Large-scale low earth orbit satellite dynamic routing method of self-adaptive network queuing time delay

Abstract

The invention relates to the technical field of satellite communication, in particular to a large-scale low earth orbit satellite dynamic routing method of self-adaptive network queuing delay, which is based on dynamic routing of a greedy thought heuristic large-scale low earth orbit satellite network. A satellite communication distance and direction, a predicted minimum hop count from a current satellite to a receiving place, an average queuing time delay in a network and a queuing time delay parameter of a next-hop relay satellite are used, a value function used for searching the next-hop satellite is provided, and the function value reflects the distance of advancing to the receiving place in single-hop transmission. And the queuing time delay of the relay satellites. By drawing a changing curve of the weight corresponding to the shortest transmission path time delay and the satellite average queuing time delay, when the average queuing time delay changes, the corresponding optimal weight is found, so that the self-adaption to the network load condition is achieved. Through verification, compared with other routing modes which allow non-adjacent satellites to use ISL communication, the method has the advantage of short total time delay for networks with different on-satellite queue load conditions.

Description

Technical Field

[0001] The present invention relates to the field of satellite communication technology, and in particular to a large-scale low-orbit satellite dynamic routing method with adaptive network queuing delay. Background Art

[0002] In a large-scale low Earth Orbit (LEO) satellite network, as the number of satellites launched increases, the distance between adjacent satellites is much smaller than the maximum communication distance of LEO. Using only adjacent inter-satellite links (ISL) for routing will increase the number of forwarding and queuing times, resulting in a higher total link latency, affecting the user experience of enjoying low-latency services.

[0003] Existing low-orbit satellite routing methods can be divided into static routing and dynamic routing according to whether queuing delay is considered. However, due to the large number of low-orbit satellite networks and the high-speed movement characteristics of the ground, the network topology is changeable. When the load of a satellite is too large, the static routing cannot be adjusted in time. The dynamic routing method converts the satellite's delay into a dynamic cost, and sets the minimum hop path and flooding area in the process of route discovery to reduce unnecessary waste of resources. When the orbital plane and the number of launched satellites increase, the routing method that only allows adjacent orbital intersatellite links will increase the number of link relays, resulting in longer delays. However, even by reducing the minimum hop count and shortening the link distance, the delay of the constellation at a certain altitude is reduced. Since the queuing delay factor of the satellite's task storage and forwarding in actual situations is not considered, when the satellite needs to forward a large amount of data at the same time, the forwarding task will generate a certain queuing delay. Summary of the invention

[0004] The purpose of the present invention is to provide a large-scale low-orbit satellite dynamic routing method with adaptive network queuing delay, which maximizes the single-hop transmission distance to reduce the number of relays while satisfying the line-of-sight communication constraints, and minimizes the phase difference between the link and the ideal path to reduce the total link delay.

[0005] To achieve the above object, the present invention provides a large-scale low-orbit satellite dynamic routing method with adaptive network queuing delay, comprising the following steps:

[0006] Step 1: Select a constellation network model that matches the satellite position coordinates and link delay;

[0007] Step 2: Determine the coordinates of the sending location and the receiving location;

[0008] Step 3: Get the communicable satellites, use the current coordinates as the center of the circle and the maximum communication distance as the radius to detect the communicable range;

[0009] Step 4: Discard the relay satellite whose distance from the received coordinates is greater than the current coordinates;

[0010] Step 5: Calculate the minimum number of hops K between the current coordinates and the receiving location min , calculate the communication value function for the remaining communicable satellites;

[0011] Step 6: Select the relay satellite with the greatest communication value as the next-hop relay satellite;

[0012] Step 7: If the next-hop satellite cannot communicate directly with the receiving station, the coordinates of the next-hop relay satellite are changed to the sending location, and steps 2 to 6 are repeated.

[0013] Optionally, the satellite position coordinates in step 1 satisfy the following quantitative relationship:

[0014]

[0015] Among them, C (p,n,t) is the position parameter matrix of the low-orbit satellite at any time t. The x-axis is the direction from the center of the earth O to the vernal equinox, the z-axis is perpendicular to the equatorial plane and points to the North Pole; the y-axis is perpendicular to the plane formed by the x-axis and the z-axis and forms a left-hand rectangular coordinate system with the x-axis and the z-axis. The default satellite S 0,0 The initial position is on the positive semi-axis of x;

[0016] The satellite is located at point H, the origin O coincides with the center of the earth, the satellite's altitude relative to the center of the earth is h, and h = |OH|; point H' is the projection of point H on the xOy plane, and HH' is perpendicular to the plane xOy; angle θ z is the angle between line segment OH and plane xOy; angle θ x is the angle between the line segment OH′ and the x-axis; ω e is the angular velocity of the satellite relative to the Earth’s rotation, and U(t) is a step function.

