A method for updating satellite resource data in network management based on deep learning

Abstract

The present invention relates to the field of satellite communication networks, and discloses a method for updating network management satellite resource data based on deep learning. The key points of the technical solution are: a network management system prepares task resource data, and sends the task resource data to a satellite communication terminal station system; the satellite communication terminal station system searches for matching communication equipment and sends task parameters according to the received task resource data; the communication equipment executes the task parameters, and reports behavior data to the satellite communication terminal station system in real time; the satellite communication terminal station system pushes the behavior data reported by the communication equipment to the network management system in real time; the network management system prepares the behavior data and pushes it to a page through a full-duplex communication protocol; a deep reinforcement model is established to optimize the update speed and update sequence of four types of data, thereby solving the problem of slow update speed of satellite resources, antenna resources, earth station resources, and link data in large satellite communication networks, and improving the communication efficiency of satellite communication networks.

Description

Technical Field

[0001] The present invention relates to the field of satellite communication networks, and more specifically, to a method for updating network management satellite resource data based on deep learning. Background Art

[0002] The real-time changes in the location, status, and communication conditions of communication equipment require real-time monitoring and updating of device information after the communication equipment changes its action, and device information data push interaction. Because it is necessary to actively obtain device information before this, it is usually impossible to update device information in real time. Therefore, it is necessary to solve the problem of delayed information acquisition to achieve real-time data update and interaction. Summary of the invention

[0003] The purpose of the present invention is to provide a network management satellite resource data updating method based on deep learning, so as to solve the problem that the communication terminal equipment data cannot be reported and updated in real time during satellite communication.

[0004] In order to achieve the above object, the technical solution adopted by the present invention is as follows: a method for updating satellite resource data of a network management system based on deep learning, comprising the following steps:

[0005] The network management system prepares mission resource data and sends the mission resource data to the satellite communication terminal station system;

[0006] The satellite communication terminal station system searches for matching communication equipment and issues mission parameters based on the received mission resource data;

[0007] The communication equipment executes the mission parameters and reports the behavior data to the satellite communication terminal station system in real time;

[0008] The satellite communication terminal station system pushes the behavior data reported by the communication equipment to the network management system in real time;

[0009] The network management system prepares and processes the behavior data and pushes it to the page through the full-duplex communication protocol.

[0010] A reinforcement learning model is established for four types of data that need to be updated, including satellite resource interaction, antenna resource interaction, earth station resource interaction, and link data. Let P represent the queuing state, 1 for queuing, 0 for not queuing, and V represent the update speed. Then the queuing matrix and speed matrix of satellite resources, antenna resources, earth station resources, and link data of the i-th communication link in the satellite communication system are respectively expressed as and Let S represent the current update state, then the data update state of the i-th link is Where S i represents the state space of the ith link; let the queue lengths of the ith link at time t and time t+1 be and The reward value at each moment is When the reward value is positive, it means that the action taken at time t has a positive impact on data update, making data update faster; when the reward value is negative, it means that the system's action causes the data update time to increase. Suppose the action executed at time t is Among them A i Indicates the update order of the four types of data of the i-th link;

[0011] Based on the reinforcement learning model, the Q value calculation formula of the deep reinforcement learning state of the i-th link is as follows:

[0012]

[0013] Among them, θ i and θ′ i denote the network parameters of the estimated network and the target network respectively, N denotes the set of all links in the communication satellite system, and ω(i, j) denotes the influence weight value of the update process of the jth link on the ith link.

[0014] Preferably, the task resource data includes satellite data, antenna data, beam data, transponder data, bandwidth data and basic information data required for the task. The network management system selects antennas and transponders suitable for the current task according to the satellite used by the task, and automatically plans the required communication rate options according to the transponder data, and sets the control channel. The bandwidth will be selected according to the available resources held by the transponder. After clicking, the bandwidth and center frequency will be automatically calculated. The calculation formula is: CF represents the center frequency, E represents the end frequency, and S represents the start frequency.

