CN120692557A - A distributed collaborative processing method and system for ground satellite signals - Google Patents

Abstract

The present invention provides a distributed collaborative processing method and system for ground satellite signals, relating to the field of communications technology. The method comprises: extracting a user MAC address hash value, an emergency priority code, and a spectrum occupancy matrix from a synchronous network to generate a triplet feature vector; transmitting the triplet feature vector through a two-layer mesh network; selecting two reference target points with fixed geographic coordinates on the ground in the disaster area, obtaining their three-dimensional coordinates, and constructing a spatial reference line; calculating the vertical projection distance of each node to the spatial reference line based on the real-time position data of each node to generate a node offset parameter; and dynamically generating a refraction correction coefficient for the signal propagation path based on a comparison result of the node offset parameter with a preset threshold. When the offset exceeds the threshold, a piecewise linear interpolation method is used to increase the correction weight. The present invention can achieve highly reliable and low-latency emergency communication coverage.

Description

Ground satellite signal distributed cooperative processing method and system

Technical Field

The invention relates to the technical field of communication, in particular to a distributed cooperative processing method and system for ground satellite signals.

Background

In the prior art, disaster area communication still depends on fixed ground base stations or temporarily deployed single-type communication nodes, but the following limitations exist in the schemes:

For example, the conventional node deployment generally adopts a uniform distribution strategy, and cannot be adaptively adjusted according to ruin porosity and dynamic change of obstacles in disaster areas, so that signal coverage dead zones are large, resource utilization rate is low, and particularly in complex terrains, fixed nodes are easily influenced by displacement of obstacles caused by aftershocks, so that communication links are frequently interrupted.

For example, some use static path loss models for signal compensation without taking into account real-time refraction effects of signal propagation paths in disaster area dynamic environments. For example, scattering of high-frequency signals by metal debris in ruins, refractive index differences of obstacles made of different materials and the like can cause signal strength fluctuation and time delay jitter, and the traditional method lacks a dynamic correction mechanism based on a geospatial reference.

Disclosure of Invention

The invention aims to solve the technical problem of providing a distributed cooperative processing method and system for ground satellite signals, which can realize high-reliability and low-time-delay emergency communication coverage.

In order to solve the technical problems, the technical scheme of the invention is as follows:

In a first aspect, a method for distributed cooperative processing of terrestrial satellite signals, the method comprising:

s1, extracting a user MAC address hash value, an emergency priority code and a frequency spectrum occupation matrix from a synchronous network to generate a triplet feature vector;

Step S2, transmitting the triplet feature vector through a double-layer Mesh network, selecting two reference target points with fixed geographic coordinates on the ground of a disaster area, acquiring three-dimensional coordinates of the reference target points, constructing a space reference straight line, calculating the vertical projection distance from each node to the space reference straight line according to real-time position data of each node, generating a node offset parameter, dynamically generating a refraction correction coefficient of a signal propagation path based on a comparison result of the node offset parameter and a preset threshold value, and increasing the correction weight by adopting a piecewise linear interpolation method when the offset exceeds the threshold value;

S3, receiving the corrected triplet feature vector by the regional master node, and performing association mapping on a time slot allocation scheme, a frequency band switching strategy and a correction weight through space-time interleaving coding to generate enhanced three-dimensional virtual base station configuration containing dynamic beam pointing parameters;

And S4, establishing a communication link based on the enhanced three-dimensional virtual base station configuration, adjusting the link direction in real time according to the beam pointing parameter, and dynamically optimizing the signal intensity and time delay of the communication link to finish the signal coverage of the disaster area.

In a second aspect, a terrestrial satellite signal distributed co-processing system comprises:

the generation module is used for extracting the user MAC address hash value, the emergency priority code and the frequency spectrum occupation matrix from the synchronous network and generating a triplet feature vector;

The correction module is used for transmitting the triplet feature vector through a double-layer Mesh network, selecting two reference target points with fixed geographic coordinates on the ground of a disaster area, acquiring three-dimensional coordinates of the two reference target points, constructing a space reference straight line, calculating the vertical projection distance from each node to the space reference straight line according to real-time position data of each node, generating a node offset parameter, dynamically generating a refraction correction coefficient of a signal propagation path based on a comparison result of the node offset parameter and a preset threshold value, and increasing the correction weight by adopting a piecewise linear interpolation method when the offset exceeds the threshold value;

The mapping module is used for receiving the corrected triplet feature vector by the regional master node, and performing associated mapping on a time slot allocation scheme, a frequency band switching strategy and a correction weight through space-time interleaving coding to generate enhanced three-dimensional virtual base station configuration containing dynamic beam pointing parameters;

And the optimization module is used for establishing a communication link based on the enhanced three-dimensional virtual base station configuration, adjusting the link direction in real time according to the beam pointing parameter, and dynamically optimizing the signal intensity and time delay of the communication link to finish the signal coverage of the disaster area.

In a third aspect, a computing device includes:

One or more processors;

And a storage means for storing one or more programs that, when executed by the one or more processors, cause the one or more processors to implement the method.

In a fourth aspect, a computer readable storage medium has a program stored therein, which when executed by a processor, implements the method.

The scheme of the invention at least comprises the following beneficial effects:

The method comprises the steps of calculating ruin porosity based on three-dimensional scanning data of terrains in disaster areas, carrying out differential deployment and monitoring displacement change of barriers in real time, dynamically adjusting clock synchronization period in combination with a distributed Bayesian-tolerant consensus algorithm, forming a synchronous communication network resistant to terrains and interference in a multi-hop relay mode, effectively aiming at complex terrains in disaster areas, improving communication stability, accurately extracting user MAC address hash values, emergency priority codes and spectrum occupation matrixes from the synchronous network, dynamically classifying users in combination with a disaster emergency class table through correlation of historical data of signal transmission power of user equipment, and distributing resources more reasonably to ensure life detection and rescue command communication preferentially. The method comprises the steps of constructing a space reference straight line, calculating the vertical projection distance from a node to the reference straight line to generate a node offset parameter, dynamically generating a refraction correction coefficient of a signal propagation path based on the offset parameter, carrying out convolution operation on a frequency spectrum occupation matrix, carrying out phase compensation on a user MAC address hash value, effectively correcting the signal propagation path, improving the signal transmission quality, and carrying out associated mapping on a time slot allocation scheme, a frequency band switching strategy and correction weights by a regional master node through space-time interleaving coding to generate an enhanced three-dimensional virtual base station configuration containing dynamic beam pointing parameters. The communication link is established based on the enhanced three-dimensional virtual base station configuration, the link direction is adjusted in real time, the signal intensity and the time delay of the communication link are dynamically optimized through a self-adaptive wave beam adjustment mechanism and a Kalman filter, the cooperative transmission of the whole network node is activated, the disaster area target area is effectively covered, the signal blind area is reduced, and the disaster area signal coverage rate and the communication quality are improved.

Drawings

Fig. 1 is a flow chart of a distributed cooperative processing method for terrestrial satellite signals according to an embodiment of the present invention.

Fig. 2 is a schematic diagram of a distributed cooperative processing system for terrestrial satellite signals according to an embodiment of the present invention.

Detailed Description

Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in form and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

As shown in fig. 1, an embodiment of the present invention proposes a distributed cooperative processing method for terrestrial satellite signals, including the following steps:

s1, extracting a user MAC address hash value, an emergency priority code and a frequency spectrum occupation matrix from a synchronous network to generate a triplet feature vector;

Step S2, transmitting the triplet feature vector through a double-layer Mesh network, selecting two reference target points with fixed geographic coordinates on the ground of a disaster area, acquiring three-dimensional coordinates of the reference target points, constructing a space reference straight line, calculating the vertical projection distance from each node to the space reference straight line according to real-time position data of each node, generating a node offset parameter, dynamically generating a refraction correction coefficient of a signal propagation path based on a comparison result of the node offset parameter and a preset threshold value, and increasing the correction weight by adopting a piecewise linear interpolation method when the offset exceeds the threshold value;

S3, receiving the corrected triplet feature vector by the regional master node, and performing association mapping on a time slot allocation scheme, a frequency band switching strategy and a correction weight through space-time interleaving coding to generate enhanced three-dimensional virtual base station configuration containing dynamic beam pointing parameters;

And S4, establishing a communication link based on the enhanced three-dimensional virtual base station configuration, adjusting the link direction in real time according to the beam pointing parameter, and dynamically optimizing the signal intensity and time delay of the communication link to finish the signal coverage of the disaster area.

