Wireless communication network method and system of mobile first-aid station

Abstract

The invention provides a wireless communication network method and system of a mobile first-aid station. The method comprises the following steps: monitoring and evaluating a surrounding environment in real time, selecting an optimal working frequency band, and adjusting a frequency band switching strategy to obtain a self-adaptive multi-frequency-band access mechanism; the bandwidth and the connection priority of the wireless communication network are dynamically adjusted, the task priority is evaluated, and a dynamic resource allocation strategy is generated; deploying a distributed cache and synchronization system, realizing balanced distribution and rapid positioning of data, and generating a distributed temporary data cache pool; establishing an intelligent routing algorithm of deep reinforcement learning, calculating an optimal transmission path from a source to a destination, and finally generating an intelligent routing planning scheme; a redundant path is preset on the path, and a redundant path planning scheme is generated according to a dynamic resource allocation strategy and an intelligent routing planning result. According to the technical scheme provided by the invention, the reliability and the response speed of the wireless communication network of the mobile first-aid station are improved.

Description

Technical Field

[0001] The embodiments of the present application relate to the technical field of wireless communication of mobile emergency stations, and more particularly to a wireless communication network method and system of mobile emergency stations. Background Art

[0002] The wireless communication network approach for mobile emergency stations is designed to meet the needs for efficient, reliable and real-time data transmission in emergency rescue scenarios. In complex emergency environments, such as natural disaster sites, traffic accidents or large-scale public events, the communication network must be able to be quickly deployed and adapt to changing environmental conditions. Specific requirements include: ensuring stable transmission of mission-critical data under harsh conditions. Supporting high-definition video transmission, real-time voice communication and other large data volume applications. Ability to dynamically adjust the working frequency band and resource allocation according to environmental changes.

[0003] At present, the wireless communication solutions commonly used in mobile emergency stations mainly include the following: using multiple preset frequency bands to improve anti-interference capabilities, but lacking an adaptive adjustment mechanism and unable to optimize frequency band selection in real time. Presetting bandwidth and connection priorities makes it difficult to cope with resource competition and demand changes in emergencies. Relying on central servers for data management and synchronization leads to single point failure risks and slow response speeds. Based on fixed path selection rules, it is impossible to predict network change trends, and redundant path planning is insufficient, affecting fault recovery speed.

[0004] However, although the existing solutions have met some of the needs to a certain extent, they still have the following major defects: the existing multi-band radio system and static resource allocation strategy cannot monitor and adjust the frequency bands and resources in real time, and it is difficult to cope with complex and changing emergency environments. The centralized cache and synchronization system is prone to become a single point of failure. Once the central server has a problem, the performance of the entire system will be greatly reduced. Traditional routing algorithms fail to fully consider the changes in network topology and node status, lack effective planning and continuous optimization mechanisms for redundant paths, resulting in slow switching speed and low efficiency when the network fails.

[0005] In summary, the existing wireless communication solutions have obvious deficiencies in flexibility, reliability and intelligent path planning, and cannot fully meet the high-performance communication requirements of mobile emergency stations in complex emergency scenarios. Therefore, a more intelligent, adaptive and reliable wireless communication network method is urgently needed to improve the efficiency and stability of data transmission and provide strong technical support for emergency rescue. Summary of the invention

[0006] The embodiments of the present application provide a wireless communication network method and system for a mobile emergency station, which are used to solve the problems of low reliability and slow response speed of the wireless communication network of the mobile emergency station in the prior art.

[0007] In a first aspect, an embodiment of the present application provides a wireless communication network method of a mobile emergency station, including:

[0008] Real-time monitoring and evaluation of the signal strength, interference level and channel occupancy in the surrounding environment, selection of the optimal operating frequency band, and adjustment of the frequency band switching strategy to obtain an adaptive multi-band access mechanism;

[0009] According to the optimal working frequency band determined by the adaptive multi-band access mechanism, the bandwidth and connection priority of the wireless communication network are dynamically adjusted, the task priority is evaluated, and the bandwidth demand in the future is estimated, and resource competition between different users is simulated to generate a dynamic resource allocation strategy;

[0010] Based on the task priorities and bandwidth allocation defined in the dynamic resource allocation strategy, a distributed cache and synchronization system is deployed to achieve balanced distribution and rapid positioning of data, and concurrent updates are managed using version vector clocks. At the same time, cross-site data synchronization is achieved, reducing dependence on central servers and generating a distributed temporary data cache pool.

[0011] Using the data in the distributed cache pool, an intelligent routing algorithm based on deep reinforcement learning is established. The intelligent routing algorithm takes into account the current network topology and node status, predicts the network change trend, calculates the optimal transmission path from the source to the destination, and finally generates an intelligent routing planning solution;

[0012] Based on the optimal path information determined in the intelligent routing planning scheme, a redundant path is pre-set on the path, and the maximum flow minimum cut theorem in the graph theory algorithm is used to ensure rapid switching to the best backup path in the event of a network failure. The ant colony optimization algorithm is used to continuously optimize the path selection, and a redundant path planning scheme is generated based on the dynamic resource allocation strategy and the results of intelligent routing planning.

[0013] Optionally, the data in the distributed cache pool is used to establish an intelligent routing algorithm of deep reinforcement learning, wherein the intelligent routing algorithm takes into account the current network topology and node status, predicts the network change trend, calculates the optimal transmission path from the source to the destination, and finally generates an intelligent routing planning scheme, including:

[0014] Utilizing the multi-source heterogeneous data of historical communication records, node status information, and network topology structure in the distributed cache pool, the multi-source heterogeneous data is cleaned, converted, and normalized to obtain feature vectors suitable for deep learning model input, thereby generating an optimized feature data set;

[0015] Based on the optimized feature data set, a hybrid deep reinforcement learning model is constructed that combines a convolutional neural network for spatial feature extraction and a recurrent neural network for time series analysis. The hybrid deep reinforcement learning model can capture the spatial correlation of network topology and the trend of changes over time, and generate an intelligent decision-making framework suitable for dynamic network environments;

[0016] According to the intelligent decision-making framework suitable for dynamic network environment, the proximal policy optimization algorithm is adopted as the core of the reinforcement learning framework, and the hybrid deep learning model is trained in combination with the reward mechanism in the actual network environment, so that the hybrid deep learning model can be continuously optimized through interaction with the simulation environment, learn the ability to select the optimal path in different situations, and finally generate a trained intelligent model;

[0017] The trained intelligent model is used to calculate the optimal transmission path from the source to the destination based on the current network status and the prediction of the network change trend that may occur in the future, and finally generate an intelligent routing planning solution.

[0018] Optionally, based on the optimized feature data set, a hybrid deep reinforcement learning model is constructed that combines a convolutional neural network for spatial feature extraction and a recurrent neural network for time series analysis. The hybrid deep reinforcement learning model can capture the spatial correlation of network topology and the trend of change over time, and generate an intelligent decision-making framework suitable for a dynamic network environment, including:

[0019] The optimized feature data set is used to model the spatial distribution of network topology and node status, and a convolutional neural network is used to perform multi-layer convolution operations on the feature data set to capture the spatial correlation and local patterns between network nodes and obtain spatial feature representation;

[0020] Based on the spatial feature representation, the time series information of the node status is further combined, and a recurrent neural network is used to process the network status data that changes over time to capture the temporal dependency and dynamic change trend, and generate a time series feature representation;

[0021] Based on the spatial feature representation and the time series feature representation, a hybrid model is designed, which can process information in spatial and temporal dimensions at the same time, and an attention mechanism is introduced to enable the hybrid model to focus on important spatiotemporal features, thereby improving the accuracy and efficiency of decision-making and obtaining an intelligent decision-making framework that integrates spatiotemporal features;

[0022] An intelligent decision-making system suitable for dynamic network environments is constructed by using an intelligent decision-making framework that integrates spatiotemporal features. The intelligent decision-making system takes into account the spatial layout and node status of the current network topology, predicts the network change trend in the future, provides strong support for intelligent routing algorithms, and generates an intelligent decision-making framework suitable for dynamic network environments.

[0023] Optionally, according to the intelligent decision-making framework suitable for the dynamic network environment, the proximal policy optimization algorithm is used as the core of the reinforcement learning framework, and the hybrid deep learning model is trained in combination with the reward mechanism in the actual network environment, so that the hybrid deep learning model is continuously optimized through interaction with the simulation environment, and learns the ability to select the optimal path in different situations, and finally generates a trained intelligent model, including:

[0024] Based on an intelligent decision-making framework suitable for dynamic network environments, a reinforcement learning framework with a proximal policy optimization algorithm as the core is designed. The reinforcement learning framework continuously adjusts and optimizes model parameters through interaction with the simulation environment to ensure that the model can make the best path selection in a complex and changing network environment.

[0025] Based on the intelligent decision-making framework, a reward mechanism that meets the communication characteristics of the mobile emergency station is defined. The reward mechanism includes minimizing transmission delay, reducing packet loss rate, and maximizing bandwidth utilization. It is used to evaluate the performance of the model in different scenarios and guide the learning direction of the intelligent model to obtain the optimization objective function.

[0026] Using the optimization objective function, a simulation environment is constructed, so that the hybrid deep learning model can interact with various possible network conditions in the simulation environment. In each interaction, the deep learning model takes actions according to the current network state and receives corresponding reward feedback to accumulate experience data.

[0027] Based on the empirical data, the model parameters are updated using a proximal strategy optimization algorithm. The proximal strategy optimization algorithm avoids the performance instability problem caused by excessive update amplitude by limiting the pace of parameter update, thereby ensuring the stability and convergence of the deep learning model during training, and generating a preliminarily trained intelligent model;

[0028] By using the initially trained intelligent model to continue interacting with the simulation environment, the model gradually learns how to select the optimal path in different scenarios, giving priority to low-latency paths under high load conditions, and ultimately generates a trained intelligent model.

[0029] Optionally, the trained intelligent model is used to calculate the optimal transmission path from the source to the destination according to the current network status and the prediction of the network change trend that may occur in the future, and finally generate an intelligent routing planning solution, including:

[0030] Based on the trained intelligent model, real-time monitoring and data collection of the current network status are performed, and the data includes node load, connection quality, and traffic pattern to obtain the latest network status information;

[0031] Using the latest network status information and the time series analysis capability embedded in the intelligent model, the change trend of the network in the future is predicted to generate a network change trend prediction result;

[0032] According to the network change trend prediction result, the intelligent model evaluates multiple transmission paths, calculates the expected performance index of each path by comprehensively considering the node status on the path, the expected network change trend and the actual transmission demand, and obtains the path performance evaluation result;

[0033] Based on the path performance evaluation results, the intelligent model selects a path that meets the transmission requirements and has the best performance as the main path, giving priority to paths that can maintain efficient transmission under high load or network congestion conditions, and generates the optimal transmission path;

[0034] For the optimal transmission path, the intelligent model starts a dynamic monitoring mechanism to continuously monitor the real-time performance of the optimal transmission path, including node status, link quality, and traffic changes, to obtain dynamic monitoring results;

[0035] Using the dynamic monitoring results, the intelligent model evaluates the stability and reliability of the existing transmission path. When it is detected that the performance of the transmission path is degraded or there is a potential risk, the intelligent model automatically re-evaluates other transmission paths and obtains an updated optimal transmission path based on the latest network status and prediction results;

[0036] Based on the updated optimal transmission path, an intelligent routing planning scheme is finally generated. The intelligent routing planning scheme includes the optimal transmission path from the source to the destination and provides a mechanism that can dynamically adapt to network changes to ensure the continuity and reliability of data transmission.

[0037] Optionally, based on the optimal path information determined in the intelligent routing planning scheme, a redundant path is pre-set on the path, and the maximum flow minimum cut theorem in the graph theory algorithm is used to ensure rapid switching to the best backup path in the event of a network failure, and the ant colony optimization algorithm is used to continuously optimize the path selection, and a redundant path planning scheme is generated based on the dynamic resource allocation strategy and the results of the intelligent routing planning, including:

[0038] Analyze the current network topology and node status using the optimal path information determined in the intelligent routing planning solution, and pre-select one or more potential redundant paths as backup options according to the characteristics of the optimal path to obtain a pre-selected redundant path list;

[0039] Based on the pre-selected redundant path list, the maximum flow minimum cut theorem in the graph theory algorithm is applied to identify the key nodes and links that cause network interruption by calculating the maximum flow of the network, so as to ensure that the redundant path can be quickly switched when the main path fails, and generate a redundant path switching strategy;

[0040] According to the redundant path switching strategy, combined with the dynamic resource allocation strategy, the bandwidth availability and connection priority on the redundant path are evaluated, and according to the task priority and the bandwidth demand forecast in the future period, the resource allocation of the redundant path is adjusted to generate an optimized redundant path configuration;

[0041] Using the optimized redundant path configuration, continuously optimizing the redundant path using an ant colony optimization algorithm, wherein the ant colony optimization algorithm finds the best backup path from the source to the destination through iterative search, and generates an optimized backup path list;

[0042] Based on the optimized backup path list, according to the results of the dynamic resource allocation strategy and the intelligent routing planning solution, the network change trend and real-time performance monitoring data are comprehensively considered, and the configuration of the redundant path is continuously adjusted. When it is detected that the performance of the main path has degraded or there is a potential risk, the intelligent model automatically activates the redundant path, and re-evaluates and optimizes the path selection according to the latest network status to generate a redundant path planning solution.

