefficients for the overall score are adjusted based on these deviation values. For example, if the latency deviation exceeds 10%, the growth rate of the handover process value is increased by 20%; if the packet loss rate deviation exceeds 30%, the weighting of the corresponding path's overall score is adjusted by 15%. This dynamic adjustment mechanism continuously optimizes the path switching process, ensuring that data flows smoothly and efficiently migrate from the old path to the new path.

[0116] The progressive path switching mechanism proposed in this invention enables smooth migration of data stream transmission, significantly improving the stability and reliability of network routing switching. Through techniques such as optimal switching duration calculation, resource utilization efficiency evaluation, non-linear switching process control, differentiated service priority processing, and real-time performance monitoring and feedback, the switching strategy can be adaptively adjusted according to the real-time network status, effectively avoiding data jitter and service interruption problems in traditional routing switching.

[0117] In one optional implementation, the network area load distribution status is calculated based on the working status parameters. When uneven load distribution is detected, the nodes on the data flow transmission path are differentiated and dynamically adjusted according to the energy level value of the current node. The steps include:

[0118] A load feature vector is constructed based on the node's working status parameters, and the load feature vector is weighted based on the load weight coefficient to obtain the node's comprehensive load value.

[0119] Calculate the standard deviation and mean of the comprehensive load value within the network area. When the ratio of the standard deviation to the mean exceeds a first preset threshold, and the deviation of the comprehensive load value of any node within the area from the mean exceeds a second preset threshold, it is determined that the load distribution is uneven. In response to the uneven load distribution, obtain the current energy level value of each node on the data stream transmission path, and calculate the energy weight coefficient based on the current energy level value.

[0120] Based on the processing priority of the calculation node according to the comprehensive load value and the energy weight coefficient, the data processing sequence is sorted according to the processing priority; based on the processing priority and the forwarding ratio of the calculation node according to the comprehensive load value, data traffic is allocated according to the forwarding ratio.

[0121] A scheduling effect evaluation index is constructed, which includes the ratio of regional load standard deviation to mean, node processing priority distribution, and energy weight coefficient. When the deviation between the calculated result of the scheduling effect evaluation index and the preset target threshold exceeds a preset range, the load weight coefficient is dynamically updated.

[0122] For example, the working status parameters of each node within the network area are obtained, including multi-dimensional indicators such as CPU utilization, memory utilization, cache utilization, number of currently processed tasks, and network interface load rate. For each node within the network area, a load feature vector is constructed, which contains the above-mentioned working status parameters. For example, for node A, its load feature vector is [75%, 68%, 62%, 156, 83%], indicating that the CPU utilization is 75%, the memory utilization is 68%, the cache utilization is 62%, the number of currently processed tasks is 156, and the network interface load rate is 83%. For node B, its load feature vector is [45%, 52%, 38%, 87, 56%]. For node C, its load feature vector is [35%, 30%, 25%, 65, 42%]. Different weight coefficients are assigned to different working status parameters to reflect the degree of influence of each parameter on the node load. For example, CPU utilization has a weight of 0.3, memory utilization has a weight of 0.25, cache utilization has a weight of 0.15, the number of currently processed tasks has a weight of 0.1, and network interface load has a weight of 0.2. Multiplying the load feature vector by the weight coefficients yields the node's overall load value. For node A, its overall load value is 75%×0.3+68%×0.25+62%×0.15+156 / 200×0.1+83%×0.2=71.48%, where the number of currently processed tasks is normalized by dividing by the preset maximum number of tasks, 200. Similarly, node B's overall load value is 47.36%, and node C's overall load value is 33.28%.

[0123] Calculate the mean and standard deviation of the overall load value for all nodes within the network area. For a network area containing nodes A, B, and C, the mean overall load value is (71.48% + 47.36% + 33.28%) / 3 = 50.71%. The standard deviation is calculated by taking the square root of the sum of the squares of the differences between the overall load value of each node and the mean, which is 19.43%. The ratio of the standard deviation to the mean is calculated as 19.43% / 50.71% = 0.383. A preset first threshold is 0.35. When the ratio of the standard deviation to the mean exceeds 0.35, it indicates a high degree of dispersion in the load distribution within the area. The deviation of each node's overall load value from the mean is also calculated. The deviation for node A is (71.48% - 50.71%) / 50.71% = 0.41, for node B it is (47.36% - 50.71%) / 50.71% = -0.066, and for node C it is (33.28% - 50.71%) / 50.71% = -0.344. A second threshold is preset to 0.4. When the deviation of any node's overall load value from the mean exceeds 0.4, it indicates that the node's load significantly deviates from the regional average. In this example, node A's deviation of 0.41 exceeds the second threshold of 0.4, and the ratio of its standard deviation to the mean of 0.383 exceeds the first threshold of 0.35. Therefore, the load distribution in this network region is determined to be uneven.

