Satellite communication method and device supporting fire rescue

Abstract

The invention provides a satellite communication method and device supporting fire rescue, and the method comprises the steps: carrying out the time alignment processing of real-time multi-source data collected from a fire rescue site and a pre-stored multi-source scene database, forming a time synchronization data set, and generating a terminal state data set and a scene feature data set; performing spatial geometric analysis based on the stored satellite orbit data and terminal position data, and generating available channel data in combination with meteorological monitoring data; task priority data and resource allocation demand data are generated through service feature analysis and task feature calculation; generating network configuration data and terminal access data based on slice configuration and resource allocation; and generating real-time resource data and terminal parameter data according to load statistical analysis and service quality calculation. According to the invention, through multi-source data fusion and dynamic resource adjustment, the reliability and service quality of satellite communication in a fire rescue scene are improved.

Description

Technical Field

[0001] The invention relates to a satellite communication method, in particular to a satellite communication method and device for supporting firefighting and rescue. Background Art

[0002] In modern firefighting and rescue work, a stable and reliable communication guarantee system is of great significance to improving rescue efficiency and ensuring personnel safety. The firefighting and rescue scene environment is complex and changeable, and often faces challenges such as damage to ground communication infrastructure or insufficient signal coverage. Satellite communications, with its wide coverage and flexible deployment, have become a key technical means to ensure rescue communications. Especially in large-scale disasters such as forest fires and earthquake rescue, satellite communications are often the only reliable communication method. Its support for key businesses such as command and dispatch, situational awareness, and personnel positioning has become increasingly prominent, and has become an indispensable part of the modern firefighting and rescue system.

[0003] At present, the research on fire rescue satellite communication technology mainly focuses on fixed bandwidth allocation, static service priority division and simple link quality assessment. The traditional bandwidth allocation method mainly adopts a fixed time slot reservation mechanism to pre-allocate fixed bandwidth resources for different types of services; link quality assessment mainly relies on basic parameters such as signal-to-noise ratio and bit error rate for simple judgment; terminal management adopts a unified parameter configuration scheme, lacking targeted optimization and adjustment. In terms of rescue site data processing, independent processing is mainly adopted, and there is a lack of effective fusion analysis between various types of data, making it difficult to give full play to the synergy of multi-source data.

[0004] However, the current technical solutions have many specific problems in practical applications: First, in terms of multi-source data processing, the inconsistent sampling rates of different data sources lead to low time alignment accuracy, especially in the processing of high-bandwidth data such as video and audio, there is a significant lag; second, the traditional link state evaluation method does not fully consider Doppler frequency shift compensation, resulting in unstable communication quality in high-speed mobile scenarios; third, the existing slice isolation mechanism lacks dynamic adjustment capabilities, and the processing of time slot overlap and frequency interference is relatively rough, making it difficult to meet the differentiated service quality requirements of different services; fourth, the terminal parameter configuration lacks an adaptive optimization mechanism, especially in extreme cases such as temperature changes and power shortages, it is easy to cause a sharp decline in terminal performance; finally, the resource scheduling strategy lacks an accurate quantitative assessment of the urgency of the task, making it difficult to ensure the communication guarantee of key services under fierce resource competition. These technical problems have seriously restricted the application effect of satellite communication systems in fire rescue scenarios. Summary of the invention

[0005] The purpose of the invention is to provide a satellite communication method and device for supporting firefighting and rescue, in order to solve at least one technical problem existing in the prior art.

[0006] A technical solution, a satellite communication method for supporting firefighting and rescue, comprises the following steps:

[0007] Step S1: Collect multi-source real-time data and pre-stored scene data from the fire rescue scene, and form a synchronized data set through time alignment processing; combine terminal status data and preset thresholds to evaluate data quality; generate scene feature and terminal feature data sets through feature extraction;

[0008] Step S2: Perform spatial geometry analysis based on satellite orbit data and terminal position to generate satellite visibility data; calculate propagation loss based on meteorological monitoring data; perform link status analysis based on terminal characteristics; and output available channel data based on spectrum resource evaluation;

[0009] Step S3: Analyze business characteristics based on scenario characteristics and link status; evaluate task urgency in combination with task templates; analyze resource requirements based on historical resource data; generate task priority and resource allocation requirement data;

[0010] Step S4: select slice configuration parameters according to task priority and available channels; determine allocation method based on resource requirements; calculate isolation parameters between slices; formulate access control strategy, and output network configuration and terminal access data;

[0011] Step S5: perform load analysis based on real-time traffic data; dynamically adjust resource allocation; calculate service quality parameters based on task priority; generate real-time resource data and terminal parameter data.

[0012] According to another aspect of the present application, a satellite communication device for supporting firefighting and rescue is also provided, comprising:

[0013] at least one processor; and

[0014] a memory communicatively connected to at least one of the processors; wherein,

[0015] The memory stores instructions that can be executed by the processor, and the instructions are used to be executed by the processor to implement the satellite communication method for supporting firefighting and rescue as described in any of the above technical solutions.

[0016] Beneficial effects: through real-time multi-source data collection, time alignment processing and feature extraction, efficient synchronization and analysis of rescue scene data are achieved; dynamic evaluation based on spatial geometry and meteorological environment ensures the stability and reliability of satellite communication links; timely response to key rescue tasks is ensured through business feature analysis and task priority evaluation; network slicing configuration and resource allocation mechanism based on task requirements are adopted to improve the utilization efficiency of communication resources; combined with dynamic adjustment of real-time load status and service quality assurance, adaptive optimization of the communication system is achieved, effectively supporting the smooth implementation of fire rescue tasks. The overall solution has strong practicality and reliability, and is suitable for various complex rescue scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a flow chart of the present invention.

[0018] Figure 2 It is a flow chart of step S1 of the present invention.

[0019] Figure 3 It is a flow chart of step S2 of the present invention.

[0020] Figure 4 It is a flow chart of step S3 of the present invention.

[0021] Figure 5 It is a flow chart of step S4 of the present invention.

[0022] Figure 6 It is a flow chart of step S5 of the present invention. DETAILED DESCRIPTION

[0023] like Figure 1 As shown, a detailed description and explanation are given based on the implementation scheme provided in the invention content. According to one aspect of the present application, a satellite communication method for supporting fire rescue is provided, and a more specific process includes the following steps:

[0024] Step S1: The real-time multi-source data collected from the fire rescue scene and the pre-stored multi-source scene database are processed by time alignment to form a time synchronization data set; a terminal status data set is generated according to a preset terminal parameter threshold table and the collected terminal real-time status data; a data quality assessment result is calculated using the time synchronization data set; feature extraction is performed on the time synchronization data set, the terminal status data set and the data quality assessment result to generate a scene feature data set and a terminal feature data set;

[0025] Step S2: Generate satellite visibility data through spatial geometry analysis based on the stored satellite orbit data and terminal position data; obtain propagation loss data through atmospheric propagation calculation using the acquired meteorological monitoring data; combine the satellite visibility data, propagation loss data and the terminal feature data set obtained in step S1 to form link status data through channel status analysis; output available channel data based on the link status data and pre-stored spectrum resource data after channel availability assessment;

[0026] Step S3: Based on the scenario feature data set obtained in step S1 and the link status data obtained in step S2, generate business feature data through business feature analysis; obtain task urgency data through task feature calculation based on the business feature data and pre-stored task template data; form resource demand data through resource demand analysis based on the business feature data and pre-stored historical resource data; combine the task urgency data and the resource demand data, and generate task priority data and resource allocation demand data through priority evaluation;

[0027] Step S4: Based on the task priority data obtained in step S3 and the available channel data obtained in step S2, select appropriate slice configuration parameters and generate slice template data; use the slice template data and the resource allocation requirement data obtained in step S3 to determine the resource allocation method and obtain resource configuration data; combine the resource configuration data and the link status data obtained in step S2 to calculate the isolation parameters between slices and form slice isolation data; formulate an access control strategy by analyzing the resource configuration data and slice isolation data, and output network configuration data and terminal access data.

[0028] According to one aspect of the present application, it also includes:

[0029] Step S5: Generate load status data through load statistical analysis based on the network configuration data, terminal access data and real-time traffic data obtained in step S4; obtain resource adjustment data through dynamic resource adjustment based on the load status data and the resource configuration data obtained in step S4; form service quality data through service quality calculation based on the resource adjustment data and the task priority data obtained in step S3; generate real-time resource data and terminal parameter data through parameter optimization calculation in combination with the service quality data and the terminal feature data set obtained in step S1.

[0030] This technical solution builds a complete fire rescue satellite communication guarantee system through multi-level data processing, resource scheduling and parameter optimization. First, at the data processing level, through the time synchronization and feature extraction of multi-source data, a reliable data basis is provided for system decision-making. In terms of link evaluation, accurate evaluation of communication links is achieved by comprehensively considering satellite visibility, propagation loss and link status. At the resource scheduling level, communication guarantee for key services is ensured through task priority evaluation and network slicing technology. Finally, continuous optimization of system performance is achieved through dynamic resource adjustment and parameter optimization. This multi-level technical solution is particularly suitable for the characteristics of fire rescue scenarios with high requirements for communication reliability, diverse business types and complex environmental conditions, and can effectively ensure the stability and reliability of rescue communications. The innovation and practicality of this solution lies in its strong adaptability and good scalability. It can dynamically adjust system parameters according to actual conditions to ensure the optimal performance of the communication system.

[0031] like Figure 2 As shown, according to one aspect of the present application, step S1 is specifically:

[0032] Step S11, real-time video data, real-time audio data, real-time sensor data and terminal location data collected from the fire rescue scene, as well as historical scene data extracted from the pre-stored multi-source scene database, are processed by time alignment to form a time-synchronized data set;

[0033] Step S12: Read the antenna parameters, power parameters and frequency parameters in the preset terminal parameter threshold table, collect the real-time power data, real-time temperature data and real-time working status data of the terminal, and generate a terminal status data set through status feature calculation;

[0034] Step S13: Calculate the data integrity index and reliability index based on the time synchronization data set using the pre-stored data quality assessment model to generate a data quality assessment result;

[0035] The data quality assessment model is as follows: data quality score QM = ω1·C(t) + ω2·R(t) +ω3·V(t); where: C(t) = 1 - (Nm / Nt) 2, C(t) is the data integrity index, Nm is the number of lost data packets, and Nt is the total number of data packets; R(t) = exp(-λ·BER)·exp(-μ·PER), R(t) is the data reliability index, BER is the bit error rate, PER is the packet error rate, λ and μ are attenuation coefficients; V(t) = exp(-|Td - Te| / Ts), V(t) is the data validity index, Td is the data generation timestamp, Te is the data use timestamp, and Ts is the system tolerance delay; ω1, ω2, ω3 are adaptive weight coefficients and satisfy ω1 + ω2 + ω3 = 1; QM ranges from [0,1], and the larger the QM, the better the data quality.