[0043] Optionally, based on the optimized backup path list, according to the dynamic resource allocation strategy and the result of the intelligent routing planning solution, the configuration of the redundant path is continuously adjusted by comprehensively considering the network change trend and real-time performance monitoring data. When it is detected that the performance of the primary path is degraded or there is a potential risk, the intelligent model automatically activates the redundant path, and re-evaluates and optimizes the path selection according to the latest network status, and generates a redundant path planning solution, including:

[0044] Based on the optimized backup path list, combined with the dynamic resource allocation strategy and the results of the intelligent routing planning solution, predict the network change trend to obtain a network change trend prediction result;

[0045] Using the network change trend prediction result, pre-adjusting the redundant path to ensure that the redundant path has sufficient communication capacity when needed, and generating a pre-adjusted redundant path configuration;

[0046] According to the pre-adjusted redundant path configuration, start a real-time performance monitoring mechanism, continuously monitor the performance of the primary path and the redundant path, and collect real-time performance monitoring data;

[0047] Based on the real-time performance monitoring data, the intelligent model evaluates the health status of the primary path. If it is detected that the performance of the primary path is degraded or there is a potential risk, the best backup path is immediately activated to obtain the activated redundant path;

[0048] Using the activated redundant path, the intelligent model continues to monitor the performance of the redundant path, and further optimizes the path in combination with the ant colony optimization algorithm to ensure the high efficiency and stability of the redundant path, and generates an optimized redundant path configuration;

[0049] Finally, a redundant path planning scheme is generated, which includes preset redundant paths and their switching strategies, as well as the optimized activated redundant path configuration, to ensure that efficient communication capabilities can be maintained when network conditions change.

[0050] In a second aspect, an embodiment of the present application provides a wireless communication network system for a mobile emergency station, including:

[0051] The monitoring and evaluation module is used to monitor and evaluate the signal strength, interference level and channel occupancy rate in the surrounding environment in real time, select the optimal working frequency band, and adjust the frequency band switching strategy to obtain an adaptive multi-band access mechanism;

[0052] An adjustment and estimation module is used to dynamically adjust the bandwidth and connection priority of the wireless communication network according to the optimal working frequency band determined by the adaptive multi-band access mechanism, evaluate the task priority, and estimate the bandwidth demand in the future period of time, while simulating the resource competition between different users and generating a dynamic resource allocation strategy;

[0053] Deployment synchronization module, which is used to deploy distributed cache and synchronization system based on the task priority and bandwidth allocation defined in the dynamic resource allocation strategy, realize balanced distribution and rapid positioning of data, and use version vector clock to manage concurrent updates, while achieving cross-site data synchronization, reducing dependence on central servers, and generating distributed temporary data cache pools;

[0054] A prediction and calculation module is used to use the data in the distributed cache pool to establish an intelligent routing algorithm of deep reinforcement learning. The intelligent routing algorithm takes into account the current network topology and node status, predicts the network change trend, calculates the optimal transmission path from the source to the destination, and finally generates an intelligent routing planning solution;

[0055] The optimization generation module is used to pre-set a redundant path on the path based on the optimal path information determined in the intelligent routing planning scheme, and ensure rapid switching to the best backup path in the event of a network failure through the maximum flow minimum cut theorem in the graph theory algorithm, and use the ant colony optimization algorithm to continuously optimize the path selection, and generate a redundant path planning scheme based on the dynamic resource allocation strategy and the results of intelligent routing planning.

[0056] In a third aspect, an embodiment of the present application provides a computing device, comprising a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a wireless communication network method for a mobile emergency station as described in any one of the first aspects.

[0057] In a fourth aspect, an embodiment of the present application provides a computer storage medium storing a computer program, wherein when the computer program is executed by a computer, a wireless communication network method for a mobile emergency station as described in any one of the first aspect is implemented.

[0058] In the embodiment of the present application, the signal strength, interference level and channel occupancy rate in the surrounding environment are monitored and evaluated in real time, the optimal working frequency band is selected, and the frequency band switching strategy is adjusted to obtain an adaptive multi-band access mechanism; according to the optimal working frequency band determined by the adaptive multi-band access mechanism, the bandwidth and connection priority of the wireless communication network are dynamically adjusted, the task priority is evaluated, and the bandwidth demand in the future period is estimated, and the resource competition between different users is simulated to generate a dynamic resource allocation strategy; based on the task priority and bandwidth allocation defined in the dynamic resource allocation strategy, a distributed cache and synchronization system is deployed to achieve balanced distribution and rapid positioning of data, and a version vector clock is used to manage concurrent updates, while achieving data consistency across sites. According to synchronization, the dependence on the central server is reduced, and a distributed temporary data cache pool is generated; using the data in the distributed cache pool, an intelligent routing algorithm of deep reinforcement learning is established, and the intelligent routing algorithm takes into account the current network topology and node status, predicts the network change trend, calculates the optimal transmission path from the source to the destination, and finally generates an intelligent routing planning scheme; based on the optimal path information determined in the intelligent routing planning scheme, a redundant path is pre-set on the path, and the maximum flow minimum cut theorem in the graph theory algorithm is used to ensure rapid switching to the best backup path in the event of a network failure, and the ant colony optimization algorithm is used to continuously optimize the path selection, and a redundant path planning scheme is generated based on the dynamic resource allocation strategy and the results of intelligent routing planning.

[0059] The technical solution of this application has the following beneficial effects:

[0060] By real-time monitoring and evaluating the signal strength, interference level and channel occupancy in the surrounding environment, selecting the optimal working frequency band and adjusting the frequency band switching strategy, an adaptive multi-band access mechanism is implemented. This greatly enhances the anti-interference ability and stability of the network and reduces the risk of communication interruption. According to the optimal working frequency band determined by the adaptive multi-band access mechanism, the bandwidth and connection priority of the wireless communication network are dynamically adjusted, the task priority is evaluated, and the bandwidth demand in the future is estimated. At the same time, the resource competition between different users is simulated to generate a dynamic resource allocation strategy. This method effectively improves the bandwidth utilization, ensures the priority processing of key tasks, and avoids resource waste. Based on the task priority and bandwidth allocation defined in the dynamic resource allocation strategy, a distributed cache and synchronization system is deployed to achieve balanced distribution and rapid positioning of data, and use the version vector clock to manage concurrent updates, while achieving cross-site data synchronization, reducing dependence on the central server, and generating a distributed temporary data cache pool. This design not only speeds up data access, but also improves the fault tolerance and scalability of the system. Using the data in the distributed cache pool, an intelligent routing algorithm of deep reinforcement learning is established, taking into account the current network topology and node status, predicting the trend of network changes, calculating the optimal transmission path from the source to the destination, and finally generating an intelligent routing planning scheme. This intelligent path selection mechanism can flexibly respond to network changes and ensure the efficiency and reliability of data transmission. Based on the optimal path information determined in the intelligent routing planning scheme, a redundant path is pre-set on the path, and the maximum flow minimum cut theorem in the graph theory algorithm is used to ensure that the network is quickly switched to the best backup path in the event of a network failure, and the ant colony optimization algorithm is used to continuously optimize the path selection, and a redundant path planning scheme is generated based on the dynamic resource allocation strategy and the results of intelligent routing planning. This redundant path planning scheme can immediately enable the backup path when the main path fails, greatly shortening the fault recovery time and ensuring the continuity of communication.

[0061] Furthermore, the multi-source heterogeneous data of historical communication records, node status information and network topology in the distributed cache pool are cleaned, converted and normalized to generate feature vectors suitable for deep learning model input, forming an optimized feature data set. Based on the optimized feature data set, a hybrid deep reinforcement learning model that integrates convolutional neural network for spatial feature extraction and recurrent neural network for time series analysis is constructed. The model can capture the spatial correlation and time-varying trends of network topology and generate an intelligent decision-making framework suitable for dynamic network environments. The proximal policy optimization algorithm is used as the core of the reinforcement learning framework, and the hybrid deep learning model is trained in combination with the reward mechanism in the actual network environment. Through continuous optimization through interaction with the simulation environment, the model learns to select the optimal path in different situations, and finally generates a trained intelligent model. Using the trained intelligent model, according to the current network status and the prediction of the network change trend that may occur in the future, the optimal transmission path from the source to the destination is calculated, and an intelligent routing planning scheme is generated. Based on the optimal path information determined in the intelligent routing planning scheme, one or more potential redundant paths are pre-selected as backup options to obtain a pre-selected redundant path list. By applying the maximum flow minimum cut theorem in graph theory algorithms, the maximum flow of the network is calculated to identify the key nodes and links that cause network interruption, ensure that the redundant path can be quickly switched to when the main path fails, and generate a redundant path switching strategy. According to the redundant path switching strategy, combined with the dynamic resource allocation strategy, the bandwidth availability and connection priority on the redundant path are evaluated. According to the task priority and the bandwidth demand forecast in the future, the resource allocation of the redundant path is adjusted to generate an optimized redundant path configuration. Using the optimized redundant path configuration, the ant colony optimization algorithm is used to continuously optimize the redundant path, and the best backup path from the source to the destination is found through iterative search to generate an optimized backup path list. Based on the optimized backup path list, the configuration of the redundant path is continuously adjusted by comprehensively considering the network change trend and real-time performance monitoring data. When the performance of the main path is detected to be degraded or there is a potential risk, the intelligent model automatically activates the redundant path, re-evaluates and optimizes the path selection according to the latest network status, and generates a redundant path planning solution.

[0062] Through the introduction of deep reinforcement learning models, the system can automatically generate the optimal transmission path in a complex and changeable network environment, significantly improving the intelligence level and adaptability of path selection. This not only improves the efficiency of data transmission, but also enhances the flexibility and response speed of the system. By using redundant path planning and fast switching mechanisms, the system can quickly switch to the best backup path when the main path fails, greatly shortening the fault recovery time and ensuring the continuity and reliability of communication. At the same time, by continuously optimizing path selection, the system can cope with various emergencies and reduce the risk of network interruption. Combined with dynamic resource allocation strategies, the system can reasonably allocate bandwidth resources according to task priorities and bandwidth demand predictions, avoid resource waste, and ensure priority processing of critical tasks. In addition, the deployment of distributed caching and synchronization systems further improves data access speed, reduces dependence on central servers, and enhances the overall performance and stability of the system.

[0063] These and other aspects of the present application will become more clearly understood in the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0065] Figure 1 A flowchart of a wireless communication network method for a mobile emergency station provided in an embodiment of the present application;

[0066] Figure 2 A schematic diagram of the structure of a wireless communication network system of a mobile emergency station provided in an embodiment of the present application;

[0067] Figure 3 A schematic diagram of the structure of a computing device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0068] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0069] In some of the processes described in the specification and claims of this application and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this article or executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., do not represent the order of precedence, and do not limit the "first" and "second" to be different types.

[0070] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.

[0071] Figure 1 A flowchart of a wireless communication network method for a mobile emergency station is provided in an embodiment of the present application, such as Figure 1 As shown, the method includes:

[0072] 101. Real-time monitoring and evaluation of the signal strength, interference level and channel occupancy in the surrounding environment, selection of the optimal working frequency band, and adjustment of the frequency band switching strategy to obtain an adaptive multi-band access mechanism;

[0073] In this step, real-time monitoring and evaluation refers to the continuous collection of data on signal strength, interference level, and channel occupancy through a sensor network deployed in and around the mobile emergency station. These data include, but are not limited to, the received power of the wireless signal, the degree of interference from adjacent frequencies, and the usage of each communication channel. This process is used to identify the most stable and efficient communication frequency band in the current environment, thereby ensuring the reliability and low latency of data transmission. By dynamically adjusting the frequency band switching strategy, the system can automatically select the optimal operating frequency band under different environmental conditions and realize an adaptive multi-band access mechanism.

[0074] First, the system collects environmental data in real time through sensors installed on the emergency station and its peripheral equipment. Then, the data is processed using data analysis algorithms to evaluate the quality of each frequency band and select the most suitable working frequency band for the current environment. Finally, the frequency band switching strategy is adjusted based on the evaluation results, so that the system can respond quickly to environmental changes and ensure the optimization of communication quality. This step lays a solid foundation for subsequent resource allocation and path planning.