According to the embodiment of the invention, the synchronization problem of dynamic node position change is solved through spatial reference straight line modeling, high-precision coordination among nodes can be maintained under a GPS-free scene, the signal demodulation error rate is reduced, the signal correction weight is dynamically adjusted based on node offset, signal fading caused by topography such as ruins and metal structures is effectively compensated, signal attenuation of a shielding area is reduced, and a communication blind area is eliminated. The method combines emergency priority and spectrum occupation to dynamically allocate channel resources, preferentially guarantees real-time service transmission such as rescue, improves multi-source signal separation accuracy, reduces network interruption risk, dynamically adjusts beam pointing to expand signal coverage, focuses on a high-demand area, reduces edge node power consumption, prolongs equipment endurance, reduces data packet loss, guarantees stable data transmission, and can still maintain stable communication in a strong interference environment by dual processing of phase compensation and convolution operation to inhibit electromagnetic interference and multipath fading influence.

In a preferred embodiment of the present invention, in step S1, the method further includes, before extracting the user MAC address hash value, the emergency priority code and the spectrum occupation matrix from the synchronous network and generating the triplet feature vector:

calculating ruin porosity based on three-dimensional scanning data of the topography of the disaster area, dividing a high porosity area and a low porosity area, and executing differential deployment strategies according to the porosity difference to form a node network;

Monitoring the displacement change of the obstacle in real time based on the node network, and if the local porosity change exceeds a preset threshold value due to the detection of the displacement of the obstacle, hovering to a new pore position again according to a pore distribution thermodynamic diagram fed back by the individual nodes so as to obtain the adjusted node distribution density of the hybrid topology network;

based on the adjusted distribution density of the nodes of the hybrid topology network, a dynamic voting mechanism in a distributed Bayesian-busy-court fault-tolerant consensus algorithm is adopted, the clock synchronization period is dynamically adjusted according to the signal propagation delay between the nodes, and a synchronization signal is broadcast to the nodes shielded by the barrier in a multi-hop relay mode, so that a synchronous communication network resistant to the topographic interference is formed.

In the embodiment of the present invention, when applied specifically, the above steps may be implemented in the following manner, for example:

Dividing three-dimensional scanning data of disaster areas into cube grid units according to spatial resolution (such as 1m multiplied by 1 m), extracting volume proportion of ruin materials (concrete, reinforcing steel bars, masonry and the like) in each grid and space distribution, calculating pore volume by 'void volume=unit total volume-solid material volume', and forming a global porosity matrix by porosity=pore volume/unit total volume multiplied by 100 percent for each grid unit.

Region division and differential deployment:

Setting a porosity critical value (such as 30%), defining a region with porosity of >30% as a 'high porosity region' (with better signal penetrability but changeable terrain), deploying by adopting unmanned aerial vehicle hovering nodes, wherein hovering points are randomly selected at a safe height (15-30 m) according to the area of the region and the signal coverage radius (such as 500 m), and ensuring that the overlapping rate of the signal coverage edges of adjacent nodes is less than or equal to 10% so as to avoid interference.

The area with the porosity less than or equal to 30% is a low-porosity area (signal shielding is serious), the vehicle-mounted mobile node encryption deployment is adopted, nodes are arranged on a road or an open path according to grid intervals (such as 200m multiplied by 200 m), the position with higher topography or fewer obstacles is preferentially selected, the signal transmission loss in the range of 50m of the periphery is scanned in real time through a vehicle-mounted sensor, and the node intervals are dynamically adjusted until no blind area is covered by the signal.

The individual soldier nodes acquire triaxial motion data in real time through an inertial navigation module (accelerometer and gyroscope), and calculate three-dimensional displacement (delta x, delta y and delta z) of the obstacle once every 2 seconds by combining barometer height information, and trigger early warning when the displacement amount is more than 20cm for 3 continuous times.

The displacement obstacle is taken as the center, an influence area with the radius of 50m is defined, the porosity data of the three-dimensional grid in the area is called, the porosity change value (delta rho=rho new-rho original) before and after displacement is calculated, and if the local area delta rho is more than 15% (namely, the porosity is obviously reduced or increased), a pore distribution thermodynamic diagram (the porosity change gradient is marked by different colors) is generated.

Unmanned aerial vehicle node relocation strategy:

After the unmanned aerial vehicle receives thermodynamic diagram data, a new position (namely an unblocked high-porosity channel) with the porosity more than or equal to an original threshold value of +5% in an influence area is preferentially selected, an obstacle avoidance path is planned based on Dijkstra algorithm, an obstacle after displacement is avoided (the safety distance is kept to be more than or equal to 10 m), after hovering to a target point, a coverage effect is verified through inter-node signal mutual detection (the strength of a received signal is required to be more than or equal to-80 dBm), and if the coverage effect does not reach the standard, a suboptimal position is reselected within a range of 50m nearby.

Signal propagation delay measurement and synchronization period adjustment:

Each node periodically (initial period 10 s) sends ping packets containing time stamps to adjacent nodes, a receiver records the receiving time, single-pass propagation delay is calculated through a bidirectional delay formula ((receiving time-sending time)/2), and the average value is continuously measured for 5 times to be taken as the inter-node delay tau.

When the time delay tau of more than 30% of nodes in the network is more than 50ms (indicating that the network topology is complex or has shielding), triggering a dynamic voting mechanism, wherein the nodes participate in voting according to own time delay values, if a plurality of nodes (2/3) support to shorten the synchronization period, the period is adjusted to be 5s, if the time delay is generally less than 20ms, the period is recovered for 10s, and the synchronization precision and the energy consumption are balanced.

Multi-hop relay anti-occlusion broadcasting:

the synchronized beacon node preferentially selects the adjacent node with the signal strength of-70 dBm and the time delay of <30ms as a relay node to construct a multi-hop path.

For the nodes shielded by the barrier (continuous 3 times of failure of receiving the synchronous signals), a blind relay mode is started, namely the beacon node broadcasts an enhanced signal (power is improved by 3 dB) to the direction of a shielded area, the relay node reversely pushes the shielded position through the RSSI signal strength, and bypasses the barrier (such as bypassing building ruins) to forward the signal, so that the arrival rate of the synchronous signal is ensured to be more than or equal to 95%.

According to the embodiment of the invention, the unmanned aerial vehicle and the vehicle-mounted node are deployed in a differentiated mode according to the signal propagation characteristics of high/low pore areas through quantitative analysis of three-dimensional porosities, the unmanned aerial vehicle is used for flexible coverage in the low shielding areas, dense relay is carried out through the vehicle-mounted node in the strong shielding areas, the signal coverage balance of the whole network is improved, the porosity mutation caused by obstacle displacement is perceived in real time, the position of the unmanned aerial vehicle node is dynamically adjusted based on thermodynamic diagram, the nodes are always positioned on a signal transmission optimal path in dynamic scenes such as aftershocks and collapse, local communication blind areas caused by environmental changes are avoided, the synchronization period is dynamically adjusted according to the signal propagation delay, the synchronization interval is shortened under complex topology to reduce clock deviation, the multi-hop relay and blind relay strategies are combined, the shielding limit of the obstacle on the synchronization signal is broken, the time consistency of the distributed nodes is guaranteed, and the collaborative processing efficiency and the system reliability are improved.

In a preferred embodiment of the present invention, step S1 of extracting the user MAC address hash value, the emergency priority code and the spectrum occupation matrix from the synchronous network, generating the triplet feature vector may include:

s1-1, capturing broadcast signals of user equipment through a synchronous network, extracting an MAC address of the broadcast signals, and generating a unique hash identification code with a fixed length by adopting an SHA-256 hash algorithm;

S1-2, dynamically classifying users based on hash identification codes and associated signal transmission power historical data of user equipment in combination with a preset disaster emergency level table, wherein if the signal transmission power is continuously higher than a threshold value and the identification codes belong to a white list of rescue equipment, the signals are marked as life detection types;

S1-3, sampling a frequency band where a life detection user is located according to emergency priority coding, sparse sampling is carried out on the frequency band where a common communication user is located, and signal intensities of different frequency bands are integrated into a frequency spectrum occupation matrix according to time sequences;

And S1-4, packaging data of the hash identification code, the step emergency priority code and the spectrum occupation matrix, and arranging the data into a triplet feature vector according to a preset binary format, wherein in the data structure of the triplet feature vector, the hash identification code is used as a head field, the emergency priority code is used as an intermediate control field, the spectrum occupation matrix is used as a tail load field, and the seamless splicing matrix is realized by the length identifier.

In the embodiment of the present invention, when applied specifically, the above steps may be implemented in the following manner, for example:

The implementation process of the step S1-1 can be as follows:

broadcast signal capturing, namely continuously monitoring broadcast signals (such as Beacon frames of Wi-Fi and broadcast packets of Bluetooth) in a frequency band of 2.4GHz/5GHz by an access node in a synchronous network, screening effective signals through signal strength detection (RSSI > -90 dBm), and extracting a source MAC address (48-bit hexadecimal character string) in the signals.

