Intelligent auxiliary and artificial intelligence visual gateway control method and system for station building

Abstract

The invention relates to the technical field of power distribution station supervision, in particular to an intelligent auxiliary and artificial intelligence visual gateway control method and system for a station building, which monitors network state parameters in real time, calculates a network quality score and dynamically selects an optimal transmission path based on a reinforcement learning algorithm. Comprising a high bandwidth mode, a low delay mode and a disaster recovery backup mode, stage processing and dynamic compression are performed on transmission data according to data types and network quality scores, link switching and a local caching mechanism are automatically triggered when a network is abnormal, and finally, multi-gateway load balancing and visual monitoring are realized through a software defined network architecture. The problem of transmission delay or interruption caused by the fact that a static data transmission framework cannot adapt to network state changes in the prior art is effectively solved, and the reliability and the real-time performance of data transmission are remarkably improved through dynamic path optimization, intelligent hierarchical transmission and a fault self-healing mechanism.

Description

Technical Field

[0001] The present invention relates to the technical field of power distribution station supervision, and in particular to a method and system for controlling a power distribution station intelligent assistance and artificial intelligence visualization gateway. Background Art

[0002] Currently, the intelligent auxiliary systems in power distribution stations across various cities exhibit significant differences in data access methods, making it difficult to achieve efficient and unified data transmission and integration between these systems. Power distribution stations are widely distributed and numerous, each housing a wide variety of equipment, which generates vast amounts of data during operation. Managing such a complex data environment is extremely challenging, and the existing construction and operation model suffers from numerous shortcomings. Specifically, the existing system is unable to provide timely and effective early warning and monitoring of station environmental safety, posing a high safety risk and hindering ongoing maintenance and management.

[0003] In order to solve the above problems, the existing publication number CN113433882A discloses a station building intelligent assistance and artificial intelligence visualization gateway control method, which establishes an initial data transmission architecture by acquiring the equipment information and data transmission configuration information of each station building, and imports the equipment information into a preset classification model to generate data classification information. Subsequently, the initial data transmission architecture is adjusted to form a target data transmission architecture. On this basis, the equipment operation data of each station building is acquired and imported into the classification model in real time to generate real-time classification information, and based on this information, the operation data of various types of equipment are transmitted to the IOT platform or the artificial intelligence visualization platform. This method can achieve effective management of distribution station equipment data to a certain extent, improve data transmission efficiency, and thereby achieve effective supervision of each station building.

[0004] However, the existing methods of intelligent station assistance and artificial intelligence visualization gateway control have limitations in terms of the flexibility of the data transmission architecture. Specifically, when adjusting the initial architecture, this method is based only on static device information and configuration information, and does not fully consider real-time changes in network status, such as bandwidth fluctuations and node failures. When the transmission environment changes, such as sudden high loads, the target architecture may not be able to dynamically optimize the transmission path or switch channels, resulting in transmission delays or even interruptions. Under this static data transmission architecture, facing a complex network environment, it is difficult to ensure the efficiency and reliability of data transmission, which in turn affects the overall performance of the system and user experience. Therefore, how to improve the flexibility of the data transmission architecture so that it can dynamically adapt to changes in network status has become an urgent problem to be solved. Summary of the Invention

[0005] The purpose of the present invention is to provide a station building intelligent assistance and artificial intelligence visualization gateway control method and system, which solves the problem of transmission delay or interruption caused by network status changes in the prior art.

[0006] To achieve the above objectives, the present invention provides a station building intelligent assistance and artificial intelligence visualization gateway control method, comprising the following steps:

[0007] Monitor network status parameters in real time and calculate network quality scores;

[0008] Based on the network quality score, the optimal transmission path is dynamically selected through the reinforcement learning algorithm;

[0009] Classify the transmitted data according to the data type and network quality score, and dynamically adjust the data compression rate and transmission priority;

[0010] When a network anomaly is detected, the link switching mechanism is automatically triggered and local data caching is started until the network is restored;

[0011] Achieve load balancing among multiple gateways through software-defined network architecture, and display the network topology status in real time on a visualization platform.

