CN116886176A - Predictable inter-satellite routing method based on link utility function - Google Patents
Predictable inter-satellite routing method based on link utility function Download PDFInfo
- Publication number
- CN116886176A CN116886176A CN202311069788.7A CN202311069788A CN116886176A CN 116886176 A CN116886176 A CN 116886176A CN 202311069788 A CN202311069788 A CN 202311069788A CN 116886176 A CN116886176 A CN 116886176A
- Authority
- CN
- China
- Prior art keywords
- link
- satellite
- inter
- utility function
- node
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 39
- 230000006870 function Effects 0.000 claims abstract description 70
- 238000010586 diagram Methods 0.000 claims abstract description 11
- 238000013135 deep learning Methods 0.000 claims abstract description 9
- 230000008859 change Effects 0.000 claims abstract description 4
- 230000005540 biological transmission Effects 0.000 claims description 14
- 238000012549 training Methods 0.000 claims description 10
- 238000012545 processing Methods 0.000 claims description 9
- 230000003044 adaptive effect Effects 0.000 claims description 7
- 239000011159 matrix material Substances 0.000 claims description 7
- 230000008569 process Effects 0.000 claims description 7
- 238000004364 calculation method Methods 0.000 claims description 3
- 238000013507 mapping Methods 0.000 claims description 2
- 230000009471 action Effects 0.000 claims 8
- 239000000654 additive Substances 0.000 claims 2
- 230000000996 additive effect Effects 0.000 claims 2
- 238000005315 distribution function Methods 0.000 claims 2
- 241000854291 Dianthus carthusianorum Species 0.000 claims 1
- 239000002355 dual-layer Substances 0.000 claims 1
- 230000007613 environmental effect Effects 0.000 claims 1
- 230000003993 interaction Effects 0.000 claims 1
- 238000012804 iterative process Methods 0.000 claims 1
- 230000007774 longterm Effects 0.000 claims 1
- 230000007704 transition Effects 0.000 claims 1
- 238000004891 communication Methods 0.000 abstract description 6
- 230000033001 locomotion Effects 0.000 abstract description 5
- 238000011160 research Methods 0.000 description 6
- 230000004927 fusion Effects 0.000 description 5
- 238000013461 design Methods 0.000 description 4
- 238000013136 deep learning model Methods 0.000 description 3
- 238000011156 evaluation Methods 0.000 description 3
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 230000008447 perception Effects 0.000 description 2
- 230000002787 reinforcement Effects 0.000 description 2
- 235000008694 Humulus lupulus Nutrition 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 238000013139 quantization Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B7/00—Radio transmission systems, i.e. using radiation field
- H04B7/14—Relay systems
- H04B7/15—Active relay systems
- H04B7/185—Space-based or airborne stations; Stations for satellite systems
- H04B7/18521—Systems of inter linked satellites, i.e. inter satellite service
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L45/00—Routing or path finding of packets in data switching networks
- H04L45/02—Topology update or discovery
- H04L45/08—Learning-based routing, e.g. using neural networks or artificial intelligence
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L45/00—Routing or path finding of packets in data switching networks
- H04L45/12—Shortest path evaluation
- H04L45/121—Shortest path evaluation by minimising delays
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L45/00—Routing or path finding of packets in data switching networks
- H04L45/12—Shortest path evaluation
- H04L45/124—Shortest path evaluation using a combination of metrics
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L45/00—Routing or path finding of packets in data switching networks
- H04L45/44—Distributed routing
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- Astronomy & Astrophysics (AREA)
- Aviation & Aerospace Engineering (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Radio Relay Systems (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention discloses a predictable inter-satellite routing method based on a link utility function, and belongs to the technical field of wireless communication. The method provides a predictable inter-satellite routing method aiming at the problem of frequent network topology change and intermittent connection of inter-satellite links caused by relative motion among satellite nodes in a low-orbit satellite network. Giving out an inter-satellite link comprehensive utility function according to the signal-to-noise ratio, the duration and the overall time delay, and obtaining an inter-satellite link weighted space-time diagram based on the utility function by utilizing a channel perception-oriented self-adaptive algorithm; and obtaining a predictable optimal inter-satellite route by using a distributed deep learning algorithm according to the comprehensive utility value of the inter-satellite link, so that the packet loss rate, the congestion value and the end-to-end delay of the inter-satellite link are reduced on the basis of the route predictability.
