CN-122001448-A - Ground-moon communication link self-adaptive switching method based on intelligent risk perception

CN122001448ACN 122001448 ACN122001448 ACN 122001448ACN-122001448-A

Abstract

The invention belongs to the technical field of earth-moon deep space communication and intelligent communication control, and particularly relates to an earth-moon communication link self-adaptive switching method based on intelligent risk perception, which comprises the steps of establishing an earth-moon communication system model based on discrete time slices; the method comprises the steps of calculating an end-to-end signal-to-noise ratio based on an end-to-end total propagation loss model, constructing a communication mode availability indication function, constructing judging characteristic features of a link structure, environment risks and time stability, processing by a machine learning method based on risk perception time sequence modeling and risk weighting judgment based on the judging characteristic features to obtain a communication mode judgment result, and effectively reducing the probability of error switching caused by short-time fluctuation based on comprehensive judgment of the long-term stability and risk trend of the link.

Inventors

XU CHUAN
ZHOU HANG
ZHAO GUOFENG
HAN ZHENZHEN
CHEN HEJI
ZHU XINGYU

Assignees

重庆邮电大学

Dates

Publication Date: 20260508
Application Date: 20260210

Claims (6)

1. The earth-moon communication link self-adaptive switching method based on intelligent risk perception is characterized by comprising the following steps of: s1, establishing a ground-moon communication system model based on discrete time slices, which comprises the following steps: S11, constructing a total communication link from a ground station to a lunar base station through a synchronous orbit satellite and a ground-month relay satellite, and further dividing the total communication link into 3 segmented communication links from the ground station to the synchronous orbit satellite, the synchronous orbit satellite to the ground-month relay satellite and the ground-month relay satellite to the lunar base station; s12, dividing the communication process according to time slices, and establishing a discrete time sequence ={1,2,...,t,...}; S13, calculating the time slice t epsilon of each segmented communication link End-to-end total propagation distance, and constructing an end-to-end total propagation loss model; s2, calculating an end-to-end signal-to-noise ratio based on an end-to-end total propagation loss model, and constructing a communication mode availability indication function; S3, constructing judgment characteristic features of a link structure, environmental risks and time stability, wherein: The judging characteristic features in the aspect of the link structure comprise a length ratio, a dominant index, a structure dispersion index and a visibility mark; the judging characteristic features in the aspect of environmental risk comprise an environmental disturbance risk accumulation intensity feature, a link structure amplification risk feature, an evolution trend risk feature and a directional stability risk feature; The judging characteristic features in the aspect of time stability comprise transient communication performance features, dynamic features and time window statistical stability features; S4, processing by a machine learning method based on time sequence modeling and risk weighting judgment based on risk perception based on judgment characteristic features to obtain a communication mode judgment result, wherein the method comprises the following steps: S41, constructing an enhanced link state feature vector x (t) of the current time slice t based on the judgment characteristic features, wherein the enhanced link state feature vector x (t) is expressed as: x(t)=[ρ(t), x inst (t), R env (t), x dyn (t), x stat (t)], Wherein ρ (t) = [ ρ GS (t), ρ SR (t), ρ RL (t) ] represents the total communication link structural feature; S42, selecting the first T continuous time slices of the current time slice T, and constructing an input sequence X seq (T) = [ X (T-T+1), X (T-T+2) and X (T) ] according to the enhanced link state feature vector; S43, constructing a corresponding link physical risk vector z (t) aiming at each time slice of the input sequence, wherein the link physical risk vector z (t) is expressed as: z(t)=[ρ GS (t), ρ SR (t), ρ RL (t),R env (t),R point (t),V(t)], Wherein ρ GS (t)、ρ SR (t)、ρ RL (t) represents the length ratio of the ground station G to the segmented communication link of the geosynchronous orbit satellite S, the segmented communication link of the geosynchronous orbit satellite S to the earth-month relay satellite R, and the segmented communication link of the earth-month relay satellite R to the moon base station L at the time slice t, R env (t) represents the cumulative intensity characteristic of the environmental disturbance risk, R point (t) represents the directional stability risk characteristic, and V (t) represents the visibility mark of the total communication link at the time slice t; S44, inputting the input sequence and the corresponding link physical risk vector sequence into a risk modulation gating long-term memory network to obtain a time sequence evolution feature vector h LSTM (t); S45, performing risk weighted voting based on a random forest based on the time sequence evolution feature vector h LSTM (t) to obtain a communication mode judgment result.