[0012] The network status parameters are monitored in real time and the network quality score is calculated. The specific steps include:

[0013] Periodically collect real-time bandwidth data of the transmission link, response delay data of each communication node, and signal-to-noise ratio of the wireless channel;

[0014] Perform sliding window smoothing on the collected data to eliminate instantaneous fluctuation interference;

[0015] Calculate the bandwidth stability coefficient, node health index and channel quality index respectively;

[0016] The calculated indicators are weighted and integrated to generate a comprehensive network quality score.

[0017] Among them, based on the network quality score, the optimal transmission path is dynamically selected through the reinforcement learning algorithm.

[0018] The transmission paths include high-bandwidth mode, low-latency mode, and disaster recovery backup mode.

[0019] Based on the network quality score, the optimal transmission path is dynamically selected through a reinforcement learning algorithm. The specific steps include:

[0020] The network quality score is used as the state input of the reinforcement learning model, and the state vector is constructed by combining the type characteristics of the current data to be transmitted and the service priority.

[0021] Select a transmission path from a preset action space based on the state vector;

[0022] The reward value is calculated based on the actual performance indicators after each transmission is completed, including latency improvement, data integrity rate and energy efficiency;

[0023] Use the reward value to update the parameters of the reinforcement learning model through the policy gradient algorithm;

[0024] When the network quality score changes significantly, re-evaluate and adjust the transmission path selection strategy.

[0025] The transmission data is graded based on the data type and network quality score, and the data compression rate and transmission priority are dynamically adjusted. The specific steps include:

[0026] Data to be transmitted is divided into three processing levels: Level 1 data includes real-time alarm signals and equipment failure information, uses a lossless compression algorithm and is assigned the highest transmission priority; Level 2 data includes periodic environmental monitoring data, uses a lossy compression algorithm based on differential coding and is assigned a medium transmission priority; Level 3 data includes video surveillance streams, and the H.265 encoding parameters are dynamically adjusted based on the network quality score;

[0027] Configure dynamic adjustment strategies for each processing level;

[0028] Deploy hierarchical buffer queues at the edge gateway and implement differentiated scheduling according to processing levels.

[0029] When a network anomaly is detected, the link switching mechanism is automatically triggered and local data caching is started until the network is restored. The specific steps include:

[0030] Continuously monitor the changing trends of network quality scores. When the score drops by more than 20% for three consecutive cycles, it is determined to be a network anomaly. Combined with the output of the reinforcement learning model, the reliability degradation degree of the current transmission path is predicted.

[0031] Implement hierarchical switching strategies based on the divided data processing levels;

[0032] Adopting hierarchical buffer queue architecture to implement differentiated caching;

[0033] When the network quality score recovers to above the threshold, the cached first-level data is transmitted first, the second-level data transmission is gradually restored according to the transmission quota, and the encoding parameters of the third-level data are reinitialized according to the current score.

[0034] The software-defined network architecture is used to achieve load balancing among multiple gateways, and the network topology status is displayed in real time on the visualization platform. The specific steps include:

[0035] Dynamically calculate the real-time load rate of each gateway node and generate a load balancing decision table based on the network quality score. According to the data classification strategy, the transmission resource allocation of the first-level data is prioritized. When the load of a single gateway exceeds the threshold, the task is migrated to the adjacent gateway according to the path selection strategy.

[0036] Collect network quality scores, transmission path status, and cache usage of each gateway node, build a dynamic topology map, and mark it with different colors;

[0037] The identified abnormal nodes and their impact range are marked on the visualization platform, and a manual adjustment interface is provided to allow operation and maintenance personnel to override the automatic path selection strategy, record data hierarchical transmission efficiency indicators, and generate load balancing optimization suggestions.