[0036] Step S14: input the time synchronization dataset, the terminal status dataset and the data quality assessment result into the pre-trained feature extraction model, extract key features, and generate a scene feature dataset and a terminal feature dataset respectively.

[0037] Among them, the feature extraction model is specifically as follows: feature vector F = α·Fv + β·Fa + γ·Fs; where: Fv =CNN(V)·RNN(V), Fv is the video feature vector, V is the input video data, CNN() is the improved convolutional neural network, and RNN() is the recurrent neural network; Fa = MFCC(A)·VAD(A), Fa is the audio feature vector, A is the input audio data, MFCC() is the Mel-frequency cepstral coefficient, and VAD() is voice activity detection; Fs = PCA(S)·WPT(S), Fs is the sensor feature vector, S is the sensor data, PCA() is the principal component analysis, and WPT() is the wavelet packet transform; α, β, and γ are dynamic weight coefficients, which are calculated through the attention mechanism.

[0038] Through the time alignment processing of multi-source data and the generation of terminal status data, efficient integration and analysis of fire rescue scene data are achieved. This step first ensures the temporal consistency of data from different sources by time synchronization of video, audio, sensor and location data, which is crucial for accurately understanding the real-time situation at the rescue site. At the same time, through the preset terminal parameter threshold table and the collected terminal real-time status data, the system can timely grasp the working status of communication equipment, including power, temperature and operating status, thereby preventing the risk of communication interruption. In particular, in the data quality assessment link, a quantitative evaluation mechanism for data quality is established by calculating integrity indicators and reliability indicators, which enables the system to accurately judge the availability of data. Finally, through feature extraction, scene feature data sets and terminal feature data sets are generated, providing a reliable decision-making basis for subsequent communication resource allocation. This multi-dimensional data processing method significantly improves the efficiency and accuracy of data processing in complex rescue environments and ensures the stable operation of the communication system.

[0039] like Figure 3 As shown, according to one aspect of the present application, step S2 is specifically:

[0040] Step S21, based on the stored satellite orbit data and the terminal position data obtained in step S1, the elevation angle and azimuth angle data are obtained through spatial coordinate conversion calculation, and the satellite visibility data is comprehensively generated;

[0041] Step S22: According to the temperature data, humidity data, air pressure data obtained from the weather station and the pre-stored geographical location data, the frequency-related attenuation value is obtained through atmospheric propagation attenuation calculation to form propagation loss data;

[0042] Step S23, based on the satellite visibility data, the propagation loss data and the terminal feature data set obtained in step S1, by analyzing the uplink and downlink links, the signal strength and bit error rate parameters are calculated to generate link status data;

[0043] Step S24: Based on the link status data and the pre-stored spectrum resource data, the frequency band availability parameters are calculated through channel quality evaluation, and qualified frequency bands are screened to obtain available channel data.

[0044] Comprehensive analysis and optimization have been achieved in the evaluation of satellite communication link status. First, satellite visibility data is obtained through spatial geometry analysis, and propagation loss is calculated in combination with real-time meteorological data. This method ensures that the system can accurately evaluate the link quality under the current communication environment. In particular, when calculating satellite visibility, an accurate spatial coordinate conversion algorithm is used, and key parameters such as elevation and azimuth are considered, so that the system can accurately predict the reception of satellite signals. In the propagation loss calculation, an accurate evaluation of signal attenuation is achieved by comprehensively considering meteorological factors such as temperature, humidity, and air pressure. Through uplink and downlink analysis, the system can calculate signal strength and bit error rate parameters in real time to evaluate the actual performance of the link. Finally, based on link status data and spectrum resource data, the system can screen out the most suitable frequency band to ensure the reliability of the communication link. This method is particularly suitable for fire rescue scenarios because it can maintain stable communication quality under complex environmental conditions.

[0045] like Figure 4 As shown, according to one aspect of the present application, step S3 is specifically:

[0046] Step S31, based on the scene feature data set obtained in step S1 and the link state data obtained in step S2, extract the service scale characteristics through service classification analysis, calculate the timeliness parameters, and form service feature data;

[0047] Step S32: According to the business feature data and the pre-stored task template data, by matching the task type, calculating the response time parameter and the completion deadline parameter, and obtaining the task urgency data;

[0048] Step S33: Based on the service feature data and the pre-stored historical resource data, the bandwidth requirement value and the delay requirement value are calculated through resource consumption analysis to generate resource requirement data;

[0049] Step S34: Based on the task urgency data and resource requirement data, a multi-dimensional scoring calculation is performed to obtain a task ranking value and a resource allocation value, and the task priority data and the resource allocation requirement data are output respectively.

[0050] The multi-dimensional scoring calculation is as follows: task priority score PS = Wu·U(t) + Wr·R(t) +Wc·C(t); where: U(t) = exp(-(Tc - Tn) / Tu), U(t) is the task urgency, Tc is the current time, Tn is the task deadline, and Tu is the time unit; R(t) = Br / Ba + Dr / Da, R(t) is the resource demand, Br is the requested bandwidth, Ba is the available bandwidth, Dr is the requested delay, and Da is the available delay; C(t) = Nc / Nt + Pc / Pt, C(t) is the business importance, Nc is the current number of businesses, Nt is the total business capacity, Pc is the current processing capacity, and Pt is the total processing capacity; Wu, Wr, and Wc are the weight coefficients calculated by the hierarchical analysis method.

[0051] Through the comprehensive analysis of business characteristics and task urgency, the intelligent allocation and priority management of communication resources are realized. This step first analyzes the business characteristics based on the scene feature data set and link status data, and accurately evaluates the communication needs of different types of business through business classification statistics and traffic prediction. In the task feature calculation link, the system combines the business feature data and the pre-stored task template data to establish a scientific task urgency evaluation mechanism, which is particularly important for ensuring the communication guarantee of key rescue tasks. In particular, in the resource demand analysis process, through the analysis of historical resource data, the system can accurately predict the bandwidth and latency requirements, and generate a reasonable resource allocation strategy based on this. Finally, through multi-dimensional scoring calculation, different tasks are prioritized to ensure that the communication needs of important rescue tasks are prioritized under resource constraints. This resource allocation method based on business characteristics and task urgency significantly improves the utilization efficiency of communication resources and ensures the smooth implementation of key rescue tasks.

[0052] like Figure 5 As shown, according to one aspect of the present application, step S4 is specifically:

[0053] Step S41: Calculate the slice parameters and slice configuration requirements of each service type based on the task priority data obtained in step S3 and the available channel data obtained in step S2, determine the slice function configuration scheme, and comprehensively form slice template data;

[0054] Step S42: Calculate the time slot allocation scheme, frequency point allocation scheme and power allocation parameters according to the slice template data and the resource allocation requirement data obtained in step S3, and generate resource configuration data through resource optimization processing;

[0055] Step S43: Analyze the interference between slices according to the resource configuration data and the link status data obtained in step S2, evaluate the degree of resource overlap, determine the isolation protection requirements, and output the slice isolation data;

[0056] Step S44: Comprehensive resource configuration data and slice isolation data, determine terminal access conditions, calculate access priority requirements, formulate resource reservation plans, and obtain network configuration data and terminal access data respectively.

[0057] The intelligent configuration and resource isolation protection of network slices are realized, which significantly improves the service quality and reliability of the communication system. This step first selects appropriate slice configuration parameters according to the task priority data and available channel data to ensure that different service types obtain suitable network resources. In the resource allocation link, the system realizes the refined management of network resources through the coordinated optimization of time slot allocation, frequency point allocation and power allocation. Especially in terms of slice isolation, by analyzing resource configuration data and link status data, the system can accurately calculate the isolation parameters between slices and effectively prevent mutual interference between different services. Finally, by formulating a reasonable access control strategy, the system not only guarantees the service quality of high-priority services, but also realizes the efficient use of network resources. This resource management method based on network slicing is particularly suitable for the complex communication needs of multiple concurrent services in fire rescue scenarios, and can effectively guarantee the service quality of different types of services.

[0058] like Figure 6 As shown, according to one aspect of the present application, step S5 is specifically:

[0059] Step S51, according to the network configuration data, terminal access data and collected real-time traffic data obtained in step S4, calculate the resource utilization rate, count the service load parameters, and obtain the load status data;

[0060] Step S52: Based on the load status data and the resource configuration data obtained in step S4, a resource adjustment range is determined by dynamic threshold calculation to generate resource adjustment data;

[0061] Step S53: Calculate quality assurance parameters based on the resource adjustment data and the task priority data obtained in step S3, and output service quality data through service level classification;

[0062] Step S54: Combining the service quality data with the terminal feature data set obtained in step S1, the resource allocation parameters and the terminal operating parameters are obtained through parameter adaptive calculation, and real-time resource data and terminal parameter data are generated respectively.

[0063] Dynamic adjustment of communication resources and real-time optimization of service quality are realized. This step first monitors the network load status in real time through load statistical analysis, and dynamically adjusts resources based on this. In the service quality calculation link, the system combines resource adjustment data and task priority data to establish a complete service quality evaluation mechanism. In particular, during the parameter optimization calculation process, the system can adaptively adjust resource allocation parameters and terminal operating parameters based on service quality data and terminal feature data sets to ensure that the system always runs in the optimal state. Through this dynamic adjustment mechanism, the system can respond to network load changes and service quality fluctuations in a timely manner, and realize intelligent scheduling of communication resources. This method is particularly suitable for rapidly changing communication needs in fire rescue scenarios, and can ensure the continuous and stable operation of the communication system.

[0064] According to one aspect of the present application, step S11 is specifically:

[0065] Step S111: perform frame extraction processing on the real-time video data collected at the fire rescue scene, extract the video frame timestamp information, and obtain video timestamp data; perform segmentation processing on the real-time audio data, extract the audio segment timestamp information, and obtain audio timestamp data; perform sampling processing on the real-time sensor data and terminal position data, extract the sampling timestamp information, and obtain sensor timestamp data and position timestamp data;

[0066] Step S112: Generate time reference data by calculating the minimum time granularity according to the video time stamp data, the audio time stamp data, the sensor time stamp data and the position time stamp data; divide various types of data into time windows based on the time reference data to obtain time window data;

[0067] Among them, the minimum time granularity Tmin = Max(Tv, Ta, Ts, Tl)·K(n); Among them: Tv = 1 / fv, Tv is the video sampling period, fv is the video frame rate; Ta = 1 / fa, Ta is the audio sampling period, fa is the audio sampling rate; Ts = 1 / fs, Ts is the sensor sampling period, fs is the sensor sampling rate; Tl = 1 / fl, Tl is the position sampling period, fl is the position update rate; K(n) = 1 + log(1 + σt 2), K(n) is the adaptive adjustment coefficient, σt 2 is the timestamp variance; Max() is the maximum value function.