Description
Technical Field
The invention belongs to the technical field of wireless communication, and particularly relates to a predictable inter-satellite routing method based on a link utility function.
Background
Satellite-to-earth fusion network (STIN) is an emerging framework, and uses a Satellite network and a ground network as backbones, and is jointly formed by a deep space network and a near space network, so as to be used for low-delay long-distance high-throughput data transmission. The satellite-ground fusion network utilizes the advantages of the satellite network to provide bandwidth access for the whole world and realize the global three-dimensional seamless coverage network service. However, as satellites fly continuously, there is a difference in mobility between different nodes in the network, and the higher the satellite orbit, the slower the movement speed. Thus, connections established between pairs of nodes in space are sporadic, resulting in constantly changing inter-satellite links and topologies of the satellite network. Extensive research has been conducted in order to find optimal routes based on link functions in inter-star dynamic topologies, as well as adaptive predictable routes. The routing design of the low-orbit satellite network can be based on the multi-attribute decision and the self-adaptive predictive aspects of the inter-satellite links.
The multiple attributes of the inter-satellite links include signal-to-noise ratio, link duration, and overall latency. The utility values of these parameters are different even though the inter-satellite links have the same state. To more accurately characterize the properties of inter-satellite links, these parameters need to be considered in combination in determining link utility to support reliable routing path selection. The multi-attribute decision is specially designed for solving the problem of selecting an optimal scheme considering a plurality of attributes, then a multi-attribute utility function of the link is constructed, the specific influence of three influencing factors is comprehensively quantized, and corresponding weights are set according to actual conditions, so that quantization indexes of the inter-satellite link are more reliable and specific.
Adaptive predictability refers to interpretable route prediction. The current giant constellation networks of low orbit satellites are becoming increasingly important, and for those systems with inter-satellite links, the large number of satellites increases the routing complexity and the number of hops required for inter-satellite link paths. In addition, the large number of satellites in a giant constellation network not only increases system throughput, but also presents challenges for inter-satellite link use and routing issues. Because of the high satellite density in the network, more link relays are required to connect the paths of two fixed terrestrial users, thus requiring additional processing costs and increasing routing complexity. Thus, improving network performance using artificial intelligence techniques is a promising approach. Through extensive data training, the AI model may enable high-precision fast path allocation and generation of adaptive predictive routing strategies. To avoid complex spatial routing computations, adding AI models for efficient inter-satellite terrestrial integrated network flow control is becoming a future direction.
In order to solve the one-sided problem of the joint intelligent route optimization, the invention provides a predictive route method based on an inter-satellite link utility function to dequantize the indexes. Determining satellite coverage global by establishing a Walker constellation of a low orbit satellite, and extracting all satellite nodes and links in the constellation; dividing the multiple attributes of the inter-satellite links into inter-satellite link signal-to-noise ratios, duration of the inter-satellite links and overall delay of the inter-satellite links; combining adaptive predictive routing with deep reinforcement learning; and sequentially acquiring and analyzing parameters of each index. Finally, a distributed deep reinforcement learning algorithm based on fuzzy learning is provided to obtain an optimal route. Therefore, not only can the evaluation indexes of all aspects of the inter-satellite links be clearly and quantitatively obtained, but also the unique optimal predictable inter-satellite routing result based on the link utility function can be generated in a key way through a weight setting method.
Disclosure of Invention
The present invention is directed to solving the above problems of the prior art. A predictable inter-star routing method based on a link utility function is provided. The technical scheme of the invention is as follows:
a predictable inter-satellite routing method based on a link utility function firstly needs to establish a low orbit Walker constellation, determines that satellites can cover the world, establishes three equatorial high orbit satellites and corresponding equatorial ground computing centers, and then acquires corresponding data through STK software, comprising the following steps:
step 101, constructing a link incidence matrix IM of a low-orbit satellite, and mapping out a link relation between satellites; then collecting the data of the low-orbit inter-satellite links, and obtaining a signal-to-noise ratio utility function U (s (t)), a duration utility function U (l (t)), and an overall delay utility function U (b (t)) according to the data; finally integrating the three utility functions to obtain a multi-attribute utility function of the link
Step 102, performing multiple iterative computations through a channel-aware-oriented adaptive algorithm to obtain a multi-attribute utility functionWeight of each influence factor in the database; then, according to the utility value of each link, a weighted time-space diagram is obtained;
step 103, introducing a deep learning algorithm and adding fuzzy logic according to the result, and finally obtaining the predictable inter-satellite route based on the link utility function.