[0015] Preferably, after the satellite communication terminal station system receives the task resource data issued by the network management system, it first interacts with the task result issued by the network management system; the interaction process is: when the networking task parameters issued by the network management system are all correct, a successful task issuance response will be returned first; when the parameters are erroneous, a task issuance failure response will be returned; after the task is successfully issued, the satellite communication terminal station system will perform networking operations according to the issued networking task parameters, and after the networking is completed, the networking results will be returned to the network management system.

[0016] Preferably, according to the task resource data sent by the network management system, the satellite communication terminal station system first agrees on an interface protocol with the network management system, then automatically finds the communication hardware equipment, and sends the satellite data, beam data and site data carried by the networking task to the communication equipment.

[0017] Preferably, after receiving the task networking instruction issued, the communication device will automatically enter the network, establish a link, communicate and report the behavior according to the behavior identifier carried by the networking task parameters.

[0018] Preferably, the automatic network access process of the communication device is as follows: after the task parameters are sent to the communication device, the communication device first writes data according to the task parameters. After the data writing is completed, a success mark will be returned, and the communication device completes the network access action. After the network access is successful, the communication device will notify the upper-layer service of the satellite communication terminal station system and report the return parameters of the successful network access to the network management system.

[0019] Preferably, after receiving the reported data, the network management system will first translate, convert and persist the data, and then push the data to the network management line communication status page in real time through the IP full-duplex communication protocol for data display.

[0020] Preferably, when multiple links are updating satellite resources simultaneously, the state of the i-th link at each time t is calculated based on the reinforcement learning model. Input, calculate output value Optimal control is achieved by selecting the data update sequence corresponding to the Q value of the network output in the current state.

[0021] In summary, the present invention has the following beneficial effects: it solves the problem of slow update speed of satellite resources, antenna resources, earth station resources and link data in large satellite communication networks, and improves the communication efficiency of satellite communication networks. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 It is a flowchart of a method for updating network management satellite resource data based on deep learning provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0023] The present invention is further described in detail below in conjunction with the accompanying drawings.

[0024] The present invention provides a method for updating satellite resource data of a network management system based on deep learning, such as Figure 1 As shown, the following steps are included:

[0025] S1. The network management system prepares task resource data. The task resource data includes satellite data, antenna data, beam data, transponder data, bandwidth data and basic information data required for the task. The basic information data required for the task includes task name, working mode, task code, service type, start time, end time, communication system, contact unit, task level, contact person and task description; the satellite data includes satellite name, satellite altitude, number of beams, orbit inclination, satellite longitude and latitude, and beam data includes number of channels, coverage radius, center point longitude, center point latitude, uplink control channel, downlink control channel, uplink satellite-to-ground dedicated, downlink satellite-to-ground dedicated, satellite-to-ground channel status; in order to solve the formula for the bandwidth required for the current task, the network management system selects the antenna and transponder suitable for the current task according to the satellite used by the task, and automatically plans the required communication rate options according to the transponder data, and sets the control channel. The bandwidth will be selected according to the available resources held by the transponder. After clicking, the bandwidth and center frequency will be automatically calculated. The calculation formula is: Where CF represents the center frequency, E represents the end frequency, and S represents the start frequency; the calculation of the bandwidth required for the task is completed; after the calculation of the bandwidth required for the task is completed, the network management system sends the task resource data to the satellite communication terminal station system;

[0026] After receiving the task resource data sent by the network management system, the satellite communication terminal station system first interacts with the task results sent by the network management system; the interaction process is as follows: when the networking task parameters sent by the network management system are all correct, the satellite communication terminal station system will first return a successful task response; when the parameters are wrong, the satellite communication terminal station system will return a failed task response; after the task is successfully sent, the satellite communication terminal station system will perform networking operations according to the sent networking task parameters, and after the networking is completed, it will return the networking results to the network management system.