[0124] To address uneven load distribution, the current energy level of each node along the data flow transmission path is obtained. The energy level reflects the remaining resource status of a node and ranges from 0 to 100. The current energy level values ​​for nodes A, B, and C are read as 62, 80, and 88, respectively. Energy weighting coefficients are calculated based on these energy level values. The calculation method involves normalizing the energy level values ​​and applying a nonlinear transformation to enhance the difference in energy levels. For node A, the energy weighting coefficient is 62 / 100 = 0.62; for node B, it is 80 / 100 = 0.8; and for node C, it is 88 / 100 = 0.88. These energy weighting coefficients are applied to the load balancing strategy, prioritizing nodes with higher energy levels to bear more load, while considering the current load situation to avoid overload.

[0125] A higher processing priority value indicates a higher priority for the node in processing data. The calculation method is to subtract the overall load value from 1, multiply by the energy weighting coefficient, and then normalize. For node A, the processing priority is (1-71.48%)×0.62=0.177; for node B, it's (1-47.36%)×0.8=0.421; and for node C, it's (1-33.28%)×0.88=0.588. The data processing sequence is sorted according to the processing priority, with nodes having higher priority processing data first. In this example, the processing order is node C, node B, and node A. The forwarding weight of each node is calculated based on the processing priority and overall load value. The forwarding weight determines the proportion of data traffic allocated to each node. The calculation method is to divide the processing priority by the overall load value and then normalize. For node A, the forwarding weight is 0.177 / 71.48%=0.248; for node B, the forwarding weight is 0.421 / 47.36%=0.889; and for node C, the forwarding weight is 0.588 / 33.28%=1.767. After normalization, the forwarding weights of nodes A, B, and C are 0.248 / (0.248+0.889+1.767)=0.086, 0.889 / (0.248+0.889+1.767)=0.305, and 1.767 / (0.248+0.889+1.767)=0.609, respectively. Data traffic is allocated based on these forwarding weights. For example, for a total data stream of 1000Mbps, the traffic allocated to nodes A, B, and C is 86Mbps, 305Mbps, and 609Mbps, respectively.

[0126] The scheduling effectiveness evaluation index is used to assess the effectiveness of the load balancing strategy and make dynamic adjustments. This index includes the ratio of the regional load standard deviation to the mean, the distribution of node processing priorities, and the energy weight coefficient. The load distribution of the network region is recalculated after applying the load balancing strategy. Assuming that the combined load values ​​of nodes A, B, and C after adjustment are 65.2%, 56.8%, and 48.3%, respectively, the new mean is 56.77%, and the standard deviation is 8.46%. The ratio of the standard deviation to the mean is 8.46% / 56.77% = 0.149, lower than the previous 0.383, indicating a more balanced load distribution. The uniformity of processing priorities and the utilization efficiency of the energy weight coefficient are also calculated. These indicators are compared with preset target thresholds, such as a target of the ratio of standard deviation to mean being less than 0.2 and the maximum deviation of node load from the mean being less than 0.3. If the calculated results of the scheduling effectiveness evaluation index deviate from the preset target by a certain range, the load weight coefficient will be dynamically updated. For example, if the adjusted standard deviation to mean ratio is 0.25, exceeding the target threshold of 0.2, the load weighting coefficients will be adjusted, increasing the weight of parameters causing imbalance, such as increasing the weight of CPU utilization from 0.3 to 0.35 and the weight of network interface load rate from 0.2 to 0.25, while correspondingly decreasing the weight of other parameters. Through this feedback adjustment mechanism, the load balancing strategy can be continuously optimized, resulting in a more balanced load distribution within the network area.

[0127] This invention achieves dynamic load balancing in network areas by constructing a multi-dimensional load feature vector and its weighting mechanism, combined with a differentiated scheduling strategy based on energy level values. It can accurately identify uneven load distribution and intelligently allocate data streams based on node processing priority and forwarding ratio, effectively avoiding resource waste and performance bottlenecks.

[0128] In one optional implementation, a scheduling performance evaluation index is constructed, comprising the ratio of the regional load standard deviation to the mean, the node processing priority distribution, and the energy weight coefficient. When the calculated result of the scheduling performance evaluation index deviates from a preset target threshold by more than a preset range, the step of dynamically updating the load weight coefficient includes:

[0129] Construct a multi-scale time window, calculate the ratio of regional load standard deviation to mean in each time window, and weight the load balance assessment value by assigning different weight coefficients based on the time scale.

[0130] A priority distribution rationality score is calculated based on the dispersion of the processing priorities of the nodes. The priority distribution rationality score is obtained by calculating the entropy value of the processing priorities.

[0131] The network energy efficiency index is calculated based on the energy weighting coefficient. The network energy efficiency index reflects the data processing efficiency per unit energy consumption. The ratio of the network energy efficiency index to the historical best value is used as the energy utilization evaluation value.

[0132] The load balancing evaluation value, the priority distribution rationality score, and the energy utilization evaluation value are adaptively weighted and fused to obtain the scheduling effect evaluation index. The weight coefficients of the adaptive weighted fusion are dynamically adjusted according to the network performance requirements.

[0133] When the deviation between the scheduling effect evaluation index and the preset target threshold exceeds a preset range, the contribution of each dimension of the load feature vector is calculated based on the deviation, and the load weight coefficient is updated according to the contribution.