[0075] In an emergency rescue scenario, suppose a mobile first aid station is located near a large fire scene. Due to the large number of radio devices and other wireless communication activities at the scene, some frequency bands are severely interfered. Through sensors deployed at the first aid station and on surrounding vehicles, the system monitors in real time that the 2.4GHz band has large interference, while the 5GHz band is relatively idle. Based on this information, the system automatically switches to the 5GHz band for primary communication tasks, while retaining the 2.4GHz band as a backup band. This not only improves communication efficiency, but also reduces the possibility of data transmission errors.

[0076] 102. According to the optimal working frequency band determined by the adaptive multi-band access mechanism, dynamically adjust the bandwidth and connection priority of the wireless communication network, evaluate the task priority, and estimate the bandwidth demand in the future, while simulating the resource competition between different users and generating a dynamic resource allocation strategy;

[0077] In this step, the dynamic resource allocation strategy refers to the flexible adjustment of network resource allocation based on the currently selected working frequency band, combined with task priority and future bandwidth demand forecast. This strategy aims to maximize resource utilization while ensuring that critical tasks are given priority. The system simulates resource competition between different users and plans resource allocation in advance to avoid service interruption or performance degradation due to insufficient resources.

[0078] Based on the selected optimal working frequency band, the system analyzes the current network status and task requirements, evaluates the priority of each task, and estimates the bandwidth demand in the future. Then, by simulating the resource competition between different users, a reasonable resource allocation strategy is formulated. This process involves complex mathematical modeling and simulation technology to ensure the fairness and efficiency of resource allocation.

[0079] Continuing with the fire scene example above, the emergency station needs to support multiple communication tasks simultaneously, including high-definition video transmission, voice calls, and data reporting. The system allocates more bandwidth resources to high-priority tasks and less bandwidth to low-priority tasks based on the importance and urgency of the tasks. In addition, the system estimates the changes in bandwidth demand in the next 10 minutes and reasonably adjusts resource allocation to ensure that sufficient bandwidth is available at critical moments. When it is found that multiple users request large bandwidth resources at the same time, the system optimizes the allocation ratio of each user by simulating resource competition, ensuring smooth overall communication.

[0080] 103. Based on the task priority and bandwidth allocation defined in the dynamic resource allocation strategy, deploy a distributed cache and synchronization system to achieve balanced distribution and rapid positioning of data, and use version vector clocks to manage concurrent updates, while achieving cross-site data synchronization, reducing dependence on central servers, and generating a distributed temporary data cache pool;

[0081] In this step, the distributed cache and synchronization system is an important mechanism to improve the speed and reliability of data access. It deploys cache nodes on different nodes of the network so that data can be obtained nearby, reducing the dependence on the central server. The version vector clock is used to manage concurrent updates, ensure the consistency and integrity of data, and keep data synchronized even when the network is unstable. This design is particularly suitable for highly distributed and frequently changing working environments such as mobile emergency stations.

[0082] Based on the generated dynamic resource allocation strategy, the system deploys a distributed cache and synchronization system. The system caches commonly used data to each node according to task priority and bandwidth allocation, achieving balanced distribution and fast positioning of data. At the same time, the version vector clock is used to manage concurrent updates to ensure that the data of all nodes remains consistent. In addition, the cross-site data synchronization mechanism further enhances the system's fault tolerance and scalability.

[0083] At the fire scene, the emergency station deployed multiple distributed cache nodes, distributed between different emergency vehicles and command centers. These nodes store important data such as commonly used medical records, map information, and command and dispatch instructions. When an emergency vehicle needs to access the medical records of a specific patient, it can quickly obtain them through the nearest cache node without waiting to download from the central server. If multiple nodes update the same data at the same time (such as the latest fire spread), the version vector clock ensures that all update operations are in the correct order. Even if a node temporarily loses connection, other nodes can continue to work normally and automatically synchronize data after the connection is restored.

[0084] 104. Using the data in the distributed cache pool, establish an intelligent routing algorithm of deep reinforcement learning, wherein the intelligent routing algorithm takes into account the current network topology and node status, predicts the network change trend, calculates the optimal transmission path from the source to the destination, and finally generates an intelligent routing planning solution;

[0085] In this step, the intelligent routing algorithm of deep reinforcement learning is an advanced path selection method that uses multi-source heterogeneous data such as historical communication records, node status information, and network topology in the distributed cache pool to build a hybrid deep learning model. The model extracts spatial features through convolutional neural networks and analyzes time series data through recurrent neural networks to capture the spatial correlation of network topology and trends over time, thereby generating an intelligent decision-making framework suitable for dynamic network environments. Ultimately, the system can predict future trends based on the current network status and calculate the optimal transmission path.

[0086] The system first cleans, converts and normalizes the multi-source heterogeneous data in the distributed cache pool to generate feature vectors suitable for deep learning model input. Then, a hybrid deep reinforcement learning model that integrates convolutional neural networks and convolutional neural networks is built to train the model to capture the spatial correlation and temporal change trends of network topology. By interacting with the simulation environment, the model parameters are continuously optimized to enable it to learn the ability to choose the optimal path in different scenarios. Finally, based on the current network status and predicted future changes, the system calculates the optimal transmission path from source to destination and generates an intelligent routing planning solution.

[0087] At the fire scene, the system used the data in the previously deployed distributed cache pool, including past communication records, node status, and network topology, to establish an intelligent routing algorithm. The spatial features of the current network topology were extracted through a convolutional neural network, which analyzed time series data, and the model predicted possible network changes in the next few minutes. For example, the system predicted that the signal in a certain area might weaken due to increased smoke, so it adjusted the communication path in advance to bypass the area, ensuring the continuity and stability of data transmission. In this way, the system can always maintain efficient data transmission in a complex and changing environment.

[0088] Optionally, the step 104 uses the data in the distributed cache pool to establish an intelligent routing algorithm for deep reinforcement learning, which takes into account the current network topology and node status, predicts network change trends, calculates the optimal transmission path from the source to the destination, and finally generates an intelligent routing planning scheme, including: using the multi-source heterogeneous data of historical communication records, node status information and network topology in the distributed cache pool, cleaning, converting and normalizing the multi-source heterogeneous data to obtain feature vectors suitable for deep learning model input, and generating an optimized feature data set; based on the optimized feature data set, constructing a hybrid deep reinforcement learning model that integrates a convolutional neural network for spatial feature extraction and a recurrent neural network for time series analysis, The hybrid deep reinforcement learning model can capture the spatial correlation and time-varying trends of network topology, and generate an intelligent decision-making framework suitable for dynamic network environments; based on the intelligent decision-making framework suitable for dynamic network environments, a proximal policy optimization algorithm is used as the core of the reinforcement learning framework, and combined with the reward mechanism in the actual network environment, the hybrid deep learning model is trained, so that the hybrid deep learning model is continuously optimized through interaction with the simulated environment, and learns the ability to select the optimal path in different situations, and finally generates a trained intelligent model; using the trained intelligent model, according to the current network status and the prediction of possible network change trends in the future, the optimal transmission path from the source to the destination is calculated, and finally an intelligent routing planning solution is generated.

[0089] Among them, the step is to build a hybrid deep reinforcement learning model based on the optimized feature data set, which integrates convolutional neural network for spatial feature extraction and recurrent neural network for time series analysis. The hybrid deep reinforcement learning model can capture the spatial correlation of network topology and the trend of changes over time, and generate an intelligent decision-making framework suitable for dynamic network environments, including: using the optimized feature data set to model the spatial distribution of network topology and node status, performing multi-layer convolution operations on the feature data set through convolutional neural network, capturing the spatial correlation and local patterns between network nodes, and obtaining spatial feature representation; according to the spatial feature representation, further combining the time series information of node status, using recurrent neural network to process the network that changes over time. state data to capture temporal dependencies and dynamic change trends, and generate time series feature representations; based on the spatial feature representation and the time series feature representation, a hybrid model is designed, and the hybrid model can process information in spatial and temporal dimensions at the same time, and an attention mechanism is introduced to enable the hybrid model to focus on important spatiotemporal features, thereby improving the accuracy and efficiency of decision-making, and obtaining an intelligent decision-making framework that integrates spatiotemporal features; using the intelligent decision-making framework that integrates spatiotemporal features, an intelligent decision-making system suitable for dynamic network environments is constructed, which takes into account the spatial layout and node status of the current network topology, predicts network change trends in the future, provides strong support for intelligent routing algorithms, and generates an intelligent decision-making framework suitable for dynamic network environments.

[0090] In this step, the optimized feature data set includes multi-source heterogeneous data such as historical communication records, node status information, and network topology structure that have been cleaned, converted, and normalized. These data are used to generate feature vectors suitable for deep learning model input to ensure that the model can accurately capture the key features in the network environment. Convolutional neural networks are used for spatial feature extraction. By performing multi-layer convolution operations on the optimized feature data set, the spatial correlation and local patterns between network nodes are captured to obtain spatial feature representations. This enables the model to understand the spatial layout of the network topology and identify which nodes are closely connected or have potential bottlenecks. Recurrent neural networks are used for time series analysis. Combined with the time series information of node status, they process network status data that changes over time to capture temporal dependencies and dynamic change trends and generate time series feature representations. Recurrent neural networks are particularly suitable for processing data with time series properties and can predict future network change trends. The hybrid model refers to combining the advantages of convolutional neural networks and recurrent neural networks to design a model that can simultaneously process spatial and temporal dimension information. The model introduces an attention mechanism that enables the model to focus on important spatiotemporal features and improve the accuracy and efficiency of decision-making. This design is particularly suitable for complex routing decision tasks in dynamic network environments.

[0091] First, the network topology and spatial distribution of node states are modeled using the optimized feature data set. Multi-layer convolution operations are performed on the feature data set through a convolutional neural network to capture the spatial correlation and local patterns between network nodes and obtain spatial feature representation. This process is similar to using convolution kernels to extract edge and texture features of images in image processing, but is applied here to network topology.

[0092] Secondly, based on the spatial feature representation obtained above, the time series information of the node status is further combined. A recurrent neural network is used to process the network status data that changes over time to capture the temporal dependencies and dynamic change trends and generate a time series feature representation. This step helps the model predict the network change trend in the future and provide a basis for path selection.

[0093] Next, based on the spatial feature representation and time series feature representation, a hybrid model is designed. This model can process information in both spatial and temporal dimensions at the same time, and introduces an attention mechanism, which enables the model to focus on important spatiotemporal features and improve the accuracy and efficiency of decision-making. The attention mechanism allows the model to automatically adjust the focus on different features according to the importance of the current task, so as to better cope with the complex and changing network environment.

[0094] Finally, an intelligent decision-making framework integrating spatiotemporal features is used to build an intelligent decision-making system suitable for dynamic network environments. This system takes into account the spatial layout and node status of the current network topology, predicts the network change trend in the future, provides strong support for intelligent routing algorithms, and generates an intelligent decision-making framework suitable for dynamic network environments.

[0095] In the embodiment of the present application, in a wireless communication scenario of a mobile emergency station, it is assumed that a certain emergency station needs to transmit a large amount of high-definition video and medical data in real time when performing a rescue mission. In order to ensure the efficiency and reliability of data transmission, the system adopts the intelligent routing algorithm in the above-mentioned optional solution 104.

[0096] The system first uses the optimized feature data set to model the current network topology and node status. Through the convolutional neural network, the system captures the spatial correlation and local patterns between nodes. For example, the system recognizes that some nodes form a stable communication link due to their geographical proximity, while other nodes may have weak signals due to the presence of obstacles.

[0097] Next, the system combines the time series information of the node status and uses a recurrent neural network to process the network status data that changes over time. Through this process, the system predicts that the signal in a certain area may weaken due to the increase of smoke in the next few minutes, so it adjusts the communication path in advance to bypass the area.

[0098] Based on the above spatial and temporal feature representations, the system designs a hybrid model and introduces an attention mechanism. The model can automatically focus on important spatiotemporal features, improving the accuracy and efficiency of path selection. For example, in an emergency, the system can quickly identify the most reliable communication path to ensure that critical tasks are handled first.

[0099] Finally, the system uses an intelligent decision-making framework that integrates spatiotemporal features to build an intelligent decision-making system suitable for dynamic network environments. The system not only takes into account the spatial layout and node status of the current network topology, but also predicts the trend of network changes in the future, providing strong support for intelligent routing algorithms. During the entire rescue process, no matter how the environment changes, the system can adjust the path in time to ensure the efficiency and reliability of data transmission.

[0100] Through the above steps, the system can always maintain efficient path selection capabilities in complex and changing environments, significantly improve the reliability and stability of data transmission, and provide strong technical support for emergency rescue.