[0038] A station building intelligent assistance and artificial intelligence visualization gateway control system, comprising a network status monitoring module, a path decision module, a data hierarchical processing module, a fault self-healing module, and a visualization monitoring platform, wherein the path decision module is connected to the network status monitoring module, the data hierarchical processing module is respectively connected to the network status monitoring module and the path decision module, the fault self-healing module is respectively connected to the network status monitoring module, the path decision module, and the data hierarchical processing module, and the visualization monitoring platform is respectively connected to the path decision module, the fault self-healing module, and the data hierarchical processing module;

[0039] The network status monitoring module is used to monitor network status parameters in real time and calculate network quality scores;

[0040] The path decision module is used to dynamically select the optimal transmission path based on the network quality score through a reinforcement learning algorithm;

[0041] The data classification processing module is used to classify the transmission data according to the data type and network quality score, and dynamically adjust the data compression rate and transmission priority;

[0042] The fault self-healing module is used to automatically trigger the link switching mechanism when a network anomaly is detected, and start local data caching until the network is restored;

[0043] The visualization monitoring platform is used to achieve load balancing among multiple gateways through a software-defined network architecture and to display the network topology status in real time on the visualization platform.

[0044] The present invention provides a method and system for controlling intelligent station building assistance and artificial intelligence visualization gateways. By real-time monitoring of network status parameters and calculating network quality scores, the system dynamically selects the optimal transmission path based on a reinforcement learning algorithm, including high-bandwidth mode, low-latency mode, and disaster recovery backup mode. The system hierarchically processes and dynamically compresses transmission data according to data type and network quality score, and automatically triggers link switching and local caching mechanisms when the network is abnormal. Ultimately, the system achieves multi-gateway load balancing and visualization monitoring through a software-defined network architecture. This effectively addresses the transmission delay or interruption problem in the prior art caused by the inability of static data transmission architectures to adapt to changes in network status. Through dynamic path optimization, intelligent hierarchical transmission, and fault self-healing mechanisms, the system significantly improves the reliability and real-time nature of data transmission, achieving intelligent and adaptive management of distribution station equipment data. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art.

[0046] Figure 1 It is a step flow chart of the station building intelligent assistance and artificial intelligence visualization gateway control method of the first embodiment of the present invention.

[0047] Figure 2 It is a principle block diagram of the station building intelligent assistance and artificial intelligence visualization gateway control system of the second embodiment of the present invention.

[0048] In the figure: 201-network status monitoring module, 202-path decision module, 203-data classification processing module, 204-fault self-recovery module, 205-visual monitoring platform. DETAILED DESCRIPTION

[0049] The embodiments of the present invention are described in detail below. Examples of the embodiments are shown in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, but should not be understood as limiting the present invention.

[0050] The first embodiment of this application is:

[0051] See also Figure 1 ,in, Figure 1 It is a step flow chart of the station building intelligent assistance and artificial intelligence visualization gateway control method of the first embodiment of the present invention.

[0052] The present invention provides a station building intelligent assistance and artificial intelligence visualization gateway control method, comprising the following steps:

[0053] S101: monitor network status parameters in real time and calculate network quality score;