[0134] For example, three time windows of different scales are created: a short-term window (5 minutes), a medium-term window (30 minutes), and a long-term window (2 hours). Within each time window, network area load data is collected, and the ratio of the area load standard deviation to the mean is calculated. In the short-term window, data is collected every 10 seconds, resulting in an area load standard deviation of 12.5% ​​and a mean of 58.3%, with a ratio of 0.214. In the medium-term window, data is collected every minute, resulting in an area load standard deviation of 9.8% and a mean of 54.7%, with a ratio of 0.179. In the long-term window, data is collected every 5 minutes, resulting in an area load standard deviation of 7.2% and a mean of 52.4%, with a ratio of 0.137. Different weighting coefficients are assigned to different time windows: 0.5 for the short-term window, 0.3 for the medium-term window, and 0.2 for the long-term window, reflecting the importance of recent load conditions to current scheduling decisions. Multiplying the ratios of each time window by their corresponding weights and summing them, we get the load balancing assessment value as 0.214×0.5 + 0.179×0.3 + 0.137×0.2 = 0.186. The lower this assessment value, the more balanced the load distribution across network areas.

[0135] The rationality of the node processing priority distribution is evaluated by calculating the entropy value of the priorities. The processing priorities of all nodes within the network area are obtained, such as node A being 0.177, node B being 0.421, and node C being 0.588. These priority values ​​are normalized, resulting in normalized priorities of 0.149, 0.355, and 0.496 for nodes A, B, and C, respectively, with a sum of 1. The entropy value of the processing priorities is calculated by multiplying the normalized priority of each node by its natural logarithm, taking the negative value, and then summing the results. For node A, the calculated value is 0.149 × negative natural logarithm 0.149; for node B, it is 0.355 × negative natural logarithm 0.355; and for node C, it is 0.496 × negative natural logarithm 0.496. The sum of these three values ​​yields an entropy value of 1.01. The calculated entropy value is compared with the maximum entropy value of an ideal uniform distribution to obtain a priority distribution rationality score. For a network with 3 nodes, the maximum entropy is the natural logarithm 3 = 1.1, therefore the priority distribution rationality score is 1.01 / 1.1 = 0.918. The higher this score, the more rational the processing priority distribution and the higher the resource utilization efficiency.

[0136] Network energy efficiency index reflects the data processing efficiency per unit of energy consumption. The network energy efficiency index is calculated based on energy weighting coefficients. The energy weighting coefficients for each node are obtained, such as 0.62 for node A, 0.8 for node B, and 0.88 for node C. The data processing volume of each node is also obtained, such as 86 Mbps for node A, 305 Mbps for node B, and 609 Mbps for node C. The energy efficiency of each node is calculated by dividing the data processing volume by the reciprocal of the energy weighting coefficient, representing the data processing efficiency per unit of energy consumption. For node A, the energy efficiency is 86 / (1 / 0.62) = 53.3; for node B, it is 305 / (1 / 0.8) = 244; and for node C, it is 609 / (1 / 0.88) = 535.9. The energy efficiency index of the entire network is calculated by summing the energy efficiencies of all nodes and dividing by the number of nodes, resulting in 277.7. Comparing the current network energy efficiency index with the historical best value, assuming the historical best value is 320, the energy utilization assessment value is 277.7 / 320 = 0.868. The closer this assessment value is to 1, the closer the network energy utilization efficiency is to the historical best level.

[0137] The load balancing assessment value, priority distribution rationality score, and energy utilization assessment value are adaptively weighted and fused, with the weight coefficients dynamically adjusted according to network performance requirements. If the network is currently facing a high load challenge, the weight of the load balancing assessment value will increase; if energy is limited, the weight of the energy utilization assessment value will increase; if service quality requirements are high, the weight of the priority distribution rationality score will increase. Assuming the current network performance requirement weights are 0.4, 0.35, and 0.25 respectively, the scheduling effect assessment index is 0.186×0.4 + 0.918×0.35 + 0.868×0.25 = 0.604. Comparing this assessment index with a preset target threshold, assuming the target threshold is 0.75, the deviation is 0.75 - 0.604 = 0.146. The preset acceptable deviation range is 0.1; the current deviation of 0.146 exceeds the preset range, requiring dynamic updates to the load weight coefficients.

[0138] The contribution of each dimension of the load feature vector is calculated based on the deviation. The current load balancing status is analyzed to identify the main factors causing imbalance. Each dimension of the load feature vector (CPU utilization, memory utilization, cache utilization, number of current tasks, and network interface load rate) is examined to analyze its correlation with load imbalance. The correlation coefficient between each dimension's feature value and the overall load value is calculated for each node, and the contribution of each dimension is obtained by combining all nodes across the network. For example, the contribution of CPU utilization is 0.42, memory utilization is 0.28, cache utilization is 0.15, the number of current tasks is 0.05, and network interface load rate is 0.1. A higher contribution indicates a greater impact of that dimension on load imbalance. The load weight coefficients are updated based on the contribution, increasing the weight of dimensions with high contribution and decreasing the weight of dimensions with low contribution. The updated load weighting coefficients are as follows: CPU utilization weight adjusted from 0.3 to 0.38, memory utilization weight adjusted from 0.25 to 0.28, cache utilization weight adjusted from 0.15 to 0.15 (remaining unchanged), number of currently processed tasks weight adjusted from 0.1 to 0.05, and network interface load weight adjusted from 0.2 to 0.14. The weight adjustments are proportional to the contribution, but the total remains at 1.