[0068] Step S113: construct a time index for the historical scene data in the pre-stored multi-source scene database according to the time window data to obtain a historical data index; based on the historical data index, extract the historical data corresponding to the current time window to form matching historical data;

[0069] Step S114, resample the real-time video data according to the time window data to obtain regularized video data; resample the real-time audio data according to the time window data to obtain regularized audio data; interpolate the real-time sensor data and the terminal position data according to the time window data to obtain regularized sensor data and regularized position data respectively;

[0070] Step S115: align the regularized video data, regularized audio data, regularized sensor data, regularized position data and matching historical data in time sequence according to the time reference data to generate a time synchronization data set.

[0071] Through refined time series processing and data synchronization mechanisms, high-precision time alignment of multi-source data is achieved. This step first extracts timestamps from video, audio, sensor and location data, and ensures the accuracy of time tags for different types of data through methods such as frame extraction and segmentation. In the time base calculation link, the system uses the minimum time granularity calculation method to establish a unified time base, which provides a reliable basis for subsequent data synchronization. Especially in the processing of historical scene data, by establishing a time index structure, the system can quickly match and extract historical data of the corresponding time window. In the data resampling stage, corresponding processing methods are used for different types of data, such as video and audio resampling, sensor data interpolation calculation, etc., to ensure the alignment accuracy of all data under a unified time base. This time synchronization processing method for multi-source data effectively solves the problems of inconsistent sampling rates and inconsistent time bases of different data sources in fire rescue scenarios, and provides a high-quality data foundation for subsequent data analysis and decision-making.

[0072] According to one aspect of the present application, step S12 is specifically:

[0073] Step S121, read the antenna gain threshold and antenna polarization parameter from the preset terminal parameter threshold table to obtain antenna threshold data; read the transmission power upper limit and the power adjustment step to obtain power threshold data; read the working frequency band range and the frequency offset limit to obtain frequency threshold data; combine the antenna threshold data, the power threshold data and the frequency threshold data to form parameter threshold data;

[0074] Step S122, calculating the remaining power of the collected real-time power data to obtain remaining power data; calculating the sustainable working time based on the remaining power data to generate working time data; matching the working time data with the power parameter in the parameter threshold data to form power status data;

[0075] Step S123, calculating the temperature change rate of the collected real-time temperature data to obtain temperature change data; calculating the temperature trend characteristics based on the temperature change data to generate temperature trend data; matching the temperature trend data with the power parameters in the parameter threshold data to form temperature state data;

[0076] Step S124, performing work mode recognition on the collected real-time work status data to obtain work mode data; calculating resource occupancy based on the work mode data to generate resource occupancy data; matching the resource occupancy data with the frequency parameter in the parameter threshold data to form operation status data;

[0077] Step S125, combining the power state data, the temperature state data and the operation state data to obtain state feature data; based on the state feature data, calculating the overall working state of the current terminal to generate a terminal state data set.

[0078] A complete terminal status monitoring and evaluation mechanism has been established to achieve accurate status perception of communication terminals. This step first establishes a complete parameter constraint system including antenna, power and frequency through the organization and management of parameter threshold tables. In terms of power status monitoring, the system can accurately evaluate the energy status of the terminal through the calculation of remaining power and the prediction of sustainable working time. The temperature status monitoring link adopts the temperature change rate calculation and trend analysis method to achieve early warning of abnormal terminal temperature. Especially in terms of working status identification, through working mode identification and resource occupancy analysis, the system can fully grasp the operating status of the terminal. Finally, through the combination of state features and overall status evaluation, an accurate terminal status data set is formed. This multi-dimensional terminal status monitoring method is particularly suitable for the high requirements for terminal reliability in fire rescue scenarios, and can effectively prevent terminal failures and ensure the stable operation of the communication system.

[0079] According to one aspect of the present application, step S14 is specifically:

[0080] Step S141, input the regularized video data in the time synchronization data set into a pre-trained video feature model, extract target detection features, scene segmentation features and action recognition features, and obtain video feature data; input the regularized audio data into a pre-trained audio feature model, extract sound category features, speech recognition features and environmental noise features, and obtain audio feature data; classify the regularized sensor data according to the sensor type, extract temperature change features, pressure change features and gas concentration features, and obtain sensor feature data;

[0081] Step S142: construct a multimodal feature fusion matrix based on the video feature data, the audio feature data, and the sensor feature data to obtain a fused feature matrix; perform time series correlation analysis on the fused feature matrix and the matching historical data to generate time series correlation features; calculate the scene change trend based on the time series correlation features to obtain scene evolution data;

[0082] Among them, the fusion feature matrix, M = H·W + (1-H)·S; where: H = sigmoid(ΣWi·Xi), H is the feature fusion weight, Wi is the feature weight vector, Xi is the feature input vector; W = FFT(F)·DCT(F), W is the frequency domain feature matrix, F is the original feature matrix, FFT() is the fast Fourier transform, DCT() is the discrete cosine transform; S = SVD(F)·PCA(F), S is the spatial domain feature matrix, SVD() is the singular value decomposition, PCA() is the principal component analysis.

[0083] The calculation process of the time series correlation feature can be as follows: time series correlation R(t) = Wt•T(t) + Ws•S(t)+ Wf•F(t); where: T(t) = exp(-|Δt| / τ)•Σ(xi•yi) / sqrt(Σxi 2 •Σyi 2 ), T(t) is the time correlation function, Δt is the time difference, τ is the time constant, xi and yi are time series data; S(t) = 1 - (Σ|ri - ri-1|) / (2•N•σr), S(t) is the stability function, ri is the correlation coefficient, N is the number of samples, and σr is the standard deviation; F(t) = |FFT(x)•conj(FFT(y))| / sqrt(|FFT(x)| 2 •|FFT(y)| 2 ), F(t) is the frequency domain correlation function, FFT() is the fast Fourier transform, conj() is the conjugate complex number; Wt, Ws, Wf are the dynamic weight coefficients.

[0084] Step S143: construct a terminal state feature vector based on the real-time power data, real-time temperature data and real-time working state data in the terminal state data set to obtain terminal state features; perform weight calculation on the integrity index and reliability index in the data quality assessment result to obtain quality weight data; generate a terminal working feature matrix based on the terminal state features and the quality weight data to obtain terminal working features;

[0085] Step S144: perform multi-dimensional feature aggregation on the fused feature matrix, scene evolution data and time-series related features, screen key scene features through an adaptive feature selection algorithm, and generate a scene feature data set; perform feature combination on the terminal state features, quality weight data and terminal working features, extract key terminal features through a dynamic threshold screening algorithm, and generate a terminal feature data set.

[0086] The adaptive feature selection algorithm is as follows: feature importance I(f) = Wc•Corr(f) + Wm•MI(f)+ Wr•RF(f); where: Corr(f) = |Σ(xi - x*)(yi - y*)| / sqrt(Σ(xi - x*) 2 •Σ(yi -y*) 2 ), Corr(f) is the feature correlation, xi is the feature value, and yi is the target value; MI(f) = Σp(x,y)•log(p(x,y) / (p(x)•p(y))), MI(f) is the mutual information value, p(x,y) is the joint probability distribution; RF(f) = ΔImpurity / ΣΔImpurity, RF(f) is the random forest importance; Wc, Wm, Wr are adaptive weight coefficients.

[0087] Deep feature fusion and intelligent feature extraction of multimodal data are realized. This step first extracts features from video, audio and sensor data through pre-trained feature models to obtain multi-dimensional features including target detection, scene segmentation, and sound recognition. In the feature fusion stage, by constructing a multimodal feature fusion matrix, the effective integration of different types of features is achieved. Especially in terms of time series correlation analysis, the system can accurately grasp the trend of scene changes by comparing the fused features with historical data. In the terminal feature extraction stage, a scientific terminal work feature evaluation system is established by calculating the terminal state feature vector and quality weight. Finally, through adaptive feature selection and dynamic threshold screening, the system can accurately extract key feature information. This feature extraction method based on deep learning is particularly suitable for the analysis of complex and changeable environmental features in fire rescue scenarios, and provides a reliable decision-making basis for the intelligent scheduling of communication resources.

[0088] According to one aspect of the present application, step S21 is specifically:

[0089] Step S211, extract satellite orbit parameters and satellite real-time position parameters from the stored satellite orbit data to obtain satellite position data; convert the longitude value, latitude value and altitude value in the terminal position data to generate terminal coordinate data; establish a spatial rectangular coordinate system based on the satellite position data and the terminal coordinate data to form coordinate system data.

[0090] Step S212, convert the satellite position data from the orbital coordinate system to the geocentric coordinate system to obtain geocentric position data; convert the terminal coordinate data to the geocentric coordinate system to generate geocentric terminal data; based on the geocentric position data and the geocentric terminal data, calculate the relative position vector of the satellite and the terminal to form relative position data.

[0091] Step S213: According to the relative position data, the satellite position is projected onto the horizontal plane of the terminal station to obtain projection coordinate data; based on the projection coordinate data, the angle of the satellite relative to the north direction of the terminal is calculated to generate azimuth data; according to the relative position data, the angle of the satellite relative to the horizontal plane of the terminal is calculated to form elevation data.

[0092] Step S214, compare the azimuth data and the elevation data with the preset minimum elevation angle threshold to obtain angle determination data; based on the angle determination data, calculate the visibility status of the satellite at the current moment to generate visibility status data; according to the relative position data, calculate the straight-line distance between the satellite and the terminal to form distance data.

[0093] Step S215: Combine the visibility status data, azimuth data, elevation data and distance data to obtain visibility feature data; based on the visibility feature data, calculate the visibility duration and visibility performance index to form satellite visibility data.

[0094] High-precision satellite visibility analysis and spatial coordinate calculation are achieved. This step first establishes an accurate spatial coordinate system by processing satellite orbit parameters and terminal position data. In the coordinate conversion link, by converting the satellite position from the orbital coordinate system to the geocentric coordinate system and combining the terminal's geographic location information, the system can accurately calculate the relative position relationship between the satellite and the terminal. In particular, in terms of azimuth and elevation calculation, precise projection algorithms and angle calculation methods are used to ensure the accuracy of visibility analysis. By comparing with the preset minimum elevation threshold, the system can scientifically judge the visibility status of the satellite. Finally, through a comprehensive analysis of visibility characteristics, the visibility duration and visibility performance indicators are accurately predicted. This visibility calculation method based on precise spatial geometric analysis is particularly suitable for the high reliability requirements of satellite communication links in fire rescue scenarios, and can effectively ensure that the communication system selects the optimal satellite signal resources.