[0054] Specifically, real-time monitoring and assessment of network status is achieved through edge gateways, such as the lightweight probe module integrated into the Allwinner T3-C chip. During implementation, three key parameters are continuously collected at a 500ms cycle: the actual available bandwidth of the transmission link is measured using the ICMP protocol; heartbeat packets are sent to each communication node, including LoRa base stations and 4G modules, to record response delays; and the signal-to-noise ratio (SNR) of the wireless channel is calculated using the FFT transform to identify frequency band interference sources. To eliminate instantaneous fluctuations, the raw data is processed using a sliding window weighted average with a window size of 10 sampling periods and dynamically weighted using a time decay coefficient (0.9^(n-1)). The calculation formula is: bandwidth smoothing value = ∑(bandwidth sampling value × time decay coefficient) / window length, decay coefficient = 0.9^(n-1), where n is the current sampling sequence number. Based on preprocessed data, the system calculates three core metrics: the Bandwidth Stability Coefficient (BSC) reflects bandwidth fluctuations, the Node Health Index (NHI) assesses the status of communication nodes, and the Channel Quality Index (CQI) quantifies wireless channel conditions. The calculation formulas are: BSC = (1 - |Current Bandwidth - Historical Average| / Nominal Bandwidth) × 100%, NHI = 100 - (Maximum Delay - Baseline Delay) / Baseline Delay × 50 (NHI < 0 is forced to zero), and CQI = SNR / 20 + Spectral Purity Score × 0.3. Spectral purity is calculated based on the proportion of harmonic components. These three indicators are weighted together (with weights of 50%, 30%, and 20%, respectively) to generate a comprehensive Network Quality Score (NQS) ranging from 0 to 100. The calculation formula is: NQS = BSC × 0.5 + NHI × 0.3 + CQI × 0.2. The score is then divided into three levels: excellent (≥80 points), good (60-79 points), and poor (<60 points), with corresponding green, yellow, and red indicators, respectively. Compared to traditional single-parameter assessments, this multi-dimensional assessment approach offers a more comprehensive monitoring perspective and stronger anti-interference capabilities, providing a precise and reliable basis for subsequent intelligent path selection and data classification processing.

[0055] S102: Based on the network quality score, the optimal transmission path is dynamically selected through a reinforcement learning algorithm;

[0056] Specifically, first, a multi-dimensional reinforcement learning model is constructed, whose state space consists of three core elements: (1) the current network quality score (NQS) score, which is obtained by real-time calculation in step S101; (2) the type of data to be transmitted, which is specifically divided into three categories: real-time alarm data (such as equipment failure signals), periodic monitoring data (such as temperature and humidity readings), and video streaming data; (3) business priority indicators, which are divided into levels 0-5 according to the criticality of the data. After normalization, these state parameters together constitute a 128-dimensional state feature vector. Secondly, an action space containing three types of transmission modes is defined: (1) high bandwidth mode: using 4G network or Ethernet channel, with a maximum support of 100Mbps transmission rate, suitable for large-volume data such as video streaming; (2) low latency mode: using LoRa or Bluetooth 5.0 communication, with end-to-end latency controlled within 50ms, dedicated to real-time alarm signal transmission; (3) disaster recovery backup mode: enabling Beidou RDSS short message service under extreme network conditions. Although the bandwidth only supports 1kbps, it can ensure the accessibility of critical data. In the model training phase, the proximal policy optimization (PPO) algorithm is used for policy iteration. After each transmission is completed, a compound reward value R = 0.5Δt + 0.3η + 0.2ε is calculated, where Δt represents the delay improvement (the ratio of the current transmission delay to the baseline value), η represents the data integrity rate (the number of successfully received bytes / the number of sent bytes), and ε is the energy efficiency (the inverse of the energy consumption per unit data volume). The reward value is calculated by time difference with a discount factor of γ = 0.9 to guide the update of the policy network parameters. In order to cope with the dynamic changes in the network environment, a dual trigger mechanism is set: (1) when the NQS score changes by more than 15%; (2) when the reward value of three consecutive transmissions is lower than the threshold. If any of the conditions are met, the policy re-evaluation process is triggered, and the system will resample the network status and calculate the optimal transmission path.