[0027] S2, the satellite communication terminal station system searches for matching communication equipment and issues mission parameters based on the received mission resource data;

[0028] like Figure 1 As shown, according to the task resource data issued by the network management system, the satellite communication terminal station system first agrees on an interface protocol with the network management system. Based on the beam data and site data carried in the task resource data issued by the network management system, the satellite communication terminal station system will automatically search for matching communication site equipment that meets the conditions according to the three conditions of beam coverage, site longitude and latitude, and communication capability, and send the satellite data, beam data and site data carried by the networking task to the communication equipment; after completing the data delivery, that is, after the satellite communication terminal station system returns the site status data, site location data, and networking result data, the network management system performs data analysis, and after the analysis is completed, it is reorganized into data within the network management system, and finally pushed to the network management system status page for data display.

[0029] S3, the communication equipment executes the mission parameters and reports the behavior data to the satellite communication terminal station system in real time;

[0030] After receiving the task networking instruction, the communication equipment will automatically access the network, establish a link, communicate and report behavior according to the behavior identifier carried by the networking task parameters; the automatic network access process of the communication equipment is as follows: after the task parameters are issued to the communication equipment, the communication equipment first writes data according to the task parameters. After the data is written, it will return a success identifier, and the communication equipment completes the network access action. After the network access is successful, the communication equipment will notify the upper-layer service of the satellite communication terminal system and report the return parameters of successful network access to the network management system.

[0031] S4, the satellite communication terminal station system pushes the behavior data reported by the communication equipment to the network management system in real time;

[0032] S5, the network management system prepares and processes the behavior data and pushes it to the page via the full-duplex communication protocol;

[0033] After receiving the reported data, the network management system will first perform data translation, data conversion and data persistence preparation processing on the data; data translation compares and translates the data sent by the satellite communication terminal station system to the network management system into data that can be recognized and used by the network management system; data conversion is to convert the data type of the data returned by the satellite communication terminal station system to the data type of the unified unit of the network management system; data persistence is to write the data returned from the satellite communication terminal station system to the network management system into the physical memory to perform data persistence operations; after the data preparation processing is completed, the data is now available and displayable by the network management system, and is pushed to the network management line communication status page in real time through the IP full-duplex communication protocol for data display, so that users can view the networking tasks issued by themselves in real time.

[0034] S6. Establish a reinforcement learning model for four types of data that need to be updated. The four types of data that need to be updated include satellite resource interaction, antenna resource interaction, earth station resource interaction, and link data. Let P represent the queuing state, 1 for queuing, 0 for not queuing, and V represent the update speed. Then the queuing matrix and speed matrix of the satellite resources, antenna resources, earth station resources, and link data of the i-th communication link in the satellite communication system are respectively expressed as and Let S represent the current update state, then the data update state of the i-th link is Where S i represents the state space of the ith link; let the queue lengths of the ith link at time t and time t+1 be and The reward value at each moment is When the reward value is positive, it means that the action taken at time t has a positive impact on data update, making data update faster; when the reward value is negative, it means that the system's action causes the data update time to increase. Suppose the action executed at time t is Among them A i Indicates the update order of the four types of data of the i-th link;

[0035] Based on the reinforcement learning model, the Q value calculation formula of the deep reinforcement learning state of the i-th link is as follows:

[0036]

[0037] Among them, θ i and θ′ i They represent the network parameters of the estimated network and the target network respectively, N represents the set of all links in the communication satellite system, and ω(i, j) represents the weight value of the influence of the update process of the jth link on the ith link; usually, the smaller the change of the IP message data, the smaller the weight, and the larger the change of the IP message data, the larger the weight;

[0038] When multiple links are updating satellite resources at the same time, the order of updating multiple links is involved. Based on the reinforcement learning model, the state of the i-th link at each time t is used to calculate the update order of multiple links. Input, calculate output value The update sequence of multiple links is determined by the Q value output by the network in the current state and the data corresponding to the Q value is sorted in order, so that the speed of simultaneous update of multiple links can be optimally controlled.