[0139] The updated load weighting coefficients are applied to recalculate the overall load value for each node. For node A, the new overall load value is 75%×0.38+68%×0.28+62%×0.15+156 / 200×0.05+83%×0.14=72.49%; for node B, the new overall load value is 45%×0.38+52%×0.28+38%×0.15+87 / 200×0.05+56%×0.14=47.35%; for node C, the new overall load value is 35%×0.38+30%×0.28+25%×0.15+65 / 200×0.05+42%×0.14=32.97%. After the update, the regional load average is (72.49% + 47.35% + 32.97%) / 3 = 50.94%, with a standard deviation of 19.75%. The ratio of the standard deviation to the mean is 0.388. Although this ratio has increased slightly, adjusting the weighting coefficients highlights the monitoring of CPU utilization, a key factor, which helps subsequent scheduling decisions to more accurately identify and handle load imbalance issues. The load balancing process continues, recalculating processing priorities and forwarding ratios, and adjusting data flow allocation until the scheduling performance evaluation indicators reach the expected targets or the deviation is within acceptable limits.

[0140] The scheduling effect evaluation mechanism of this invention enables adaptive optimization of network load balancing strategies. Through the integrated application of multi-scale time window analysis, priority entropy calculation, and energy efficiency assessment, the effectiveness of scheduling strategies can be comprehensively evaluated. In particular, the dynamic update mechanism of load weight coefficients based on contribution allows for accurate identification of the main factors causing load imbalance and targeted adjustment of strategies, significantly improving network resource utilization efficiency and service quality while reducing energy consumption.

[0141] In one optional implementation, the step of calculating the routing suitability of a node based on its energy level and regional load distribution, and selecting the node with the lowest communication power consumption that meets preset conditions to form a data forwarding channel includes:

[0142] A candidate route set is constructed based on the communication range of the nodes and the network topology. The hop count between each node in the candidate route set and the target node is calculated. The link quality from each node to the target node is determined based on the hop count. The link quality is obtained by weighting the signal strength, bit error rate and packet loss rate.

[0143] The remaining lifetime of a node is predicted based on its energy level value. The remaining lifetime prediction value decreases exponentially as the energy level value decreases. The maximum load carrying capacity of the node is then set based on the remaining lifetime prediction value.

[0144] The load balancing index in the regional load distribution status is combined with the maximum load carrying capacity of the node to calculate the load capacity score of the node.

[0145] Based on the link quality, the remaining lifetime prediction value, and the load capacity score, the routing suitability of the node is calculated, and the communication power consumption per unit data transmission is calculated for nodes whose routing suitability meets the preset conditions.

[0146] The node sequence with the highest routing suitability and lowest communication power consumption is selected to construct a data forwarding channel. The reliability index of the data forwarding channel is calculated. When the reliability index is lower than a preset reliability threshold, the node sequence with the second-best routing suitability is selected to construct a backup forwarding channel.

[0147] For example, the location information and communication range parameters of all nodes in the network are obtained, such as node A having a communication range of 100 meters, node B having a communication range of 120 meters, node C having a communication range of 80 meters, node D having a communication range of 150 meters, and node E being the target node. A network topology diagram is constructed by analyzing the relative positions and overlapping communication ranges between nodes. In the topology diagram, if the distance between two nodes is less than or equal to the sum of their communication ranges, then a direct communication link can be established between these two nodes. For example, the distance between node A and node B is 180 meters, which is less than the sum of their communication ranges of 220 meters, so a direct communication link can be established between A and B; however, the distance between node A and node D is 300 meters, which is greater than the sum of their communication ranges of 250 meters, so a direct communication link cannot be established between A and D. Based on this principle, a complete network connection diagram is constructed, and the minimum hop count from each node to the target node E is calculated. Using the breadth-first search algorithm, the minimum hop count from node A to E is determined to be 3, the minimum hop count from node B to E is 2, the minimum hop count from node C to E is 2, and the minimum hop count from node D to E is 1.

[0148] The link quality, determined by the hop count, is calculated using a weighted average of signal strength, bit error rate (BER), and packet loss rate. Data on signal strength, BER, and packet loss rate of the communication links between each node are collected. For example, the average signal strength between node A and its neighboring nodes is -65 dBm, with a BER of 0.002 and a packet loss rate of 0.015; the average signal strength between node B and its neighboring nodes is -58 dBm, with a BER of 0.001 and a packet loss rate of 0.008; the average signal strength between node C and its neighboring nodes is -70 dBm, with a BER of 0.003 and a packet loss rate of 0.02; and the average signal strength between node D and its neighboring nodes is -50 dBm, with a BER of 0.0005 and a packet loss rate of 0.005. The signal strength is normalized to the range of 0-1, such as -50 dBm corresponding to 1 and -100 dBm corresponding to 0. Therefore, the normalized signal strengths of nodes A, B, C, and D are 0.7, 0.84, 0.6, and 1, respectively. Weights are assigned to these three indicators: signal strength (0.4), bit error rate (0.3), and packet loss rate (0.3). The link quality score for each node is calculated as follows: for node A, the link quality score is 0.7×0.4+(1-0.002)×0.3+(1-0.015)×0.3=0.85; for node B, the link quality score is 0.84×0.4+(1-0.001)×0.3+(1-0.008)×0.3=0.91; for node C, the link quality score is 0.6×0.4+(1-0.003)×0.3+(1-0.02)×0.3=0.81; and for node D, the link quality score is 1×0.4+(1-0.0005)×0.3+(1-0.005)×0.3=0.96. Considering the impact of hop count on link quality, a higher hop count generally results in lower link quality. A hop count attenuation factor is calculated, which is the reciprocal of the hop count. Specifically, the hop count attenuation factors for nodes A, B, C, and D are 1 / 3, 1 / 2, 1 / 2, and 1, respectively. Multiplying the link quality score by the hop count attenuation factor yields the overall link quality. The overall link quality scores for nodes A, B, C, and D are 0.85 × 1 / 3 = 0.283, 0.91 × 1 / 2 = 0.455, 0.81 × 1 / 2 = 0.405, and 0.96 × 1 = 0.96, respectively.