The invention has the advantages and beneficial effects as follows:
the invention provides a predictable inter-satellite routing method based on a link utility function, aiming at the problems that in a low-orbit constellation network, relative motion among low-orbit satellite nodes causes frequent network topology change and intermittent connection of inter-satellite links. The main innovation of the invention is to provide a multi-attribute utility function of the inter-satellite link and a probability and a calculation formula of a deep learning model based on fuzzy logic, so as to accurately calculate the optimal route of the inter-satellite link. In the existing researches, most of the researches start from a static space-time diagram, a most basic shortest path algorithm is introduced, and an evaluation result is more one-sided and is difficult to be applied to a low-orbit satellite constellation. Therefore, the intelligent optimized routing scheme provided by the invention is not easy to think of the prior art. Furthermore, according to the characteristics of the current low-orbit satellite constellation, the invention invents a highly reliable deep learning model, a targeted inter-satellite link utility function and a corresponding deep learning algorithm through a weighted space-time diagram, so that an intelligent route optimization design with low-orbit constellation characteristics can be generated. The invention is thus unique and inventive. Due to the particularities and complexity of the satellite trajectories and inter-satellite links, both satellite energy consumption and link loss of the satellite constellation make the connection of the inter-satellite links a significant impact. The invention fully analyzes the states and parameters of each satellite node and the inter-satellite links, and develops a routing algorithm with high accuracy so as to meet the design of the routing table between the existing dynamic inter-satellite links. In the existing research, the research model set by researchers is too ideal, and the changeable limiting conditions in extreme scenes are not considered. Therefore, the invention has the characteristics of creativity and difficult realization in solving means. And intelligent calculation of all inter-satellite links is performed through a distributed routing deep learning algorithm, so that an autonomous and predictable routing design result is generated in real time. The automatic update of the inter-satellite route is comprehensively and efficiently realized. The real-time performance and the reliability of the satellite constellation to ground communication transmission are improved. Meanwhile, according to the actual situation, the inter-satellite routing is better subjected to specialized treatment.
Drawings
FIG. 1 is a diagram of a distributed deep learning model in accordance with a preferred embodiment of the present invention;
fig. 2 is a flow chart of a predictive routing method based on a link utility function in accordance with the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and specifically described below with reference to the drawings in the embodiments of the present invention. The described embodiments are only a few embodiments of the present invention.
The technical scheme for solving the technical problems is as follows:
an integrated utility function of an inter-satellite link is proposed to dequantize the above-mentioned index. The utility value of the inter-satellite link starts from the three aspects of signal-to-noise ratio, duration and overall delay. The overall delay includes a processing delay and a transmission delay. And then, obtaining the comprehensive utility value of the inter-satellite link by using a weighted summation and a channel perception oriented adaptive algorithm. Finally, obtaining an optimal routing result by a distributed deep learning algorithm based on fuzzy learning. The method comprises the following specific steps:
and quantitatively evaluating the comprehensive utility value of the inter-satellite link by using an integral weighting formula.
A deep learning algorithm is provided, and an optimal inter-satellite link routing result is automatically generated.
The first step: the Walker constellation was set up using the STK software and contained 64 low orbit satellites, 8 satellites each, and corresponding parameters were derived by the software.
And a second step of: and quantifying the signal-to-noise ratio, the transmission time and the overall delay among the satellites to obtain corresponding utility functions respectively. And then introducing an overall utility function to quantify the overall utility effect of the inter-satellite link.
And a third step of: and importing actual satellite constellation specific parameters such as satellite orbit parameters, satellite transmission time, satellite connection time and the like. And calculating the comprehensive utility function of the inter-satellite link by using the utility functions 1-4 respectively. In particular, the overall latency of the inter-satellite link includes processing latency and transmission latency.
Fourth step: and (3) carrying out iterative updating on the comprehensive utility function obtained in the third step by using a self-adaptive algorithm facing channel perception, and calculating the comprehensive utility value of the inter-satellite link. And introducing a weighted time-space diagram, wherein the weight value of each link is the obtained comprehensive utility value.