[0095] According to one aspect of the present application, step S23 is specifically:

[0096] Step S231, based on the satellite visibility data, calculate the satellite relative motion speed and acceleration to obtain motion state data; based on the motion state data, calculate through the relative motion model to obtain Doppler frequency shift data; based on the Doppler frequency shift data, generate carrier frequency compensation parameters to form frequency compensation data.

[0097] In the calculation of Doppler frequency shift data, Doppler frequency shift compensation Δf = fc·(v·cosθ) / c + K·a·t; where: fc is the carrier frequency, v is the relative velocity, θ is the incident angle, c is the speed of light; a is the relative acceleration, t is the time interval; K = 1 - exp(-|v| / v0), K is the nonlinear compensation coefficient, v0 is the reference velocity; the final carrier frequency f = f0 +Δf, f0 is the original carrier frequency.

[0098] Step S232, the gas absorption loss data, rainfall loss data and cloud loss data in the propagation loss data are combined and calculated to obtain comprehensive attenuation data; based on the frequency compensation data and the comprehensive attenuation data, the channel transmission characteristics are calculated to generate channel characteristic data; according to the terminal working characteristics in the terminal characteristic data set, the transmission power and antenna gain parameters are extracted to obtain the terminal RF data.

[0099] Step S233: Based on the terminal radio frequency data and the channel characteristic data, the uplink power data is calculated through the uplink budget model to obtain the uplink power data; the uplink power data is calculated with the system noise temperature to generate the uplink signal-to-noise ratio data; according to the uplink signal-to-noise ratio data, the uplink performance data is calculated through the bit error rate model.

[0100] The uplink budget optimization model is as follows: Uplink signal-to-noise ratio SNRup = EIRP + Gr - Lfs- La - k - T - B + Gc; where: EIRP is the equivalent isotropic radiated power, Gr is the receiving antenna gain, Lfs is the free space loss, La is the atmospheric loss, k is the Boltzmann constant, T is the system noise temperature, B is the signal bandwidth, Gc = α·exp(-β·θ 2 ), Gc is the improved atmospheric correction factor, θ is the elevation angle, and α and β are adaptive coefficients.

[0101] Step S234: According to the terminal radio frequency data and the channel characteristic data, the downlink power data is calculated by the downlink budget model to obtain the downlink power data; the downlink power data is calculated with the system noise temperature to generate the downlink signal-to-noise ratio data; according to the downlink signal-to-noise ratio data, the downlink performance data is calculated by the bit error rate model.

[0102] Step S235: Comprehensively analyze the uplink performance data and the downlink performance data to obtain bidirectional performance data; calculate the link availability index based on the bidirectional performance data to obtain availability data; and perform feature combination on the availability data, the uplink performance data, and the downlink performance data to generate link status data.

[0103] A complete link status evaluation system has been established to achieve accurate performance prediction of the communication link. This step first calculates the Doppler frequency shift and generates carrier frequency compensation parameters through satellite relative motion analysis, effectively solving the frequency offset problem in high-speed mobile scenarios. In terms of channel transmission characteristics analysis, the system comprehensively considers multiple attenuation factors such as gas absorption, rainfall and cloud layer to achieve accurate evaluation of propagation loss. Especially in the uplink and downlink analysis link, by establishing a complete link budget model, the system can accurately calculate key performance indicators such as signal-to-noise ratio and bit error rate. Finally, through a comprehensive analysis of two-way performance, a scientific link status evaluation result is formed. This multi-factor comprehensive link status evaluation method is particularly suitable for the complex and changeable propagation environment in fire rescue scenarios, and can provide accurate link quality prediction for communication systems.

[0104] According to one aspect of the present application, step S24 is specifically:

[0105] Step S241, extracting uplink signal-to-noise ratio data and downlink signal-to-noise ratio data from the link status data to obtain link quality data; reading frequency band division parameters and frequency band restriction parameters in the pre-stored spectrum resource data to generate frequency band parameter data; correlating the link quality data and the frequency band parameter data to form frequency band evaluation data.

[0106] Step S242: based on the frequency band evaluation data, calculate the bit error rate index of each frequency band to obtain the bit error rate data; according to the bit error rate data, calculate the transmission reliability value of each frequency band to generate reliability data; compare the reliability data with the system bit error rate requirement to form reliability evaluation data.

[0107] Step S243, perform interference analysis on the frequency band evaluation data, calculate the carrier-to-interference ratio of each frequency band, and obtain carrier-to-interference ratio data; based on the carrier-to-interference ratio data, calculate the anti-interference capability value of each frequency band to generate anti-interference data; perform weighted combination of the anti-interference data and the reliability evaluation data to form frequency band quality data.

[0108] Step S244: Calculate the transmission capacity index of each frequency band according to the frequency band quality data to obtain capacity data; obtain the frequency band availability parameter by comprehensive calculation based on the capacity data and the frequency band quality data; compare the frequency band availability parameter with the preset frequency band quality threshold to screen out qualified frequency bands.

[0109] Step S245, matching bandwidth requirements for qualified frequency bands to obtain bandwidth matching data; based on the bandwidth matching data, calculating the available time window of each frequency band to generate time window data; integrating the qualified frequency bands, bandwidth matching data and time window data to form available channel data.

[0110] Scientific frequency band quality assessment and channel availability analysis are achieved. This step first establishes a complete frequency band assessment system through correlation analysis of link quality data and frequency band parameters. In the reliability assessment link, the scientific nature of frequency band selection is ensured through bit error rate calculation and reliability value evaluation. Especially in terms of interference analysis, the system provides reliable quality assurance for frequency band selection through carrier-to-interference ratio calculation and anti-interference capability evaluation. In the frequency band screening process, the scientific selection of the optimal frequency band is achieved by comprehensively considering the transmission capacity and frequency band quality parameters. Finally, through bandwidth demand matching and time window analysis, complete available channel data is formed. This multi-dimensional frequency band quality assessment method is particularly suitable for the high requirements for communication quality in fire rescue scenarios, and can ensure that the system selects the optimal frequency band resources.

[0111] According to one aspect of the present application, step S31 is specifically:

[0112] Step S311, extract target detection features, scene segmentation features and action recognition features from the scene feature data set, perform scene type identification, and obtain scene type data; based on the scene type data, generate business type data through preset scene-business mapping rules; according to the business type data, extract communication requirement parameters of each type of business to form business requirement data.

[0113] Step S312: Analyze the bidirectional performance data and availability data in the link status data to obtain link resource data; based on the link resource data, calculate the current available bandwidth and delay indicators to generate performance indicator data; match and analyze the performance indicator data with the business demand data to obtain resource matching data.

[0114] Step S313: Perform data volume statistics for each type of business in the business type data to obtain business scale data; based on the business scale data, calculate the traffic demand in the future time window through the data traffic prediction model to generate traffic prediction data; compare and analyze the traffic prediction data with the resource matching data to obtain business capacity data.

[0115] The data traffic prediction model is as follows: predicted traffic F(t) = μ·AR(t) + (1-μ)·LSTM(t); AR(t) = Σai·x(ti) + Σbj·ε(tj), AR(t) is the predicted value of the autoregressive model, x(t) is the historical traffic, and ε(t) is the prediction error; LSTM(t) is the predicted value of the long short-term memory network; μ = σ 2 AR / (σ 2 AR + σ 2 LSTM), μ is the adaptive weight, σ 2 is the prediction variance.

[0116] Step S314: extract the service response time requirement from the service demand data to obtain the delay requirement data; based on the delay requirement data, combined with the real-time delay indicator in the performance indicator data, calculate the delay margin parameter to generate timeliness data; correlate the timeliness data with the service capacity data to obtain service quality data.

[0117] Step S315: Combine the service type data, service capacity data, timeliness data and service quality data to obtain comprehensive feature data; based on the comprehensive feature data, calculate key business indicators through a service feature extraction model to generate service feature data.

[0118] The service feature extraction model is specifically as follows: service feature vector B(t) = λ1·V(t) + λ2·D(t) +λ3·Q(t); where: V(t) = Vr / Vm·exp(-|Td-Tr| / T0), V(t) is the service volume feature, Vr is the real-time service volume, Vm is the maximum service volume capacity, Td is the current delay, Tr is the required delay, and T0 is the reference delay; D(t) = (Dr·Pr) / (Dm·Pm), D(t) is the service density feature, Dr is the current data rate, Dm is the maximum data rate, Pr is the current processing rate, and Pm is the maximum processing rate; Q(t) = exp(-ε·BER)·exp(-η·PLR), Q(t) is the service quality feature, BER is the bit error rate, PLR is the packet loss rate, ε and η are weight coefficients; λ1, λ2, and λ3 are adaptive feature weights.

[0119] Accurate business feature analysis and service demand prediction are achieved. This step first realizes accurate scene type identification and business type mapping through in-depth analysis of scene feature data. In the business demand analysis link, the system accurately evaluates the service capabilities of the current network through matching analysis of link resource data and performance indicators. In particular, in terms of business scale prediction, the data traffic prediction model is used to accurately predict communication needs in future time windows. In the timeliness analysis process, by comparing delay requirements and real-time delay indicators, the system can accurately evaluate the timeliness requirements of the business. Finally, through comprehensive feature extraction, complete business feature data is formed. This business analysis method based on multi-dimensional features is particularly suitable for the diverse business needs in fire rescue scenarios. It can accurately identify and predict the communication needs of different types of businesses and provide a reliable basis for resource allocation.

[0120] According to one aspect of the present application, step S34 is specifically:

[0121] Step S341, normalize the response time parameters and completion deadline parameters in the task urgency data to obtain time-normalized data; normalize the bandwidth requirement value and delay requirement value in the resource requirement data to obtain resource-normalized data; construct a scoring dimension matrix based on the time-normalized data and the resource-normalized data to generate scoring dimension data.

[0122] Step S342: Calculate the timeliness urgency of the task based on the time-normalized data to obtain timeliness score data; calculate the resource demand intensity based on the resource-normalized data to obtain resource score data; perform weighted combination of the timeliness score data and the resource score data to generate comprehensive score data.

[0123] Step S343: perform importance analysis on each dimension in the scoring dimension data, and obtain dimension weight data through hierarchical analysis calculation; based on the dimension weight data, perform weighted sum operation on the comprehensive scoring data to generate weighted score data; based on the weighted score data, perform numerical sorting processing to form sorting result data.