[0057] S103: performing hierarchical processing on the transmitted data according to the data type and network quality score, and dynamically adjusting the data compression rate and transmission priority;

[0058] Specifically, a scientific three-level data processing system was established, with differentiated processing strategies adopted at each level: (1) First-level critical alarm data: including key information such as equipment failure signals, fire alarms, and flooding alarms, using the LZMA lossless compression algorithm (compression level set to 9) to ensure that the data is 100% complete and not lost, and giving it the highest transmission priority (priority 5, preemptive transmission); (2) Second-level periodic monitoring data: including equipment operating parameters such as temperature, humidity, voltage, and current, using the optimized Delta coding lossy compression algorithm, and using adaptive quantization technology to intelligently control the compression ratio between 3:1 and 5:1 (error range ±0.5%), giving it a medium priority (priority 5, preemptive transmission); 3, weighted fair queue scheduling); (3) Level 3 video surveillance data stream: adopt the dynamically adjustable H.265 / HEVC encoding scheme, and implement three-level quality control according to the real-time NQS score: NQS ≥ 80 (high-quality network): adopt CRF18 high-quality mode, GOP structure is IBP, quantization parameter QP = 26; 60 ≤ NQS < 80 (good network): adopt CRF23 balanced mode, GOP structure is IBBP, QP = 32; NQS < 60 (poor network): enable CRF28 high compression mode, QP = 38, and supplemented by intelligent frame extraction technology, the frame rate is gradually reduced from 30fps to 15fps (every 5fps reduction interval is 30 seconds). An intelligent hierarchical buffer queue system is deployed at the edge gateway, adopting the following management strategies: (1) First-level data: 20% of the cache space is exclusively occupied (the default configuration is 256MB), strict FIFO queue management is adopted, and end-to-end transmission delay is ensured to be <100ms (99.9% percentile) through hardware acceleration; (2) Second-level data: 50% of the space is allocated (640MB), and an optimized batch packaging transmission mechanism is used (every 10 data are packaged, with a header compression rate of 50%), and the packaging timeout is set to 200ms; (3) Third-level data: 30% of the space is occupied (384MB), and a dynamic traffic shaping algorithm based on TCP-Friendly is adopted. When the NQS decreases, the bandwidth occupancy rate is dynamically adjusted according to the formula BW = BaseBW × (NQS / 100)^2. The following enhancements have also been implemented: (1) Dynamic priority promotion mechanism: When an abnormal event is detected, the relevant monitoring data can be temporarily promoted to the first level of processing; (2) Compression parameter self-learning: Automatically optimize the parameter configuration of each compression algorithm through historical data analysis; (3) Cache intelligent prefetching: Pre-load the monitoring video clips that may be needed based on the access pattern prediction.

[0059] S104: When a network anomaly is detected, a link switching mechanism is automatically triggered, and local data caching is started until the network is restored;

[0060] Specifically, a multi-index composite criterion is used to detect network anomalies: when the network quality score (NQS) drops by more than 20% for three consecutive cycles, or when node latency increases by more than 300%, or when bandwidth availability drops below 10% for five seconds, it is considered a network anomaly. Upon detecting an anomaly, the system immediately initiates a three-level emergency response mechanism: First, based on the data classification system established in step S103, a differentiated link switching strategy is implemented: for first-level critical alarm data, the system immediately switches to the pre-configured BeiDou RDSS satellite link. Although this link has limited bandwidth (1.2kbps), it ensures the accessibility of critical alarm information; second-level monitoring data is switched to the LoRa self-organizing network backup channel, which uses forward error correction (FEC) encoding to enhance reliability; and third-level video data is suspended from transmission and enters the local cache state. Secondly, the system starts the intelligent cache management module, which includes the following key technologies: adopting a three-level cache architecture: the first level uses SRAM to cache critical alarm data (capacity 8MB, access time <10μs); the second level uses SLC NAND flash memory to store monitoring data (capacity 256MB, write speed 50MB / s); the third level uses TLC SSD to cache video data (capacity 1TB, support parallel reading and writing); implementing a dynamic cache strategy: automatically adjust the cache time window according to the data type and NQS prediction value, cache critical data for 72 hours, cache monitoring data for 24 hours, and the video data cache duration is dynamically calculated according to the formula T = 12 × (100-NQS) hours; enabling compressed cache technology: implementing H.265 lightweight recompression (CRF32) on video data to improve cache efficiency by 40%. When network recovery conditions are met (NQS rises to above 70 and remains so for 3 minutes), the system performs orderly data return: the return is based on the priority order of "critical data → latest monitoring data → historical video data"; a bandwidth-adaptive transmission algorithm is used to ensure that the return process does not affect normal business; data integrity verification (SHA-256) is implemented, and automatic retransmission occurs in the event of failure.