2. The method for adaptive switching of a terrestrial-moon communication link based on intelligent risk awareness according to claim 1, wherein step S13 comprises: S131, under a unified reference coordinate system, calculating position vectors of a ground station G, a synchronous orbit satellite S, a ground-moon relay satellite R and a moon base station L in a time slice t; S132, calculating the geometric distance of each segmented communication link in a time slice t according to the position vector, thereby obtaining an end-to-end total propagation distance d total (t), which is expressed as: , , , , Wherein d GS (t) represents the geometric distance between the ground station G and the geostationary satellite S in the time slot t, d SR (t) represents the geometric distance between the geostationary satellite S and the terrestrial month relay satellite R in the time slot t, d SR (t) represents the geometric distance between the geostationary satellite S and the terrestrial month relay satellite R in the time slot t, and I; S133, calculating free space path loss of the communication mode mE { RF, FSO } in the time slice t according to the total end-to-end propagation distance d total (t) Expressed as: , Wherein lambda m represents the operating wavelength of communication mode m, RF represents the radio frequency communication mode, FSO represents the laser communication mode; s134, calculating an end-to-end total propagation loss model, wherein the model is expressed as: , in the formula, Representing the environmental impact of the ground segment, Representing the trade-off caused by the deep space plasma disturbance, Representing the environmental impact of the near month segment.
3. The method for adaptive switching of a terrestrial-moon communication link based on intelligent risk awareness according to claim 1, wherein step S2 comprises: S21, when the communication mode m is adopted in the calculation time slice t, the received power at the lunar base station is calculated Expressed as: , in the formula, Representing the transmit power at the transmitting end when communication mode m is employed, 、 Equivalent gains of a transmitting end and a receiving end are respectively represented; S22, calculating an end-to-end signal-to-noise ratio when the communication mode m is adopted in the time slice t, wherein the signal-to-noise ratio is expressed as: , Where k represents a Boltzmann constant, Representing equivalent system noise temperature converted to lunar base station, B m representing the working bandwidth when adopting communication mode m; S23, defining a communication mode availability indication function, wherein the communication mode availability indication function is expressed as: , In the formula, 1{ · } represents an indication function, gamma min represents a minimum signal-to-noise threshold, when a m (t) =1, it represents that the communication mode m satisfies the basic communication condition within the time slice t and has physical feasibility, and when a m (t) =0, it represents that the communication mode m does not satisfy the basic communication condition within the time slice t.
4. The method for adaptive switching of a terrestrial-moon communication link based on intelligent risk awareness according to claim 1, wherein step S3 comprises: S31, constructing judgment characteristic features in the aspect of a link structure, wherein the judgment characteristic features comprise: S311, calculating the length ratio of each segmented communication link in the total end-to-end propagation distance, and taking the maximum length ratio as a dominant index of the total communication link; S312, calculating a structural dispersion index of the total communication link based on the duty ratio, wherein the structural dispersion index is expressed as: , Wherein D ρ (t) represents the structural dispersion index of the total communication link at time slice t, ρ l (t) represents the length ratio of the segmented communication links L ε { GS, SR, RL } at time slice t, GS represents the segmented communication link from the ground station G to the geostationary orbit satellite S, SR represents the segmented communication link from the geostationary orbit satellite S to the terrestrial month relay satellite R, RL represents the segmented communication link from the terrestrial month relay satellite R to the lunar base station L; S313, constructing a visibility mark V (t) of the total communication link in a time slice t according to the elevation angle and shielding relation of the total communication link, when all 3 segmented communication links meet the condition that the sight is not shielded and the elevation angle is higher than a preset threshold value, V (t) =1, otherwise V (t) =0, S32, constructing characteristics in the aspect of environmental risk, wherein the method comprises the following steps: S321, constructing an environment disturbance risk accumulation intensity characteristic R env (t), which is expressed as follows: , Wherein T w represents the risk accumulation sliding window length, lambda 1 、λ 2 、λ 3 、λ 4 represents the weight coefficient, E turb (u) represents the atmospheric turbulence risk, E weather (u) represents the weather attenuation risk, E space (u) represents the solar background noise risk, and E dust (u) represents the moon dust light depth risk; S322, constructing a link structure amplification risk feature R str (t), which is expressed as: , Where ω 1 、ω 2 represents the structural risk weight, ρ max represents the dominant indicator of the total communication link, and D ρ (t) represents the structural dispersion indicator of the total communication link; S323, constructing an evolution trend risk feature R trend (t), which is expressed as: , , in the formula, Represents a weight coefficient, max () represents taking the maximum value, Representing the change rate of the received signal-to-noise ratio in the communication mode m, wherein Deltat represents the interval of adjacent