[0124] The specific calculation process of dimension weight is as follows: dimension weight W = Em·AHP(t); where: Em = -Σ(pi·ln(pi)) / ln(n), Em is the entropy weight, pi is the normalized value of the i-th indicator, and n is the number of indicators; AHP(t) = Σ(Cij·Rij) / Σ(Cij·Sij), AHP(t) is the hierarchical analysis weight, Cij is the judgment matrix element, Rij is the relative importance, and Sij is the consistency index; the final weight Wi = Em·AHP(t) / Σ(Em·AHP(t)).

[0125] Step S344, classify the sorting result data according to the score threshold to obtain task grade data; based on the weighted score data, calculate the standardized score value of each task to obtain the task ranking value; generate a task priority identifier according to the task grade data and the task ranking value to form task priority data.

[0126] Step S345: perform initial allocation calculation on bandwidth resources based on the allocation ratio data to obtain bandwidth allocation data; perform initial allocation calculation on delay resources based on the allocation ratio data to obtain delay allocation data; standardize the bandwidth allocation data and delay allocation data to obtain resource configuration values; generate a final resource allocation plan based on the resource configuration values ​​to form resource allocation demand data.

[0127] A scientific task priority evaluation and resource allocation mechanism has been established. This step first establishes a unified scoring standard by normalizing time parameters and resource requirements. In the scoring process, a comprehensive evaluation of the urgency of the task is achieved through the weighted combination of timeliness score and resource score. In particular, in terms of dimension weight calculation, the hierarchical analysis method is adopted to ensure the scientificity and rationality of the scoring dimensions. In the task priority determination link, an accurate task sorting mechanism is formed through hierarchical processing and standardized calculation. Finally, a scientific resource allocation plan is generated through initial resource allocation and standardized processing. This multi-dimensional task evaluation and resource allocation method is particularly suitable for the characteristics of large differences in task urgency in fire rescue scenarios, and can ensure that critical tasks are given priority in communication resource guarantees.

[0128] According to one aspect of the present application, step S42 is specifically:

[0129] Step S421: extract the time slot requirement parameters and priority parameters of each slice according to the slice template data to obtain slice parameter data; calculate the burst traffic characteristics and average traffic characteristics of each service based on the resource allocation requirement data to obtain traffic characteristic data; match and analyze the slice parameter data and traffic characteristic data to generate time slot requirement data.

[0130] Step S422, divide the time slot demand data into time periods to obtain period segmentation data; based on the period segmentation data, calculate the load distribution of each time period to generate load distribution data; according to the load distribution data, obtain the time slot mapping data through dynamic time slot allocation operation; optimize the time slot mapping data and the priority parameters of each slice to form a time slot allocation plan.

[0131] The process of dynamic time slot allocation operation is as follows: time slot allocation ratio P(i) = α·L(i) + β·Q(i) +γ·U(i); where: L(i) = Li / ΣLj, L(i) is the load ratio, Li is the load size of service i; Q(i) = exp(qi / qmax), Q(i) is the service quality factor, qi is the service level of service i, qmax is the highest service level; U(i) = 1- (Ti / Tmax) 2 , U(i) is the time urgency, Ti is the remaining processing time, Tmax is the maximum allowed time; α, β, γ are dynamic adjustment coefficients and satisfy α+β+γ=1.

[0132] Step S423, extract frequency resource constraint conditions from the slice template data to obtain frequency constraint data; classify the bandwidth requirements in the resource allocation requirement data according to the frequency band characteristics to obtain frequency band requirement data; based on the frequency constraint data and the frequency band requirement data, calculate the available frequency point range of each slice to generate frequency point range data.

[0133] Step S424, perform interference analysis on the frequency range data, calculate the interference matrix between the frequencies, and obtain interference matrix data; determine the specific frequency of each slice through frequency allocation operation according to the interference matrix data, and generate frequency configuration data; verify and optimize the frequency configuration data and the adjacent frequency interference constraint to form a frequency allocation plan.

[0134] Step S425: Based on the time slot allocation scheme and the frequency point allocation scheme, the power requirements of each slice at different time slots and frequencies are calculated to obtain power requirement data; according to the system power limit and slice priority, power allocation calculation is performed to generate power allocation parameters.

[0135] Among them, the power demand Pd(t) = Pb·M(t)·G(t)·L(t); where: Pb is the reference power; M(t) = 1+k1·log(1 + SNRr / SNRt), M(t) is the modulation factor, SNRr is the required signal-to-noise ratio, SNRt is the current signal-to-noise ratio, and k1 is the modulation coefficient; G(t) = exp(k2·|θr - θc| / θmax), G(t) is the gain compensation factor, θr is the required beam angle, θc is the current beam angle, θmax is the maximum angle deviation, and k2 is the gain coefficient; L(t) =1+α·BER +β·PER, L(t) is the loss compensation factor, BER is the bit error rate, PER is the packet loss rate, and α and β are compensation coefficients.

[0136] Step S426: perform resource mapping processing on the time slot allocation plan, frequency allocation plan and power allocation parameters to obtain resource mapping data; perform resource conflict detection between slices based on the resource mapping data to generate conflict detection data; optimize the resource allocation result through resource adjustment calculation according to the conflict detection data to form resource configuration data.

[0137] Refined time-frequency resource allocation and power control are achieved. This step first establishes a scientific time slot demand assessment mechanism through the analysis of slice parameters and traffic characteristics. In the time slot allocation link, dynamic time slot resource scheduling is achieved through cycle segmentation and load distribution analysis. Especially in terms of frequency point allocation, the system effectively avoids mutual interference between frequency points through interference matrix analysis and frequency point configuration optimization. In the power allocation process, reasonable power resource allocation is achieved by comprehensively considering slice priority and system power limit. Finally, resource mapping and conflict detection are used to ensure the feasibility of resource allocation. This refined resource allocation method is particularly suitable for the high communication quality requirements in fire rescue scenarios, and can achieve the optimal configuration of network resources.

[0138] According to one aspect of the present application, step S43 is specifically:

[0139] Step S431, extract the time slot allocation scheme, frequency allocation scheme and power allocation parameters of each slice from the resource configuration data to obtain slice resource data; based on the link status data, calculate the channel transmission characteristics of each slice to generate transmission characteristic data; classify the slice resource data according to the resource type to form resource classification data.

[0140] The process of slice resource analysis may also be included, for example: resource utilization U(s) = α•B(s) + β•T(s)+ γ•P(s); where: B(s) = (Br / Bt)•exp(-|SNRr - SNRt| / SNR0), B(s) is the bandwidth utilization function, Br is the actual bandwidth, Bt is the total bandwidth, SNRr is the actual signal-to-noise ratio, SNRt is the target signal-to-noise ratio, and SNR0 is the reference signal-to-noise ratio; T(s) = (1 - Dr / Dt)•(1 - Jr / Jt), T(s) is the time slot utilization function, Dr is the actual delay, Dt is the allowable delay, Jr is the jitter rate, and Jt is the jitter threshold; P(s) = (Pr / Pt)•exp(-|Gr - Gt| / G0), P(s) is the power utilization function, Pr is the actual power, Pt is the total power, Gr is the actual gain, Gt is the target gain, G0 is the reference gain; α, β, γ are weight coefficients.

[0141] Step S432: perform overlapping analysis on the time slot resources in the resource classification data, calculate the time slot overlapping interval, and obtain the time slot overlapping data; based on the time slot overlapping data, calculate the duration of each overlapping interval and generate overlapping duration data; based on the overlapping duration data and the transmission characteristic data, evaluate the degree of time slot overlap impact and form time slot interference data.

[0142] Step S433, perform spectrum analysis on the frequency resources in the resource classification data, calculate the frequency overlap range, and obtain frequency overlap data; based on the frequency overlap data, calculate the frequency interval and bandwidth overlap, and generate spectrum overlap data; based on the spectrum overlap data and transmission characteristic data, evaluate the degree of frequency overlap impact and form frequency interference data.

[0143] The calculation process of the frequency overlap data is: spectrum overlap O = Σ(Wi·Si·Ii) / Σ(Wi·Si); where: Wi = exp(-|fi - fc| / B), Wi is the frequency weight, fi is the current frequency, fc is the center frequency, and B is the bandwidth; Si= Pi / Pmax, Si is the power weight, Pi is the signal power, and Pmax is the maximum power; Ii = 1 - exp(-|Δf| / f0), Ii is the interference factor, Δf is the frequency interval, and f0 is the reference frequency interval.

[0144] Step S434: cross-analyze the time slot interference data and the frequency point interference data, calculate the time-frequency overlapping area, and obtain the time-frequency overlapping data; based on the time-frequency overlapping data, evaluate the inter-slice interference between each slice; calculate the overlapping ratio according to the time and frequency dimensions of slice resource allocation to form the degree of resource overlap.

[0145] The time-frequency overlap analysis can be specifically expressed as follows: overlap index O(t,f) = Wt•Ot(t) + Wf•Of(f) +Wc•Oc(t,f); where: Ot(t) = Σ(ti∩tj) / Σ(ti∪tj), Ot(t) is the time overlap function, ti and tj are time intervals; Of(f) = Σ(|Fi•Fj|) / sqrt(Σ|Fi| 2 •Σ|Fj| 2 ), Of(f) is the frequency overlap function, Fi and Fj are the spectrum distributions; Oc(t,f) = exp(-Δt / T0)•exp(-Δf / F0), Oc(t,f) is the time-frequency coupling function, Δt is the time interval, Δf is the frequency interval, T0 and F0 are reference values; Wt, Wf, Wc are adaptive weight coefficients.

[0146] Step S435: Calculate the minimum isolation bandwidth and protection time slot required for each slice according to the interference situation and resource overlap degree between slices to obtain the isolation protection requirements; based on the isolation protection requirements, generate time slot protection interval parameters and frequency point protection bandwidth parameters to form slice isolation data.

[0147] An efficient slice isolation and interference control mechanism is implemented. This step first establishes a complete resource analysis framework by classifying and processing slice resource data. In the time slot overlap analysis link, the system can accurately evaluate the interference impact caused by time slot overlap by calculating the overlapping interval and duration. Especially in the frequency overlap analysis, the accurate evaluation of interference between frequency points is achieved through spectrum analysis and bandwidth overlap calculation. In the process of time-frequency overlap analysis, the system comprehensively evaluates the interference between slices through a cross-analysis method. Finally, an effective slice isolation solution is formed by calculating the isolation bandwidth and protection time slot. This multi-dimensional slice isolation method is particularly suitable for the characteristics of multi-service concurrency in fire rescue scenarios, and can effectively ensure that different services do not interfere with each other and ensure communication quality.

[0148] According to one aspect of the present application, step S52 is specifically:

[0149] Step S521, segment the business load parameters in the load status data according to the time series to obtain load segmentation data; based on the load segmentation data, calculate the load change trend in each time period to generate load trend data; compare and analyze the load trend data with the historical data of the same period to form load characteristic data.