[0101] Optionally, in step 104, according to the intelligent decision-making framework suitable for dynamic network environments, a proximal policy optimization algorithm is used as the core of the reinforcement learning framework, and the hybrid deep learning model is trained in combination with the reward mechanism in the actual network environment, so that the hybrid deep learning model is continuously optimized through interaction with the simulated environment, and learns the ability to select the optimal path in different situations, and finally generates a trained intelligent model, including: based on the intelligent decision-making framework suitable for dynamic network environments, a reinforcement learning framework with the proximal policy optimization algorithm as the core is designed, and the reinforcement learning framework continuously adjusts and optimizes model parameters through interaction with the simulated environment to ensure that the model can make the optimal path selection in a complex and changeable network environment; according to the intelligent decision-making framework, a reward mechanism that conforms to the communication characteristics of the mobile emergency station is defined, and the reward mechanism includes minimizing transmission delay, reducing packet loss rate, and maximizing bandwidth utilization, which is used to evaluate the model in different situations. , and guide the learning direction of the intelligent model to obtain the optimization objective function; using the optimization objective function, construct a simulation environment, so that the hybrid deep learning model can interact with various possible network conditions in the simulation environment, in each interaction, the deep learning model takes action according to the current network status, and receives corresponding reward feedback to accumulate experience data; based on the experience data, the model parameters are updated using a proximal policy optimization algorithm, and the proximal policy optimization algorithm avoids the performance instability problem caused by excessive update amplitude by limiting the pace of parameter update, thereby ensuring the stability and convergence of the deep learning model during training, and generating a preliminarily trained intelligent model; using the preliminarily trained intelligent model to continue to interact with the simulation environment, the model gradually learns how to select the optimal path in different situations, and gives priority to low-latency paths under high load conditions, and finally generates a trained intelligent model.

[0102] Among them, the use of the trained intelligent model, according to the current network status and the prediction of the network change trend that may occur in the future, calculates the optimal transmission path from the source to the destination, and finally generates an intelligent routing planning solution, including: based on the trained intelligent model, real-time monitoring and data collection of the current network status, the data includes node load, connection quality, traffic pattern, to obtain the latest network status information; using the latest network status information, combined with the time series analysis capability embedded in the intelligent model, predicting the network change trend in the future, and generating network change trend prediction results; based on the network change trend prediction results, the intelligent model evaluates multiple transmission paths, and calculates the expected performance indicators of each path by comprehensively considering the node status on the path, the expected network change trend and the actual transmission demand, to obtain the path performance evaluation result; based on the path performance evaluation result, the intelligent model The model selects a path that meets the transmission requirements and has the best performance as the main path, gives priority to the path that can maintain efficient transmission under high load or network congestion, and generates the optimal transmission path; for the optimal transmission path, the intelligent model starts a dynamic monitoring mechanism to continuously monitor the real-time performance of the optimal transmission path, including node status, link quality, and traffic changes, and obtains dynamic monitoring results; using the dynamic monitoring results, the intelligent model evaluates the stability and reliability of the existing transmission path. When it is detected that the performance of the transmission path has degraded or there is a potential risk, the intelligent model automatically re-evaluates other transmission paths and obtains an updated optimal transmission path based on the latest network status and prediction results; based on the updated optimal transmission path, an intelligent routing planning scheme is finally generated, which includes the optimal transmission path from the source to the destination, and provides a mechanism that can dynamically adapt to network changes to ensure the continuity and reliability of data transmission.

[0103] In this step, the proximal policy optimization algorithm is a reinforcement learning algorithm that avoids performance instability caused by excessive update amplitude by limiting the pace of parameter updates, ensuring stability and convergence during model training. In this solution, proximal policy optimization is used as the core algorithm to adjust and optimize the parameters of the hybrid deep learning model, so that the model can make the best path selection in a complex and changeable network environment. The reward mechanism includes indicators such as minimizing transmission delay, reducing packet loss rate, and maximizing bandwidth utilization, which are used to evaluate the performance of the model in different scenarios and guide the learning direction of the intelligent model. These indicators reflect the key requirements of mobile emergency station communication and ensure that the system can provide efficient and reliable communication services in practical applications. The simulation environment is a virtualized network simulation platform that allows the hybrid deep learning model to interact with various possible network conditions. In each interaction, the model takes actions based on the current network status and receives corresponding reward feedback to accumulate experience data. This interactive method helps the model gradually learn how to choose the optimal path in different scenarios. The dynamic monitoring mechanism refers to real-time monitoring of the selected optimal transmission path and continuous collection of information such as node status, link quality, and traffic changes. These data are used to evaluate the stability and reliability of existing transmission paths. When path performance degradation or potential risks are detected, the intelligent model will automatically re-evaluate other transmission paths and dynamically adjust the optimal path based on the latest network status and prediction results.

[0104] First, based on an intelligent decision-making framework suitable for dynamic network environments, a reinforcement learning framework with a proximal policy optimization algorithm as the core is designed. This framework continuously adjusts and optimizes model parameters through interaction with the simulation environment to ensure that the model can make the best path selection in a complex and changing network environment.

[0105] Secondly, based on the intelligent decision-making framework, a reward mechanism that meets the communication characteristics of the mobile emergency station is defined. These reward mechanisms include minimizing transmission delay, reducing packet loss rate, maximizing bandwidth utilization, etc., which are used to evaluate the performance of the model in different scenarios and guide the learning direction of the intelligent model to obtain the optimization objective function.

[0106] Next, the optimization objective function is used to construct a simulation environment so that the hybrid deep learning model can interact with various possible network conditions in the simulation environment. In each interaction, the model takes actions based on the current network state and receives corresponding reward feedback to accumulate experience data.

[0107] Furthermore, based on the accumulated empirical data, the proximal policy optimization algorithm is used to update the model parameters. By limiting the pace of parameter updates, the proximal policy optimization algorithm ensures the stability and convergence of the model training process and generates a preliminarily trained intelligent model. By continuing to use the preliminarily trained intelligent model to interact with the simulation environment, the model gradually learns how to choose the optimal path in different situations, giving priority to low-latency paths under high load conditions, and finally generating a fully trained intelligent model.

[0108] Finally, the current network status is monitored in real time and data is collected to obtain the latest network status information, including node load, connection quality, traffic pattern, etc. Combined with the time series analysis capabilities embedded in the intelligent model, the change trend of the network in the future is predicted to generate network change trend prediction results. According to the network change trend prediction results, multiple transmission paths are evaluated, and the node status on the path, the expected network change trend and the actual transmission demand are comprehensively considered to calculate the expected performance indicators of each path to obtain the path performance evaluation results. A path that meets the transmission requirements and has the best performance is selected as the main path, and the path that can still maintain efficient transmission under high load or network congestion is given priority to generate the optimal transmission path. The dynamic monitoring mechanism is started to continuously monitor the real-time performance of the optimal transmission path and evaluate its stability and reliability. When the path performance is detected to be degraded or there is a potential risk, the intelligent model automatically re-evaluates other transmission paths, and dynamically adjusts the optimal path based on the latest network status and prediction results, and finally generates an intelligent routing planning solution.

[0109] In the embodiment of the present application, in a wireless communication scenario of a mobile emergency station, it is assumed that a certain emergency station needs to transmit high-definition video streams and a large amount of medical data in real time when performing emergency rescue tasks. In order to ensure the efficiency and reliability of data transmission, the system adopts the intelligent routing algorithm.

[0110] The system first designs a reinforcement learning framework with a proximal policy optimization algorithm as the core based on an intelligent decision-making framework suitable for dynamic network environments. The framework continuously adjusts and optimizes model parameters through interaction with the simulation environment to ensure that the model can make the best path selection in a complex and changing network environment.

[0111] Secondly, the system defines a set of reward mechanisms that are consistent with the communication characteristics of mobile emergency stations, including minimizing transmission delays, reducing packet loss rates, maximizing bandwidth utilization, etc. These reward mechanisms are used to evaluate the performance of the model in different scenarios and guide the learning direction of the intelligent model to obtain the optimized objective function.

[0112] Then, using the above optimization objective function, the system built a simulation environment, which enabled the hybrid deep learning model to interact with various possible network conditions in the simulation environment. In each interaction, the model takes actions based on the current network state and receives corresponding reward feedback to accumulate experience data.

[0113] Furthermore, based on the accumulated empirical data, the system uses the proximal policy optimization algorithm to update the model parameters. By limiting the pace of parameter updates, the proximal policy optimization algorithm ensures stability and convergence during model training and generates a preliminarily trained intelligent model. By continuing to use the preliminarily trained intelligent model to interact with the simulation environment, the model gradually learns how to choose the optimal path in different situations, giving priority to low-latency paths under high load conditions, and finally generates a fully trained intelligent model.

[0114] Furthermore, during the actual rescue process, the system monitors the current network status in real time and collects data to obtain the latest network status information, including node load, connection quality, traffic patterns, etc. Combined with the time series analysis capabilities embedded in the intelligent model, the system predicts the change trend of the network in the future and generates network change trend prediction results. Based on these prediction results, the system evaluates multiple transmission paths, comprehensively considers the node status on the path, the expected network change trend and the actual transmission demand, calculates the expected performance indicators of each path, and obtains the path performance evaluation results. In the end, the system selects a path that meets the transmission requirements and has the best performance as the main path, giving priority to paths that can maintain efficient transmission under high load or network congestion, and generating the optimal transmission path.

[0115] Finally, the system starts a dynamic monitoring mechanism to continuously monitor the real-time performance of the optimal transmission path and evaluate its stability and reliability. When it detects that the path performance has degraded or there is a potential risk, the system automatically re-evaluates other transmission paths and dynamically adjusts the optimal path based on the latest network status and prediction results to ensure the continuity and reliability of data transmission.

[0116] Through the above steps, the system can always maintain efficient path selection capabilities in complex and changing environments, significantly improve the reliability and stability of data transmission, and provide strong technical support for emergency rescue.

[0117] This application takes into account that in the wireless communication environment of the mobile emergency station, the efficiency and reliability of data transmission are crucial. Due to the complex and changeable network environment in the emergency scenario, traditional static routing algorithms are difficult to cope with dynamically changing needs. To this end, the research team developed an intelligent routing model based on the proximal policy optimization algorithm, which aims to enable the model to automatically select the optimal path in a complex and changeable network environment through a reinforcement learning framework. Therefore, a new optional solution is proposed, which includes:

[0118] Optionally, in step 104, according to the intelligent decision-making framework suitable for the dynamic network environment, a proximal policy optimization algorithm is used as the core of the reinforcement learning framework, and the hybrid deep learning model is trained in combination with the reward mechanism in the actual network environment, so that the hybrid deep learning model is continuously optimized through interaction with the simulation environment, and learns the ability to select the optimal path in different situations, and finally generates a trained intelligent model, including:

[0119] Based on an intelligent decision framework suitable for dynamic network environments, a reinforcement learning framework with a proximal policy optimization algorithm as the core is designed; the reinforcement learning framework continuously adjusts and optimizes model parameters through interaction with the simulation environment to ensure that the model can make the optimal path selection in a complex and changing network environment;

[0120]

[0121] Where L(θ) is the loss function; represents the expected value at time step t; clip(r t (θ), 1-∈, 1+∈) is the clipping operation, limiting r t The range of (θ) is within [1-∈, 1+∈] to prevent the performance from being unstable due to excessive parameter updates; ∈ is the clipping range of the proximal policy optimization algorithm to control the pace of parameter updates; A t is the action advantage function, which means that in state s t Take action a t The advantage of taking a particular action over other actions is measured by the degree of advantage of taking a particular action over other actions and is defined as:

[0122] A t =Q(s t , a t )-V(s t )

[0123] Where Q(s t , a t ) is to take action a t The expected return after s t ) is the state s t The value function of

[0124] r t (θ) is the ratio of the new and old strategies, indicating that action a is selected under the current parameter θ t The probability of the old parameter θ old The ratio of the probabilities of choosing the same action under , the formula is:

[0125]

[0126] where π θ (a t ∣s t )and Respectively represent the action a taken under the new and old strategies t The probability of

[0127] According to the requirements of the intelligent decision-making framework, a reward mechanism that meets the communication characteristics of the mobile emergency station is defined; the reward mechanism includes minimizing transmission delay, reducing packet loss rate, and maximizing bandwidth utilization, which is used to evaluate the performance of the model in different scenarios and guide the learning direction of the intelligent model to obtain the optimization objective function;

[0128] R(s t , a t )=w1·(1-d t )+w2·(1-p t )+w3·u t

[0129] Among them, R(s t , a t ) is the reward mechanism, s t is the current network status; a t is the action at the current moment; d t is the transmission delay; p t is the packet loss rate; u t is the bandwidth utilization; w1, w2, and w3 are the weights of various indicators, which are used to balance the importance of different goals;

[0130] Using the optimization objective function, a simulation environment is constructed to enable the hybrid deep learning model to interact with various possible network conditions in the simulation environment; in each interaction, the deep learning model takes actions according to the current network state and receives corresponding reward feedback to accumulate experience data;

[0131]

[0132] Among them, τ is a complete trajectory; π θ is the policy distribution with parameter θ; J(θ) is the policy gradient objective function, which represents the expected cumulative reward obtained from a series of interactions from the initial state to the terminal state under a given policy parameter θ; represents the expected value of all possible trajectories τ, where τ is the value of the current strategy π θ generated; represents the cumulative discounted reward from time step t = 0 to the termination time step T; γ is the discount factor, ranging from [0, 1]; R(s t , a t ) is in state s at time step t t and take action a t The immediate rewards received;

[0133] Based on the accumulated empirical data, the model parameters are updated using a proximal strategy optimization algorithm; the proximal strategy optimization algorithm avoids the performance instability problem caused by excessive update amplitude by limiting the pace of parameter update, thereby ensuring the stability and convergence of the deep learning model training process and generating a preliminarily trained intelligent model;

[0134]

[0135] Among them, α is the learning rate, which controls the step size of each parameter update; is the gradient of the loss function L(θ) with respect to θ; θ new is the parameter update formula; θ old is the model parameter before updating; θ new are the updated model parameters;

[0136] The preliminarily trained intelligent model is used to continue interacting with the simulation environment. The model gradually learns how to select the optimal path in different scenarios, giving priority to low-latency paths under high load conditions, and eventually generates a fully trained intelligent model.