The hash algorithm processing is that the MAC address is converted into byte stream, the byte stream is input into SHA-256 hash algorithm, and through 64 rounds of iterative operation (including message filling, grouping and compression function processing), 256-bit hexadecimal hash values with fixed length (such as 64-bit character strings) are generated and used as unique identification codes of user equipment, so that the hash values of different MAC addresses are ensured to be collision-free and irreversibly restored.

The implementation process of the step S1-2 can be as follows:

Historical data association, each node maintains a signal transmitting power log of the user equipment (transmitting power is recorded every 10 seconds, and data of the last 5 minutes are reserved), and average transmitting power is calculated (for example, default transmitting power of a mobile phone is 20dBm, and professional equipment can reach 30 dBm).

And (3) judging classification conditions:

life detection class, if the average transmitting power lasts for 5 minutes >25dBm (threshold value), and the hash identification code is matched with a preset rescue equipment white list (such as a life detector and the MAC prefix of the unmanned aerial vehicle control terminal), the highest priority (code 001) is marked.

Rescue command class, if the signal transmission delay (round trip time RTT) <50ms (threshold value), and the hash identification code exists in a command center special equipment database (such as a vehicle-mounted command terminal and a MAC list of a satellite phone), marking as sub-priority (code 010).

Common communication class, user equipment (such as mobile phones of people suffering from disaster) which does not meet the above conditions is marked as basic priority (code 100).

And (3) allocating a 3-bit binary emergency priority code (001/010/100) to each user equipment according to the classification result for subsequent resource scheduling.

The implementation process of the step S1-3 can be as follows:

And the differentiated sampling strategy is that a life detection user performs high-density sampling in a communication frequency band (such as a special rescue frequency band 400-470 MHz), and signal intensity is acquired every 10ms (200 frequency points are covered, each frequency point is spaced by 100 kHz) for 1 second to form a 200 multiplied by 100 time-intensity matrix. The common communication users perform sparse sampling in a public frequency band (such as 2.4 GHzWi-Fi), acquire signal strength (covering 50 frequency points) every 100ms, and form a 50×10 time-strength matrix for 1 second.

And (3) matrix integration, namely splicing the frequency band sampling data of different users according to the dimension of frequency band sequence number multiplied by time stamp, wherein rows represent frequency points (in ascending order), columns represent time (granularity is 10 ms), unit grid values are signal strength (unit dBm), and finally a frequency spectrum occupation matrix with dimension of NxM is formed (N is the total frequency point number, and M is the time sequence length).

The implementation process of the step S1-4 can be as follows:

Field definition:

header field (hash identification code) the 256-bit hash value is converted into 32-byte binary data, and the first 2 bytes are added with field length identification (0×0020), so that the receiving end can identify the header boundary.

Intermediate field (priority coding) 3-bit binary coding is extended to 1 byte (e.g., 001→0× 01,010 →0× 02,100 →0×04), with the first 1 byte added with a length identification (0×0001).

The tail field (spectrum occupation matrix) is to convert the matrix into binary stream, store it in the order of "row priority" (the first row frequency point 1-10ms intensity is stored first, the second row frequency point 1-10ms intensity is stored again), the matrix dimension information (N x M) is used as prefix (4 bytes N+4 bytes M), the total length is identified as 8+N xM bytes.

Splicing according to the sequence of head length identification, head data, middle length identification, middle data, tail length identification and tail data to form a complete binary feature vector, and enabling each field to realize unambiguous analysis through a length identifier (for example, a receiving end reads the head length of 0 multiplied by 0020 and then reads the 32-byte hash value).

In the embodiment of the invention, the device identification code is generated through SHA-256 hash, so that the uniqueness and data security of different devices are ensured, privacy risks caused by direct exposure of MAC addresses are avoided, meanwhile, reliable identity references are provided for subsequent user classification, the life detection, rescue command and common communication users are accurately distinguished by combining transmission power, signal delay and device whitelist, the resource hierarchical scheduling under emergency scenes is realized, the response efficiency of key services is improved, high-density spectrum sampling is implemented for high-priority users, the signal occupation condition of the high-priority users is monitored in real time, the smoothness of rescue frequency bands is ensured, sparse sampling is adopted for the common users, the calculation load of the system is reduced, and the spectrum monitoring precision and the resource consumption are balanced. By means of triple packaging of fixed field formats and length identifiers, efficient splicing and analysis of different types of data are achieved, unified input interfaces are provided for subsequent distributed cooperative processing, and system compatibility and processing efficiency are improved.

In a preferred embodiment of the present invention, step S2 includes transmitting the triplet feature vector through a double-layer Mesh network, selecting two reference target points with fixed geographic coordinates on the disaster area ground, obtaining three-dimensional coordinates thereof, and constructing a spatial reference line, which may include:

S2-1, selecting the top end of a building with a stable foundation at the entrance of a disaster area as a first target point according to the topological characteristics of a disaster area rescue map, and simultaneously selecting a directional antenna tower base deployed by a rescue center as a second target point to acquire the longitude and latitude and altitude data of a geodetic coordinate system of the two target points;

s2-2, converting the acquired geodetic coordinates into a space rectangular coordinates, wherein the method specifically comprises the steps of calculating a conversion matrix from the geodetic coordinates of the target points to the space rectangular coordinates based on WGS-84 ellipsoid parameters;

And S2-3, constructing a space reference straight line by adopting a two-point type straight line equation based on three-dimensional space coordinates of the two target points.

In the embodiment of the present invention, when applied specifically, the above steps may be implemented in the following manner, for example:

the implementation process of the step S2-1 can be as follows:

target point screening criteria:

The first target point is to preferentially select more than 3 layers of building tops (such as hotels and office buildings) with no cracks or inclination of foundations at the entrance of the disaster area based on the road network and building distribution in the disaster area rescue map, and the surrounding 100m range of the first target point is required to have no obstacle higher than 1/2 height of the first target point, so that the Beidou signal is ensured to be received without shielding;

The second target point is that a directional antenna tower base fixedly deployed by the rescue command center is the first choice, if no selectable open area (such as a square center and a mountain top) is available, the condition that the terrain is flat and the ground surface has no dynamic displacement (such as soft sandy soil area removal) is required, and the visibility between the two points is required to be ensured (namely, the height of an obstacle in the connecting line direction is not more than 10% of the altitude difference between the two points).

And (3) coordinate acquisition flow:

Through the Beidou terminal (positioning accuracy +/-2 cm) with RTK differential positioning function, longitude and latitude (B, L) and altitude (H) data of 1 minute (1 second interval) are continuously collected for each target point, abnormal values with deviation exceeding 5cm are removed, and then average values are obtained, so that stable geodetic coordinates (B ₁,L₁,H₁) and (B ₂,L₂,H₂) are finally obtained.

The implementation process of the step S2-2 can be as follows:

Initializing ellipsoid parameters:

the WGS-84 ellipsoid basic parameters are configured to be a long half shaft a= 6378137m, a short half shaft b= 6356752.3142m and a flat rate f= (a-b)/a=1/298.257223563, and are used for geometric modeling of subsequent coordinate transformation.

Coordinate conversion is implemented:

For each target point, firstly calculating a meridian curvature radius N and a mortise circle curvature radius M (based on latitude B) as geometric parameters of coordinate transformation;

Through a coordinate conversion engine (such as an API interface supporting NMEA-0183 protocol) arranged in the Beidou positioning module, the module automatically completes conversion from a geodetic coordinate system (longitude and latitude+altitude) to a space rectangular coordinate system (X, Y, Z), and outputs three-dimensional coordinates (X ₁,Y₁,Z₁) and (X ₂,Y₂,Z₂), wherein the conversion accuracy is limited by module hardware to be within +/-5 cm.

The implementation process of the step S2-3 may be:

Determining a linear direction vector:

taking the first target point P ₁(X₁,Y₁,Z₁) as a starting point and the second target point P ₂(X₂,Y₂,Z₂) as an end point, calculating the three-dimensional space distance between the two points (the coordinate difference of the XYZ axes is synthesized through the Pythagorean theorem), requiring the distance to be more than or equal to 500 meters to ensure the space representativeness of the reference straight line, and if the distance is insufficient, reselecting the target point;

The coordinate differences are extracted as direction vectors (DeltaX, deltaY, deltaZ) for describing the spatial course of the reference line.