[0150] Step S522: extract the time slot allocation plan, frequency allocation plan and power allocation parameters from the resource configuration data to obtain resource utilization data; based on the resource utilization data, calculate the current utilization rate of each type of resources to generate resource occupancy data; compare the resource occupancy data with the system capacity upper limit to form resource margin data.

[0151] Step S523: compare the load change trend in the load characteristic data with the preset load fluctuation range to obtain load assessment data; calculate the load adjustment coefficient of each business according to the load assessment data to generate adjustment coefficient data; calculate the upper and lower limit thresholds of resource adjustment based on the adjustment coefficient data and resource margin data to form threshold range data.

[0152] The calculation process of the threshold range can be: threshold Th(t) = μ·Ma(t) + σ·Md(t); where: Ma(t) = α·avg(L) + (1-α)·Ma(t-1), Ma(t) is the moving average value of the load, and L is the real-time load; Md(t) = β·dev(L) + (1-β)·Md(t-1), Md(t) is the dynamic value of the load standard deviation; μ and σ are adaptive coefficients, which are dynamically adjusted according to the load change rate; avg() is the average value function, and dev() is the standard deviation function.

[0153] Step S524, classify the threshold range data, divide it according to time slot resources, frequency resources and power resources, and obtain classified threshold data; based on the classified threshold data, calculate the adjustment step size of each type of resources and generate adjustment step size data; match the adjustment step size data with the current resource allocation status to form adjustment plan data.

[0154] Among them, the adjustment step size S(t) = Sb•A(t)•D(t)•Q(t); where: Sb is the benchmark step size; A(t) = 1 +tanh(|Lt - Lt-1| / L0), A(t) is the amplitude factor, Lt is the current load, and L0 is the reference load; D(t) = exp(-|Rt- Tt| / T0), D(t) is the direction factor, Rt is the actual resource amount, Tt is the target resource amount, and T0 is the benchmark deviation; Q(t) = (1+ η•SNR) / (1 + η•SNR0), Q(t) is the quality factor, SNR is the current signal-to-noise ratio, SNR0 is the target signal-to-noise ratio, and η is the sensitivity coefficient.

[0155] Step S525, perform constraint check on each parameter in the adjustment plan data to obtain constraint check data; based on the constraint check data, screen the adjustment plans that meet the system constraints and generate feasible plan data; optimize and sort the feasible plan data, select the optimal adjustment plan, and form resource adjustment data.

[0156] A dynamic resource adjustment mechanism and load balancing system have been established. This step firstly realizes the accurate grasp of load changes through load segmentation analysis and trend prediction. In the resource utilization evaluation link, the system accurately grasps the current resource usage status by calculating resource occupancy and margin analysis. Especially in terms of threshold calculation, scientific resource adjustment standards are established through load evaluation and adjustment coefficient calculation. In the process of formulating the adjustment plan, the system ensures the accuracy of resource adjustment through classification processing and step length calculation. Finally, the optimal adjustment plan is selected through constraint verification and optimization sorting. This dynamic resource adjustment method is particularly suitable for the characteristics of rapid changes in communication needs in fire rescue scenarios, and can realize real-time optimization of network resources.

[0157] According to one aspect of the present application, step S54 is specifically:

[0158] Step S541, decompose the quality assurance parameters in the service quality data to obtain service quality data; extract terminal working characteristics and terminal status characteristics from the terminal feature data set to generate terminal status data; perform adaptability analysis on the service quality data and terminal status data to form parameter matching data.

[0159] Step S542: Based on the parameter matching data, calculate the service transmission rate requirement and delay requirement to obtain transmission requirement data; prioritize each service according to the transmission requirement data to generate service priority data; constrain the service priority data and system resource limitations to form resource limitation data.

[0160] It is also possible to predict the demand data and then compare them. The transmission demand can be predicted by: resource demand R(t) = μ•H(t) + (1-μ)•P(t) + K•V(t); where: H(t) = Σwi•ri(ti), H(t) is the historical trend function, wi is the time weight, and ri is the historical record; P(t) = a•exp(b•t) / (1 + exp(b•t)), P(t) is the prediction function, a is the growth limit, and b is the growth rate; V(t) = σ 2 / μt, V(t) is the fluctuation correction function, σ 2 is the variance, μt is the mean; K = 1- exp(-|dR / dt| / R0), K is the dynamic correction coefficient, dR / dt is the rate of change, and R0 is the reference rate of change.

[0161] Step S543, dynamically divide the parameters in the resource restriction data to obtain parameter range data; based on the parameter range data, calculate the time slot allocation ratio, frequency allocation ratio and power allocation ratio to generate allocation ratio data; convert the allocation ratio data into specific resource allocation parameters to form real-time resource data.

[0162] Among them, the allocation ratio P(r) = α·D(r) + β·E(r) + γ·Q(r); where: D(r) = (Dr / Dt)·exp(-|Lr - L0| / Ls), D(r) is the demand function, Dr is the requested resource amount, Dt is the total resource amount, Lr is the current load, L0 is the optimal load, and Ls is the load standard deviation; E(r) = (1 - Ur / Um)·(1 - Cr / Cm), E(r) is the efficiency function, Ur is the resource utilization, Um is the maximum utilization, Cr is the conflict rate, and Cm is the maximum conflict rate; Q(r) = exp(-BER / BER0)·exp(-PLR / PLR0), Q(r) is the quality function, BER is the bit error rate, PLR is the packet loss rate, BER0 and PLR0 are reference values; α, β, γ are dynamic weights.

[0163] Step S544, extract the power status and temperature status of the terminal from the terminal status data to obtain the operating status data; based on the operating status data, calculate the maximum transmission power and maximum operating time of the terminal to generate working constraint data; according to the working constraint data, determine the power control parameters of the terminal to form power control data.

[0164] Among them, the calculation process of the working constraint data can be: working constraint index C(t) = We•E(t) + Wt•T(t) + Ws•S(t); among them: E(t) = (Er / Em)•exp(-λ•t), E(t) is the power constraint function, Er is the remaining power, Em is the maximum power, λ is the power consumption coefficient, and t is the working time; T(t) = exp(-(|Tc - To|) / Tm), T(t) is the temperature constraint function, Tc is the current temperature, To is the optimal working temperature, and Tm is the maximum temperature deviation; S(t) = (1 - Lr / Lm)•(1 - Pr / Pm), S(t) is the state constraint function, Lr is the resource load rate, Lm is the maximum load rate, Pr is the power consumption rate, and Pm is the maximum power rate; We, Wt, and Ws are dynamic weight coefficients.

[0165] Step S545, combine the power control data with the terminal's transmission frequency parameters to obtain RF parameter data; based on the RF parameter data, calculate the antenna pointing angle and polarization mode to generate antenna parameter data; according to the antenna parameter data, adjust the communication protocol parameters and coding parameters to form protocol parameter data; integrate the RF parameter data, antenna parameter data and protocol parameter data to generate terminal operating parameters.

[0166] Among them, the terminal working parameters can also be optimized through the model: parameter optimization objective function O(p) = α•P(p) + β•Q(p) - γ•C(p); where: P(p) = η•(Pt / Pm)•exp(-|Gt - Go| / Gm), P(p) is the performance function, Pt is the transmission power, Pm is the maximum power, Gt is the antenna gain, Go is the optimal gain, Gm is the maximum gain deviation, and η is the efficiency coefficient; Q(p) = (1 - BER)•(1 - PER)•exp(-D / Do), Q(p) is the quality function, BER is the bit error rate, PER is the packet loss rate, D is the delay, and Do is the target delay; C(p) = (Ec / Em)•(Tc / Tm), C(p) is the cost function, Ec is the energy consumption, Em is the energy consumption upper limit, Tc is the processing time, and Tm is the time upper limit; α, β, and γ are balance coefficients.

[0167] Step S546, decomposing the terminal operating parameters into operating parameters to obtain parameter decomposition data; based on the parameter decomposition data, verifying the validity of the parameters to generate verification result data; and finally adjusting the terminal operating parameters according to the verification result data to form terminal parameter data.

[0168] Among them, parameter verification can be adopted: parameter validity V(p) = Ws•S(p) + Wr•R(p) + We•E(p); among them: S(p) = (1 - |Pc - Pt| / Pmax)•exp(-|fc - ft| / B), S(p) is the stability function, Pc is the current power, Pt is the target power, Pmax is the maximum power, fc is the current frequency, ft is the target frequency, and B is the bandwidth; R(p) = (SNR / SNR0)•(1 - PER)•exp(-D / D0), R(p) is the reliability function, SNR is the signal-to-noise ratio, PER is the packet error rate, and D is the delay; E(p) = (Er / Em)•(1 - Tr / Tm)•η, E(p) is the efficiency function, Er is the residual energy, Em is the maximum energy, Tr is the temperature rise, Tm is the maximum temperature rise, and η is the efficiency coefficient; Ws, Wr, and We are weight coefficients.

[0169] Adaptive optimization and performance guarantee of terminal parameters are achieved. This step first establishes a parameter matching mechanism through adaptability analysis of service quality data and terminal status. In the transmission demand analysis link, the system ensures the service quality of key services through service priority sorting and resource constraint processing. Especially in terms of parameter range division, reasonable allocation of resources is achieved through dynamic range calculation and allocation ratio optimization. In the process of terminal parameter adjustment, the system comprehensively considers the terminal's power status, temperature status and other operating constraints, and ensures the optimal working state of the terminal through the coordinated optimization of RF parameters, antenna parameters and protocol parameters. Finally, a reliable terminal parameter configuration plan is formed through parameter verification and final adjustment. This adaptive parameter optimization method is particularly suitable for the complex working environment of terminals in fire rescue scenarios, and can ensure the stable operation of terminals under various conditions.

[0170] According to another aspect of the present application, a satellite communication device for supporting firefighting and rescue is also provided, comprising:

[0171] at least one processor; and

[0172] a memory communicatively connected to at least one of the processors; wherein,

[0173] The memory stores instructions that can be executed by the processor, and the instructions are used to be executed by the processor to implement the satellite communication method for supporting firefighting and rescue as described in any of the above technical solutions.

[0174] In another embodiment of the present application, the overall composition of the communication system is described:

[0175] Satellite ground terminal equipment, including:

[0176] Satellite portable station and / or satellite communication vehicle equipped with:

[0177] Foldable parabolic antenna system, high-precision antenna control system, integrated transmitting and receiving RF system, baseband signal processing unit, intelligent terminal controller, power management system, and environmental monitoring unit.