[0017] Optionally, in step 1, S p1,n1 Send data to S p2,n2 The total delay of a single-hop link consists of three parts: propagation delay D pro , transmission delay D t , and the queuing delay D queue , the expression is as follows:

[0018] Delay(S p1,n1 ,S p2,n2 )=D pro (S p1,n1 ,S p2,n2 )+D t +D queue

[0019] Among them, p1, n1, p2, and n2 represent the orbital plane numbers of satellites S1 and S2 and the satellite numbers within the orbital plane respectively.

[0020] Optionally, in step 3, when the satellite communicates at the maximum communication distance, the following quantitative relationship exists:

[0021]

[0022] θ max Indicates that when the satellite communication distance is min(d max ,d' max ), relative to a circle with radius h and center O, the maximum angular span, d max is the maximum communication range of the satellite itself, d′ max It is the maximum satellite communication range when not blocked by the ground.

[0023] Optionally, the cost function for the next hop of the dynamically loaded satellite network in step 5 is defined as follows:

[0024] f(S pm,nm ,S p(m+1),n(m+1) ,t)=α·f vector +β·f queue

[0025] f vector =cosΔp·(d(S pm,nm ,Des,t)-d(S p(m+1),n(m+1) ,Des,t)) / c

[0026] f queue =(E[D queue ]-D queue )

[0027] Among them, f vector It reflects the distance that the mission data travels in the direction of the receiving location during this hop propagation. Δp represents the distance that the low-orbit satellite S p(m+1),n(m+1) The phase difference with the ideal relay satellite position of this hop, f queue Reflects the relationship between the current satellite queuing delay and the expected value, E[D queue ] is the expectation of queuing delay, reflecting the network load status, α is the distance factor, β is the queuing factor, and α∈[0,1], β∈[0,1], and α+β=1.

[0028] Optionally, when calculating the specific values ​​of the optimal α and β, the cases of α∈[0,1] are traversed to find the corresponding α value under the minimum delay.

[0029] The present invention provides a large-scale low-orbit satellite dynamic routing method with adaptive network queuing delay. First, a large-scale circular polar orbit low-orbit constellation model in continuous time is constructed, and the dynamic routing of a large-scale low-orbit satellite network is heuristic based on the greedy idea. The satellite communication distance and direction, the expected minimum number of hops from the current satellite to the receiving location, the average queuing delay in the network, and the queuing delay parameters of the next-hop relay satellite are used to propose a value function for finding the next-hop satellite. The function value reflects the distance to the receiving location in a single-hop transmission, and the size of the queuing delay of the relay satellite. By drawing the weight corresponding to the shortest transmission path delay and the change curve of the satellite average queuing delay, the corresponding optimal weight is found when the average queuing delay changes to achieve an adaptive network load condition. Experimental data show that compared with other routing methods that allow non-adjacent satellites to use ISL communication, the present invention has the advantage of shorter total delay for networks with different on-board queue load conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0031] Figure 1 The present invention is a schematic flow chart of the steps of a large-scale low-orbit satellite dynamic routing method with adaptive network queuing delay.

[0032] Figure 2 It is a schematic diagram of the environmental architecture of the large-scale circular polar orbit low-orbit constellation model of the present invention.

[0033] Figure 3 It is a schematic diagram of the satellite orbit rectangular coordinate system in a specific embodiment of the present invention.

[0034] Figure 4 It is a schematic diagram of the cross-section of a satellite orbit in a specific embodiment of the present invention.

[0035] Figure 5 It is a schematic diagram of the maximum transmission angle in a specific embodiment of the present invention.

[0036] Figure 6 It is a schematic diagram comparing the number of iterations in a specific embodiment of the present invention with a comparison algorithm.

[0037] Figure 7 It is a schematic diagram of the relationship between delay and number of hops changing with distance factor in a specific embodiment of the present invention.

[0038] Figure 8It is a schematic diagram of the numerical relationship of the optimal distance factor for an average queuing delay of less than 40 ms in a specific embodiment of the present invention.

[0039] Fig. 9 It is a schematic diagram of propagation delay comparison in a specific embodiment of the present invention.

[0040] Fig.10 It is a schematic diagram of queuing delay comparison in a specific embodiment of the present invention.

[0041] Fig.11 It is a schematic diagram of the total delay with an average queuing delay of less than 40 ms in a specific embodiment of the present invention.