Straight line equation construction and storage:

defining a space reference straight line by adopting a parameterized representation method, wherein the coordinate of any point P on the straight line can be expressed as P=P ₁ +t (delta X, delta Y, delta Z), wherein t is a real parameter;

the starting point coordinates and the direction vectors of the reference straight line are packaged into a standard data structure (such as a structure body containing (X ₁,Y₁,Z₁), deltaX, deltaY and DeltaZ), and are synchronized to a local coordinate system database of all nodes through a double-layer Mesh network to serve as reference of subsequent space positioning.

In the embodiment of the invention, the standard coordinates are ensured not to be affected by local terrain changes in dynamic environments such as aftershocks and building collapse by screening target points with stable foundation and good visibility and combining with RTK differential positioning technology, a long-term reliable space standard is provided for the whole disaster area communication network, the traditional geodetic coordinates are converted into rectangular coordinates suitable for three-dimensional space calculation by utilizing WGS-84 ellipsoid parameters and a Beidou module built-in conversion algorithm, the coordinate system difference of different positioning systems is eliminated, the position calculation of the whole network nodes under the unified space standard is realized, and the subsequent three-dimensional signal propagation model construction and cooperative processing are facilitated. The constructed space reference straight line can be used as the axis of the Mesh division of the disaster area, the geometric reference of the flight path of the unmanned aerial vehicle and the constraint condition of the three-dimensional coordinate calibration of the nodes, the positioning precision of the distributed nodes in the complex ruin environment is improved, accurate space coordinate support is provided for tasks such as the positioning of a life detector, the accurate throwing of the unmanned aerial vehicle and the like, the reference straight line parameters are synchronized to all the nodes through a double-layer Mesh network, the relative position calibration of each node in the multi-hop transmission of signals can be carried out based on the reference straight line, the collaborative processing deviation caused by the coordinate misalignment of the local nodes is reduced, and the anti-interference capability of the network under extreme topography is enhanced.

In a preferred embodiment of the present invention, calculating a vertical projection distance from each node to the spatial reference line according to real-time position data of each node, and generating a node offset parameter may include:

S2-4, acquiring three-dimensional coordinates of a real-time position of a node, projecting the three-dimensional coordinates to a plane where a space reference straight line is located, and calculating Euclidean distance between the coordinates of the node and the nearest point on the straight line as a vertical projection distance;

And S2-5, adding a terrain correction factor for the vertical projection distance according to the influence of the disaster area ground surface material on the signal refractive index, and generating a normalized node offset parameter.

In the embodiment of the present invention, when applied specifically, the above steps may be implemented in the following manner, for example:

the implementation process of the step S2-4 may be:

And (3) obtaining node coordinates:

Each node acquires three-dimensional space coordinates (X, Y and Z) of the node in real time through a built-in Beidou/GPS module, the acquisition frequency is 1 time/second, multipath effect and instantaneous noise are eliminated by adopting sliding window filtering (reserving the latest 5 times of data to average value), and stable real-time position coordinates (X _n,Y_n,Z_n) are obtained.

Reference straight line parameter call:

The starting point coordinates (X ₁,Y₁,Z₁) and direction vectors (delta X, delta Y, delta Z) of the space reference straight line are read from the node local database, wherein the direction vectors are two reference point coordinate differences (X ₂-X₁,Y₂-Y₁,Z₂-Z₁).

Calculating projection points:

Constructing a vector (X _n-X₁,Y_n-Y₁,Z_n-Z₁) from the starting point of the reference straight line to the node, and determining the nearest point of the node on the base alignment line through a vector projection principle:

The projection ratio of the included angle between the vector and the reference straight line direction vector is calculated, and the point coordinate (X _p,Y_p,Z_p) closest to the node on the straight line is found.

Vertical distance calculation:

For the node coordinates (X _n,Y_n,Z_n) and the nearest point coordinates (X _p,Y_p,Z_p), calculating Euclidean distances in a three-dimensional space, namely respectively solving coordinate differences along X, Y, Z axes, and synthesizing a final distance by Pythagorean theorem to be used as a vertical projection distance D from the node to a reference straight line.

The implementation process of the step S2-5 can be as follows:

classification of surface materials:

based on three-dimensional scanning data of disaster areas and image recognition results returned by on-site individual soldier nodes, the ground surface materials at the positions of the nodes are divided into three types:

High reflection materials (such as metal ruins and glass curtain walls), medium refraction materials (such as concrete and masonry), low refraction materials (such as soil and vegetation), and initial correction factors of 1.5, 1.0 and 0.8 are respectively set.

Correction factor application:

and (3) carrying out weighted correction on the vertical projection distance D according to the earth surface material type of the real-time position of the node, wherein if the node is positioned in a high reflection material area, the corrected distance D ' =D multiplied by 1.5, the medium material area is not corrected (D ' =D), and the low refraction material area D ' =D multiplied by 0.8.

Normalization:

The corrected distances of the nodes of the whole network are counted, a maximum distance threshold Dmax (such as 500 meters) is determined, and D' of each node is converted into an offset parameter delta in the range of 0-1:

δ=d '/Dmax (if D' > Dmax, δ=1 is taken), it is ensured that the offsets of different regions have a uniform dimension, and subsequent algorithm processing is facilitated.

According to the embodiment of the invention, through calculation of the vertical projection distance, the spatial offset of the node relative to the reference straight line is quantized, geometric constraint is provided for multi-hop path planning of the double-layer Mesh network, diffraction loss of signals in complex terrains is reduced, transmission efficiency is improved, and earth surface material correction factors are introduced to compensate influences of different materials on signal refraction and reflection, so that offset parameters are closer to actual signal transmission path deviation, and accurate environment parameters are provided for subsequent frequency spectrum allocation and interference avoidance algorithms. The normalization processing of the offset parameters eliminates the influence of the geographic range difference of different disaster areas, so that the algorithm can uniformly process node distribution data in different scenes such as plain, mountain areas and the like, the cross-scene adaptation capability of the system is improved, the node position and the ground surface material information are updated in real time, the offset parameters are ensured to be dynamically adjusted along with the change of ruin environments, and data support is provided for real-time optimization of an emergency communication network.

In a preferred embodiment of the present invention, dynamically generating the refraction correction coefficient of the signal propagation path based on the comparison result between the node offset parameter and the preset threshold value, and increasing the correction weight by using piecewise linear interpolation when the offset exceeds the threshold value may include:

S2-6, setting a first threshold value and a second threshold value based on the statistical distribution of the node offset parameter, dividing the node into a low offset region, namely an offset less than or equal to the first threshold value, a middle offset region, namely a first threshold value < offset less than or equal to the second threshold value and a high offset region, namely offset > the second threshold value, and initializing a dynamic compensation mechanism according to the region type;

S2-7, executing a differential correction strategy on the type of the region, namely adopting a fixed correction coefficient for the nodes in the low-offset region and superposing the topographic refraction compensation quantity, calculating a linear scale weight factor according to the gradient of the offset quantity for the nodes in the middle-offset region, and fusing the offset trend of the adjacent nodes to carry out weight smoothing;

And S2-8, triggering local reconstruction of the hybrid topology network when three or more nodes continuously exist in a high-offset area, namely generating an offset hot spot diagram according to the spatial distribution of the high-offset nodes, calculating the hover position and height of the unmanned aerial vehicle nodes to be adjusted, controlling the unmanned aerial vehicle nodes to move to a new position to reduce the whole offset after confirming a reconstruction scheme through a distributed consensus algorithm, and recalculating a correction weight matrix based on the reconstructed node distribution and updating the correction weight matrix to all associated nodes to obtain a final correction weight matrix.

In the embodiment of the present invention, when applied specifically, the above steps may be implemented in the following manner, for example:

The implementation process of the step S2-6 may be:

Threshold setting:

Based on the statistical distribution of the full-network node offset parameters (e.g., 50% of the quantiles as a first threshold T ₁ and 80% of the quantiles as a second threshold T ₂), the nodes are divided into:

The low offset region (delta is less than or equal to T ₁) is that the signal propagation path is close to a reference straight line, so that the influence of topography is small;

a middle offset region (T ₁<δ≤T₂) in which the linear influence of the topography refraction on the signal is taken into account;

The high offset region (delta > T ₂) is that the signal deviates seriously from the reference straight line, and the compensation needs to be enhanced.

The compensation mechanism is initialized:

The initial correction coefficients are allocated to different areas, namely a low offset area K ₁ =1.0 (basic value), a medium offset area K ₂ =1.2 (slight compensation), a high offset area K ₃ =1.5 (severe compensation), and an area-compensation mapping table is established and stored in each node.

The implementation process of the step S2-7 can be as follows:

low offset region processing:

The fixed correction factor K ₁ =1.0 is used, the topography refraction compensation quantity (which is adjusted according to the earth surface material correction factors of the step S2-5, such as metal area +0.3 and soil area-0.1) is superimposed, and the final factor k=k ₁ + topography factor.