decision time slices, and gamma m (t) represents the received signal-to-noise ratio in the time slices t in the communication mode m; s324, constructing a directional stability risk feature R point (t), which is expressed as follows: , Wherein mu 1 、μ 2 represents a weight coefficient, theta err (t) represents an instantaneous pointing error angle, and theta bean represents a half-power beam width of the laser terminal; s33, constructing characteristics in time stability, including: S331, constructing an instantaneous communication performance characteristic x inst (t), which is expressed as: , , Wherein, gamma RF (t)、γ FSO (t) respectively represents the received signal-to-noise ratio of the radio frequency communication mode and the laser communication mode at the time slice t, and LM RF (t)、LM FSO (t) respectively represents the link margin of the radio frequency communication mode and the laser communication mode at the time slice t; 、 respectively represents the equivalent loss brought by the ground section atmospheric attenuation in the radio frequency communication mode and the laser communication mode, Represents equivalent loss brought by the deep space plasma disturbance in a radio frequency communication mode, Θ err (t) represents the pointing error of the communication platform, and V (t) represents the visibility mark of the total communication link at time slice t; s332, constructing dynamic characteristics x dyn (t), which are expressed as: , in the formula, 、 The change rates of the received signal-to-noise ratios of the radio frequency communication mode and the laser communication mode in the time slice t are respectively shown, 、 Respectively represents the link margin change rates of the radio frequency communication mode and the laser communication mode in a time slice t, Indicating the pointing error change rate of the communication platform; S333, constructing a time window statistical stability characteristic x stat (t), which is expressed as follows: , in the formula, 、 Respectively represent the average value of the received signal-to-noise ratios of the radio frequency communication mode and the laser communication mode in the risk accumulation sliding window length T w , 、 The standard deviation of the received signal-to-noise ratio of the radio frequency communication mode and the laser communication mode in the risk accumulation sliding window length T w is respectively shown, 、 Representing the minimum of the received signal-to-noise ratios of the radio frequency communication mode and the laser communication mode in the risk accumulation sliding window length T w , The sliding standard deviation of pointing error is indicated.
5. The method for adaptively switching a ground-month communication link based on intelligent risk awareness according to claim 1, wherein the processing procedure of the risk modulation gating long-term memory network comprises the following steps: ; ; ; ; ; ; Wherein f t represents the output value of the forgetting gate, i t represents the output value of the input gate, o t represents the output value of the output gate, sigma (·) represents a Sigmoid function, h t-1 represents the hidden layer state of the time slice t-1, W f 、W i 、W o represents the weight matrix corresponding to the enhanced link state feature vector x (t), b f 、b i 、b o represents the bias vector, U f 、U i 、U o represents the recursive weight matrix corresponding to the hidden layer state h t-1 , M f 、M i 、M o represents the risk weight matrix of the link physical risk vector z (t) modulating each gating mechanism, alpha (t) represents the risk driving memory attenuation coefficient obtained by mapping the link physical risk vector z (t) through the projection weight vector a and the bias term c, and c t-1 represents the unit state of the time slice t-1; representing the candidate cell state for time slice t-1, tan h(s) represents the hyperbolic tangent activation function, Representing the hadamard product of the matrix.
6. The method for adaptive switching of a terrestrial-moon communication link based on intelligent risk awareness according to claim 1, wherein step S45 comprises: S451, constructing random forest input u (t) = [ x (t), h LSTM (t) ] based on a time sequence evolution feature vector h LSTM (t); S452, constructing a comprehensive risk index R (t), which is expressed as: , Wherein k 1 、k 2 、k 3 、k 4 represents a weight coefficient, R env (t) represents an environment disturbance risk accumulated intensity characteristic, R str (t) represents a multi-section link structure amplified risk characteristic, R trend (t) represents an evolution trend risk characteristic, and R point (t) represents a directional stability risk characteristic; S453, calculating a dynamic voting weight w i of the ith decision tree according to the comprehensive risk index, wherein the dynamic voting weight is expressed as: , where η i represents the risk sensitivity coefficient of the ith decision tree; S454, calculating the supporting probability p FSO (t) of the laser communication mode, wherein the supporting probability is expressed as: , in the formula, 1 {.cndot } represents an indication function, and y i (t) represents a communication mode judgment result output by an ith decision tree in a random forest at a time slice t; s455, calculating a communication mode judgment result according to the support probability p FSO (t) and the availability indication function A FSO (t) of the laser communication mode Expressed as: , Wherein τ represents a preset decision threshold, when =1, Indicating that the target communication method is the laser communication method, when =0, Indicating that the target communication scheme is a radio frequency communication scheme.