[0042] Fig.12 It is a schematic diagram of the variation of total delay with the number of satellite launches in a specific embodiment of the present invention. DETAILED DESCRIPTION

[0043] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.

[0044] See also Figure 1 The present invention provides a large-scale low-orbit satellite dynamic routing method with adaptive network queuing delay, comprising the following steps:

[0045] S1: Select a constellation network model that matches the satellite position coordinates and link delay;

[0046] S2: Determine the coordinates of the sending location and the receiving location;

[0047] S3: Acquire communicable satellites, and detect the communicable range with the current coordinates as the center and the maximum communication distance as the radius;

[0048] S4: Discard the relay satellite whose distance from the receiving coordinates is greater than the current coordinates;

[0049] S5: Calculate the minimum number of hops K between the current coordinates and the receiving location min , calculate the communication value function for the remaining communicable satellites;

[0050] S6: Select the relay satellite with the greatest communication value as the next-hop relay satellite;

[0051] S7: If the next-hop satellite cannot communicate directly with the receiving station, the coordinates of the next-hop relay satellite are changed to the sending location, and steps S2 to S6 are repeated.

[0052] The following is a further explanation of the implementation process, model construction and problem modeling:

[0053] See also Figure 2 In the low-orbit satellite network of the polar-orbit constellation, satellites are allowed to communicate with each other using cross-orbit links. The low-orbit satellite with the forwarding task calculates the value function for all low-orbit satellites within its range with its own maximum communication range, selects the satellite with the largest value function as the next-hop low-orbit satellite, and then the selected relay satellite repeats the above steps until the satellite can communicate directly with the ground receiving station. The routing method of the present invention makes full use of the maximum communication range of the satellite, allows cross-orbit satellite communication, and takes into account the queuing delay of the satellite. Finally, the relay satellite whose propagation path distance is much greater than the ideal path distance is excluded to achieve the purpose of reducing the delay.

[0054] Using a circular polar orbit constellation (N,P,h (p,n) ) as the constellation configuration, N is the number of satellites on a single orbital plane; P is the number of orbital planes; p is a positive integer orbital plane number, and p∈[1,P]; n represents the nth low-orbit satellite in a single orbital plane, and n∈[1,N]; h (p,n) is the operating altitude of the nth satellite in the pth orbital plane. The nth low-orbit satellite in the pth orbital plane is denoted as S p,n .

[0055] 1.1 Satellite Position Coordinate Model

[0056] like Figure 3 As shown, the satellite orbital plane is perpendicular to the equatorial plane and is evenly distributed in the earth's hemisphere from the p=0 orbital plane along the direction of the earth's rotation; the n=0 satellites in the single orbital plane are evenly distributed on the entire orbital circle on the equatorial plane toward the south.

[0057] Assume that the origin O coincides with the center of the earth, the x-axis is the intersection of the equatorial plane and the p=0 orbital plane, and the direction points to the satellite S 0,0 Initial position; the z-axis is perpendicular to the equatorial plane and points to the North Pole; the y-axis is perpendicular to the plane formed by the x-axis and the z-axis and forms a left-hand rectangular coordinate system with the x-axis and the z-axis.

[0058] Suppose the satellite is located at point H, the satellite's altitude above the Earth's center is h, and h = |OH|; point H' is the projection of point H on the xOy plane, and HH' is perpendicular to the xOy plane; angle θ z is the angle between line segment OH and plane xOy; angle θ x is the angle between line segment OH′ and the x-axis.

[0059] Assume that the period of the low-orbit satellite is T, and h is a constant. When determining θ x With θ zBy understanding the relationship between the angle of the satellite and time, the three-dimensional coordinates of the satellite at any time can be determined. Suppose the position parameter matrix of the low-orbit satellite at any time t is C (p,n) (t) = [x (p,n) (t),y (p,n) (t),z (p,n) (t)] According to the figure, the coordinates of the satellite and θ x With θ z The angles have the following quantitative relationship:

[0060]

[0061] Any orbital plane of a polar orbiting constellation is always perpendicular to the equatorial plane xOy, through Figure 4 It can be seen that when a satellite on the same orbital plane moves within a hemisphere, θ x remains unchanged and only changes when crossing the South Pole and the North Pole; z changes at a constant speed within the period. So we can get θ x With θ z The expression at t∈[0,T] is:

[0062]

[0063] where ω e is the angular velocity of the satellite relative to the Earth’s rotation, and U(t) is a step function.