[0178] Data acquisition and processing module for:

[0179] Collect multi-source data on site: video data acquisition unit, audio data acquisition unit, environmental sensor data acquisition unit, terminal status data acquisition unit.

[0180] Perform data preprocessing: data time synchronization processing unit, data quality assessment unit and feature extraction processing unit.

[0181] Link analysis and management module, used for: satellite orbit calculation unit, link prediction analysis unit, propagation loss calculation unit, spectrum resource management unit.

[0182] The business scheduling and resource allocation module includes: a business feature analysis unit, a task priority evaluation unit, a resource demand analysis unit and a resource allocation strategy unit.

[0183] The network slicing configuration module includes: a slice template configuration unit, a resource configuration management unit, a slice isolation control unit and an access control management unit.

[0184] The service quality monitoring module includes: a load analysis unit, a resource adjustment unit, a service quality evaluation unit and a parameter optimization unit.

[0185] The general workflow of the system is as follows:

[0186] System initialization phase: the satellite ground terminal equipment starts up and performs self-test; the antenna system automatically unfolds and orients; and each functional module is initialized and configured.

[0187] Data collection stage: real-time collection of multi-source data; data time synchronization processing; feature extraction and quality assessment.

[0188] Link establishment phase: satellite orbit tracking calculation; automatic adjustment of antenna pointing; link quality assessment and optimization.

[0189] Business operation stage: business demand analysis; dynamic resource allocation; service quality assurance; real-time monitoring and adjustment.

[0190] The portable satellite station adopts modular design, which is easy to carry and quickly deploy. The satellite communication vehicle has the ability to adapt to all terrains and can be quickly maneuvered. It adopts redundant design, has environmental adaptability, and supports fault self-diagnosis. It supports automatic link establishment, has intelligent resource scheduling capabilities, and can perform adaptive adjustment of service quality. It supports multiple business concurrency, has differentiated service guarantee capabilities, and can realize priority guarantee of key businesses.

[0191] This system is suitable for communication support at natural disaster sites, communication support for forest fire rescue, emergency rescue command communication, communication support at major accident sites, and remote rescue communication support.

[0192] In summary, the present invention provides a satellite communication method for supporting consumer rescue, and the brief process is as follows:

[0193] Multi-source real-time data is collected from the fire rescue scene, and combined with the pre-stored scene database through time alignment processing to form a time-synchronized data set; terminal status data is collected and feature calculations are performed against preset parameter thresholds to generate a terminal status data set; the time-synchronized data set is passed through a quality assessment model to obtain data quality results; the above data is input into a feature extraction model to generate scene feature data and terminal feature data.

[0194] Visibility data is obtained by performing spatial analysis based on satellite orbit data and terminal location data; propagation loss is calculated in combination with meteorological data; link analysis is performed based on comprehensive terminal feature data to obtain link status data; and available channel data is output based on the evaluation of pre-stored spectrum resource data.

[0195] The scenario feature data and link status data are subjected to business analysis to obtain business feature data; combined with pre-stored task templates and historical resource data, task priority data and resource requirement data are generated through priority evaluation.

[0196] Generate slice template data based on task priority data and available channel data; optimize resource configuration data based on resource demand data; calculate slice isolation data based on link status data; and finally output network configuration data and terminal access data.

[0197] The network configuration data, terminal access data and real-time traffic data are load analyzed to obtain load status data; resource configuration data is combined for adjustment to obtain resource adjustment data; service quality data is calculated based on task priority data; and finally real-time resource data and terminal parameter data are generated.

[0198] In the present invention, in order to solve the problem of low time alignment accuracy of multi-source data, the timestamp information of each type of data is extracted by performing frame extraction processing on video data, segmentation processing on audio data, and sampling processing on sensor and position data. Unified time reference data is generated by calculating the minimum time granularity. Through the resampling mechanism, different types of data are regularized and time-series aligned, ensuring high-precision synchronization of multi-source data.

[0199] To address the problem of unstable communication quality in high-speed mobile scenarios: By calculating the relative speed and acceleration of the satellite, the relative motion model is used to accurately calculate the Doppler shift. By generating carrier frequency compensation parameters, real-time compensation for frequency deviation is achieved. Combining comprehensive attenuation data and terminal RF data, accurate link quality assessment is achieved through the uplink and downlink link budget model.

[0200] To address the lack of dynamic adjustment of the slice isolation mechanism: Accurately analyze the overlap of time slots and frequencies. Evaluate the interference between slices by cross-analyzing the overlap of time and frequency. Implement dynamic slice isolation protection by calculating the minimum isolation bandwidth and protection time slot.

[0201] To address the problem of lack of adaptive optimization of terminal parameter configuration: Generate working constraint data by analyzing the terminal power status and temperature status. Optimize RF parameters, antenna parameters and protocol parameters comprehensively. Ensure the reliability of terminal working parameters through parameter validity verification.

[0202] To address the problem of lack of accurate quantification of task urgency in resource scheduling: By implementing standardized scoring of time parameters and resource requirements, using the hierarchical analysis method to calculate dimension weights, and by grading tasks based on weighted scores, an accurate resource allocation plan is generated.

[0203] The uniqueness of this technical solution lies in the construction of a complete closed-loop system, in which each step closely cooperates and supports each other, and can maintain stable and reliable satellite communication support in complex firefighting and rescue scenarios.

[0204] In another embodiment of the present application, step S6 is further included:

[0205] Step S6: Based on the real-time resource data and terminal parameter data obtained in step S5, performance indicator data is generated through performance indicator collection. Based on the performance indicator data and the pre-stored performance benchmark data, abnormal event data is obtained through abnormal detection calculation. Based on the abnormal event data and the link status data obtained in step S2, fault diagnosis data is formed through fault location analysis. Combined with the fault diagnosis data and the performance indicator data, performance optimization data and terminal adjustment data are generated through optimization strategy calculation.

[0206] Step S61: Based on the real-time resource data and terminal parameter data obtained in step S5, performance monitoring parameters are collected through key indicator statistics to obtain performance indicator data.

[0207] Step S62: Based on the performance indicator data and the pre-stored performance benchmark data, the abnormal pattern characteristics are identified through deviation analysis calculation to generate abnormal event data.

[0208] Step S63: Based on the abnormal event data and the link status data obtained in step S2, fault correlation parameters are calculated through fault feature analysis to form fault diagnosis data.

[0209] Step S64: Combine the fault diagnosis data and the performance index data, calculate through the optimization scheme, obtain the performance improvement parameters and the terminal adjustment parameters, and generate the performance optimization data and the terminal adjustment data respectively.

[0210] By establishing a performance indicator collection mechanism for real-time resource data and terminal parameter data, combined with pre-stored performance benchmark data for anomaly detection, early warning of satellite communication system anomalies is achieved; the fault location analysis method assisted by link status data is adopted to improve the accuracy and efficiency of fault diagnosis; based on the dual constraints of fault diagnosis data and performance indicator data, performance optimization and terminal adjustment strategies are dynamically generated to form an adaptive performance optimization closed loop. This fault diagnosis and performance optimization method based on multi-dimensional data can quickly identify and handle communication anomalies in complex fire rescue scenarios and improve system reliability; through real-time monitoring and adaptive adjustment of performance indicators, the continuity of rescue communications is guaranteed and the impact of communication interruptions on rescue work is reduced; at the same time, based on the anomaly detection mechanism of historical benchmark data, potential faults are discovered in advance, providing rescue personnel with sufficient warning time and ensuring the communication guarantee capability at critical moments.

[0211] In another embodiment of the present application, steps S22 to S24 may also be:

[0212] Step S22, according to the temperature data, humidity data, air pressure data, rainfall intensity data, cloud cover data and pre-stored geographical location data obtained from the meteorological station, obtain the atmospheric data through atmospheric stratification calculation; according to the atmospheric data, calculate the oxygen absorption attenuation value and the water vapor absorption attenuation value to obtain the gas absorption loss data; based on the rainfall intensity data, calculate the rainfall attenuation coefficient to obtain the rainfall loss data; according to the cloud cover data, calculate the cloud attenuation coefficient to obtain the cloud loss data; superimpose the gas absorption loss data, rainfall loss data and cloud loss data to form the propagation loss data.

[0213] By introducing comprehensive analysis of multi-level meteorological parameters (temperature, humidity, air pressure, rainfall intensity, cloud cover) and geographic location data, an accurate atmospheric stratification model was established; based on atmospheric data, oxygen absorption, water vapor absorption, rainfall attenuation and cloud attenuation were calculated respectively to achieve accurate modeling of propagation loss. This multi-factor superposition propagation loss calculation method accurately reflects the impact of different meteorological conditions on satellite communication links; through layered calculation and step-by-step superposition, the accuracy of propagation loss prediction is improved, providing a reliable basis for subsequent link budget and channel evaluation; especially in fire rescue scenarios under severe weather conditions, it can dynamically evaluate the quality of communication links based on real-time meteorological data, provide a scientific basis for parameter adjustment of rescue communication systems, and effectively reduce the risk of communication interruption caused by meteorological factors.

[0214] Step S23, based on the satellite visibility data obtained in step S21, calculate the satellite relative motion speed to obtain relative speed data; based on the relative speed data, calculate the carrier frequency offset to obtain Doppler frequency shift data; based on the Doppler frequency shift data, generate frequency compensation parameters to obtain frequency compensation data; comprehensively analyze the frequency compensation data with the propagation loss data obtained in step S22 and the terminal feature data set obtained in step S1, calculate the uplink budget and the downlink budget to obtain link budget data; based on the link budget data, calculate the signal-to-noise ratio and bit error rate parameters to generate link status data.

[0215] The introduction of satellite relative motion speed calculation and Doppler frequency shift compensation mechanism solves the frequency offset problem in high-speed motion scenarios; through the comprehensive analysis of frequency compensation data, propagation loss data and terminal characteristic data, accurate budgeting of uplink and downlink links is achieved. This link budgeting method based on Doppler effect compensation significantly improves the communication reliability in high-dynamic scenarios; through the accurate calculation of carrier-to-noise ratio and bit error rate, it provides a quantitative basis for link status evaluation; in complex dynamic scenarios in fire rescue, it can track and compensate for frequency offset in real time to ensure the stability of the communication link, which is particularly suitable for the communication guarantee needs of high-speed mobile scenarios such as rapid maneuvers of rescue vehicles or helicopter rescue.

[0216] Step S24: Based on the link status data obtained in step S23 and the pre-stored spectrum resource data, the frequency band availability parameters are calculated through channel quality assessment, and qualified frequency bands are screened to obtain available channel data.