[0137]

[0138] Among them, f θ (s t ) is the unnormalized probability value output by the policy network; softmax converts the output into a probability distribution so that the model selects the optimal path based on the current network state; According to the new parameters In status t Next select action a t The probability distribution of θ is the model parameter, which remains the same in all cases involving model updates.

[0139] The following are detailed comments on the parameters:

[0140] In the loss function L(θ) formula, E t A represents the expected value at time step t, which is used to measure the average performance of the model at different time points; tThe definition of A t =Q(s t , a t )-V(s t ), indicating that in state s t Take action a t The advantage of taking a specific action relative to other actions helps the model choose the optimal path; min is used to limit the scope of parameter updates to prevent large adjustments from causing unstable performance. It ensures that the updated strategy does not deviate too far from the old strategy; clip limits r t The range of (θ) is in [1-∈, 1+∈] to control the pace of parameter update and avoid excessive update amplitude;

[0141] In the ratio of new and old strategies t (θ) In the formula, π θ (a t ∣s t ) means taking action a under the current parameter θ t The probability of Indicates that the old parameter θ old The probability of taking the same action under

[0142] In the reward mechanism R(s t , a t ) formula, s t is the current network status; a t is the action at the current moment; d t Indicates the transmission time of data from source to destination; u t Indicates the efficiency of network bandwidth usage; p t Indicates the proportion of data packets lost during data transmission; w1, w2, and w3 are used to balance the importance of different goals, and these weights can be adjusted according to specific application scenarios;

[0143] In the policy gradient objective function J(θ) formula, τ is a complete trajectory; π θ is the policy distribution with parameter θ; J(θ) is the policy gradient objective function, which represents the expected cumulative reward obtained from a series of interactions from the initial state to the terminal state under a given policy parameter θ; represents the expected value of all possible trajectories τ, where τ is the value of the current strategy π θ generated; represents the cumulative discounted reward from time step t = 0 to the termination time step T; γ is the discount factor, ranging from [0, 1]; R(s t , a t ) is in state s at time step t t and take action a t The immediate rewards received;

[0144] In the parameter update θ new In the formula, α controls the step size of each parameter update and determines the speed of model learning; the loss function The gradient of θ indicates the direction of parameter update; θ old are the model parameters before updating;

[0145] The probability of action selection In the formula, f θ (s t ) is the unnormalized probability value output by the policy network; softmax converts the output into a probability distribution so that the model selects the optimal path based on the current network state;

[0146] The following is a brief introduction to the design reasons of each sub-item of the formula:

[0147] In the reward mechanism R(s t , a t ) formula, w1·(1-d t ) The reason for the design of this sub-item is the transmission delay d t It is a key indicator to measure the speed of data transmission. In order to encourage the model to choose a low-latency path, 1-d t To calculate the reward for this part. t When it is close to 0, 1-d t is close to 1, indicating that the path has lower latency and thus gives a higher reward. On the contrary, if d t Higher, then 1-d t The lower the value, the smaller the reward. The weight w1 controls the impact of transmission delay on the total reward; w2·(1-p t ) The reason for designing this sub-item is the packet loss rate p t Indicates the proportion of packets lost during data transmission. In order to encourage the model to choose reliable paths, 1-p is used. t To calculate the reward for this part. t When it is close to 0, 1-p t is close to 1, indicating that the path has a lower packet loss rate, thus giving a higher reward. On the contrary, if p t Higher, then 1-p t The lower the packet loss rate, the smaller the reward. The weight w2 controls the impact of the packet loss rate on the total reward; w3·u t The design reason for this sub-item is that the bandwidth utilization u t Indicates the efficiency of network bandwidth usage. In order to encourage the model to make full use of the available bandwidth, u t To calculate the reward for this part. tWhen it is close to 1, it means that the bandwidth utilization is high and a higher reward is given; on the contrary, if u t The lower the bandwidth, the lower the reward. The weight w3 controls the impact of bandwidth utilization on the total reward.

[0148] Reasons for adding the sub-items together:

[0149] By adding them together, we can comprehensively consider the three key indicators of transmission delay, packet loss rate and bandwidth utilization. Each sub-item represents an independent but important performance indicator, and the result of the addition can form a comprehensive reward value to comprehensively evaluate the overall performance of the model in taking a certain action in the current state.

[0150] Through the above formula, a comprehensive reward mechanism can be constructed to guide the learning direction of the intelligent routing model. Each sub-item is designed to optimize a specific network performance indicator, and adding them together provides a comprehensive evaluation of the overall system performance. This approach not only improves the accuracy of model selection, but also ensures its efficiency and reliability in practical applications.

[0151] The following is a brief introduction to the design reasons of each sub-item of the formula:

[0152] After updating the model parameters θ new In the formula, The reason for designing this sub-item is that the learning rate α controls the pace of parameter updates to prevent instability caused by excessive gradients. By multiplying, the magnitude of each update can be flexibly adjusted to ensure that the model can maintain an appropriate update speed at different stages. Different parameters may have different gradient scales, and directly using gradients may cause some parameters to update too quickly or too slowly. By introducing a learning rate, these updates can be standardized so that each parameter can be adjusted at an appropriate speed.

[0153] Reasons for adding the sub-items together:

[0154] By replacing the old parameters θ old Add an increment proportional to the gradient The parameter update is realized. This update method enables the model to move in the direction of reducing the loss function in each iteration, thereby gradually optimizing the model performance.

[0155] Through the above formula, the model parameters can be effectively updated. This method not only ensures that the model can gradually reduce the value of the loss function, but also provides a mechanism to control the update amplitude through the learning rate α, thereby ensuring the stability and efficiency of the training process. This method is widely used in the training process of various machine learning and deep learning models, and is one of the core optimization algorithms.

[0156] Assume that at a large fire scene, a mobile emergency station needs to support high-definition video streaming, voice calls, and the transmission of a large amount of medical data at the same time. In order to ensure the efficiency and reliability of data transmission for these critical tasks, the above formula scheme is used to develop and train the intelligent routing model. The following are the implementation steps and calculation examples:

[0157] Set the transmission delay d t is 0.5 seconds, and the packet loss rate is p t is 0.05, bandwidth utilization u t is 0.9.

[0158] The weights w1=0.4, w2=0.3, and w3=0.3 represent the importance attached to transmission delay, packet loss rate, and bandwidth utilization.

[0159] Calculate the instant reward R(s t , a t ):

[0160] R(s t , a t )=w1·(1-d t )+w2·(1-p t )+w3·u t

[0161] R(s t , a t )=0.4·(1-0.5)+0.3·(1-0.05)+0.3·0.9

[0162] R(s t , a t )=0.4·0.5+0.3·0.95+0.3·0.9

[0163] R(s t , a t )=0.2+0.285+0.27=0.755

[0164] A simulation environment is built to enable the hybrid deep learning model to interact with various possible network conditions. In each interaction, the model takes actions based on the current network state and receives corresponding reward feedback to accumulate experience data.

[0165] Assume the discount factor γ = 0.9, the time step T = 10, and the cumulative reward for:

[0166]

[0167] If the immediate reward R(s) at each time step t , a t ) are all 0.755, then:

[0168]

[0169] Using the geometric series summation formula Where r = 0.9 and n = 10:

[0170]

[0171] J(θ)≈0.755·6.862≈5.178

[0172] Assume that the old parameter θ old The probability of taking action is , the probability of taking the same action under the new parameter θ is π θ (a t ∣s t )=0.8, clipping range∈=0.2, then:

[0173]

[0174] Calculate the loss function L(θ):

[0175] L(θ)=E t [A t ·min(r t (θ)·A t ,clip(r t (θ), 1-∈, 1+∈)·A t )]

[0176] Assume that the action advantage function A t =0.5:

[0177] L(θ)=E t [0.5·min(1.333·0.5, clip(1.333, 0.8, 1.2)·0.5)]

[0178] L(θ)=E t [0.5 min (0.6665, 0.6 0.5)]

[0179] L(θ)=E t [0.5 0.3] = 0.15

[0180] Assuming the learning rate α = 0.01, then:

[0181]

[0182] If θ old = 0.5, and

[0183] θnew =0.5+0.01·(-0.3)=0.5-0.003=0.497

[0184] By using the initially trained intelligent model to continue interacting with the simulation environment, the model gradually learns how to select the optimal path in different scenarios, giving priority to low-latency paths under high load conditions, and ultimately generates a fully trained intelligent model.

[0185] Through the above calculation, we can draw the following conclusions: The immediate reward R(s t , a t )=0.755, indicating that the current path selection has a good balance between transmission delay, packet loss rate and bandwidth utilization. The cumulative reward J(θ)≈5.178 shows that the model performs well in this simulation environment and is expected to obtain a higher cumulative reward. The new parameter θ new = 0.497, which is slightly less than the old parameter, indicating that the model is gradually optimizing its decision-making process and avoiding instability caused by large updates.

[0186] Overall, this solution effectively improves the adaptability and stability of the intelligent routing model by introducing a proximal strategy optimization algorithm, ensuring that the optimal path can always be found in a complex and changing network environment, thereby significantly improving the efficiency and reliability of data transmission and providing strong technical support for emergency rescue.

[0187] 105. Based on the optimal path information determined in the intelligent routing planning scheme, a redundant path is pre-set on the path, and the maximum flow minimum cut theorem in the graph theory algorithm is used to ensure rapid switching to the best backup path in the event of a network failure. The ant colony optimization algorithm is used to continuously optimize the path selection, and a redundant path planning scheme is generated based on the dynamic resource allocation strategy and the results of intelligent routing planning.

[0188] In this step, redundant path planning refers to presetting one or more backup paths in addition to the main path so that the best backup path can be quickly switched to in the event of a network failure. By applying the maximum flow minimum cut theorem in graph theory algorithms, the system can identify key nodes and links that may cause network interruptions, ensuring that it can quickly switch to redundant paths when the main path fails. In addition, the ant colony optimization algorithm is used to continuously optimize redundant paths to ensure that path selection is always in the optimal state. This method greatly improves the fault tolerance and reliability of the system.

[0189] Based on the optimal path information determined in the intelligent routing planning scheme, the system pre-selects one or more potential redundant paths as backup options. By applying the maximum flow minimum cut theorem, the maximum flow of the network is calculated, and key nodes and links are identified to ensure that the redundant path can be quickly switched when the main path fails. According to the redundant path switching strategy, combined with the dynamic resource allocation strategy, the bandwidth availability and connection priority on the redundant path are evaluated, the resource allocation of the redundant path is adjusted, and the optimized redundant path configuration is generated. Finally, the ant colony optimization algorithm is used to continuously optimize the redundant path to ensure that the path selection is always in the optimal state.

[0190] At the fire scene, the system pre-sets several redundant paths based on the optimal path determined by the intelligent routing planning solution. Through the maximum flow minimum cut theorem, the system identifies key nodes and links that may cause network interruptions, ensuring that it can quickly switch to redundant paths when the main path fails. For example, when a main path is interrupted due to the collapse of a building, the system immediately activates the pre-set redundant path to ensure the continuity of communication. In addition, the system uses the ant colony optimization algorithm to continuously optimize redundant paths to ensure that path selection is always in the optimal state. During the entire rescue process, no matter how the environment changes, the system can adjust the path in time to ensure the efficiency and reliability of data transmission.