Middle offset region processing:

Calculating an offset gradient (the difference between the current node offset and the adjacent node offset) to generate a linear scale weight factor (e.g., weight+0.2 when the gradient is > 0.1);

The offset trends (up/down) of the adjacent 3 nodes are fused, and the weight change is smoothed by a moving average (weight is additionally +0.1 if continuously up), and the final coefficient k=k ₂ × (1+gradient weight+trend weight).

High offset region processing:

Starting an exponential compensation function, wherein the larger the offset is, the exponentially increasing compensation intensity is;

And dynamically adjusting the compensation intensity by combining the current hovering height (such as compensation coefficient multiplied by 1.2 when the height is more than 20m and compensation coefficient multiplied by 0.8 when the height is less than 10 m) of the unmanned plane node, wherein the final coefficient K=exponential compensation multiplied by the height factor.

The implementation process of the step S2-8 can be as follows:

Hotspot graph generation and analysis:

When 3 or more nodes continuously appear in the high offset region are detected, three-dimensional coordinates of the nodes are extracted, and a space thermodynamic diagram (color depth represents offset) is generated;

calculating the geometric center of the hot spot area, and predicting the optimal propagation path of the signal by combining the topographic data (such as building height and porosity) of the disaster area.

Scheduling unmanned aerial vehicle nodes:

planning a new hovering position of the unmanned plane node according to the analysis result of the hotspot graph (preferably selecting a region with a height of 50-100m above the hotspot and a porosity of >40 percent);

the reconstruction scheme is validated by consensus among the relevant nodes via a distributed bayer fault-tolerance algorithm (e.g., PBFT).

Network reconstruction and coefficient update:

Controlling the unmanned aerial vehicle to move to a new position, and recalculating the projection distance and the offset of the whole network node to the reference straight line;

And (3) re-executing the steps S2-6 to S2-7 based on the reconstructed node distribution, generating a new correction weight matrix, and synchronizing to all nodes through multi-hop broadcasting.

According to the embodiment of the invention, the refraction compensation intensity of the signal propagation path is dynamically adjusted according to the node offset degree through threshold division and regional differentiation correction, the communication quality is ensured, meanwhile, the resource waste caused by excessive compensation is avoided, the network reconstruction mechanism triggered by the high offset region effectively reduces the signal offset of the local region through the active scheduling of the unmanned plane nodes, the network connectivity and the anti-interference capability under complex terrains are improved, and the linear interpolation and trend fusion strategy of the middle offset region avoids the mutation of the compensation coefficient, so that the signal path adjustment is smoother, and the network fluctuation caused by the tiny change of the environment is reduced. The network reconfiguration is triggered only in the high-offset area, limited unmanned aerial vehicle resources are concentrated to be used in the area needing optimization, balance of resource utilization efficiency and communication quality is achieved, a reconfiguration scheme is confirmed through a distributed consensus algorithm, and reliability and consistency of a decision process are guaranteed.

In a preferred embodiment of the present invention, performing convolution operation on the correction weight and the spectrum occupation matrix in the triplet feature vector, and performing phase compensation on the hash value of the user MAC address, to generate a corrected triplet feature vector may include:

s2-9, carrying out dynamic convolution processing on the frequency spectrum occupation matrix in the triplet feature vector based on the final correction weight matrix, wherein the correction weight matrix is used as a convolution kernel, and carrying out two-dimensional sliding window convolution operation along a time axis and a frequency axis of the frequency spectrum occupation matrix;

s2-10, calculating the phase compensation quantity of the user MAC address hash value based on the frequency band occupation probability in the corrected frequency spectrum matrix, namely extracting a time slot number corresponding to the frequency band with high occupation probability in the frequency spectrum matrix, performing modular operation on the time slot number and the hash value length to obtain a reference bit number of cyclic shift, performing cyclic left shift operation on a binary sequence of the user MAC address hash value, and dynamically adjusting the shift bit number by the product of the reference bit number and the correction weight;

And S2-11, carrying out data reconstruction on the corrected frequency spectrum matrix, the phase compensated hash value and the emergency priority code, carrying out normalized scaling on the numerical value of the emergency priority code according to the correction amplitude of the frequency spectrum matrix, and repackaging into corrected triplet feature vectors according to the sequence of a head field, namely the hash value, an intermediate control field, namely the scaled priority code, and a tail field, namely the frequency spectrum matrix.

In the embodiment of the present invention, when applied specifically, the above steps may be implemented in the following manner, for example:

the implementation process of the step S2-9 can be as follows:

convolution kernel configuration:

The final modified weight matrix is adjusted to a two-dimensional convolution kernel matching the spectrum occupancy matrix, ensuring that the convolution kernel size (e.g., 3 x 3 or 5 x 5) covers at least 3 consecutive frequency points and 3 time windows.

Sliding window convolution:

Sliding the convolution kernels along the time axis (column direction) and the frequency axis (row direction) of the spectrum matrix, covering one local area at a time;

and multiplying the signal intensity value in each local area with the correction weight of the corresponding position, and then summing to obtain a convolved value, and reflecting the comprehensive correction result of the area signal affected by the terrain.

Updating probability distribution:

normalizing the convolution result, and recalculating the occupation probability of each frequency band in different time slots (for example, the original probability is 0.8, and is adjusted to be 0.92 after convolution);

And generating a corrected frequency spectrum matrix, and highlighting the frequency band where the high-priority user is located (such as the increase of the occupied probability of the life detection user frequency band).

The implementation process of the step S2-10 can be as follows:

calculating the reference bit number:

Extracting time slot numbers (such as time slot 5 and time slot 12) corresponding to the frequency band with the occupation probability more than or equal to 0.7 from the corrected frequency spectrum matrix;

the slot numbers are accumulated and the hash value length (256 bits) is modulo to obtain the reference bit number of the cyclic shift (i.e. the sum is 345 and the modulo 256 is 89).

Dynamic phase adjustment:

Converting the user MAC address hash value into a binary sequence, and determining an actual shift bit number (e.g., weight 1.5×benchmark 89=133.5, rounding 134 bits) according to the product of the benchmark bit number and the correction weight of the corresponding node;

A cyclic shift left operation (e.g., shift to the tail before the original sequence 134) is performed on the binary sequence to generate a phase compensated hash value.

The implementation process of the step S2-11 may be:

Priority encoding scaling:

Normalized scaling of the emergency priority code is performed according to the modified amplitude (e.g., maximum probability boost value) of the spectrum matrix:

if the correction amplitude is >20%, the priority code value is multiplied by 1.2 (such as 010-011);

if the correction amplitude is less than 5%, the original value is kept unchanged.

And (3) data reorganization and encapsulation:

Repackaging the modified triplet feature vector in the following order:

Header field phase compensated hash value (256-bit binary sequence);

intermediate control field-scaled emergency priority coding (extended to 4 bits, e.g. 010→0011);

tail field, corrected spectrum matrix (two-dimensional array stored by row priority);

each field adds a length identifier (e.g., the first 32 bytes, the middle 1 byte, and the tail nxm bytes) to ensure that the receiving end can parse correctly.

In the embodiment of the invention, the topography refraction correction is directly blended into the frequency spectrum matrix through convolution operation, the occupation probability of each frequency band is dynamically adjusted, the penetrability and the reliability of signals in complex topography are improved, the influence of multipath fading and signal shielding is reduced, the dynamic phase compensation is implemented on the MAC address hash value, the hash value is changed in real time along with the service condition of the frequency spectrum, the replay attack and the falsification of identities are effectively resisted, and the safety of an emergency communication network is improved. The method comprises the steps of scaling priority codes according to frequency spectrum correction amplitude, realizing further resource inclination of high-priority users, ensuring smoothness of key communication in severe environments, keeping the basic structure of triplet feature vectors unchanged, realizing enhancement functions through dynamic adjustment of values in fields, reducing upgrading cost, combining terrain correction and priority scheduling, improving the anti-interference capability of a system, avoiding excessively occupying frequency spectrum resources, and realizing optimal balance of communication quality and resource consumption.