Fifth step: and a deep learning algorithm is utilized, a trigonometric function model based on fuzzy logic is utilized, the limitation of Boolean logic is purposefully eliminated, and a better predictable inter-satellite route is obtained.
Preferably, in the third step, the signal-to-noise ratio, duration and overall delay of the inter-satellite link are defined and calculated respectively. Comprising utility functions 1-3:
utility function 1: the signal-to-noise ratio is used to ensure the accuracy of data transmission between two nodes. The statistical properties of the inter-satellite laser link are reflected in the link loss factor. Therefore, to ensure the transmission efficiency of the data packet, the utility function U (s (t)) of the signal-to-noise ratio may be defined as the probability of successful transmission of the data packet, expressed as:
utility function 2: the duration of the inter-satellite link is determined by the law of motion of the satellite entering or leaving the polar region, based on the dynamics of the LEO satellite network. To characterize the stability of a link, the utility function of the link duration at time t can be defined as:
utility function 3: under the condition of establishing a reliable transmission path, when forwarding a data packet, the influence of satellite data processing capacity on link path selection needs to be considered, so that the data packet loss caused by network congestion is avoided. Thus, to characterize the packet processing capability of a satellite, the utility function of the buffer queue can be defined as:
wherein D is p (t) is the inter-satellite transmission delay, as follows:
preferably, the weighted ensemble evaluation formula is applied. And quantizing the signal-to-noise ratio, the duration and the overall delay of the inter-satellite link as a whole to obtain the comprehensive utility function of the inter-satellite link. Comprising utility function 4:
utility function 4: with p reachable paths from satellite S to D, link l k Utility values on reachable paths.
The target route isThe smaller the utility value of the link, the more superior and stable the link.
The concepts and models to which the present disclosure relates are as follows:
1. network model:
the main research scenario of the invention is a star-to-ground fusion network (STIN). The STIN comprises a terrestrial gateway, a low-orbit satellite and a high-orbit satellite. The satellite network and the ground network are used as backbones, and the satellite network, the ground network, the deep space network and the near space network are combined together to form the system for low-delay long-distance high-throughput data transmission. The satellite-ground fusion network utilizes the advantages of the satellite network to provide bandwidth access for the whole world and realize the global three-dimensional seamless coverage network service. The star-ground fusion network can not only enlarge the coverage capacity of the network, ensure the continuity of the service and improve the reliability of the network. In addition, the method can be widely applied to earth observation, sea area communication, airborne communication, emergency communication and the like.
2. The technical scheme of the invention is as follows:
the invention provides a predictable inter-satellite routing method based on a link utility function. The method aims at the problems that in a low orbit constellation network, relative motion among low orbit satellite nodes causes frequent network topology change and intermittent connection of inter-satellite links, and provides a predictable inter-satellite routing method. Establishing a corresponding inter-satellite link comprehensive utility function according to the quantized signal-to-noise ratio, duration and overall delay, and obtaining a weighted space-time diagram by using a channel perception-oriented self-adaptive algorithm; based on the obtained weighted space-time diagram, a predictable optimal inter-satellite route is obtained by utilizing a distributed deep learning algorithm according to the comprehensive utility value of each link, so that the packet loss rate, the congestion value and the end-to-end delay of the inter-satellite link are reduced on the basis of the predictable route.
3. Objective function:
Max Q new (S,a)＝(1-δ)Q old (S,a)+δ(R+max(S,a)) (6)
the system, apparatus, module or unit set forth in the above embodiments may be implemented in particular by a computer chip or entity, or by a product having a certain function.
It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article or apparatus that comprises the element.
The above examples should be understood as illustrative only and not limiting the scope of the invention. Various changes and modifications to the present invention may be made by one skilled in the art after reading the teachings herein, and such equivalent changes and modifications are intended to fall within the scope of the invention as defined in the appended claims.
Claims (9)
1. A predictable inter-star routing method based on a link utility function, comprising the steps of:
step 101, constructing a link incidence matrix IM of a low-orbit satellite, and mapping out a link relation between satellites; then collecting the data of the low-orbit inter-satellite links, and obtaining a signal-to-noise ratio utility function U (s (t)), a duration utility function U (l (t)), and an overall delay utility function U (b (t)) according to the data; finally integrating the three utility functions to obtain a multi-attribute utility function of the link
Step 102, performing multiple iterative computations through a channel-aware-oriented adaptive algorithm to obtain a multi-attribute utility functionWeight of each influence factor in the database; then, according to the utility value of each link, a weighted time-space diagram is obtained;
step 103, introducing a deep learning algorithm, adding fuzzy logic, and finally obtaining the predictable inter-satellite routing based on the link utility function.