[0064] 1.2 Link Delay Model

[0065] Assume that the maximum communication distance of low-orbit satellite is d max , the radius of the earth is r, and the satellite storage and forwarding space is η. Define the satellite phase Φ p,n ∈[0,π], is the low-orbit satellite S p,n The angle formed by the line connecting the center of the earth O and the positive half axis of the z axis; define the phase difference ΔΦ(S (p1,n1) ,S (p2,n2) )=Φ (p1,n1) -Φ (p2,n2) , where p1, n1, p2, n2 represent the orbital plane numbers of satellites S1 and S2 and the satellite numbers within the orbital plane respectively

[0066] Defined by S p1,n1 Send data to S p2,n2 The total delay of a single-hop link is:

[0067]

[0068] The total delay of a single-hop link consists of three parts: propagation delay D pro , transmission delay D t , and the queuing delay D queue.

[0069] For S p1,n1 To S p2,n2 The propagation delay D pro (S p1,n1 ,S p2,n2 ) has the following definition: D pro (S p1,n1 ,S p2,n2 )=d(S p1,n1 ,S p2,n2 ,t) / c, where d(S p1,n1 ,S p2,n2 ,t) indicates that at time t, the low-orbit satellite S p1,n1 To S p2,n2 distance, c is the speed of light in vacuum, and:

[0070]

[0071] Definition D t is the data transmission delay, which is the time it takes for the satellite to send the mission data from the beginning to the end. t =pac / R, where pac is the size of the task data and R is the transmission rate, which has the following definition:

[0072]

[0073] ρ t is the transmitted power of the signal, ρ n is the noise power, A t is the transmitting antenna gain, L(S p1,n1 ,S p2,n2 ,t) is the low-orbit satellite S p1,n1 To S p2,n2 The free space path loss in a vacuum at time t is:

[0074]

[0075] Define the queuing delay D queue is the queuing time before the mission data is sent by the current satellite. Before sending the mission data, the satellite needs to forward the data received from other satellites. Therefore, when the mission data is forwarded by each relay satellite, it needs to wait in line for the current relay satellite to process other tasks. In the waiting queue, the mission data that arrives first is sent first. The satellite data acceptance and storage and forwarding conform to the Poisson process and exponential distribution respectively.

[0076] The arrival of mission data in the satellite network satisfies the following conditions: the arrival of missions is independent of each other and conforms to the conditions of Poisson distribution with parameter λ; there is only one service satellite and the processing time of data is independent of each other, the satellite processing speed is constant and the processing time of missions obeys the negative exponential distribution with parameter u.

[0077] Therefore, the task queue on the satellite can be described by the M / M / 1 / K queue model, and the queue time D queue Calculated by the M / M / 1 / K queue model. Let the waiting time of the task data in the satellite storage and forwarding queue be the random variable D queue It conforms to the exponential distribution, so the average waiting time D for data on the satellite is queue There are the following expressions:

[0078]

[0079] 1.3 Problem Modeling

[0080] Let any K-hop transmission path be: path(S p0,n0 ,S pK,nK ,ξ1,...,ξ k ,...,ξ K |k=1,2,...,K), where S p0,n0 Indicates the launch of a satellite, S pK,nK To finally receive the satellite, k (k=1,2,......K) represents the kth path. The task data transmission process is first performed by S p0,n0 Send, through the first path ξ1 to S p1,n1 , and then transmitted to S via the second path ξ2 p2,n2 , and so on, in the last section of the entire link, the data is transmitted by S p(K-1),n(K-1) Through the Kth path ξ K Transmitted to the final receiving satellite S pK,nK , transmission path path(S p0,n0 ,S pK,nK ,ξ1,...,ξ k ,...,ξ K |k=1,2,......K) There are K paths in total.

[0081] According to the definition of single link delay, the total delay of a K-hop transmission path is the sum of the delays of each path segment, so the total delay of a K-hop transmission path is T(path(S p0,n0 ,S pK,nK ,ξ1,...,ξ k ,...,ξ K |k=1,2,......K) can be expressed as:

[0082]

[0083] Therefore, the problem of minimizing the total delay of a transmission path can be expressed as follows:

[0084]

[0085] The following constraints should be met when solving this problem:

[0086] 1) Maximum satellite communication distance limit

[0087] Assume that at time t, the path path (S p0,n0 ,S pK,nK ,ξ1,...,ξ k ,...,ξ K The distance of a single-hop low-orbit satellite communication link in |k=1,2,......K) is: d(S pm,nm ,S p(m+1),n(m+1) ,t)m∈N * ,m<K-1

[0088] At this time, it is necessary to consider whether the signals transmitted by satellites with maximum communication range will be blocked by the ground when low-orbit satellites at a certain altitude communicate with each other.