In a preferred embodiment of the present invention, step S3, the area master node receives the modified triplet feature vector, and maps the time slot allocation scheme, the frequency band switching strategy and the modification weight in an associated manner through space-time interleaving coding, so as to generate an enhanced three-dimensional virtual base station configuration including dynamic beam pointing parameters, which includes:

S3-1, mapping a weight value to a time domain through space-time interleaving coding based on the corrected weight in the corrected triplet feature vector, wherein a time slot allocation priority table is generated according to the size of the corrected weight, and user equipment corresponding to high weight is preferentially allocated with continuous time slots;

s3-2, performing spectrum hole detection on the corrected spectrum occupation matrix based on a time slot allocation weight table, namely identifying unoccupied spectrum hole areas in the frequency band range of the allocated time slots;

S3-3, generating dynamic beam pointing parameters through a beam forming algorithm based on a step frequency band switching path and a time slot allocation weight table, wherein the correction weights are subjected to weighted fusion with space position coordinates of nodes to calculate phase gradients of beam forming;

And S3-4, performing three-dimensional association mapping on a time slot allocation scheme, a frequency band switching strategy and dynamic beam pointing parameters, namely constructing a three-dimensional configuration matrix by taking the time slot allocation scheme as time dimension constraint, the frequency band switching path as frequency dimension constraint and the beam pointing parameters as space dimension constraint, and associating the three-dimensional configuration matrix with correction weights through space-time interleaving coding to generate the enhanced three-dimensional virtual base station configuration containing the dynamic beam pointing parameters.

In the embodiment of the present invention, when applied specifically, the above steps may be implemented in the following manner, for example:

The implementation process of the step S3-1 can be as follows:

priority table generation:

And arranging the correction weights in the corrected triple feature vectors in a descending order, constructing a time slot allocation priority table, and preferentially allocating continuous time slots for user equipment with higher weight (for example, 3 continuous time slots are preferentially obtained for users with weight of more than 1.5).

Dynamic update period adjustment:

dynamically adjusting the update frequency of the time slot allocation table according to the scaling value of the emergency priority code:

the slot allocation table for the high priority user (e.g., scaled to 0011) is updated every 100 ms;

the slot allocation table for the normal user (coded 1000) is updated every 500ms, balancing real-time with computational overhead.

The implementation process of the step S3-2 can be as follows:

and (3) identifying a cavity area:

and in the frequency band range of the allocated time slot, scanning the corrected frequency spectrum occupation matrix line by line, and identifying a continuous frequency point region with the signal strength of <0.3 (namely unoccupied), wherein the continuous frequency point region is marked as a frequency spectrum hole.

Dynamic path planning:

for the time slot requirement of high-priority users (such as life detection), preferentially searching spectrum holes near the current frequency band;

If the local frequency band has no available hole, a cross-frequency band switching path (such as switching from 2.4GHz to 5 GHz) is planned, so that the signal interruption time is ensured to be less than 10ms in the switching process.

The implementation process of the step S3-3 can be as follows:

phase gradient calculation:

carrying out weighted fusion on the correction weight and the three-dimensional space coordinates of the nodes, wherein the higher the weight is, the larger the influence of the nodes on the beam direction is;

Phase differences among the nodes are calculated, and phase gradients required by beam forming are generated.

Beam direction adjustment:

according to the time-frequency characteristic of the frequency band switching path, the direction of the main lobe of the wave beam is dynamically adjusted:

For the area where the high-priority user is located (such as a rescue command center), the main lobe gain is improved by 3dB;

A null region with the depth of-15 dB is generated in the direction of an interference source (such as the ruins of the strong reflection metal), and interference signals are restrained.

The implementation process of the step S3-4 can be as follows:

three-dimensional matrix construction:

constructing a three-dimensional configuration matrix by taking a time slot allocation scheme as a time dimension (X axis), a frequency band switching path as a frequency dimension (Y axis) and a beam pointing parameter as a space dimension (Z axis);

Each element of the matrix stores parameters such as beam gain, phase and the like of the corresponding space-time position.

Space-time interleaving coding:

Associating the three-dimensional configuration matrix with the correction weight, and performing space-time interleaving coding on matrix elements:

in the time domain, the configuration parameters of adjacent time slots are differentially encoded, so that transmission redundancy is reduced;

on the frequency domain, the frequency band switching path is convolutionally encoded, so that the anti-interference capability is enhanced;

An enhanced three-dimensional virtual base station configuration is generated that includes dynamic beam pointing parameters.

In the embodiment of the invention, the time slot allocation, the frequency band switching and the beam pointing three-dimensional resource scheduling depth are fused through space-time interleaving coding, so that the three-dimensional collaborative optimization of a full space domain, a full time domain and a full frequency domain is realized, the spectrum utilization rate is improved, and the time slot priority allocation, the frequency band special guarantee and the beam pointing enhancement are implemented for the high priority user based on the correction weight and the emergency priority coding, so that the reliability of key communication in a complex environment is ensured. The method effectively avoids obstacle reflection and multipath interference, improves signal transmission quality, reduces error rate through frequency spectrum cavity detection and dynamic null wave beam formation, updates configuration parameters in real time along with correction weights, supports quick response to disaster area environment changes, maintains dynamic stability of a communication network, breaks through space limitation of a traditional physical base station through three-dimensional virtual base station configuration, achieves omnibearing signal coverage of complex disaster area environments, and reduces communication blind areas.

In a preferred embodiment of the present invention, step S4 includes establishing a communication link based on the enhanced three-dimensional virtual base station configuration, adjusting a link direction in real time according to beam pointing parameters, dynamically optimizing signal strength and time delay of the communication link, and completing disaster area signal coverage, including:

S4-1, establishing an initial communication link based on dynamic beam pointing parameters in the enhanced three-dimensional virtual base station configuration, and detecting link quality in real time through received signal strength indication;

step S4-2, triggering an adaptive beam adjustment mechanism if the detected signal intensity is lower than a preset threshold value, namely determining a signal attenuation direction according to a signal intensity distribution diagram, gradually adjusting a beam pointing angle according to a spiral search mode, and re-detecting the signal intensity after each adjustment;

Step S4-3, dynamically optimizing the link delay based on the adjusted beam pointing parameter and completing signal coverage, namely predicting delay jitter trend through a Kalman filter, marking the current link as an unstable state if a predicted value exceeds a tolerance range, switching to a pre-configured redundant link according to a marking result, recalculating a transmission path based on an updated spectrum occupation matrix, carrying out matching verification on the optimized transmission path and a correction weight to generate a final communication link configuration, activating cooperative transmission of all network nodes through the communication link configuration, covering a disaster area target area according to the beam pointing parameter, and monitoring a signal blind area blind compensation state in real time until a preset coverage rate threshold is reached.

In the embodiment of the present invention, when applied specifically, the above steps may be implemented in the following manner, for example:

the implementation process of the step S4-1 can be as follows:

Link initialization:

And the regional master node sends a link establishment request to the target node according to dynamic beam pointing parameters (including phase gradient, main lobe direction and null position) in the enhanced three-dimensional virtual base station configuration, wherein the link establishment request carries initial phase and gain parameters required by beam forming.

After receiving, the target node adjusts the phase offset of its own antenna array according to the parameters to form an initial communication link (such as a high priority area near the alignment reference line of the main lobe of the directional antenna).

And (3) real-time quality monitoring:

the link quality is detected by a Received Signal Strength Indicator (RSSI) with a period of 50ms, and a real-time signal strength value (in dBm) is recorded.

And setting a signal strength threshold (such as-85 dBm), if the detection value is lower than the threshold for 3 times continuously, judging that the link quality is deteriorated, and triggering the beam adjustment mechanism of the step S4-2.

The implementation process of the step S4-2 can be as follows:

Attenuation direction positioning:

and collecting signal intensity distribution data of the nodes of the whole network, generating a three-dimensional signal intensity heat map, and locating the direction of signal attenuation exceeding 10dB (for example, the signal intensity of a certain area is 15dB lower than that of the periphery, and the attenuation direction is marked).

Spiral search adjustment:

And adjusting beam pointing by taking the current beam pointing angle as a center according to the range of azimuth angle +/-15 degrees and pitch angle +/-10 degrees, adopting a spiral increment step length (first 5 degrees and gradually increasing by 2 degrees), waiting for 20ms of stabilizing time after each adjustment, and detecting signal intensity.

If the signal strength is improved by more than or equal to 5dB after adjustment, the fine adjustment is continued along the direction, and if the adjustment is not improved for 5 times continuously, the reverse search is performed.

Pointing lock and parameter update:

when the signal strength reaches or exceeds a threshold (such as-80 dBm) and the continuous detection is stable for 5 times, the current beam pointing angle (azimuth angle theta, pitch angle) is locked ) And synchronizing the new parameters to the dynamic beam pointing parameter table of the regional master node.

The implementation process of the step S4-3 may be:

and (3) predicting delay jitter:

modeling the link delay by using a Kalman filter, inputting historical delay data (such as round trip time RTT within the past 100 ms), and predicting the delay jitter trend of 50ms in the future;

A delay tolerance range (e.g., mean ± 20 ms) is set, and if the predicted value exceeds the tolerance, the link is marked as "unstable".