[0149] Obtain the current energy level value of each node, such as 72 for node A, 85 for node B, 60 for node C, and 45 for node D. The maximum energy level value is 100, indicating the node's optimal resource status. Calculate the predicted remaining lifetime of each node based on its energy level value. The predicted value has an exponential relationship with the energy level value; the lower the energy level value, the faster the predicted remaining lifetime decays. The calculation method is to multiply the base lifetime by an exponential decay factor, which is the difference between the energy level value and the baseline value (the natural constant e) divided by the adjustment coefficient raised to the power of the result. Assuming a base lifetime of 100 hours, a baseline value of 80, and an adjustment coefficient of 20, then the predicted remaining lifetime of node A is 100 × (72-80) / 20 = 100 × e -0.4 = 100 × 0.67 = 67 hours; the predicted remaining lifetime of node B is 100 × (85-80) / 20 = 100 × e 0.25 = 10 hours. 0 × 1.28 = 128 hours; the predicted remaining lifetime of node C is 100 × e^(60-80) / 20 = 100 × e^(-1) = 100 × 0.368 = 36.8 hours; the predicted remaining lifetime of node D is 100 × e^(45-80) / 20 = 100 × e^(-1.75) = 100 × 0.174 = 17.4 hours. The maximum load capacity of the nodes is set based on the predicted remaining lifetime values, calculated by normalizing the predicted remaining lifetime values ​​and multiplying them by the full load value. If the full load is 100Mbps, then the maximum load capacity of nodes A, B, C, and D are 67 / 128×100=52.3Mbps, 128 / 128×100=100Mbps, 36.8 / 128×100=28.8Mbps, and 17.4 / 128×100=13.6Mbps, respectively.

[0150] Obtain the regional load distribution status, including load balancing metrics such as the ratio of the load standard deviation to the mean. Assuming the current regional load balancing metric is 0.25, it indicates an uneven load distribution. Adjust the load allocation strategy based on the load balancing metric; when the load distribution is uneven, it tends to allocate more load to nodes with lower loads. Obtain the current load level of each node, such as node A at 30Mbps, node B at 45Mbps, node C at 20Mbps, and node D at 8Mbps. Calculate the remaining load capacity of each node, i.e., the maximum load capacity minus the current load level. The remaining load capacities of nodes A, B, C, and D are 52.3-30=22.3Mbps, 100-45=55Mbps, 28.8-20=8.8Mbps, and 13.6-8=5.6Mbps, respectively. Calculate the node's load capacity score by dividing the remaining load capacity by the maximum load capacity and then multiplying by the balance factor. The balance factor is related to the regional load balancing metric; when the load distribution is uneven, a larger balance factor tends to favor nodes with larger remaining capacity. Assuming the balance factor is 1 + 0.25 × 2 = 1.5, the load capacity scores for nodes A, B, C, and D are 22.3 / 52.3 × 1.5 = 0.64, 55 / 100 × 1.5 = 0.825, 8.8 / 28.8 × 1.5 = 0.458, and 5.6 / 13.6 × 1.5 = 0.618, respectively.

[0151] Route suitability is a comprehensive score that reflects how well a node is suited to be a routing node. Weights are assigned to the scores for link quality, predicted remaining lifetime, and load capacity, such as link quality with a weight of 0.4, predicted remaining lifetime with a weight of 0.35, and load capacity with a weight of 0.25. Calculate the routing suitability of each node. For node A, the routing suitability is 0.283×0.4+67 / 128×0.35+0.64×0.25=0.49; for node B, it is 0.455×0.4+128 / 128×0.35+0.825×0.25=0.69; for node C, it is 0.405×0.4+36.8 / 128×0.35+0.458×0.25=0.36; and for node D, it is 0.96×0.4+17.4 / 128×0.35+0.618×0.25=0.59. Set preset conditions for routing suitability, such as a minimum suitability threshold of 0.4. Therefore, nodes A, B, and D meet the conditions, while node C does not. For nodes that meet the conditions, calculate the communication power consumption per unit of data transmission. Obtain the transmit and receive power parameters of each node. For example, node A has a transmit power of 20mW and a receive power of 10mW; node B has a transmit power of 25mW and a receive power of 12mW; and node D has a transmit power of 30mW and a receive power of 15mW. Calculate the communication power consumption per unit data transmission, which is the sum of the transmit and receive power divided by the link data rate. Assuming the link data rates of nodes A, B, and D are 10Mbps, 12Mbps, and 15Mbps respectively, the power consumption per unit data transmission is (20+10) / 10 = 3mW / Mbps, (25+12) / 12 = 3.08mW / Mbps, and (30+15) / 15 = 3mW / Mbps respectively.