[0039] The above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims

1. A method for updating network management satellite resource data based on deep learning, characterized in that: The steps include: The network management system prepares mission resource data and sends the mission resource data to the satellite communication terminal station system; The satellite communication terminal station system searches for matching communication equipment and issues mission parameters based on the received mission resource data; The communication equipment executes the mission parameters and reports the behavior data to the satellite communication terminal station system in real time; The satellite communication terminal station system pushes the behavior data reported by the communication equipment to the network management system in real time; The network management system processes the behavior data and pushes it to the page through the full-duplex communication protocol; A reinforcement learning model is established for four types of data that need to be updated, including satellite resource interaction, antenna resource interaction, earth station resource interaction, and link data. Let P represent the queuing state, 1 for queuing, 0 for not queuing, and V represent the update speed. Then the queuing matrix and speed matrix of satellite resources, antenna resources, earth station resources, and link data of the i-th communication link in the satellite communication system are respectively expressed as and Let S represent the current update state, then the data update state of the i-th link is Where S i represents the state space of the ith link; let the queue lengths of the ith link at time t and time t+1 be and The reward value at each moment is When the reward value is positive, it means that the action taken at time t has a positive impact on data update, making data update faster; when the reward value is negative, it means that the system's action causes the data update time to increase. Suppose the action executed at time t is Among them A i Indicates the update order of the four types of data of the i-th link; Based on the reinforcement learning model, the Q value calculation formula of the deep reinforcement learning state of the i-th link is as follows: Among them, θ i and θ′ i denote the network parameters of the estimated network and the target network respectively, N denotes the set of all links in the communication satellite system, and ω(i, j) denotes the influence weight value of the update process of the jth link on the ith link.

2. The method for updating network management satellite resource data based on deep learning according to claim 1, characterized in that: The mission resource data includes satellite data, antenna data, beam data, transponder data, bandwidth data and basic information data required for the mission. The network management system selects the antenna and transponder suitable for the current mission according to the satellite used by the mission, and automatically plans the required communication rate options and sets the control channel according to the transponder data. The bandwidth will be selected according to the available resources held by the transponder. After clicking, the bandwidth and center frequency will be automatically calculated. The calculation formula is: CF represents the center frequency, E represents the end frequency, and S represents the start frequency.

3. The method for updating network management satellite resource data based on deep learning according to claim 1, characterized in that: After receiving the task resource data sent by the network management system, the satellite communication terminal station system first interacts with the task result sent by the network management system; the interaction process is: when all the networking task parameters sent by the network management system are correct, a task sending success response will be returned first; If there is an error in the parameters, a task failure response will be returned; after the task is successfully sent, the satellite communication terminal station system will perform networking operations according to the sent networking task parameters, and return the networking results to the network management system after the networking is completed.

4. The method for updating network management satellite resource data based on deep learning according to claim 1, characterized in that: According to the mission resource data sent by the network management system, the satellite communication terminal station system first agrees on the interface protocol with the network management system, then automatically finds the communication hardware equipment, and sends the satellite data, beam data and site data carried by the networking mission to the communication equipment.

5. The method for updating network management satellite resource data based on deep learning according to claim 1, characterized in that: After receiving the task networking instruction, the communication device will automatically enter the network, establish a link, communicate and report the behavior according to the behavior identifier carried by the networking task parameters.

6. The method for updating network management satellite resource data based on deep learning according to claim 5, characterized in that: The automatic network access process of communication equipment is as follows: after the task parameters are sent to the communication equipment, the communication equipment first writes data according to the task parameters. After the data writing is completed, it will return a success mark, and the communication equipment completes the network access action. After the network access is successful, the communication equipment will notify the upper-layer service of the satellite communication terminal system and report the return parameters of successful network access to the network management system.

7. The method for updating network management satellite resource data based on deep learning according to claim 1, characterized in that: After receiving the reported data, the network management system will first translate, convert and persist the data, and then push it to the network management line communication status page in real time through the IP full-duplex communication protocol for data display.

8. The method for updating network management satellite resource data based on deep learning according to claim 1, characterized in that: When multiple links are updating satellite resources at the same time, the reinforcement learning model is used to calculate the state of the i-th link at each time t. Input, calculate output value Optimal control is achieved by selecting the data update sequence corresponding to the Q value of the network output in the current state.

Images