[0191] Optionally, in step 105, based on the optimal path information determined in the intelligent routing planning scheme, a redundant path is pre-set on the path, and the maximum flow minimum cut theorem in the graph theory algorithm is used to ensure that the network fails and quickly switches to the best backup path, and the ant colony optimization algorithm is used to continuously optimize the path selection, and a redundant path planning scheme is generated according to the dynamic resource allocation strategy and the results of the intelligent routing planning, including: using the optimal path information determined in the intelligent routing planning scheme, analyzing the current network topology and node status, and pre-selecting one or more potential redundant paths as backup options according to the characteristics of the optimal path to obtain a pre-selected redundant path list; based on the pre-selected redundant path list, applying the maximum flow minimum cut theorem in the graph theory algorithm, by calculating the maximum flow of the network, identifying the key nodes and links that cause network interruption, ensuring that the redundant path can be quickly switched to when the main path fails, and generating a redundant path switching strategy; According to the redundant path switching strategy, combined with the dynamic resource allocation strategy, the bandwidth availability and connection priority on the redundant path are evaluated, and the resource allocation of the redundant path is adjusted according to the task priority and the bandwidth demand forecast in the future period, so as to generate an optimized redundant path configuration; using the optimized redundant path configuration, the redundant path is continuously optimized by using the ant colony optimization algorithm, and the ant colony optimization algorithm module finds the best backup path from the source to the destination through iterative search to generate an optimized backup path list; based on the optimized backup path list, according to the results of the dynamic resource allocation strategy and the intelligent routing planning scheme, the network change trend and real-time performance monitoring data are comprehensively considered, and the configuration of the redundant path is continuously adjusted. When it is detected that the performance of the main path is degraded or there is a potential risk, the intelligent model automatically activates the redundant path, and re-evaluates and optimizes the path selection according to the latest network status to generate a redundant path planning scheme.

[0192] Optionally, in step 105, based on the optimized backup path list, according to the results of the dynamic resource allocation strategy and the intelligent routing planning scheme, the network change trend and real-time performance monitoring data are comprehensively considered, and the configuration of the redundant path is continuously adjusted. When it is detected that the performance of the main path is degraded or there is a potential risk, the intelligent model automatically activates the redundant path, and re-evaluates and optimizes the path selection according to the latest network status to generate a redundant path planning scheme, including: based on the optimized backup path list, combined with the results of the dynamic resource allocation strategy and the intelligent routing planning scheme, predicting the network change trend to obtain a network change trend prediction result; using the network change trend prediction result, pre-adjusting the redundant path to ensure that the redundant path has sufficient communication capacity when needed, and generating a pre-adjusted redundant path. path configuration; according to the pre-adjusted redundant path configuration, start the real-time performance monitoring mechanism, continuously monitor the performance of the main path and the redundant path, and collect real-time performance monitoring data; based on the real-time performance monitoring data, the intelligent model evaluates the health of the main path, and if it is detected that the performance of the main path is degraded or there is a potential risk, immediately activate the best backup path to obtain the activated redundant path; using the activated redundant path, the intelligent model continues to monitor the performance of the redundant path, and further optimizes the path in combination with the ant colony optimization algorithm to ensure the high efficiency and stability of the redundant path, and generate the optimized redundant path configuration; finally generate a redundant path planning scheme, the redundant path planning scheme includes the preset redundant path and its switching strategy, and the optimized activated redundant path configuration, to ensure that efficient communication capabilities can be maintained when network conditions change.

[0193] In this step, the pre-selected redundant path list includes one or more potential redundant paths pre-selected as backup options based on the optimal path information determined in the intelligent routing planning scheme. These paths are used to quickly switch in the event of a network failure to ensure the continuity and reliability of communication. The maximum flow minimum cut theorem is a graph theory algorithm that identifies the key nodes and links that cause network interruptions by calculating the maximum flow of the network. This theorem ensures that the redundant path can be quickly switched to when the primary path fails, and a redundant path switching strategy is generated, thereby improving the fault tolerance of the network. The dynamic resource allocation strategy refers to the reasonable adjustment of bandwidth availability and connection priority on the redundant path based on the task priority and the bandwidth demand forecast for a period of time in the future. This strategy ensures the effective use of resources and can respond quickly in emergencies. The ant colony optimization algorithm is a heuristic search algorithm that iteratively searches for the best backup path from the source to the destination by simulating the behavior of ants looking for food. The algorithm can continuously optimize path selection in a complex and changing environment and improve the stability and efficiency of the path. The real-time performance monitoring mechanism refers to the continuous monitoring of the primary and redundant paths and the collection of real-time performance monitoring data, such as node status, link quality, and traffic changes. This data is used to assess the stability and reliability of existing transmission paths and guide the decision-making of intelligent models.

[0194] First, the optimal path information determined in the intelligent routing planning solution is used to analyze the current network topology and node status, and one or more potential redundant paths are pre-selected as backup options based on the characteristics of the optimal path to obtain a pre-selected redundant path list. This step ensures that the system has alternative paths available when needed.

[0195] Secondly, based on the pre-selected redundant path list, the maximum flow minimum cut theorem in graph theory is applied to identify the key nodes and links that may cause network interruption by calculating the maximum flow of the network. This step ensures that the redundant path can be quickly switched to when the main path fails, generates a redundant path switching strategy, and enhances the fault tolerance of the system.

[0196] Next, according to the redundant path switching strategy, combined with the dynamic resource allocation strategy, the bandwidth availability and connection priority on the redundant path are evaluated. According to the task priority and the bandwidth demand forecast in the future, the resource allocation of the redundant path is adjusted to generate the optimized redundant path configuration. This process ensures the effective use of resources and can respond quickly to emergencies.

[0197] Furthermore, the ant colony optimization algorithm is used to continuously optimize the redundant paths using the optimized redundant path configuration. The best backup path from the source to the destination is found through iterative search, and an optimized backup path list is generated. This step ensures that the path selection is always in the optimal state, improving the stability and efficiency of the path.

[0198] Furthermore, based on the optimized backup path list, combined with the results of the dynamic resource allocation strategy and the intelligent routing planning solution, the network change trend is predicted to obtain the network change trend prediction results. Using these prediction results, the redundant paths are pre-adjusted to ensure that the redundant paths have sufficient communication capabilities when needed, and the pre-adjusted redundant path configuration is generated. The real-time performance monitoring mechanism is started to continuously monitor the performance of the primary path and the redundant path, and collect real-time performance monitoring data. Based on this data, the intelligent model evaluates the health of the primary path. If the performance of the primary path is detected to be degraded or there is a potential risk, the best backup path is immediately activated to obtain the activated redundant path. Using the activated redundant path, the intelligent model continues to monitor its performance, and combines the ant colony optimization algorithm to further optimize the path, ensure the efficiency and stability of the redundant path, and generate the optimized redundant path configuration.

[0199] Finally, a redundant path planning scheme is generated, which includes preset redundant paths and their switching strategies, as well as optimized activated redundant path configurations, to ensure efficient communication capabilities when network conditions change.

[0200] In an embodiment of the present application, in a wireless communication scenario of a mobile emergency station, it is assumed that a certain emergency station needs to ensure the continuity and reliability of communication when performing an emergency rescue mission. In order to cope with possible network failures, the system adopts the redundant path planning and continuous optimization method in the above-mentioned optional solution 105.

[0201] The system first uses the optimal path information determined in the intelligent routing planning solution to analyze the current network topology and node status, pre-selects one or more potential redundant paths as backup options, and obtains a pre-selected redundant path list. For example, at a fire scene, the system selects several different backup paths to ensure that there are backup paths available when the main path fails.

[0202] Secondly, based on the pre-selected redundant path list, the system applies the maximum flow and minimum cut theorem in graph theory to calculate the maximum flow of the network and identify the key nodes and links that may cause network interruption. For example, the system identifies certain key nodes and links to ensure that it can quickly switch to the redundant path when the main path fails, thus enhancing the fault tolerance of the system.

[0203] Next, based on the redundant path switching strategy, the system combines the dynamic resource allocation strategy to evaluate the bandwidth availability and connection priority on the redundant path. For example, the system adjusts the resource allocation of the redundant path based on the task priority and the bandwidth demand forecast in the future period, and generates an optimized redundant path configuration. If a redundant path is expected to carry a large amount of video data in the next few minutes, the system will prioritize allocating more bandwidth resources to it.

[0204] Furthermore, using the optimized redundant path configuration, the system uses the ant colony optimization algorithm to continuously optimize the redundant paths. For example, the best backup path from the source to the destination is found through iterative search, and an optimized backup path list is generated. When a redundant path is found to have higher transmission efficiency, the system automatically updates the path selection to ensure that the path selection is always in the optimal state.

[0205] Furthermore, based on the optimized backup path list, the system combines the results of the dynamic resource allocation strategy and the intelligent routing planning scheme to predict the trend of network changes and obtain the prediction results of network change trends. For example, the system predicts that the signal in a certain area may weaken due to the increase of smoke, so it adjusts the resource allocation of the redundant path in advance. After starting the real-time performance monitoring mechanism, the system continuously monitors the performance of the main path and redundant path and collects real-time performance monitoring data. When the performance of the main path is detected to be degraded or there is a potential risk, the intelligent model immediately activates the best backup path to ensure the continuity of communication. For example, when the main path is interrupted due to the collapse of a building, the system automatically switches to the pre-set redundant path to ensure uninterrupted communication. After that, the system continues to monitor the performance of the redundant path, and further optimizes the path in combination with the ant colony optimization algorithm to ensure the efficiency and stability of the redundant path, and finally generates a redundant path planning scheme.

[0206] Through the above steps, the system can not only maintain efficient path selection capabilities in complex and changing environments, but also significantly improve the reliability and stability of data transmission, providing strong technical support for emergency rescue.

[0207] This application takes into account that in the wireless communication environment of the mobile first aid station, it is crucial to ensure the continuity and reliability of data transmission. Due to the complex and changeable network environment in the first aid scenario, traditional static routing algorithms are difficult to cope with dynamically changing needs. To this end, the research team has developed an intelligent redundant path planning scheme based on graph theory and ant colony optimization algorithm, which aims to enable the system to quickly switch to the optimal redundant path when the main path fails and continuously optimize the path selection through dynamic resource allocation strategies and real-time performance monitoring. Therefore, a new optional scheme is proposed, which includes:

[0208] Optionally, the configuration of the redundant path is continuously adjusted based on the optimized backup path list in step 105 according to the result of the dynamic resource allocation strategy and the intelligent routing planning solution, taking into account the network change trend and real-time performance monitoring data. When it is detected that the performance of the primary path is degraded or there is a potential risk, the intelligent model automatically activates the redundant path, re-evaluates and optimizes the path selection according to the latest network status, and generates a redundant path planning solution, including:

[0209] The optimal path information determined in the intelligent routing planning scheme is used to analyze the current network topology and node status, and one or more potential redundant paths are pre-selected as backup options according to the characteristics of the optimal path to obtain a pre-selected redundant path list P. redundant ;

[0210] P redundant ={p1, p2, ..., p i}

[0211] Among them, P redundant : Pre-selected redundant path list; p i is the i-th potential redundant path;

[0212] Based on the pre-selected redundant path list, the maximum flow minimum cut theorem in graph theory algorithm is applied to calculate the maximum flow C of the network. max , identify the key nodes and links that cause network interruption, ensure rapid switching to redundant paths when the primary path fails, and generate redundant path switching strategies;

[0213]

[0214] Among them, C max The maximum flow of the network; S is a set of nodes in the network; is the complement of S, that is, the set of nodes not in S; c(u, v) is the capacity of edge (u, v);

[0215] According to the redundant path switching strategy, combined with the dynamic resource allocation strategy, the bandwidth availability on the redundant path is evaluated. i and connection priority P i , based on task priority and bandwidth demand prediction in the future future , adjust the resource allocation of redundant paths and generate optimized redundant path configuration;

[0216] R i =f(B i , P i , D future )

[0217] Among them, R i is a redundant path p i resource allocation; B i is a redundant path p i Bandwidth availability on P i is a redundant path p i Connection priority of D future It is the bandwidth demand forecast for a period of time in the future;

[0218] Using the optimized redundant path configuration, the ant colony optimization algorithm is used to continuously optimize the redundant paths; the ant colony optimization algorithm finds the best backup path from the source to the destination through iterative search and generates an optimized backup path list;

[0219] Update the path quality using the following formula:

[0220] q(t+1)=Q(t)+ΔQ(t)

[0221] Where Q(t+1) is the updated path quality at time step t+1; Q(t) is the path quality at time step t; ΔQ(t) is the path quality increment between time step t and t+1;

[0222] Update the pheromone concentration using the following formula:

[0223] τ ij (t+1)=(1-ρ)·τ ij (t)+Δτ ij (t)

[0224] Among them, τ ij (t+1) is the updated pheromone concentration at time step t+1; τ ij (t) is the pheromone concentration on edge (i, j) at time step t; ρ is the pheromone evaporation rate; Δτ ij (t) is the amount of pheromone added on edge (i, j) between time steps t and t+1;

[0225] Based on the optimized backup path list, the configuration of redundant paths is continuously adjusted by comprehensively considering network change trends and real-time performance monitoring data. When it is detected that the performance of the primary path has degraded or there is a potential risk, the intelligent model automatically activates the redundant path, and re-evaluates and optimizes the path selection based on the latest network status, and finally generates a redundant path planning solution.