Redundant link switching and path computation:

3 standby links (based on different projection directions of a reference straight line) are preconfigured, and a path with the node offset less than or equal to a first threshold value and the signal strength more than or equal to-75 dBm is preferentially selected as a redundant link;

And (3) adopting Dijkstra algorithm to recalculate the transmission path by combining the updated frequency spectrum occupation matrix (containing corrected frequency band occupation probability), and avoiding the high-interference frequency band (such as the frequency band with the occupation probability of more than 0.9).

Matching verification and cooperative transmission:

Matching the optimized transmission path with the node offset parameter and the correction weight, wherein the average correction weight of the nodes on the path is required to be less than or equal to 1.2 (an excessive compensation area is avoided);

Activating the cooperative transmission of the whole network nodes, adjusting the transmission direction by the main node according to the new beam pointing parameter, synchronously adjusting the hovering height (such as the down to 15m enhanced signals when the coverage blind area is reduced) by the unmanned plane nodes, and monitoring the signal blind area in real time (the area with the coverage rate of less than 90% is marked as the blind area);

and repeating the beam adjustment and the path optimization until the signal coverage rate of the disaster area is more than or equal to 95 percent (a preset threshold value).

According to the embodiment of the invention, signal attenuation caused by terrain shielding or obstacle movement is rapidly compensated through real-time RSSI monitoring and spiral search adjustment, link signal intensity is ensured to be stable in a reliable interval, communication success rate in a disaster area complex environment is improved, a Kalman filter predicts delay jitter in real time, an unstable link is identified in advance and is switched to a redundant path, data transmission delay and packet loss rate are reduced, continuity of real-time services such as rescue command and the like is ensured, beam pointing is dynamically adjusted by combining correction weight and node distribution, unmanned plane nodes actively complement blindness, signal coverage of disaster area edge areas and low-porosity shielding areas is ensured, and the problem of blind areas which are difficult to deploy in a traditional base station is solved. The method has the advantages of realizing the linkage optimization of beam pointing, transmission paths and correction weights, automatically adapting to dynamic environment changes such as aftershocks, obstacle displacement and the like, reducing manual intervention, improving the self-maintenance capability of an emergency communication network, preferentially guaranteeing high-priority user links (such as life detection equipment), reducing the energy consumption of the whole network through cooperative emission and dynamic power adjustment while meeting communication requirements, prolonging the endurance time of the equipment and adapting to the power shortage scene of a disaster area.

As shown in fig. 2, an embodiment of the present invention further provides a terrestrial satellite signal distributed co-processing system 20, including:

A generating module 21, configured to extract a user MAC address hash value, an emergency priority code, and a spectrum occupation matrix from the synchronous network, and generate a triplet feature vector;

The correction module 22 is used for transmitting the triplet feature vector through a double-layer Mesh network, selecting two reference target points with fixed geographic coordinates on the ground of a disaster area, acquiring three-dimensional coordinates of the two reference target points, constructing a space reference straight line, calculating the vertical projection distance from each node to the space reference straight line according to real-time position data of each node, generating a node offset parameter, dynamically generating a refraction correction coefficient of a signal propagation path based on a comparison result of the node offset parameter and a preset threshold value, and increasing the correction weight by adopting a piecewise linear interpolation method when the offset exceeds the threshold value;

the mapping module 23 is configured to receive the modified triplet feature vector by the regional master node, and map a time slot allocation scheme, a frequency band switching policy and a modification weight in an associated manner through space-time interleaving coding, so as to generate an enhanced three-dimensional virtual base station configuration including a dynamic beam pointing parameter;

And the optimizing module 24 is used for establishing a communication link based on the enhanced three-dimensional virtual base station configuration, adjusting the link direction in real time according to the beam pointing parameter, and dynamically optimizing the signal intensity and time delay of the communication link to complete the signal coverage of the disaster area.

While the foregoing is directed to the preferred embodiments of the present invention, it will be appreciated by those skilled in the art that various modifications and adaptations can be made without departing from the principles of the present invention, and such modifications and adaptations are intended to be comprehended within the scope of the present invention.

Claims (10)