[0089] Assume that the maximum satellite communication range when not blocked by the ground is d′ max , at this time, the line-of-sight path of the signal is exactly tangent to the surface of the earth, the length of the line segment OH is h, and there exists Obviously d′ max It is related to the actual operating altitude h of the low-orbit satellite, so it is necessary to consider not only the maximum communication range d of the satellite itself max , it is also necessary to consider the maximum communication range d′ of the satellite without surface obstruction max , so for a single link distance d(S pm,nm ,S p(m+1),n(m+1) ,t)There are the following constraints:

[0090] d(S pm,nm ,S p(m+1),n(m+1) ,t)<min(d max ,d′ max ) (10)

[0091] 2) Maximum satellite waiting queue length limit

[0092] Assume that the maximum storage and forwarding space of the low-orbit satellite is η. When the size of the data to be forwarded exceeds the maximum storage and forwarding space, the low-orbit satellite will no longer accept new forwarding tasks until the remaining forwarding space can allow the forwarding of the current task. This needs to satisfy: D queue ·R+pac<η, so for the waiting time Dqueue The following limitations apply:

[0093]

[0094] In P1, the decision variable is ξ k ,ξ k It can reflect the path(S p0,n0 ,S pK,nK ,ξ k ) in the relay forwarding low-orbit satellite. The selection of these nodes needs to consider the propagation delay between adjacent nodes, the queuing waiting time, and the link ξ k After completion, the distance the data has traveled in the direction of the final receiving location is jointly judged.

[0095] Therefore, the problem of finding the link with the shortest delay is:

[0096]

[0097] When a low-orbit satellite transmits at the maximum communication distance, it may select the next-hop relay satellite with a high queuing delay; while the relay satellite with a small queuing delay may be far away from the ideal path. To address this problem, the present invention combines the greedy idea and takes into account that the on-board queuing delay and the satellite position are independent of each other, and the satellite position variable in continuous time is continuous, while the queuing delay is a discrete variable with an integer value. Therefore, the corresponding problem is a mixed integer nonlinear programming problem.

[0098] Based on this, the subsequent steps of the present invention consider a low-latency low-orbit satellite routing method that maximizes the communication distance of each hop of the relay satellite under different satellite queue delays.

[0099] Steps S2 to S6 also involve the determination of the theoretical minimum number of hops, the evaluation function and the optimal distance factor.

[0100] 1) Theoretical minimum number of hops

[0101] There is only one least delay path (LDP) in the low-orbit satellite network (if the constellation is completely symmetrical and the network load is not considered, the number of least delay links is two), but there are more than one least hop path (LHP). We try to prove that the least delay link is a subset of the least hop link, that is: It has been proven that the lowest latency link is not a subset of the minimum hop link. However, the counterexample has the characteristic of severely uneven satellite coverage. Considering the overall coverage effect of a constellation, the counterexample scenario is unlikely to occur in reality. So it can still be assumed This theory is valid in reality. Figure 5As shown, when the satellite communicates at the maximum communication distance, the following quantitative relationship exists:

[0102]

[0103] θ max Indicates that when the satellite communication distance is min(d max ,d' max ), relative to a circle with radius h and center O, the maximum circumferential angle span. Let the circumferential angle formed by the sending gateway, the receiving gateway and the center O on the earth be θ centre , then assuming that there is a relay satellite at any ideal location, the theoretical minimum number of hops is However, there are gaps in the actual constellation, so it is impossible for each hop distance to reach the maximum communication distance, which leads to the actual number of hops being greater than the theoretical minimum number of hops K. min .

[0104] 2) Evaluation function

[0105] In the low-orbit satellite S pm,nm When searching for the next relay satellite, the evaluation function should include the single-hop communication distance d(S pm,nm ,S p(m+1),n(m+1) ,t), the queuing time of the next hop satellite D queue It also reflects the consistency between the direction of the single-hop receiving satellite and the direction of the final ground receiving station. Therefore, the value function of the next hop of the dynamic load satellite network is defined as follows:

[0106] f(S pm,nm ,S p(m+1),n(m+1) ,t)=α·f vector +β·f queue (14)

[0107] For formula (14), f vector With f queue There are the following definitions:

[0108]

[0109] f vector It can reflect the distance that the mission data travels in the direction of the receiving location during this hop propagation. Δp represents the distance that the low-orbit satellite S p(m+1),n(m+1) The calculation of the phase difference Δp with the ideal relay satellite position of this hop is related to the theoretical minimum number of hops:

[0110]

[0111] f queue =(E[D queue ]-D queue), which reflects the relationship between the current satellite queuing delay and the expected value. queue ] is the expectation of the queuing delay, which can reflect the network load state, α is the distance factor, β is the queuing factor, and α∈[0,1], β∈[0,1], and α+β=1. The unit of the value function is milliseconds, which can provide a reference for the current satellite to select the next-hop relay satellite. According to the numerical result of the value function, the next-hop relay satellite with the largest value value is selected within the communication range. vector With f queue The maximum value of is in the order of milliseconds, so the value function will not be greatly affected by one of the two being too large. The distance factor α and the queuing factor β can be adjusted according to the network load. Different values ​​of α and β will also affect the selection of relay satellites, thereby affecting the final delay result of the entire link. The function value reflects the distance to the receiving location in a single-hop transmission and the size of the queuing delay of the relay satellite. At the same time, it can be seen from formula (15) and formula (16) that the larger the distance to the receiving location in a single-hop transmission, the greater the delay of the relay satellite. vector The larger the value is; the smaller the current relay satellite forwarding queuing experiment is than the overall satellite queuing delay in the network, f queue The larger the value of , the satellite with the maximum value of the value function is selected as the next-hop relay satellite.

[0112] α and β represent the emphasis of the satellite when searching for the next hop satellite. When there is a queuing delay in the satellite, the total delay of the transmission process can be reduced by reducing the total number of hops in the transmission process or selecting a relay satellite with a smaller queuing delay. When the single-hop distance increases, the number of hops in the transmission process can be reduced, thereby reducing the number of queuing times and the total link delay. When a relay satellite with a smaller queuing delay is directly selected, the position of the satellite may not be around the maximum transmission distance from the satellite to the destination. Therefore, the values ​​of α and β represent two strategies for reducing the queuing delay. When the value of α becomes larger, it is more inclined to reduce the number of queuing times and reduce the delay. On the contrary, it is more inclined to reduce the delay by directly selecting a relay satellite with a lower queuing delay.

[0113] 3) Determination of the optimal distance factor

[0114] By changing the values ​​of α and β, under the overall network state of a certain average queuing delay, when α and β are the optimal factors, the total delay of a K-hop path should be the minimum value. However, since the amount of tasks received by the satellite in the future is uncertain, the average queuing delay of the network varies, and the queuing delay of each satellite will also change at different times, resulting in the inability to directly calculate the total delay. When calculating the specific values ​​of the optimal α and β, the case of α∈[0,1] will be traversed to find the corresponding α value under the minimum delay. After the optimal distance factor α in the current network load is determined, considering the applicability of other queuing delay conditions, it is also necessary to change the average queuing delay E[D queue ] is taken, and the optimal distance factor under other common network queuing delay conditions is calculated in this way, and the function change curve of the average queuing delay and the optimal distance factor α is obtained, thereby meeting the demand for finding the minimum delay path under different network queuing delays.

[0115] The method of the present invention is converted into an algorithm and compared with other algorithms for complexity analysis:

[0116] The algorithm steps are as follows:

[0117] 1. Input: transmitting position Sta, receiving position Des, maximum transmission rate R, transmission task size pac, optimal distance factor α

[0118] 2. Search for all communicable satellites of the current node

[0119] 3. Calculate the minimum hop value from the current node to the receiving location

[0120] 4. Select satellite nodes that are closer to the receiving location than the current satellite node and calculate the value function of these nodes

[0121] 5. Select the satellite node with the largest value function as the relay satellite

[0122] 6. The selected relay satellite node is replaced with the current node. If the current node cannot communicate directly with the receiving location, jump to step 27 and the routing algorithm ends.

[0123] There are two loops in the algorithm, the first is the While loop, and the second is the For loop in the While loop. When each single-hop data transmission is completed, the While loop completes a cycle. When the value of the optimal distance factor α is given, the number of hops can converge to the theoretical minimum number of hops K min , so the maximum number of loops of the While loop body is the minimum number of jumps K min .

[0124] The maximum number of loops of the For loop body should be the number of satellites within the maximum communication range of the satellite. When the satellite is in the polar regions, the satellite nodes are the densest, so the number of satellites within the communication range is the largest at this time. When it is at the pole, the number of satellites that can communicate within the satellite communication range is L, where So the maximum number of loops for the For loop body should be L.

[0125] Therefore, the maximum number of cycles for the algorithm of the present invention is K min L, in summary, the complexity of the algorithm can be obtained as O(K min L).