Description

Ground-moon communication link self-adaptive switching method based on intelligent risk perception Technical Field The invention belongs to the technical field of earth-moon deep space communication and intelligent communication control, and particularly relates to an earth-moon communication link self-adaptive switching method based on intelligent risk perception. Background As lunar exploration tasks evolve from short-term exploration to long-term residence and continuous operation, the terrestrial-lunar communication system is undergoing a multi-hop relay architecture evolution from a traditional single-hop direct-connect mode to "earth-synchronous orbit relay-terrestrial-lunar relay-lunar surface". Although the complex topological structure effectively expands the communication window and coverage area, the heterogeneous physical environments such as the ground, deep space, the near month and the like are spanned, so that each section presents obvious differences in geometric configuration, propagation characteristics and environmental disturbance, and the link state presents strong time variability. At present, radio frequency and laser communication have the advantages of high stability, strong environmental adaptability and limited bandwidth, and ultra-high bandwidth and ultra-high speed, but are extremely sensitive to pointing accuracy, link visibility and environmental conditions. In a multi-hop link, performance fluctuation of any key node is amplified in cascade, and end-to-end transmission quality is affected. Therefore, how to reasonably switch the radio frequency and laser communication modes becomes a key for improving the performance of the terrestrial communication system. However, the existing switching strategy depends on an empirical threshold or a single instantaneous index (such as signal-to-noise ratio and link margin), so that the coupling influence of the multi-hop link is not fully considered, and the robustness to the dynamic change of the environment is also lacking, which is easy to cause unnecessary frequent switching and damage the communication continuity. Although some studies have introduced machine learning, often focus only on transient state analysis, ignoring the cumulative effects of the timing features of link evolution and environmental risks, and lacking risk constraints based on physical mechanisms. Therefore, it is needed to propose a new adaptive switching method of communication modes, which can combine the time sequence information of link structure features, environmental risks and link states, and implement intelligent judgment and switching of communication modes through an intelligent risk-aware machine learning model, so as to improve the stability and reliability of the lunar communication system. Disclosure of Invention In order to solve the problems, the invention provides a ground month communication link self-adaptive switching method based on intelligent risk perception, which comprises the following steps: s1, establishing a ground-moon communication system model based on discrete time slices, which comprises the following steps: S11, constructing a total communication link from a ground station to a lunar base station through a synchronous orbit satellite and a ground-month relay satellite, and further dividing the total communication link into 3 segmented communication links from the ground station to the synchronous orbit satellite, the synchronous orbit satellite to the ground-month relay satellite and the ground-month relay satellite to the lunar base station; s12, dividing the communication process according to time slices, and establishing a discrete time sequence ={1,2,...,t,...}; S13, calculating the time slice t epsilon of each segmented communication linkEnd-to-end total propagation distance, and constructing an end-to-end total propagation loss model; s2, calculating an end-to-end signal-to-noise ratio based on an end-to-end total propagation loss model, and constructing a communication mode availability indication function; S3, constructing judgment characteristic features of a link structure, environmental risks and time stability, wherein: The judging characteristic features in the aspect of the link structure comprise a length ratio, a dominant index, a structure dispersion index and a visibility mark; the judging characteristic features in the aspect of environmental risk comprise an environmental disturbance risk accumulation intensity feature, a link structure amplification risk feature, an evolution trend risk feature and a directional stability risk feature; The judging characteristic features in the aspect of time stability comprise transient communication performance features, dynamic features and time window statistical stability features; S4, processing by a machine learning method based on time sequence modeling and risk weighting judgment based on risk perception based on judgment characteristic features to obtain a commun