[0226] The following are detailed comments on the parameters:

[0227] In the pre-selected redundant path list P redundant In the formula, P redundant is a pre-selected redundant path list, which represents a set of pre-selected potential redundant paths; p i is the ith potential redundant path. Each path is determined based on the network topology and node status analysis as a backup option when the primary path fails;

[0228] In computing the maximum flow C of the network max In the formula, C max is the maximum flow of the network, indicating the maximum possible flow from source to destination; S is a set of nodes in the network; is the complement of S, that is, the set of nodes not in S; c(u, v) is the capacity of edge (u, v), which indicates the maximum amount of data that the link connecting nodes u and v can carry;

[0229] In redundant path resource allocation R i In the formula, R i is a redundant path p i resource allocation as a function of bandwidth availability, connection priority, and future bandwidth demand forecasts; i is a redundant path p i The bandwidth availability on the path indicates the bandwidth resources that the path can currently provide; i is a redundant path p i The connection priority reflects the importance and priority of the path among all redundant paths; future It is a forecast of bandwidth demand in the future, which is used to evaluate the bandwidth resources required for future tasks and ensure the forward-looking allocation of resources;

[0230] In the formula for updating the path quality Q(t+1), Q(t+1) is the updated path quality at time step t+1, which indicates the overall performance of the path after one iteration; Q(t) is the path quality at time step t, which indicates the performance level of the current path; ΔQ(t) is the path quality increment between time step t and t+1, which indicates the performance improvement or degradation brought about by adjusting the path through the optimization algorithm;

[0231] The following is a brief introduction to the design reasons of each sub-item of the formula:

[0232] In the updated pheromone concentration τ ij (t+1) In the formula, (1-ρ)·τ ij (t) This sub-item is designed because the evaporation rate of pheromones can be precisely controlled through the multiplication operation. Larger ρ values ​​will cause more pheromones to evaporate, while smaller ρ values ​​will retain more pheromones. This flexibility allows the persistence of pheromones to be adjusted in different application scenarios to meet different needs. The multiplication operation ensures the standardization of the pheromone renewal process, so that each renewal can be carried out within a controllable range, avoiding instability problems caused by too high or too low pheromone concentrations. Δτ ij (t) The reason for designing this item is that every time an ant successfully finds a valid path, it leaves pheromones on the path to mark the quality of the path. ij (t), can strengthen those paths that have been proven to be effective, making subsequent ants more likely to choose these paths. The increase in pheromones provides a positive feedback mechanism, making paths with good performance more popular in future choices, thereby gradually optimizing path selection.

[0233] Here are the reasons for adding the sub-items together:

[0234] By adding together the pheromone, both the evaporation and the new additions can be considered. This not only keeps the memory of the historical path, but also introduces the latest feedback on the path quality. This helps the algorithm find a balance between exploring new paths and using existing paths. After each iteration, the pheromone concentration is updated based on the current path quality and historical performance. This approach allows the algorithm to dynamically respond to network changes and continuously optimize path selection.

[0235] The above formula can effectively update the pheromone concentration. This method not only simulates the behavior of pheromones in nature, but also provides a mechanism to balance the memory of historical paths and the exploration of new paths, thereby ensuring that the ant colony optimization algorithm can continuously optimize path selection in a complex and changing environment. This method is widely used in various path optimization problems, such as routing selection, logistics distribution, etc., and significantly improves the accuracy and efficiency of path selection.

[0236] Assume that at a large fire scene, a mobile emergency station needs to support high-definition video streaming, voice calls, and the transmission of a large amount of medical data at the same time. In order to ensure the efficiency and reliability of data transmission for these critical tasks, the above formula scheme is used to develop and train intelligent redundant path planning.

[0237] The following are the implementation steps and calculation examples:

[0238] Assume that at a large fire scene, a mobile emergency station needs to support high-definition video streaming, voice calls, and the transmission of a large amount of medical data at the same time. In order to ensure the efficiency and reliability of data transmission for these critical tasks, the above formula scheme is used to develop and train intelligent redundant path planning.

[0239] Analyze the current network topology and node status, pre-select one or more potential redundant paths as backup options, and obtain the pre-selected redundant path list P redundant ={p1, p2, ..., p i}.

[0240] Assume that three potentially redundant paths are selected: p1, p2, and p3.

[0241] Calculate the maximum flow C of the network max :

[0242]

[0243] Assume that after calculation, we get C max =50Mbps. This means that at any split point in the network, the maximum possible flow from source to destination is 50Mbps.

[0244] According to the redundant path switching strategy, combined with the dynamic resource allocation strategy, the bandwidth availability on the redundant path is evaluated. i and connection priority P i , based on task priority and bandwidth demand prediction in the future future , adjust the resource allocation of redundant paths and generate the optimized redundant path configuration R i :

[0245] R i =f(B i , P i , D future )

[0246] Assume that for path p1, B1 = 40 Mbps, P1 = 0.8 (high priority), D future =35Mbps, then R1 may be set to allocate a bandwidth of 35Mbps.

[0247] The ant colony optimization algorithm is used to continuously optimize redundant paths, find the best backup path from the source to the destination through iterative search, and generate an optimized backup path list.

[0248] Update path quality Q(t):

[0249] Q(t+1)=Q(t)+ΔQ(t)

[0250] Assuming that the initial path quality Q(0) = 0.7, and the path quality increment ΔQ(0) = 0.1 after one iteration, the updated path quality is:

[0251] Q(1)=0.8.

[0252] Update pheromone concentration τ ij (t):

[0253] τ ij (t+1)=(1-ρ)·τ ij (T)+Δτ ij (T)

[0254] Assuming the pheromone evaporation rate ρ = 0.1, the initial pheromone concentration τ ij (0) = 0.5, the amount of pheromone increased after one iteration Δτ ij (0) = 0.2, then the updated pheromone concentration τ ij (1) = (1 - 0.1) · 0.5 + 0.2 = 0.65.

[0255] When it is detected that the performance of the main path has degraded or there is a potential risk, the intelligent model automatically activates redundant paths, re-evaluates and optimizes path selection based on the latest network status, and finally generates a redundant path planning solution.

[0256] Assuming that the performance of the primary path degrades, the system automatically activates the redundant path p1 and re-evaluates other paths, such as p2 and p3, based on the latest network status and prediction results to ensure that the optimal path is selected.

[0257] Through the above calculations, a group of potential redundant paths can be pre-selected by analyzing the network topology and node status to ensure that there are alternative paths available when the main path fails. max =50Mbps indicates the maximum carrying capacity of the network and helps identify key nodes and links that may cause network interruptions. By evaluating the bandwidth availability and connection priority of redundant paths and combining future bandwidth demand forecasts, resources are reasonably allocated to ensure the effective use of redundant paths. By continuously updating path quality and pheromone concentration, the ant colony optimization algorithm can iteratively search for the best backup path, improving the accuracy and efficiency of path selection. When the performance of the main path is detected to be degraded or there is a potential risk, the intelligent model can automatically activate the redundant path and re-evaluate and optimize the path selection based on the latest network status to ensure the continuity and reliability of data transmission.

[0258] Overall, the solution effectively improves the adaptability and stability of intelligent redundant path planning by introducing graph theory and ant colony optimization algorithm, ensuring that the optimal path can always be found in a complex and changing network environment, thereby significantly improving the efficiency and reliability of data transmission and providing strong technical support for emergency rescue.

[0259] Figure 2 A structural diagram of a wireless communication network system of a mobile emergency station is provided in an embodiment of the present application. Figure 2 As shown, the system includes:

[0260] The monitoring and evaluation module 21 is used to monitor and evaluate the signal strength, interference level and channel occupancy rate in the surrounding environment in real time, select the optimal working frequency band, and adjust the frequency band switching strategy to obtain an adaptive multi-band access mechanism;

[0261] An adjustment and estimation module 22 is used to dynamically adjust the bandwidth and connection priority of the wireless communication network according to the optimal working frequency band determined by the adaptive multi-band access mechanism, evaluate the task priority, and estimate the bandwidth demand in the future period, while simulating the resource competition between different users and generating a dynamic resource allocation strategy;

[0262] Deployment synchronization module 23, which is used to deploy a distributed cache and synchronization system based on the task priority and bandwidth allocation defined in the dynamic resource allocation strategy, realize balanced distribution and rapid positioning of data, and use version vector clock to manage concurrent updates, while achieving cross-site data synchronization, reducing dependence on the central server, and generating a distributed temporary data cache pool;

[0263] The prediction and calculation module 24 is used to use the data in the distributed cache pool to establish an intelligent routing algorithm of deep reinforcement learning, which takes into account the current network topology and node status, predicts the network change trend, calculates the optimal transmission path from the source to the destination, and finally generates an intelligent routing planning solution;

[0264] The optimization generation module 25 is used to pre-set a redundant path on the path based on the optimal path information determined in the intelligent routing planning scheme, and ensure rapid switching to the best backup path in the event of a network failure through the maximum flow minimum cut theorem in the graph theory algorithm, and use the ant colony optimization algorithm to continuously optimize the path selection, and generate a redundant path planning scheme based on the dynamic resource allocation strategy and the results of the intelligent routing planning.

[0265] Figure 2 The wireless communication network system of the mobile emergency station can perform Figure 1 The implementation principle and technical effect of the wireless communication network method of the mobile emergency station described in the embodiment are not described in detail. The specific way in which each module and unit performs operations in the wireless communication network system of the mobile emergency station in the above embodiment has been described in detail in the embodiment of the method, and will not be described in detail here.

[0266] In one possible design, Figure 2 The wireless communication network system of the mobile emergency station of the embodiment shown can be implemented as a computing device, such as Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32;

[0267] The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 32 .

[0268] The processing component 32 is used to: monitor and evaluate the signal strength, interference level and channel occupancy rate in the surrounding environment in real time, select the optimal working frequency band, and adjust the frequency band switching strategy to obtain an adaptive multi-band access mechanism; dynamically adjust the bandwidth and connection priority of the wireless communication network according to the optimal working frequency band determined by the adaptive multi-band access mechanism, evaluate the task priority, and estimate the bandwidth demand in the future, while simulating the resource competition between different users to generate a dynamic resource allocation strategy; based on the task priority and bandwidth allocation defined in the dynamic resource allocation strategy, deploy a distributed cache and synchronization system to achieve balanced distribution and rapid positioning of data, and use the version vector clock to manage concurrent updates, while achieving cross-site data synchronization, reducing dependence on central servers, and generating a distributed temporary data cache pool; using the data in the distributed cache pool, establishing an intelligent routing algorithm of deep reinforcement learning, the intelligent routing algorithm takes into account the current network topology and node status, predicts the trend of network changes, calculates the optimal transmission path from the source to the destination, and finally generates an intelligent routing planning scheme; based on the optimal path information determined in the intelligent routing planning scheme, a redundant path is pre-set on the path, and the maximum flow minimum cut theorem in the graph theory algorithm is used to ensure rapid switching to the best backup path in the event of a network failure, and the ant colony optimization algorithm is used to continuously optimize the path selection, and a redundant path planning scheme is generated based on the dynamic resource allocation strategy and the results of the intelligent routing planning.

[0269] The processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above method. Of course, the processing component may also be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components to perform the above method.

[0270] The storage component 31 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0271] Of course, the computing device may also include other components, such as input / output interfaces, display components, communication components, etc.

[0272] The input / output interface provides an interface between the processing component and the peripheral interface module, which may be an output device, an input device, etc.

[0273] The communication component is configured to facilitate, among other things, wired or wireless communications between the computing device and other devices.

[0274] Among them, the computing device can be a physical device or an elastic computing host provided by a cloud computing platform, etc. In this case, the computing device can refer to a cloud server, and the above-mentioned processing components, storage components, etc. can be basic server resources rented or purchased from the cloud computing platform.

[0275] The present application also provides a computer storage medium storing a computer program, wherein the computer program can achieve the above-mentioned Figure 1 A wireless communication network method for a mobile emergency station according to the illustrated embodiment.