[0152] The data forwarding channel is constructed by selecting the node sequence with the highest routing suitability and lowest communication power consumption. Among nodes A, B, and D that meet the preset conditions, node B has the highest routing suitability at 0.69, while nodes A and D have the same and lower communication power consumption than node B. Considering both routing suitability and communication power consumption, suitability is prioritized, followed by power consumption. Therefore, node B is selected as the primary forwarding node to construct the data forwarding channel. The reliability index of this data forwarding channel is calculated, based on a comprehensive evaluation of link quality, node energy status, and load stability. Assuming the calculated reliability index is 0.82, which is lower than the preset reliability threshold of 0.85, a node sequence with the second-best routing suitability needs to be selected to construct a backup forwarding channel. Among the remaining nodes, node D has a routing suitability of 0.59, which is lower than node B; therefore, node D is selected to construct the backup forwarding channel. The status of the primary and backup forwarding channels is continuously monitored. When the primary channel experiences problems or performance degradation, the system automatically switches to the backup channel to ensure the reliability and stability of data transmission.

[0153] The system collects operational status data from each network node every 30 seconds, including CPU utilization, memory usage, data queue length, network throughput, and energy consumption. Preset ranges for monitoring parameters are: CPU utilization <85%, memory usage <80%, data queue length <200, network throughput fluctuation <20%, and energy consumption <25mW / s. If any parameter exceeds the preset range, such as CPU utilization reaching 87% or data queue length increasing to 220, the node's operational status parameters are immediately updated, triggering a recalculation of the energy level value and executing subsequent routing adjustments.

[0154] The routing method proposed in this invention, based on node energy level values ​​and regional load distribution, achieves intelligent optimization of data forwarding paths in communication networks. By comprehensively considering link quality, remaining node lifetime, and load capacity, it can select the most suitable nodes to form an efficient and stable data forwarding channel; the backup forwarding channel mechanism significantly improves network reliability and fault tolerance, effectively solving the problems of uneven energy consumption and network congestion in traditional routing methods.

[0155] A second aspect of the present invention provides a low-power wireless ad hoc network dynamic configuration system based on adaptive routing, comprising:

[0156] The first unit is used to obtain the working status parameters of each node in the wireless ad hoc network;

[0157] The second unit is used to calculate the energy level value of the node according to the working status parameters, determine the data flow transmission path based on the energy level value, and set multi-level data cache occupancy thresholds; when the node data cache occupancy rate is detected to reach different level thresholds, the energy level value of the node is increased accordingly, triggering different degrees of network collaborative adjustment, and the data flow transmission path is recalculated based on the updated energy level value.

[0158] The third unit is used to calculate the network area load distribution status based on the working status parameters. When uneven load distribution is detected, it performs differentiated scheduling of nodes on the data flow transmission path in combination with the energy level value of the current node, and dynamically adjusts the data processing order and forwarding ratio.

[0159] The fourth unit is used to calculate the routing suitability of nodes based on their energy level and regional load distribution, and select nodes with routing suitability that meet preset conditions and have the lowest communication power consumption to form a data forwarding channel.

[0160] The fifth unit is used to monitor the network operation status in real time. When the monitored parameters exceed the preset range, it updates the working status parameters of the nodes and returns the energy level value of the computing nodes.

[0161] A third aspect of the present invention provides an electronic device, comprising:

[0162] processor;

[0163] Memory used to store processor-executable instructions;

[0164] The processor is configured to invoke instructions stored in the memory to execute the aforementioned method.

[0165] A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the aforementioned method.

[0166] This invention can be a method, apparatus, system, and / or computer program product. The computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for performing various aspects of the invention.