1. A method for distributed collaborative processing of terrestrial satellite signals, comprising: Step S1: extracting the user MAC address hash value, emergency priority code and spectrum occupancy matrix from the synchronization network to generate a triplet feature vector; Step S2: Transmit the triplet feature vector through a two-layer Mesh network, select two reference target points with fixed geographic coordinates on the ground in the disaster area, obtain their three-dimensional coordinates, and construct a spatial reference line; calculate the vertical projection distance of each node to the spatial reference line based on the real-time position data of each node, and generate a node offset parameter; based on the comparison result of the node offset parameter with a preset threshold, dynamically generate the refraction correction coefficient of the signal propagation path; when the offset exceeds the threshold, use piecewise linear interpolation to increase the correction weight; perform a convolution operation on the correction weight and the spectrum occupancy matrix in the triplet feature vector, and perform phase compensation on the user MAC address hash value to generate a corrected triplet feature vector; Step S3: The regional master node receives the modified triplet feature vector, associates and maps the time slot allocation scheme, the frequency band switching strategy and the modified weights through time-space interleaving coding, and generates an enhanced three-dimensional virtual base station configuration including dynamic beam pointing parameters; Step S4: establishing a communication link based on the enhanced three-dimensional virtual base station configuration, adjusting the link direction in real time according to the beam pointing parameters, dynamically optimizing the signal strength and delay of the communication link, and completing signal coverage of the disaster area. 2. The method for distributed collaborative processing of terrestrial satellite signals according to claim 1, characterized in that before step S1, extracting the user MAC address hash value, emergency priority code, and spectrum occupancy matrix from the synchronization network to generate a triplet feature vector, the method further comprises: The porosity of the ruins is calculated based on 3D scanning data of the disaster area terrain, and the areas are divided into high-porosity and low-porosity areas. Differentiated deployment strategies are implemented based on the porosity differences to form a node network. Based on the node network, the obstacle displacement changes are monitored in real time. If it is detected that the obstacle displacement causes the local porosity change to exceed the preset threshold, the node will be re-hovered to the new pore position based on the pore distribution heat map fed back by the individual node to obtain the adjusted node distribution density of the hybrid topology network. Based on the adjusted node distribution density of the hybrid topology network, the dynamic voting mechanism in the distributed Byzantine fault-tolerant consensus algorithm is adopted to dynamically adjust the clock synchronization period according to the signal propagation delay between nodes, and broadcast the synchronization signal to nodes blocked by obstacles through multi-hop relay, forming a synchronous communication network that is resistant to terrain interference. 3. The distributed collaborative processing method for terrestrial satellite signals according to claim 2, wherein step S1 extracts a user MAC address hash value, an emergency priority code, and a spectrum occupancy matrix from the synchronization network to generate a triplet feature vector, comprising: Capture the broadcast signal of the user device through the synchronization network, extract its MAC address and use the SHA-256 hash algorithm to generate a unique hash identification code of fixed length; Based on the hash identification code, the user device's signal transmission power history data is associated with the preset disaster emergency level table to dynamically classify the user: if the signal transmission power is continuously above the threshold and the identification code belongs to the rescue equipment whitelist, the user is marked as life detection; if the signal delay is below the threshold and the identification code belongs to the command center database, the user is marked as rescue command; the rest are marked as ordinary communication, and each category is assigned a corresponding emergency priority code. Based on the emergency priority code, the frequency bands where life detection users are located are sampled, and the frequency bands where general communication users are located are sparsely sampled. The signal strengths of different frequency bands are integrated into a spectrum occupancy matrix in time series. The hash identification code, the emergency priority code and the spectrum occupancy matrix are encapsulated and arranged into a triple feature vector in a preset binary format. In the data structure of the triple feature vector, the hash identification code is used as the header field, the emergency priority code is used as the middle control field, and the spectrum occupancy matrix is used as the tail payload field. The three are seamlessly spliced together through a length identifier. 4. The distributed collaborative processing method for ground satellite signals according to claim 3, wherein step S2: transmitting the triplet feature vector through a two-layer mesh network, selecting two reference target points with fixed geographic coordinates on the ground in the disaster area, obtaining their three-dimensional coordinates, and constructing a spatial reference line, comprises: Based on the topological characteristics of the disaster area rescue map, the top of a building with a stable foundation at the entrance to the disaster area was selected as the first target point. The base of the directional antenna tower deployed by the rescue center was selected as the second target point. The geodetic coordinate system latitude, longitude and altitude data of the two target points were obtained. Convert the acquired geodetic coordinate system coordinates into a spatial rectangular coordinate system, specifically including: calculating the conversion matrix from the geodetic coordinates of the target point to the spatial rectangular coordinates based on the WGS-84 ellipsoid parameters; outputting the three-dimensional spatial coordinates of the two target points through the built-in coordinate conversion interface of the Beidou positioning module; Based on the three-dimensional spatial coordinates of two target points, a two-point straight line equation is used to construct a spatial reference line. 5. The distributed collaborative processing method for ground satellite signals according to claim 4, wherein the step of calculating the vertical projection distance of each node to the spatial reference line based on the real-time position data of each node to generate the node offset parameter comprises: Obtain the three-dimensional coordinates of the node's real-time position, project them onto the plane where the spatial reference line is located, and calculate the Euclidean distance between the node coordinates and the nearest point on the line as the vertical projection distance; According to the influence of the surface material in the disaster area on the signal refractive index, a terrain correction factor is added to the vertical projection distance to generate the normalized node offset parameter. 6. The distributed collaborative processing method for ground satellite signals according to claim 5, characterized in that a refraction correction coefficient for the signal propagation path is dynamically generated based on a comparison result between a node offset parameter and a preset threshold value. When the offset exceeds the threshold value, a piecewise linear interpolation method is used to increase the correction weight, comprising: Based on the statistical distribution of node offset parameters, a first threshold and a second threshold are set to divide the nodes into a low offset zone (i.e., offset ≤ first threshold), a medium offset zone (i.e., first threshold < offset ≤ second threshold), and a high offset zone (i.e., offset > second threshold), and initialize a dynamic compensation mechanism according to the zone type; Differentiated correction strategies are implemented for different regional types: for nodes in low-offset areas, a fixed correction coefficient is used and terrain refraction compensation is superimposed; for nodes in medium-offset areas, a linear proportional weight factor is calculated based on the offset gradient, and the offset trends of adjacent nodes are integrated for weight smoothing; for nodes in high-offset areas, an exponential compensation function is enabled and the compensation intensity is dynamically adjusted based on the hovering height of the drone node. When three or more consecutive nodes are detected in the high-offset area, the local reconstruction of the hybrid topology network is triggered: an offset hotspot map is generated according to the spatial distribution of the high-offset nodes, and the hovering position and height that need to be adjusted for the drone node are calculated; after the reconstruction plan is confirmed by the distributed consensus algorithm, the drone node is controlled to move to the new position to reduce the overall offset; the corrected weight matrix is recalculated based on the reconstructed node distribution and updated to all associated nodes to obtain the final corrected weight matrix. 7. The distributed collaborative processing method for terrestrial satellite signals according to claim 6, wherein the step of convolving the modified weights with the spectrum occupancy matrix in the triplet feature vector and performing phase compensation on the user MAC address hash value to generate a modified triplet feature vector comprises: Based on the final modified weight matrix, the spectrum occupancy matrix in the triplet feature vector is dynamically convolved. The modified weight matrix is used as the convolution kernel, and a two-dimensional sliding window convolution operation is performed along the time axis and frequency axis of the spectrum occupancy matrix. The occupancy probability distribution of each frequency band is updated according to the convolution result to generate a modified spectrum matrix. Based on the corrected frequency band occupancy probabilities in the spectrum matrix, the phase compensation for the user's MAC address hash value is calculated. The time slot numbers corresponding to the frequency bands with high occupancy probabilities in the spectrum matrix are extracted and modulo-calculated with the hash value length to obtain the reference number of bits for cyclic shift. A cyclic left shift is then performed on the binary sequence of the user's MAC address hash value, with the shift number dynamically adjusted by the product of the reference number of bits and the correction weight. The corrected spectrum matrix, phase-compensated hash value, and emergency priority code are reconstructed: the value of the emergency priority code is normalized and scaled according to the correction amplitude of the spectrum matrix; and the header field (i.e., hash value), the middle control field (i.e., scaled priority code), and the tail field (i.e., spectrum matrix) are repackaged into a corrected triplet feature vector in this order. 8. The method for distributed collaborative processing of terrestrial satellite signals according to claim 7, wherein step S3: the regional master node receives the modified triplet feature vector, associates and maps the time slot allocation scheme, the frequency band switching strategy, and the modified weights through space-time interleaving coding, and generates an enhanced three-dimensional virtual base station configuration including dynamic beam pointing parameters, comprising: Based on the modified weights in the modified triplet feature vector, the weight values are mapped to the time domain through space-time interleaving coding. A time slot allocation priority table is generated based on the size of the modified weights, in which user equipment corresponding to higher weights are preferentially allocated consecutive time slots. The update period of the time slot allocation weight table is dynamically adjusted in combination with the scaling value of the emergency priority code. Based on the time slot allocation weight table, spectrum hole detection is performed on the modified spectrum occupancy matrix. Unoccupied spectrum holes are identified within the frequency band of the allocated time slots. Frequency switching paths are dynamically planned based on the time slot requirements of high-priority users. Based on the stepped frequency switching path and the time slot allocation weight table, a beamforming algorithm generates dynamic beam pointing parameters. The modified weights are weighted and fused with the node's spatial position coordinates to calculate the beamforming phase gradient. Based on the time-frequency characteristics of the frequency switching path, the beam main lobe direction is adjusted to cover the area where high-priority users are located, and a null zone is generated in the direction of the interference source. The time slot allocation scheme, frequency band switching strategy and dynamic beam pointing parameters are mapped in a three-dimensional manner: a three-dimensional configuration matrix is constructed with the time slot allocation scheme as the time dimension constraint, the frequency band switching path as the frequency dimension constraint, and the beam pointing parameters as the space dimension constraint. The three-dimensional configuration matrix is associated with the correction weights through time-space interleaving coding to generate an enhanced three-dimensional virtual base station configuration containing dynamic beam pointing parameters. 9. The distributed collaborative processing method for terrestrial satellite signals according to claim 8, wherein step S4: establishing a communication link based on the enhanced three-dimensional virtual base station configuration, adjusting the link direction in real time according to beam pointing parameters, dynamically optimizing the signal strength and delay of the communication link, and completing signal coverage of the disaster area, comprises: Establishing an initial communication link based on dynamic beam pointing parameters in an enhanced 3D virtual base station configuration and monitoring link quality in real time using received signal strength indicators; If the detected signal strength falls below a preset threshold, the adaptive beam adjustment mechanism is triggered. The system determines the direction of signal attenuation based on the signal strength distribution graph and gradually adjusts the beam pointing angle in a spiral search pattern. After each adjustment, the signal strength is re-checked. When the signal strength reaches the threshold, the system locks the current beam pointing angle and updates the dynamic beam pointing parameters. Based on the adjusted beam pointing parameters, the link delay is dynamically optimized and signal coverage is achieved: the delay jitter trend is predicted through the Kalman filter. If the predicted value exceeds the tolerance range, the current link is marked as unstable; according to the marking result, the pre-configured redundant link is switched to, and the transmission path is recalculated based on the updated spectrum occupancy matrix; the optimized transmission path is matched and verified with the corrected weights to generate the final communication link configuration; the coordinated transmission of nodes in the entire network is activated through the communication link configuration, the target area in the disaster area is covered according to the beam pointing parameters, and the signal blind spot filling status is monitored in real time until the preset coverage threshold is reached. 10. A distributed collaborative processing system for ground satellite signals, the system implementing the method according to any one of claims 1 to 9, characterized in that it comprises: A generation module is used to extract the user MAC address hash value, emergency priority code and spectrum occupancy matrix from the synchronization network to generate a triplet feature vector; A correction module is used to transmit the triplet feature vector through a double-layer mesh network, select two reference target points with fixed geographic coordinates on the ground in the disaster area, obtain their three-dimensional coordinates, and construct a spatial reference line; calculate the vertical projection distance of each node to the spatial reference line based on the real-time position data of each node, and generate a node offset parameter; based on the comparison result of the node offset parameter with a preset threshold, dynamically generate a refraction correction coefficient for the signal propagation path; when the offset exceeds the threshold, use piecewise linear interpolation to increase the correction weight; perform a convolution operation on the correction weight and the spectrum occupancy matrix in the triplet feature vector, and perform phase compensation on the user MAC address hash value to generate a corrected triplet feature vector; A mapping module is configured to receive the modified triplet feature vector at the regional master node, associate and map the time slot allocation scheme, the frequency band switching strategy, and the modified weights through time-space interleaving coding, and generate an enhanced three-dimensional virtual base station configuration including dynamic beam pointing parameters; The optimization module is used to establish a communication link based on the enhanced three-dimensional virtual base station configuration, and adjust the link direction in real time according to the beam pointing parameters, dynamically optimize the signal strength and delay of the communication link, and complete signal coverage in the disaster area.