[0061] S105: Implement load balancing among multiple gateways through a software-defined network architecture, and display the network topology status in real time on a visualization platform.

[0062] Specifically, a dynamic load scheduling engine is built based on a centralized SDN controller. Every 30 seconds, the following real-time metrics are collected: CPU utilization (weighted moving average), memory usage (calculated based on / proc / meminfo), network throughput (sampled via sFlow), and the length of the current transmission task queue for each gateway node. A genetic algorithm is used to calculate the optimal load distribution solution, with the fitness function F = 0.4 × (1-CPU_util) + 0.3 × (1-Mem_util) + 0.2 × Throughput + 0.1 × (1-Queue_len). When a node load exceeds a threshold (CPU > 80% for 1 minute), the following actions are automatically performed: Third-level video data tasks are prioritized according to the data classification policy in step S104. VXLAN tunneling technology is used to achieve seamless service migration. OpenFlow flow entries are dynamically adjusted, and forwarding paths are updated. The core functions of the visual monitoring platform 205 include: dynamic topology display: node status is indicated by three-color identification (green: NQS ≥ 80, yellow: 60 ≤ NQS < 80, red: NQS < 60), real-time display of link quality indicators (bandwidth, latency, packet loss rate), automatic highlighting of abnormal nodes (flashing warning), in-depth analysis function: load heat map display (based on D3.js visualization), historical performance trend analysis (support 7-day backtracking), intelligent early warning (based on LSTM network prediction), management and control interface: support manual adjustment of load strategy (slider adjustment weight parameter), provide emergency channel forced opening function, and can export detailed operation reports (PDF / Excel format), and ensure efficient system operation through the following technologies: use DPDK to accelerate data plane forwarding, achieve high availability of controller cluster (Raft protocol), and display actual deployment data.

[0063] By monitoring network status parameters in real time and calculating network quality scores, the system dynamically selects the optimal transmission path based on a reinforcement learning algorithm, including high-bandwidth mode, low-latency mode, and disaster recovery backup mode. Transmitted data is hierarchically processed and dynamically compressed based on data type and network quality score. Link switching and local caching mechanisms are automatically triggered in the event of network anomalies, ultimately achieving multi-gateway load balancing and visual monitoring through a software-defined network architecture. This effectively addresses the transmission delay or interruption issues inherent in existing technologies, caused by the inability of static data transmission architectures to adapt to changes in network status. Through dynamic path optimization, intelligent hierarchical transmission, and fault self-healing mechanisms, the system significantly improves the reliability and real-time nature of data transmission, enabling intelligent and adaptive management of distribution station equipment data.

[0064] The second embodiment of this application is:

[0065] Based on the first embodiment, please refer to Figure 2 ,in, Figure 2It is a principle block diagram of the station building intelligent assistance and artificial intelligence visualization gateway control system of the second embodiment of the present invention.

[0066] The station building intelligent assistance and artificial intelligence visualization gateway control system of this embodiment includes a network status monitoring module 201, a path decision module 202, a data hierarchical processing module 203, a fault self-healing module 204 and a visualization monitoring platform 205.