[0276] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0277] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0278] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0279] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A wireless communication network method for a mobile emergency station, characterized in that: include: Real-time monitoring and evaluation of the signal strength, interference level and channel occupancy in the surrounding environment, selection of the optimal operating frequency band, and adjustment of the frequency band switching strategy to obtain an adaptive multi-band access mechanism; According to the optimal working frequency band determined by the adaptive multi-band access mechanism, the bandwidth and connection priority of the wireless communication network are dynamically adjusted, the task priority is evaluated, and the bandwidth demand in the future is estimated, and resource competition between different users is simulated to generate a dynamic resource allocation strategy; Based on the task priorities and bandwidth allocation defined in the dynamic resource allocation strategy, a distributed cache and synchronization system is deployed to achieve balanced distribution and rapid positioning of data, and concurrent updates are managed using version vector clocks. At the same time, cross-site data synchronization is achieved, reducing dependence on central servers and generating a distributed temporary data cache pool. Using the data in the distributed cache pool, an intelligent routing algorithm based on deep reinforcement learning is established. The intelligent routing algorithm takes into account the current network topology and node status, predicts the network change trend, calculates the optimal transmission path from the source to the destination, and finally generates an intelligent routing planning solution; Based on the optimal path information determined in the intelligent routing planning scheme, a redundant path is pre-set on the path, and the maximum flow minimum cut theorem in the graph theory algorithm is used to ensure rapid switching to the best backup path in the event of a network failure. The ant colony optimization algorithm is used to continuously optimize the path selection, and a redundant path planning scheme is generated based on the dynamic resource allocation strategy and the results of intelligent routing planning.

2. The method according to claim 1, characterized in that The data in the distributed cache pool is used to establish an intelligent routing algorithm of deep reinforcement learning. The intelligent routing algorithm takes into account the current network topology and node status, predicts the network change trend, calculates the optimal transmission path from the source to the destination, and finally generates an intelligent routing planning solution, including: Utilizing the multi-source heterogeneous data of historical communication records, node status information, and network topology structure in the distributed cache pool, the multi-source heterogeneous data is cleaned, converted, and normalized to obtain feature vectors suitable for deep learning model input, thereby generating an optimized feature data set; Based on the optimized feature data set, a hybrid deep reinforcement learning model is constructed that combines a convolutional neural network for spatial feature extraction and a recurrent neural network for time series analysis. The hybrid deep reinforcement learning model can capture the spatial correlation of network topology and the trend of changes over time, and generate an intelligent decision-making framework suitable for dynamic network environments; According to the intelligent decision-making framework suitable for dynamic network environment, the proximal policy optimization algorithm is adopted as the core of the reinforcement learning framework, and the hybrid deep learning model is trained in combination with the reward mechanism in the actual network environment, so that the hybrid deep learning model can be continuously optimized through interaction with the simulation environment, learn the ability to select the optimal path in different situations, and finally generate a trained intelligent model; The trained intelligent model is used to calculate the optimal transmission path from the source to the destination based on the current network status and the prediction of the network change trend that may occur in the future, and finally generate an intelligent routing planning solution.

3. The method according to claim 2, characterized in that Based on the optimized feature data set, a hybrid deep reinforcement learning model is constructed that combines a convolutional neural network for spatial feature extraction and a recurrent neural network for time series analysis. The hybrid deep reinforcement learning model can capture the spatial correlation of network topology and the trend of change over time, and generate an intelligent decision-making framework suitable for dynamic network environments, including: The optimized feature data set is used to model the spatial distribution of network topology and node status, and a convolutional neural network is used to perform multi-layer convolution operations on the feature data set to capture the spatial correlation and local patterns between network nodes and obtain spatial feature representation; Based on the spatial feature representation, the time series information of the node status is further combined, and a recurrent neural network is used to process the network status data that changes over time to capture the temporal dependency and dynamic change trend, and generate a time series feature representation; Based on the spatial feature representation and the time series feature representation, a hybrid model is designed, which can process information in spatial and temporal dimensions at the same time, and an attention mechanism is introduced to enable the hybrid model to focus on important spatiotemporal features, thereby improving the accuracy and efficiency of decision-making and obtaining an intelligent decision-making framework that integrates spatiotemporal features; An intelligent decision-making system suitable for dynamic network environments is constructed by using an intelligent decision-making framework that integrates spatiotemporal features. The intelligent decision-making system takes into account the spatial layout and node status of the current network topology, predicts the network change trend in the future, provides strong support for intelligent routing algorithms, and generates an intelligent decision-making framework suitable for dynamic network environments.

4. The method according to claim 2, characterized in that: According to the intelligent decision-making framework suitable for dynamic network environment, the proximal strategy optimization algorithm is adopted as the core of the reinforcement learning framework, and the hybrid deep learning model is trained in combination with the reward mechanism in the actual network environment, so that the hybrid deep learning model can be continuously optimized through interaction with the simulation environment, learn the ability to select the optimal path in different situations, and finally generate a trained intelligent model, including: Based on an intelligent decision-making framework suitable for dynamic network environments, a reinforcement learning framework with a proximal policy optimization algorithm as the core is designed. The reinforcement learning framework continuously adjusts and optimizes model parameters through interaction with the simulation environment to ensure that the model can make the best path selection in a complex and changing network environment. Based on the intelligent decision-making framework, a reward mechanism that meets the communication characteristics of the mobile emergency station is defined. The reward mechanism includes minimizing transmission delay, reducing packet loss rate, and maximizing bandwidth utilization. It is used to evaluate the performance of the model in different scenarios and guide the learning direction of the intelligent model to obtain the optimization objective function. Using the optimization objective function, a simulation environment is constructed, so that the hybrid deep learning model can interact with various possible network conditions in the simulation environment. In each interaction, the deep learning model takes actions according to the current network state and receives corresponding reward feedback to accumulate experience data. Based on the empirical data, the model parameters are updated using a proximal strategy optimization algorithm. The proximal strategy optimization algorithm avoids the performance instability problem caused by excessive update amplitude by limiting the pace of parameter update, thereby ensuring the stability and convergence of the deep learning model during training, and generating a preliminarily trained intelligent model; By using the initially trained intelligent model to continue interacting with the simulation environment, the model gradually learns how to select the optimal path in different scenarios, giving priority to low-latency paths under high load conditions, and ultimately generates a trained intelligent model.

5. The method according to claim 2, characterized in that: The trained intelligent model is used to calculate the optimal transmission path from the source to the destination according to the current network status and the prediction of the network change trend that may occur in the future, and finally generate an intelligent routing planning solution, including: Based on the trained intelligent model, real-time monitoring and data collection of the current network status are performed, and the data includes node load, connection quality, and traffic pattern to obtain the latest network status information; Using the latest network status information and the time series analysis capability embedded in the intelligent model, the change trend of the network in the future is predicted to generate a network change trend prediction result; According to the network change trend prediction result, the intelligent model evaluates multiple transmission paths, calculates the expected performance index of each path by comprehensively considering the node status on the path, the expected network change trend and the actual transmission demand, and obtains the path performance evaluation result; Based on the path performance evaluation results, the intelligent model selects a path that meets the transmission requirements and has the best performance as the main path, giving priority to paths that can maintain efficient transmission under high load or network congestion conditions, and generates the optimal transmission path; For the optimal transmission path, the intelligent model starts a dynamic monitoring mechanism to continuously monitor the real-time performance of the optimal transmission path, including node status, link quality, and traffic changes, to obtain dynamic monitoring results; Using the dynamic monitoring results, the intelligent model evaluates the stability and reliability of the existing transmission path. When it is detected that the performance of the transmission path is degraded or there is a potential risk, the intelligent model automatically re-evaluates other transmission paths and obtains an updated optimal transmission path based on the latest network status and prediction results; Based on the updated optimal transmission path, an intelligent routing planning scheme is finally generated. The intelligent routing planning scheme includes the optimal transmission path from the source to the destination and provides a mechanism that can dynamically adapt to network changes to ensure the continuity and reliability of data transmission.

6. The method according to claim 1, characterized in that The optimal path information determined in the intelligent routing planning scheme is pre-set on the path, and the maximum flow minimum cut theorem in the graph theory algorithm is used to ensure rapid switching to the best backup path in the event of a network failure, and the ant colony optimization algorithm is used to continuously optimize the path selection. Based on the dynamic resource allocation strategy and the results of the intelligent routing planning, a redundant path planning scheme is generated, including: Analyze the current network topology and node status using the optimal path information determined in the intelligent routing planning solution, and pre-select one or more potential redundant paths as backup options according to the characteristics of the optimal path to obtain a pre-selected redundant path list; Based on the pre-selected redundant path list, the maximum flow minimum cut theorem in the graph theory algorithm is applied to identify the key nodes and links that cause network interruption by calculating the maximum flow of the network, so as to ensure that the redundant path can be quickly switched when the main path fails, and generate a redundant path switching strategy; According to the redundant path switching strategy, combined with the dynamic resource allocation strategy, the bandwidth availability and connection priority on the redundant path are evaluated, and according to the task priority and the bandwidth demand forecast in the future period, the resource allocation of the redundant path is adjusted to generate an optimized redundant path configuration; Using the optimized redundant path configuration, continuously optimizing the redundant path using an ant colony optimization algorithm, wherein the ant colony optimization algorithm finds the best backup path from the source to the destination through iterative search, and generates an optimized backup path list; Based on the optimized backup path list, according to the results of the dynamic resource allocation strategy and the intelligent routing planning solution, the network change trend and real-time performance monitoring data are comprehensively considered, and the configuration of the redundant path is continuously adjusted. When it is detected that the performance of the main path has degraded or there is a potential risk, the intelligent model automatically activates the redundant path, and re-evaluates and optimizes the path selection according to the latest network status to generate a redundant path planning solution.

7. The method according to claim 6, characterized in that Based on the optimized backup path list, according to the dynamic resource allocation strategy and the result of the intelligent routing planning solution, the configuration of the redundant path is continuously adjusted by comprehensively considering the network change trend and real-time performance monitoring data. When the performance of the primary path is detected to be degraded or there is a potential risk, the intelligent model automatically activates the redundant path, and re-evaluates and optimizes the path selection according to the latest network status, and generates a redundant path planning solution, including: Based on the optimized backup path list, combined with the dynamic resource allocation strategy and the results of the intelligent routing planning solution, predict the network change trend to obtain a network change trend prediction result; Using the network change trend prediction result, pre-adjusting the redundant path to ensure that the redundant path has sufficient communication capacity when needed, and generating a pre-adjusted redundant path configuration; According to the pre-adjusted redundant path configuration, start a real-time performance monitoring mechanism, continuously monitor the performance of the primary path and the redundant path, and collect real-time performance monitoring data; Based on the real-time performance monitoring data, the intelligent model evaluates the health status of the primary path. If it is detected that the performance of the primary path is degraded or there is a potential risk, the best backup path is immediately activated to obtain the activated redundant path; Using the activated redundant path, the intelligent model continues to monitor the performance of the redundant path, and further optimizes the path in combination with the ant colony optimization algorithm to ensure the high efficiency and stability of the redundant path, and generates an optimized redundant path configuration; Finally, a redundant path planning scheme is generated, which includes preset redundant paths and their switching strategies, as well as the optimized activated redundant path configuration, to ensure that efficient communication capabilities can be maintained when network conditions change.

8. A wireless communication network system for a mobile emergency station, characterized in that: include: The monitoring and evaluation module is used to monitor and evaluate the signal strength, interference level and channel occupancy rate in the surrounding environment in real time, select the optimal working frequency band, and adjust the frequency band switching strategy to obtain an adaptive multi-band access mechanism; An adjustment and estimation module is used to dynamically adjust the bandwidth and connection priority of the wireless communication network according to the optimal working frequency band determined by the adaptive multi-band access mechanism, evaluate the task priority, and estimate the bandwidth demand in the future period of time, while simulating the resource competition between different users and generating a dynamic resource allocation strategy; Deployment synchronization module, which is used to deploy distributed cache and synchronization system based on the task priority and bandwidth allocation defined in the dynamic resource allocation strategy, realize balanced distribution and rapid positioning of data, and use version vector clock to manage concurrent updates, while achieving cross-site data synchronization, reducing dependence on central servers, and generating distributed temporary data cache pools; A prediction and calculation module is used to use the data in the distributed cache pool to establish an intelligent routing algorithm of deep reinforcement learning. The intelligent routing algorithm takes into account the current network topology and node status, predicts the network change trend, calculates the optimal transmission path from the source to the destination, and finally generates an intelligent routing planning solution; The optimization generation module is used to pre-set a redundant path on the path based on the optimal path information determined in the intelligent routing planning scheme, and ensure rapid switching to the best backup path in the event of a network failure through the maximum flow minimum cut theorem in the graph theory algorithm, and use the ant colony optimization algorithm to continuously optimize the path selection, and generate a redundant path planning scheme based on the dynamic resource allocation strategy and the results of intelligent routing planning.

9. A computing device, characterized in that It comprises a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a wireless communication network method for a mobile emergency station as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that: A computer program is stored, and when the computer program is executed by a computer, a wireless communication network method for a mobile emergency station as claimed in any one of claims 1 to 7 is implemented.

Images