2. The method for predictable inter-satellite routing based on link utility function according to claim 1, wherein the step 101 constructs a link correlation matrix IM of the low-orbit satellites, maps out the link relationship between satellites, and specifically comprises the following steps:
(1) Obtaining a link incidence matrix IM of the satellite according to the Walker constellation;
link incidence matrix: wherein each row of the matrix corresponds to a satellite node having a different orbit, each column corresponds to a satellite in the same orbit, if node s i To s j A link exists between them, b ij 1, otherwise 0; the link incidence matrix formula is as follows:
IM＝(b ij ) n×n ∈{0，1} (1)
the average path number of any node is formulated as follows:
wherein p is i (s, r) is a path i from node s to node r,representing the sum of all paths from s to r, V representing the number of nodes in the network, i representing the ith reachable path from node s to node r, n representing the number of reachable paths from node s to node r, and if there is no path from node s to node r, p i (s, r) =0, p(s) is used to describe the importance of node s in the inter-star network.
3. The method for predictable inter-star routing based on the link utility function according to claim 2, wherein the calculating of the signal-to-noise utility function U (s (t)) in step 101 specifically includes: the signal-to-noise ratio between any two low-orbit satellite nodes is quantized to the formula:
wherein y is ij Is the final signal, n ij For signal independent additive white gaussian noise G is the transmit power, L ij For Euclidean distance between two satellites, gamma is the path loss index, the link loss factor h ij Expressed as:
wherein the tracking error angle is theta, and the beam divergence half-width is omega 0 Gaussian beam of h ij The probability density function of (2) is expressed as:
wherein sigma 2 Expressed as additive white gaussian noise n in formula (3) ij Thus, the signal-to-noise ratio of the inter-satellite link at any instant is expressed as:
wherein, the probability of the interrupt event occurrence can be deduced as:
alpha represents a threshold value, and a utility function U (s (t)) of signal-to-noise ratio is defined as probability of successful transmission of the data packet, expressed as:
U(s(t))＝1-P r {SNR ij (t)＜α} (8)。
4. a method of predictable inter-star routing based on a link utility function according to claim 3, wherein calculating the link duration U (l (t)) of the inter-star link in step 101 specifically comprises:
calculating the duration time of the link according to the switching rule of the inter-satellite link; phase difference between adjacent satellites between planes:
wherein Δω f N is the phase difference between adjacent satellites on adjacent orbit planes L Represents the number of track planes, M L Representing the number of satellites in each orbital plane, F being the phase factor of the Walker constellation; when the satellites at two ends of the inter-satellite link are v s The inter-satellite link is broken when the angular velocity of (a) enters the polar region; when satellites at two ends of the link leave the polar region, the link is re-established; thus T d The inter-satellite link outage time of (a) can be derived as:
wherein omega β The phase angle is corresponding to the orbit polar dimension beta;
the duration function of the intermittent inter-satellite link is derived as:
the utility function defining the link duration at time t is:
5. the method for predictable inter-star routing based on link utility function according to claim 4, wherein the calculating of the overall time delay U (b (t)) in step 101 is specifically: a (t) represents the number of packets arriving in the t period, and the probability function of a (t) =k is:
wherein k represents that k data packets arrive in a t period, the average speed is lambda, and the arrival rate of the data packets obeys poisson distribution;
the service time of each message obeys an independent and same distribution function and is marked as B (t); definition v n For the processing time of the nth packet,V n representing the independent processing time of the data packet; d (t) =max { n: V n Less than or equal to t } represents the number of data packets processed by the satellites within time t; n represents the nth packet processed in period t, so the probability function of D (t) =n is:
P(D(t)＝n)＝B n (t)-B n+1 (t)＝C n (t) (14)
C n and (t) represents the probability of processing the number of n data packets in the period t. In addition, inter-satellite links also have a transmission delay D p (t) the following formula:
where c represents the speed of light, the length of the packet X (t) can be expressed as:
X(t)＝m+A(t)-D(t) (16)
where m is the number of remaining packets at time t. The processing probability p of the current data queue X (t) can be obtained m,k (t)：
Defining the utility function of the buffer queue as:
U(b(t))＝p m,k (t) (18)
constructing a comprehensive utility function of the inter-satellite link according to the three sub utility functions:
the SNR and the link duration are selected, three factors of the buffer queue are listed as decision attributes to evaluate the link quality, and a multi-attribute utility function of the link utility is designed to quantify the interaction among the attributes.