[0067] In this specific embodiment, the path decision module 202 is connected to the network status monitoring module 201, the data hierarchical processing module 203 is connected to the network status monitoring module 201 and the path decision module 202 respectively, the fault self-healing module 204 is connected to the network status monitoring module 201, the path decision module 202 and the data hierarchical processing module 203 respectively, and the visual monitoring platform 205 is connected to the path decision module 202, the fault self-healing module 204 and the data hierarchical processing module 203 respectively;

[0068] The network status monitoring module 201 is used to monitor network status parameters in real time and calculate network quality scores;

[0069] The path decision module 202 is used to dynamically select the optimal transmission path based on the network quality score through a reinforcement learning algorithm;

[0070] The data classification processing module 203 is used to classify the transmission data according to the data type and network quality score, and dynamically adjust the data compression rate and transmission priority;

[0071] The fault self-healing module 204 is used to automatically trigger a link switching mechanism when a network anomaly is detected, and start local data caching until the network is restored;

[0072] The visualization monitoring platform 205 is used to achieve load balancing among multiple gateways through a software-defined network architecture and to display the network topology status in real time on the visualization platform.

[0073] Using a station house intelligent assistance and artificial intelligence visualization gateway control system of this embodiment, intelligent operation and maintenance is achieved through the collaboration of multiple modules: the network status monitoring module 201 collects and evaluates network quality parameters in real time, generates a network quality score and transmits it to the path decision module 202; the path decision module 202 dynamically selects the optimal transmission path (high bandwidth / low latency / disaster recovery mode) based on the reinforcement learning algorithm, and synchronizes the decision result to the data classification processing module 203; the data classification processing module 203 implements three-level differentiated processing (critical alarm / periodic monitoring / video stream) according to the data type and network quality; when a network anomaly is detected, the fault self-healing module 204 immediately starts the link switching and hierarchical caching mechanism; the visualization monitoring platform 205 integrates the data of each module, realizes load balancing through the SDN architecture, and displays the status of the entire network in three dimensions. It effectively solves the problems of slow response and poor reliability of traditional station house monitoring systems.

[0074] The above disclosure is merely one or more preferred embodiments of the present application and is not intended to limit the scope of the present application. A person skilled in the art will understand that all or part of the processes of the above embodiments and equivalent changes made in accordance with the claims of the present application are still within the scope of the present application.

Claims

1. A station building intelligent assistance and artificial intelligence visualization gateway control method, characterized by comprising the following steps: Monitor network status parameters in real time and calculate network quality scores; Based on the network quality score, the optimal transmission path is dynamically selected through the reinforcement learning algorithm; Classify the transmitted data according to the data type and network quality score, and dynamically adjust the data compression rate and transmission priority; When a network anomaly is detected, the link switching mechanism is automatically triggered and local data caching is started until the network is restored; Achieve load balancing among multiple gateways through software-defined network architecture, and display the network topology status in real time on a visualization platform.

2. The station building intelligent assistance and artificial intelligence visualization gateway control method according to claim 1 is characterized in that: Monitor network status parameters in real time and calculate network quality scores. The specific steps include: Periodically collect real-time bandwidth data of the transmission link, response delay data of each communication node, and signal-to-noise ratio of the wireless channel; Perform sliding window smoothing on the collected data to eliminate instantaneous fluctuation interference; Calculate the bandwidth stability coefficient, node health index and channel quality index respectively; The calculated indicators are weighted and integrated to generate a comprehensive network quality score.

3. The station building intelligent assistance and artificial intelligence visualization gateway control method according to claim 2, characterized in that: Based on the network quality score, the optimal transmission path is dynamically selected through the reinforcement learning algorithm. The transmission paths include high-bandwidth mode, low-latency mode, and disaster recovery backup mode.

4. The station building intelligent assistance and artificial intelligence visualization gateway control method according to claim 3 is characterized in that: Based on the network quality score, the optimal transmission path is dynamically selected through a reinforcement learning algorithm. The specific steps include: The network quality score is used as the state input of the reinforcement learning model, and the state vector is constructed by combining the type characteristics of the current data to be transmitted and the service priority. Select a transmission path from a preset action space based on the state vector; The reward value is calculated based on the actual performance indicators after each transmission is completed, including latency improvement, data integrity rate and energy efficiency; Use the reward value to update the parameters of the reinforcement learning model through the policy gradient algorithm; When the network quality score changes significantly, re-evaluate and adjust the transmission path selection strategy.

5. The station building intelligent assistance and artificial intelligence visualization gateway control method according to claim 4 is characterized in that: The transmission data is graded based on the data type and network quality score, and the data compression rate and transmission priority are dynamically adjusted. The specific steps include: Data to be transmitted is divided into three processing levels: Level 1 data includes real-time alarm signals and equipment failure information, uses a lossless compression algorithm and is assigned the highest transmission priority; Level 2 data includes periodic environmental monitoring data, uses a lossy compression algorithm based on differential coding and is assigned a medium transmission priority; Level 3 data includes video surveillance streams, and the H.265 encoding parameters are dynamically adjusted based on the network quality score; Configure dynamic adjustment strategies for each processing level; Deploy hierarchical buffer queues at the edge gateway and implement differentiated scheduling according to processing levels.

6. The station building intelligent assistance and artificial intelligence visualization gateway control method according to claim 5, characterized in that: When a network anomaly is detected, the link switching mechanism is automatically triggered and local data caching is started until the network is restored. The specific steps include: Continuously monitor the changing trends of network quality scores. When the score drops by more than 20% for three consecutive cycles, it is determined to be a network anomaly. Combined with the output of the reinforcement learning model, the reliability degradation degree of the current transmission path is predicted. Implement hierarchical switching strategies based on the divided data processing levels; Adopting hierarchical buffer queue architecture to implement differentiated caching; When the network quality score recovers to above the threshold, the cached first-level data is transmitted first, the second-level data transmission is gradually restored according to the transmission quota, and the encoding parameters of the third-level data are reinitialized according to the current score.

7. The station building intelligent assistance and artificial intelligence visualization gateway control method according to claim 6, characterized in that: Achieve load balancing among multiple gateways through a software-defined network architecture and display the network topology status in real time on a visualization platform. The specific steps include: Dynamically calculate the real-time load rate of each gateway node and generate a load balancing decision table based on the network quality score. According to the data classification strategy, the transmission resource allocation of the first-level data is prioritized. When the load of a single gateway exceeds the threshold, the task is migrated to the adjacent gateway according to the path selection strategy. Collect network quality scores, transmission path status, and cache usage of each gateway node, build a dynamic topology map, and mark it with different colors; The identified abnormal nodes and their impact range are marked on the visualization platform, and a manual adjustment interface is provided to allow operation and maintenance personnel to override the automatic path selection strategy, record data hierarchical transmission efficiency indicators, and generate load balancing optimization suggestions.

8. A station building intelligent assistance and artificial intelligence visualization gateway control system, applicable to the station building intelligent assistance and artificial intelligence visualization gateway control method according to any one of claims 1 to 7, characterized in that: It includes a network status monitoring module, a path decision module, a data hierarchical processing module, a fault self-healing module and a visual monitoring platform, wherein the path decision module is connected to the network status monitoring module, the data hierarchical processing module is connected to the network status monitoring module and the path decision module respectively, the fault self-healing module is connected to the network status monitoring module, the path decision module and the data hierarchical processing module respectively, and the visual monitoring platform is connected to the path decision module, the fault self-healing module and the data hierarchical processing module respectively; The network status monitoring module is used to monitor network status parameters in real time and calculate network quality scores; The path decision module is used to dynamically select the optimal transmission path based on the network quality score through a reinforcement learning algorithm; The data classification processing module is used to classify the transmission data according to the data type and network quality score, and dynamically adjust the data compression rate and transmission priority; The fault self-healing module is used to automatically trigger the link switching mechanism when a network anomaly is detected, and start local data caching until the network is restored; The visualization monitoring platform is used to achieve load balancing among multiple gateways through a software-defined network architecture and to display the network topology status in real time on the visualization platform.

Images