[0167] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention 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 or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A dynamic configuration method for low-power wireless ad hoc networks based on adaptive routing, characterized in that, include: Obtain the working status parameters of each node in the wireless ad hoc network; Calculate the node's energy level value based on the working status parameters, determine the data stream transmission path based on the energy level value, and set multi-level data cache occupancy thresholds; When the node data cache occupancy rate is detected to reach different level thresholds, the energy level value of the node is adjusted accordingly, triggering different degrees of network collaborative adjustment, and the data flow transmission path is recalculated based on the updated energy level value. The network area load distribution status is calculated based on the working status parameters. When uneven load distribution is detected, the nodes on the data flow transmission path are differentiated and scheduled in combination with the energy level value of the current node, and the data processing order and forwarding ratio are dynamically adjusted. The routing suitability of a node is calculated based on its energy level and regional load distribution. Nodes with the lowest communication power consumption that meet the preset conditions are selected to form a data forwarding channel. The system monitors the network's operational status in real time. When the monitored parameters exceed the preset range, it updates the node's operational status parameters and returns the energy level value of the computing node. The step of adjusting the energy level of a node when its data cache occupancy rate reaches different threshold levels, triggering different degrees of network collaborative adjustment, and recalculating the data flow transmission path based on the updated energy level value includes: Based on the historical average and standard deviation of the node's cache utilization rate, a dynamic adjustment mechanism for different levels of thresholds is established. The dynamic adjustment mechanism dynamically updates the threshold adjustment coefficient based on the difference between the cache processing success rate and the preset target processing rate. The cache growth trend is predicted based on historical time-series data sequences and attention weight matrices, and the cache pressure coefficient is calculated based on the prediction results. When the cache utilization rate reaches different threshold levels, the node's energy level value is updated based on the cache pressure coefficient. Based on the available resource ratio of nodes and network topology location information, select a collaborative node group and calculate the target load allocation ratio of each node in the collaborative node group. The cost of the data stream transmission path is calculated based on the updated energy level value and the target load allocation ratio. The data stream transmission path is selected according to the cost value. The optimal transmission path is determined by comprehensively scoring the reliability, latency and stability of the data stream transmission path. A gradual path switching mechanism is established to adaptively adjust the data flow allocation ratio of the old and new transmission paths according to the time process. The data flow allocation ratio is related to the comprehensive score.

2. The method according to claim 1, characterized in that, The steps of calculating the node's energy level value based on the operating status parameters, determining the data stream transmission path based on the energy level value, and setting multi-level data buffer occupancy thresholds include: The working status parameters include the node's historical performance indicators and network topology location information; Based on the data transmission success rate, performance fluctuation and data processing capability of the historical performance indicators in different time windows, a historical stability score is calculated. The historical stability score is obtained by weighting different time windows by assigning different weight coefficients to them. Based on the network topology location information, the degree centrality, betweenness centrality, and proximity centrality of nodes are calculated, and a weighted fusion is performed to obtain the topological importance coefficient. The weights of the weighted fusion are dynamically adjusted according to the network topology change rate. The basic energy value is calculated based on the resource status information of the node. The basic energy value is then adaptively weighted and fused with the historical stability score and the topology importance coefficient to obtain the energy level value. The weight coefficient of the adaptive weight fusion is dynamically adjusted according to the network stability index. The data stream transmission path is determined based on the energy level value, and warning thresholds, critical thresholds, and emergency thresholds for data cache occupancy are set.

3. The method according to claim 1, characterized in that, A gradual path switching mechanism is established to adaptively adjust the data flow allocation ratio between the old and new transmission paths according to the time progress. The steps related to the data flow allocation ratio and the comprehensive score include: Determine the sliding time window for path switching, calculate the optimal switching duration based on the total amount of data to be migrated, the current load level, and the maximum load threshold, and update the sliding time window according to the optimal switching duration; The resource utilization efficiency of the new and old transmission paths is calculated. The resource utilization efficiency is obtained by weighted fusion of various resource utilization rates. The comprehensive score is then weighted and corrected based on the resource utilization efficiency to obtain the corrected comprehensive score. The switching process value is calculated based on the time progress within the sliding time window. The switching process value increases non-linearly with the time progress within the sliding time window. The basic data flow allocation ratio of the old and new transmission paths is calculated based on the switching process value and the corrected comprehensive score. Extract the business priority of the data stream, and calculate the actual allocation ratio by weighting the business priority with the basic allocation ratio of the data stream. A performance evaluation matrix is ​​constructed based on the latency, packet loss, and bandwidth utilization metrics of the old and new transmission paths during the switching process. The deviation values ​​between the metrics in the performance evaluation matrix and the preset target metrics are calculated. The calculation parameters of the switching process value and the weighting coefficient of the comprehensive score are adjusted according to the deviation values ​​to dynamically optimize the path switching process.

4. The method according to claim 1, characterized in that, Based on the aforementioned working status parameters, the network area load distribution status is calculated. When uneven load distribution is detected, the nodes on the data flow transmission path are differentiated and scheduled according to their energy level values, dynamically adjusting the data processing order and forwarding ratio. The steps include: A load feature vector is constructed based on the node's working status parameters, and the load feature vector is weighted based on the load weight coefficient to obtain the node's comprehensive load value. Calculate the standard deviation and mean of the comprehensive load value within the network area. When the ratio of the standard deviation to the mean exceeds a first preset threshold, and the deviation of the comprehensive load value of any node within the area from the mean exceeds a second preset threshold, it is determined that the load distribution is uneven. In response to the uneven load distribution, the current energy level value of each node on the data stream transmission path is obtained, and an energy weighting coefficient is calculated based on the current energy level value. Based on the comprehensive load value and the processing priority of the energy weight coefficient calculation node, the data processing sequence is sorted according to the processing priority; The forwarding ratio of a node is calculated based on the processing priority and the overall load value, and data traffic is allocated based on the forwarding ratio. A scheduling effect evaluation index is constructed, which includes the ratio of regional load standard deviation to mean, node processing priority distribution, and energy weight coefficient. When the deviation between the calculated result of the scheduling effect evaluation index and the preset target threshold exceeds a preset range, the load weight coefficient is dynamically updated.