[0217] Based on the improved link status data evaluation mechanism, a more adaptable available channel screening method is established through the precise calculation of frequency band availability parameters and the screening of qualified frequency bands. This channel evaluation mechanism based on link quality can dynamically adjust the frequency band screening criteria according to the real-time link status; through the combination of pre-storage and real-time evaluation of spectrum resource data, the efficiency and accuracy of channel allocation are improved; in the complex electromagnetic environment of fire rescue, the optimal available channels can be quickly identified and screened, avoiding channel congestion and interference, providing reliable spectrum resource guarantee for rescue communications, and ensuring the timely transmission of key rescue information.

[0218] The preferred embodiments of the present invention are described in detail above; however, the present invention is not limited to the specific details in the above embodiments. Within the technical concept of the present invention, various equivalent transformations can be made to the technical solutions of the present invention, and these equivalent transformations all belong to the protection scope of the present invention.

Claims

1. A satellite communication method for supporting firefighting and rescue, characterized in that: The steps include: Step S1: Collect multi-source real-time data and pre-stored scene data from the fire rescue scene, and form a synchronized data set through time alignment processing; combine terminal status data and preset thresholds to evaluate data quality; generate scene feature and terminal feature data sets through feature extraction; Step S2: Perform spatial geometry analysis based on satellite orbit data and terminal position to generate satellite visibility data; calculate propagation loss based on meteorological monitoring data; perform link status analysis based on terminal characteristics; and output available channel data based on spectrum resource evaluation; Step S3: Analyze business characteristics based on scenario characteristics and link status; evaluate task urgency in combination with task templates; analyze resource requirements based on historical resource data; generate task priority and resource allocation requirement data; Step S4: select slice configuration parameters according to task priority and available channels; determine allocation method based on resource requirements; calculate isolation parameters between slices; formulate access control strategy, and output network configuration and terminal access data; Step S5: Perform load analysis based on real-time traffic data; dynamically adjust resource allocation; calculate service quality parameters based on task priority; Generate real-time resource data and terminal parameter data.

2. The satellite communication method for supporting firefighting and rescue according to claim 1, characterized in that: Step S1 is specifically as follows: Step S11, real-time video data, real-time audio data, real-time sensor data and terminal location data collected from the fire rescue scene, as well as historical scene data extracted from the pre-stored multi-source scene database, are processed by time alignment to form a time-synchronized data set; Step S12: Read the antenna parameters, power parameters and frequency parameters in the preset terminal parameter threshold table, collect the real-time power data, real-time temperature data and real-time working status data of the terminal, and generate a terminal status data set through status feature calculation; Step S13: Calculate the data integrity index and reliability index based on the time synchronization data set using the pre-stored data quality assessment model to generate a data quality assessment result; Step S14: input the time synchronization dataset, the terminal status dataset and the data quality assessment result into the pre-trained feature extraction model, extract key features, and generate a scene feature dataset and a terminal feature dataset respectively.

3. The satellite communication method for supporting firefighting and rescue according to claim 2, characterized in that: Step S2 is specifically as follows: Step S21, based on the stored satellite orbit data and the terminal position data obtained in step S1, the elevation angle and azimuth angle data are obtained through spatial coordinate conversion calculation, and the satellite visibility data is comprehensively generated; Step S22: According to the temperature data, humidity data, air pressure data obtained from the weather station and the pre-stored geographical location data, the frequency-related attenuation value is obtained through atmospheric propagation attenuation calculation to form propagation loss data; Step S23, based on the satellite visibility data, the propagation loss data and the terminal feature data set obtained in step S1, by analyzing the uplink and downlink links, the signal strength and bit error rate parameters are calculated to generate link status data; Step S24: Based on the link status data and the pre-stored spectrum resource data, the frequency band availability parameters are calculated through channel quality evaluation, and qualified frequency bands are screened to obtain available channel data.

4. The satellite communication method for supporting firefighting and rescue according to claim 3, characterized in that: Step S3 is specifically as follows: Step S31, based on the scene feature data set obtained in step S1 and the link state data obtained in step S2, extract the service scale characteristics through service classification analysis, calculate the timeliness parameters, and form service feature data; Step S32: According to the business feature data and the pre-stored task template data, by matching the task type, calculating the response time parameter and the completion deadline parameter, and obtaining the task urgency data; Step S33: Based on the service feature data and the pre-stored historical resource data, the bandwidth requirement value and the delay requirement value are calculated through resource consumption analysis to generate resource requirement data; Step S34: Based on the task urgency data and resource requirement data, a multi-dimensional scoring calculation is performed to obtain a task ranking value and a resource allocation value, and the task priority data and the resource allocation requirement data are output respectively.

5. The satellite communication method for supporting firefighting and rescue according to claim 4, characterized in that: Step S4 is specifically as follows: Step S41: Calculate the slice parameters and slice configuration requirements of each service type based on the task priority data obtained in step S3 and the available channel data obtained in step S2, determine the slice function configuration scheme, and comprehensively form slice template data; Step S42: Calculate the time slot allocation scheme, frequency point allocation scheme and power allocation parameters according to the slice template data and the resource allocation requirement data obtained in step S3, and generate resource configuration data through resource optimization processing; Step S43: Analyze the interference between slices according to the resource configuration data and the link status data obtained in step S2, evaluate the degree of resource overlap, determine the isolation protection requirements, and output the slice isolation data; Step S44: Comprehensive resource configuration data and slice isolation data, determine terminal access conditions, calculate access priority requirements, formulate resource reservation plans, and obtain network configuration data and terminal access data respectively.

6. The satellite communication method for supporting firefighting and rescue according to claim 5, characterized in that: Step S5 is specifically as follows: Step S51, according to the network configuration data, terminal access data and collected real-time traffic data obtained in step S4, calculate the resource utilization rate, count the service load parameters, and obtain the load status data; Step S52: Based on the load status data and the resource configuration data obtained in step S4, a resource adjustment range is determined by dynamic threshold calculation to generate resource adjustment data; Step S53: Calculate quality assurance parameters based on the resource adjustment data and the task priority data obtained in step S3, and output service quality data through service level classification; Step S54: Combining the service quality data with the terminal feature data set obtained in step S1, the resource allocation parameters and the terminal working parameters are obtained through parameter adaptive calculation, and real-time resource data and terminal parameter data are generated respectively.

7. The satellite communication method for supporting firefighting and rescue according to claim 1, characterized in that: Step S11 is specifically as follows: Step S111, performing frame extraction processing on the real-time video data collected at the fire rescue scene, extracting the video frame timestamp information, and obtaining video timestamp data; performing segmentation processing on the real-time audio data, extracting the audio segment timestamp information, and obtaining audio timestamp data; Sampling and processing the real-time sensor data and terminal location data, extracting sampling timestamp information, and obtaining sensor timestamp data and location timestamp data; Step S112, generating time reference data by minimum time granularity calculation according to the video time stamp data, the audio time stamp data, the sensor time stamp data and the position time stamp data; Based on the time reference data, various types of data are divided into time windows to obtain time window data; Step S113: construct a time index for the historical scene data in the pre-stored multi-source scene database according to the time window data to obtain a historical data index; based on the historical data index, extract the historical data corresponding to the current time window to form matching historical data; Step S114, resampling the real-time video data according to the time window data to obtain regularized video data; Resample the real-time audio data according to the time window data to obtain regular audio data; interpolate the real-time sensor data and the terminal position data according to the time window data to obtain regular sensor data and regular position data respectively; Step S115: align the regularized video data, regularized audio data, regularized sensor data, regularized location data and matching historical data in time sequence according to the time reference data to generate a time synchronization data set.

8. The satellite communication method for supporting firefighting and rescue according to claim 1, characterized in that: Step S12 is specifically as follows: Step S121, read the antenna gain threshold and antenna polarization parameter from the preset terminal parameter threshold table to obtain antenna threshold data; read the transmission power upper limit and the power adjustment step to obtain power threshold data; read the working frequency band range and the frequency offset limit to obtain frequency threshold data; combine the antenna threshold data, the power threshold data and the frequency threshold data to form parameter threshold data; Step S122, calculating the remaining power of the collected real-time power data to obtain remaining power data; calculating the sustainable working time based on the remaining power data to generate working time data; matching the working time data with the power parameter in the parameter threshold data to form power status data; Step S123, calculating the temperature change rate of the collected real-time temperature data to obtain temperature change data; calculating the temperature trend characteristics based on the temperature change data to generate temperature trend data; Matching the temperature trend data with the power parameters in the parameter threshold data to form temperature status data; Step S124, performing work mode recognition on the collected real-time work status data to obtain work mode data; Based on the working mode data, the resource occupancy is calculated to generate resource occupancy data; the resource occupancy data is matched with the frequency parameters in the parameter threshold data to form the operation status data; Step S125, combining the power state data, the temperature state data and the operation state data into state characteristic data to obtain state characteristic data; Based on the status feature data, the overall working status of the current terminal is calculated to generate a terminal status data set.

9. The satellite communication method for supporting firefighting and rescue according to claim 1, characterized in that: Step S14 is specifically as follows: Step S141, input the regularized video data in the time synchronization data set into the pre-trained video feature model, extract the target detection features, scene segmentation features and action recognition features, and obtain video feature data; input the regularized audio data into the pre-trained audio feature model, extract the sound category features, speech recognition features and environmental noise features, and obtain audio feature data; The regularized sensor data is classified and processed according to the sensor type, and the temperature change characteristics, pressure change characteristics and gas concentration characteristics are extracted to obtain the sensor characteristic data; Step S142: construct a multimodal feature fusion matrix based on the video feature data, the audio feature data and the sensor feature data to obtain a fusion feature matrix; Perform time series correlation analysis on the fused feature matrix and the matching historical data to generate time series related features; According to the time series related features, the scene change trend is calculated to obtain the scene evolution data; Step S143, constructing a terminal state feature vector according to the real-time power data, real-time temperature data and real-time working state data in the terminal state data set to obtain terminal state features; performing weight calculation on the integrity index and reliability index in the data quality assessment result to obtain quality weight data; Based on the terminal state characteristics and quality weight data, a terminal operation characteristic matrix is ​​generated to obtain the terminal operation characteristics; Step S144: Perform multi-dimensional feature aggregation on the fused feature matrix, scene evolution data and time-series related features, screen key scene features through an adaptive feature selection algorithm, and generate a scene feature data set; perform feature combination on the terminal state features, quality weight data and terminal working features, extract key terminal features through a dynamic threshold screening algorithm, and generate a terminal feature data set.

10. A satellite communication device for supporting firefighting and rescue, characterized in that: include: at least one processor; as well as a memory communicatively connected to at least one of the processors; wherein, The memory stores instructions that can be executed by the processor, and the instructions are used to be executed by the processor to implement the satellite communication method for supporting firefighting and rescue as described in any one of claims 1 to 9.

Images