With p reachable paths from satellite S to D, link l k Utility value on reachable path:
where R (S, D) represents a certain reachable path from satellite S to D and g represents the number of links in reachable path R (S, D).
6. The method for predictable inter-star routing based on link utility function according to claim 5, wherein said step 102 comprises the steps of:
(1) Initializing weight distribution of training;
first, the trained weight distribution D is initialized 1 ＝{w 11 ,w 12 ,...,w 1i ,...,w 1n },Each weight is assigned the same weight; then training the weak classifier, if a certain training sample point is accurately classified by the weak classifier during training, and reducing the corresponding weight when constructing the next training set; conversely, the weight is larger; the sample with updated weight is used for training the next classifier, and the whole process is iterated in this way;
the iteration number is represented by m, and D with weight distribution m Learning the training data set to obtain a basic classifier; then, the weak classifier G is recalculated m (x) Classification error rate on training dataset:e m ＝P(G m (x i )≠y i ) The error rate is the sum of the probabilities of misclassification during classification;
(2) Introducing multi-attribute utility function of inter-satellite link, and finally outputting weight distribution D m ；
Importing utility function, and then calculating G m (x) Coefficient of alpha m Represents G m (x) Importance in the final classifier;updating weight distribution of the training data set to obtain weight distribution of the sample for the next iteration; the weight distribution function is shown in the following formula:
wherein y is i Belongs to the state set, y i E { -1,1}, y when samples are misclassified i G m (x i ) -1, error sample weight:when the samples are misclassified, y i G m (x i ) =1, error sample weights: />Z m Is a normative factor, prescribing the error magnitude of the algorithm result:
finally, combining each weak classifier according to the following formula:
thus the final classifier:
finally, the weight distribution D is output M Weak classifier G M (x)。
7. The method of link utility function based predictable inter-star routing according to claim 6, wherein the optimal routing of step 103 comprises: constructing a weighted space-time diagram according to the obtained weight distribution, and importing the weighted space-time diagram into a distributed learning route model;
in the distributed learning routing model, collecting data of nodes in a cluster by a selected cluster head node, transmitting the data to an experience pool of a training center, updating related data, and generating rewards according to the state change after the last action is executed by the environment; here, selecting two adjacent nodes of the satellite node as destination nodes of the next action, selecting and using a dual-layer Q learning network DQN architecture in the distributed learning routing model, and the action is based on the current state S t Policy pi below selects action A t The environment is according to A t Transition to the next state S t+1 The method comprises the steps of carrying out a first treatment on the surface of the Rewards R for action received environmental feedback t And selects the next action according to the policy pi and adds fuzzy logic to the model.
8. The method of claim 7, wherein in step 103, only one target node parameter is required for the state in the mode, considering that the forwarding of the data packet satisfies the markov random process, as shown in the following formula:
S＝[N des ] (24)
the distributed mode reward calculation is divided into the following two cases; first, if N next Not N des Then no matter N current Whether or not it is directly reachable, rewards are the most severe penalties: r is R min The method comprises the steps of carrying out a first treatment on the surface of the The second is N current Can directly reach N next And N is next Is N des Then the prize is the maximum prize, R max Subtracting from N current To N next Multiplying the cost of (a) by a positive coefficient alpha; the following formula is shown:
the target Q network update considers two parts; one is a reward that is immediately earned by the action, the other is a long-term jackpot, calculated from the action by the neighbor's q_value network; the formula is shown below, which means N at the destination node des In the case of (2), from [ N i ]To [ N ] j ]Should be from [ N ] i ][N j ]Adding N to the direct prize of (2) j Q_value to [ N ] des ]A calculated maximum value; this is an iterative process that eventually converges to a steady state.
In the distributed algorithm, the link selection process can be regarded as a Markov process, and the focus is on the Q_value of the two-hop neighbor of the current node. Next, an update iteration is performed in the same manner as described above, with q_value being periodically updated by q_target.