5. The method according to claim 4, characterized in that, The process of constructing a scheduling performance evaluation index that includes the ratio of regional load standard deviation to mean, node processing priority distribution, and energy weight coefficient, and dynamically updating the load weight coefficient when the calculated result of the scheduling performance evaluation index deviates from a preset target threshold by more than a preset range, includes the following steps: Construct a multi-scale time window, calculate the ratio of regional load standard deviation to mean in each time window, and weight the load balance assessment value by assigning different weight coefficients based on the time scale. A priority distribution rationality score is calculated based on the dispersion of the processing priorities of the nodes. The priority distribution rationality score is obtained by calculating the entropy value of the processing priorities. The network energy efficiency index is calculated based on the energy weighting coefficient. The network energy efficiency index reflects the data processing efficiency per unit energy consumption. The ratio of the network energy efficiency index to the historical best value is used as the energy utilization evaluation value. The load balancing evaluation value, the priority distribution rationality score, and the energy utilization evaluation value are adaptively weighted and fused to obtain the scheduling effect evaluation index. The weight coefficients of the adaptive weighted fusion are dynamically adjusted according to the network performance requirements. When the deviation between the scheduling effect evaluation index and the preset target threshold exceeds a preset range, the contribution of each dimension of the load feature vector is calculated based on the deviation, and the load weight coefficient is updated according to the contribution.

6. The method according to claim 1, characterized in that, The steps for calculating the routing suitability of nodes based on their energy level and regional load distribution, and selecting nodes with the lowest communication power consumption that meet preset conditions to form a data forwarding channel, include: A candidate route set is constructed based on the communication range of the nodes and the network topology. The hop count between each node in the candidate route set and the target node is calculated. The link quality from each node to the target node is determined based on the hop count. The link quality is obtained by weighting the signal strength, bit error rate and packet loss rate. The remaining lifetime of a node is predicted based on its energy level value. The remaining lifetime prediction value decreases exponentially as the energy level value decreases. The maximum load carrying capacity of the node is then set based on the remaining lifetime prediction value. The load balancing index in the regional load distribution status is combined with the maximum load carrying capacity of the node to calculate the load capacity score of the node. Based on the link quality, the remaining lifetime prediction value, and the load capacity score, the routing suitability of the node is calculated, and the communication power consumption per unit data transmission is calculated for nodes whose routing suitability meets the preset conditions. The node sequence with the highest routing suitability and lowest communication power consumption is selected to construct a data forwarding channel. The reliability index of the data forwarding channel is calculated. When the reliability index is lower than a preset reliability threshold, the node sequence with the second-best routing suitability is selected to construct a backup forwarding channel.

7. A low-power wireless ad hoc network dynamic configuration system based on adaptive routing, used to implement the method described in claim 1 above, characterized in that, include: The first unit is used to obtain the working status parameters of each node in the wireless ad hoc network; The second unit is used to calculate the energy level value of the node based on the working status parameters, determine the data stream transmission path based on the energy level value, and set multi-level data cache occupancy thresholds. When a node's data cache occupancy rate is detected to reach different threshold levels, the node's energy level value is adjusted accordingly, triggering different degrees of network collaborative adjustment. The data flow transmission path is recalculated based on the updated energy level value, including: establishing a dynamic adjustment mechanism for different threshold levels based on the historical mean and standard deviation of the node's cache occupancy rate; dynamically updating the threshold adjustment coefficient based on the difference between the cache processing success rate and a preset target processing rate; predicting the cache growth trend based on historical time-series data sequences and attention weight matrices, and calculating the cache pressure coefficient based on the prediction results; when the cache occupancy rate reaches different threshold levels, based on... The cache pressure coefficient updates the energy level value of the node; a cooperative node group is selected based on the available resource ratio of the node and network topology location information, and the target load allocation ratio of each node in the cooperative node group is calculated; the cost value of the data flow transmission path is calculated based on the updated energy level value and the target load allocation ratio, and a data flow transmission path is selected according to the cost value; the optimal transmission path is determined by comprehensively scoring the reliability, latency and stability of the data flow transmission path; a gradual path switching mechanism is established to adaptively adjust the data flow allocation ratio of the old and new transmission paths according to the time process, and the data flow allocation ratio is related to the comprehensive score; The third unit is used to calculate the network area load distribution status based on the working status parameters. When uneven load distribution is detected, it performs differentiated scheduling of nodes on the data flow transmission path in combination with the energy level value of the current node, and dynamically adjusts the data processing order and forwarding ratio. The fourth unit is used to calculate the routing suitability of nodes based on their energy level and regional load distribution, and select nodes with routing suitability that meet preset conditions and have the lowest communication power consumption to form a data forwarding channel. The fifth unit is used to monitor the network operation status in real time. When the monitored parameters exceed the preset range, it updates the working status parameters of the nodes and returns the energy level value of the computing nodes.

8. An electronic device, characterized in that, include: processor; Memory used to store processor-executable instructions; The processor is configured to invoke instructions stored in the memory to execute the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having computer program instructions stored thereon, characterized in that, When the computer program instructions are executed by the processor, they implement the method described in any one of claims 1 to 6.

Images