[0126] The low-latency routing algorithm based on random geometry stipulates that the number of hops of the entire transmission link must be consistent with the minimum theoretical hop number. On this basis, the transmission distance is divided into equal intervals, and communication relay satellites are searched within a given range around each division point. The time complexity is O(K min ). Therefore, the time complexity of the algorithm of the present invention is greater than the time complexity of the comparison algorithm, but a lower total link delay can be obtained.

[0127] Furthermore, in order to demonstrate the benefits of the present invention, the present invention also proposes specific embodiments and conducts comparative analysis of different methods through simulation results.

[0128] 1) Simulation scene setting

[0129] With polar orbit constellation (N,P,h (p,n) ), where N = 50, P = 72, a total of 3600 low-orbit satellites, all low-orbit satellites operate at an altitude of about 550km, that is, for p∈[1,72], n∈[1,50], there is h (p,n) ≈550. Assume that the mission data size is 50MB and the transmission rate is 10GB / s. The average queuing waiting time is 25ms. The minimum number of hops between the launch location and the receiving location is 7 as calculated by formula (12). The comparison experiment object is the existing low-orbit satellite routing algorithm that allows cross-orbit communication. The unit of all delays is milliseconds.

[0130] 2) Explanation of simulation results

[0131] Figure 6 The time complexity comparison between the method of the present invention and the low-latency routing algorithm based on random geometry is shown. When the optimal distance factor is given, the number of link hops obtained by the method of the present invention can still reach the theoretical minimum number of hops. However, since the method of the present invention adds the step of finding the node with the maximum value function in each hop, the time complexity is greater than the time complexity of the low-latency routing algorithm based on random geometry.

[0132] Figure 7It shows that the change of distance factor can affect the total number of hops and the final delay of the link. From formula (14), we can see that when the distance factor gradually increases, f vector The weight of the next-hop relay satellite increases, which also leads to the selection of the next-hop relay satellite paying more attention to the single-hop propagation distance. Therefore, the number of relays of the entire transmission link decreases, and the average number of hops slowly converges to the minimum number of hops. It can be seen that the minimum delay in the network load environment with an average queuing delay of 25ms occurs when the number of hops just converges to the minimum number of hops. When the distance factor continues to increase after the total number of link hops converges, it will make f queue The weight of the relay satellite is reduced, and the queuing delay on the node is no longer concerned during the relay satellite selection process. However, the number of hops has reached the theoretical minimum number of hops, so the actual total delay will increase.

[0133] according to Figure 8 The figure shows the trend of the optimal distance factor changing with the average queuing delay when the average queuing delay in the network is below 40ms. It can be seen that when the network load level increases, which is reflected in the increase of the average queuing delay, the optimal distance factor will increase with the increase of the average queuing delay, but the slope will gradually decrease. This is because when the distance factor increases, the ratio with the queuing delay factor does not change linearly. The derivative of α / β increases with the increase of α, so when α increases by the same value, a larger α value will have a greater impact on the value of α / β.

[0134] In practical applications, the optimal distance factor value can be selected according to the fitting curve of the optimal distance factor and the average queuing delay and the current network status to achieve the effect of adaptive average queuing delay.

[0135] Fig. 9 and Fig.10 It is proved that compared with the cross-track link routing algorithm that does not consider the queuing delay, the method of the present invention significantly reduces the queuing delay in the link at the expense of a very small propagation delay, and ultimately reduces the total link delay. Moreover, as the average queuing delay in the network increases, the effect of reducing the queuing delay becomes more obvious, while the propagation delay is relatively stable and has no upward trend.

[0136] Fig.11 The total delay of the transmission path obtained by the method of the present invention is compared with the total delay of the cross-track link routing algorithm without considering the queuing delay. Figure 8 and Fig. 9 The results are consistent. The main way for the method of the present invention to reduce latency is to reduce the queuing delay in the link. The additional queuing delay is about 1-2ms, and the impact on the total latency is basically negligible. Therefore, the total latency comparison chart basically inherits the trend of the queuing delay comparison chart. The latency advantage of the method of the present invention increases with the increase of the average queuing delay.

[0137] When the number of satellite launches increases, the satellites become more dense, so there will be more satellites in ideal locations to choose from, and satellites with lower queuing delays can be selected. Since the queuing delay on the satellite is random, blindly selecting the node with the largest single-hop transmission distance as the relay satellite, the average queuing delay of the selected node should be consistent with the expected queuing delay of all nodes in the current network E(D queue ) is consistent. Fig.12 It can be seen that as the number of satellites increases, the shortest delay obtained by the method of the present invention gradually decreases, so it can better adapt to the increase in the number of satellite nodes and embodies better robustness.

[0138] What is disclosed above is only a preferred embodiment of the present invention, and it certainly cannot be used to limit the scope of rights of the present invention. Ordinary technicians in this field can understand that all or part of the processes of the above embodiment and equivalent changes made according to the claims of the present invention still fall within the scope of the invention.

Claims

1. A large-scale low-orbit satellite dynamic routing method with adaptive network queuing delay, characterized in that: The following steps are involved: Step 1: Select a constellation network model that matches the satellite position coordinates and link delay; Step 2: Determine the coordinates of the sending location and the receiving location; Step 3: Get the communicable satellites, use the current coordinates as the center of the circle and the maximum communication distance as the radius to detect the communicable range; Step 4: Discard the relay satellite whose distance from the received coordinates is greater than the current coordinates; Step 5: Calculate the minimum number of hops K between the current coordinates and the receiving location min , calculate the communication value function for the remaining communicable satellites; Step 6: Select the relay satellite with the greatest communication value as the next-hop relay satellite; Step 7: If the next-hop satellite cannot communicate directly with the receiving station, the coordinates of the next-hop relay satellite are changed to the sending location, and steps 2 to 6 are repeated.

2. The large-scale low-orbit satellite dynamic routing method with adaptive network queuing delay as claimed in claim 1, characterized in that: The satellite position coordinates in step 1 satisfy the following quantitative relationship: Among them, C (p,n,t) is the position parameter matrix of the low-orbit satellite at any time t. The x-axis is the direction from the center of the earth O to the vernal equinox, the z-axis is perpendicular to the equatorial plane and points to the North Pole; the y-axis is perpendicular to the plane formed by the x-axis and the z-axis and forms a left-hand rectangular coordinate system with the x-axis and the z-axis. The default satellite S 0,0 The initial position is on the positive semi-axis of x; The satellite is located at point H, the origin O coincides with the center of the earth, the satellite's altitude relative to the center of the earth is h, and h = |OH|; point H' is the projection of point H on the xOy plane, and HH' is perpendicular to the plane xOy; angle θ z is the angle between line segment OH and plane xOy; angle θ x is the angle between the line segment OH′ and the x-axis; ω e is the angular velocity of the satellite relative to the Earth’s rotation, and U(t) is a step function.

3. The large-scale low-orbit satellite dynamic routing method with adaptive network queuing delay as claimed in claim 2 is characterized in that: In step 1, S p1,n1 Send data to S p2,n2 The total delay of a single-hop link consists of three parts: propagation delay D pro , transmission delay D t , and the queuing delay D queue , the expression is as follows: Delay(S p1,n1 ,S p2,n2 )=D pro (S p1,n1 ,S p2,n2 )+D t +D queue Among them, p1, n1, p2, and n2 represent the orbital plane numbers of satellites S1 and S2 and the satellite numbers within the orbital plane respectively.

4. The large-scale low-orbit satellite dynamic routing method with adaptive network queuing delay as claimed in claim 3 is characterized in that: In step 3, when the satellite communicates at the maximum communication distance, the following quantitative relationship exists: θ max Indicates that when the satellite communication distance is min(d max ,d' max ), relative to a circle with radius h and center O, the maximum angular span, d max is the maximum communication range of the satellite itself, d′ max It is the maximum satellite communication range when not blocked by the ground.

5. The large-scale low-orbit satellite dynamic routing method with adaptive network queuing delay as claimed in claim 4, characterized in that: The cost function for the next hop of the dynamically loaded satellite network in step 5 is defined as follows: f(S pm,nm ,S p(m+1),n(m+1) ,t)=α·f vector +β·f queue f vector =cosΔp·(d(S pm,nm ,Des,t)-d(S p(m+1),n(m+1) ,Des,t)) / c f queue =(E[D queue ]-D queue ) Among them, f vector It reflects the distance that the mission data travels in the direction of the receiving location during this hop propagation. Δp represents the distance that the low-orbit satellite S p(m+1),n(m+1) The phase difference with the ideal relay satellite position of this hop, f queue Reflects the relationship between the current satellite queuing delay and the expected value, E[D queue ] is the expectation of queuing delay, reflecting the network load status, α is the distance factor, β is the queuing factor, and α∈[0,1], β∈[0,1], and α+β=1.

6. The large-scale low-orbit satellite dynamic routing method with adaptive network queuing delay as claimed in claim 5, characterized in that: When calculating the specific values ​​of the optimal α and β, the cases of α∈[0,1] are traversed to find the corresponding α value under the minimum delay.

Images