9. The link utility function based predictable inter-star routing method of claim 8, wherein in fuzzy logic a proposition is no longer non-true or false, it can be considered as "partial true"; the mutual independence between the two values is canceled in the fuzzy logic, and the membership degree is used for representing the state between the two values;
the fuzzy set is that elements on the domain can be partially affiliated to the set A; the degree to which an element belongs to set A is called membership, and fuzzy sets can be defined by membership functions as shown in the following formula:
wherein a, b, c are membership degrees in the fuzzy logic algorithm.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311069788.7A CN116886176A (en) | 2023-08-23 | 2023-08-23 | Predictable inter-satellite routing method based on link utility function |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311069788.7A CN116886176A (en) | 2023-08-23 | 2023-08-23 | Predictable inter-satellite routing method based on link utility function |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116886176A true CN116886176A (en) | 2023-10-13 |
Family
ID=88258916
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311069788.7A Pending CN116886176A (en) | 2023-08-23 | 2023-08-23 | Predictable inter-satellite routing method based on link utility function |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116886176A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117395188A (en) * | 2023-12-07 | 2024-01-12 | 南京信息工程大学 | Deep reinforcement learning-based heaven-earth integrated load balancing routing method |
-
2023
- 2023-08-23 CN CN202311069788.7A patent/CN116886176A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117395188A (en) * | 2023-12-07 | 2024-01-12 | 南京信息工程大学 | Deep reinforcement learning-based heaven-earth integrated load balancing routing method |
| CN117395188B (en) * | 2023-12-07 | 2024-03-12 | 南京信息工程大学 | Deep reinforcement learning-based heaven-earth integrated load balancing routing method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Oubbati et al. | Dispatch of UAVs for urban vehicular networks: A deep reinforcement learning approach | |
| CN114900225B (en) | Civil aviation Internet service management and access resource allocation method based on low-orbit giant star base | |
| CN116886176A (en) | Predictable inter-satellite routing method based on link utility function | |
| CN115173923B (en) | Low-orbit satellite network energy efficiency perception route optimization method and system | |
| CN113382060B (en) | Unmanned aerial vehicle track optimization method and system in Internet of things data collection | |
| CN116248164A (en) | Fully distributed routing method and system based on deep reinforcement learning | |
| CN106452555A (en) | Multi-path optimization algorithm planning method based on medium and low earth orbit satellite network | |
| CN117041129A (en) | Low-orbit satellite network flow routing method based on multi-agent reinforcement learning | |
| CN114567365A (en) | Routing method and system for low-earth-orbit satellite network load balancing | |
| CN116170853A (en) | QoS and load balancing cooperative guarantee satellite routing algorithm based on machine learning | |
| Pi et al. | Dynamic planning of inter-plane inter-satellite links in LEO satellite networks | |
| CN116781139A (en) | Flow prediction satellite path selection method and system based on reinforcement learning | |
| CN117014355A (en) | TSSDN dynamic route decision method based on DDPG deep reinforcement learning algorithm | |
| CN115622603A (en) | Age minimization optimization method for auxiliary transmission information | |
| CN116634498A (en) | Low orbit satellite constellation network edge calculation multistage unloading method based on reinforcement learning | |
| CN110149140B (en) | Forwarding method of satellite opportunistic network | |
| Zhao et al. | Adaptive multi-UAV trajectory planning leveraging digital twin technology for urban IIoT applications | |
| CN115941017A (en) | Internet of things communication resource joint allocation method based on multi-agent reinforcement learning | |
| Hsu et al. | A deep reinforcement learning based routing scheme for LEO satellite networks in 6G | |
| Huang et al. | A GNN-enabled multipath routing algorithm for spatial-temporal varying LEO satellite networks | |
| Chen et al. | Profit-aware cooperative offloading in uav-enabled mec systems using lightweight deep reinforcement learning | |
| Li et al. | Ant colony optimisation for the routing problem in the constellation network with node satellite constraint | |
| CN116847425A (en) | Multi-resource route optimization method based on high-dimensional data joint optimization | |
| Wang et al. | Multi-agent deep reinforcement learning for dynamic laser inter-satellite link scheduling | |
| CN116320841A (en) | FSO self-adaptive routing method based on deep reinforcement learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |