WO2022156074A1 - 无线智能决策通信方法、装置和*** - Google Patents

无线智能决策通信方法、装置和*** Download PDF

Info

Publication number
WO2022156074A1
WO2022156074A1 PCT/CN2021/086679 CN2021086679W WO2022156074A1 WO 2022156074 A1 WO2022156074 A1 WO 2022156074A1 CN 2021086679 W CN2021086679 W CN 2021086679W WO 2022156074 A1 WO2022156074 A1 WO 2022156074A1
Authority
WO
WIPO (PCT)
Prior art keywords
network node
network
node
information
application scenario
Prior art date
Application number
PCT/CN2021/086679
Other languages
English (en)
French (fr)
Inventor
刘建德
梁宏建
马显卿
黎书生
Original Assignee
深圳市科思科技股份有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 深圳市科思科技股份有限公司 filed Critical 深圳市科思科技股份有限公司
Priority to JP2021576899A priority Critical patent/JP7333839B2/ja
Priority to EP21761959.2A priority patent/EP4061048A4/en
Priority to US17/469,642 priority patent/US20220240162A1/en
Publication of WO2022156074A1 publication Critical patent/WO2022156074A1/zh

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W24/00—Supervisory, monitoring or testing arrangements
    • H04W24/02—Arrangements for optimising operational condition
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W24/00—Supervisory, monitoring or testing arrangements
    • H04W24/10—Scheduling measurement reports ; Arrangements for measurement reports
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/08—Configuration management of networks or network elements
    • H04L41/0893—Assignment of logical groups to network elements

Definitions

  • the present application relates to the technical field of wireless communication, and in particular, to a wireless intelligent decision-making communication method, device and system.
  • various wireless communication network resources are each suitable for a specific application scenario.
  • the 3GPP LTE protocol based on orthogonal frequency division multiplexing technology is suitable for terrestrial low-speed mobile communication, and the homogeneous forwarding based on time division multiplexing technology.
  • the ad hoc network protocol is suitable for scenarios with a high proportion of broadcast and multicast services.
  • the original single network protocol will present a disadvantage. Therefore, it is necessary to propose a decision-making communication method for wireless network resources to solve the problem that resources cannot be switched intelligently when the application scenarios of network nodes change, resulting in small network capacity and throughput, which cannot meet user needs.
  • the present application provides a wireless intelligent decision-making communication method, device and system to solve the problem that resources cannot be switched intelligently when the application scenario of network nodes changes, and solve the problem that network capacity and throughput are small and cannot meet user needs.
  • an embodiment of the present application provides a wireless intelligent decision-making communication method, the method includes:
  • the data information includes scheduling request information, resource reservation indication information, network status information, wireless channel information and communication capability information of the network node;
  • the decision result including the multi-domain combination is determined, and the multi-domain combination of the network node is activated at the preset time; wherein, the The multi-domain combination corresponds to the application scenario of the network node.
  • predicting the application scenario of the network node at a preset time according to the data information including:
  • the application scenario of the network node is determined according to the deterministic change information.
  • determine the decision result including the multi-domain combination including:
  • the application scenario of the network node at the preset time and the communication capability information of the network node are input into the reinforcement learning model, and the output result of the reinforcement learning model is determined as the decision result.
  • determine the decision result including the multi-domain combination including:
  • the multi-domain combination includes time domain resources and frequency domain resources of the network node; or ,
  • the decision result is to activate the time domain resources and frequency domain resources of the network node.
  • resources and airspace resources or,
  • the decision result is when the network node is activated. domain resources, frequency domain resources, spatial domain resources, and code domain resources; or,
  • the The decision result is to activate the time domain resources, frequency domain resources, space domain resources, code domain resources and delay Doppler domain resources of the network node.
  • the method is applied to a network node, and the network node predicts the application scenario where the network node is located by using a convolutional long short-term memory hybrid neural network model; the network node determines the multi-domain combination to be activated by using a reinforcement learning model. .
  • obtain data information of network nodes including:
  • predicting the application scenario of the network node at the preset time according to the data information includes:
  • the network node at the preset time determine the decision result including the multi-domain combination, and activate the multi-domain combination of the network node at the preset time, including:
  • the decision result is determined by the network autonomous decision node according to the application scenario and the communication capability information of the network node; and the decision result is sent to the network node through a unified scheduling management protocol frame, so that the network node is in the preset Multidomain Combinations in Time-Activated Decision Results.
  • obtain data information of network nodes including:
  • the data information obtained by the cloud scheduling component is the first data information
  • the data information obtained by the network autonomous decision node is the second data information
  • the first data information is long-period data information
  • the second data information is short-period data information
  • predicting the application scenario of the network node at the preset time according to the data information includes:
  • the network node at the preset time determine the decision result including the multi-domain combination, and activate the multi-domain combination of the network node at the preset time, including:
  • the network autonomous decision node determines a third decision result according to the first decision result and the second decision result, and sends the third decision result to the network node through a unified scheduling management protocol frame, so that the network node can
  • the above preset time activates the multi-domain combination in the third decision result.
  • the cloud scheduling component and the network autonomous decision-making node both include a unified scheduling management module and at least one protocol conversion module; respectively obtain the data information of the network node through the cloud scheduling component and the network autonomous decision-making node, including:
  • the cloud scheduling component or the network autonomous decision node switches the network type to be consistent with the network node through the protocol conversion module the type of network;
  • predicting the first application scenario of the network node at the preset time by the cloud scheduling component includes:
  • predicting the second application scenario of the network node at the preset time through the network autonomous decision node includes:
  • the second decision result is determined according to the second application scenario by the unified scheduling management module in the autonomous decision-making node of the network.
  • both the cloud scheduling component and the network autonomous decision node include a data storage module, and the method further includes:
  • the cloud scheduling component and the network autonomous decision node store the communication capability information of the network node in the blockchain
  • the network state information, wireless channel information, scheduling request information, scheduling request information and resource pre-emption indication information reported by the network node are stored in chronological order through the data storage module.
  • an embodiment of the present application provides a wireless intelligent decision-making communication device for implementing the method described in the first aspect, the device comprising:
  • an acquisition module configured to acquire data information of each network node;
  • the data information includes scheduling request information, resource reservation indication information, network state information, wireless channel information and communication capability information of the network node;
  • a prediction module for a network node, configured to predict the scene where the network node is located at a preset time according to the data information
  • an activation module configured to determine a decision result including a multi-domain combination according to the application scenario of the network node at a preset time and the communication capability information of the network node, and activate the multi-domain combination of the network node at the preset time ; wherein, the multi-domain combination and the application scenario of the network node are in a corresponding relationship.
  • an embodiment of the present application provides a wireless intelligent decision-making communication system for implementing the method described in the first aspect, the system comprising: a network autonomous decision-making node and a network node;
  • the network node is configured to send data information to an autonomous decision-making node of the network;
  • the data information includes scheduling request information, resource reservation indication information, network state information, wireless channel information and communication capability information of the network node;
  • the network autonomous decision node is used to obtain data information of the network node, predict the application scenario of the network node at a preset time, and determine the decision result according to the application scenario and the communication capability information of the network node;
  • the network autonomous decision node is further configured to send the decision result to the network node through a unified scheduling management protocol frame, so that the network node activates the multi-domain combination in the decision result at the preset time.
  • an embodiment of the present application provides a wireless intelligent decision-making communication system for implementing the method in the first aspect, the system comprising: a network autonomous decision-making node, a cloud scheduling component, and a network node;
  • the cloud scheduling component is configured to determine a first application scenario according to the acquired first data information, and determine a first decision result according to the first application scenario;
  • the network autonomous decision node is configured to determine a second application scenario according to the acquired second data information, and determine a second decision result according to the second application scenario;
  • the network autonomous decision node determines a third decision result according to the first decision result and the second decision result, and sends the third decision result to the network node through a unified scheduling management protocol frame, so that the network node is in the
  • the preset time activates the multi-domain combination mode in the third decision result
  • the first data information is long-period data information
  • the second data information is short-period data information
  • the wireless intelligent decision-making communication method, device, and system provided by the embodiments of the present application include: acquiring data information of a network node; the data information includes scheduling request information, resource reservation indication information, network status information, wireless channel information, and The communication capability information of the network node; predict the application scenario of the network node at the preset time according to the data information; determine the decision result including the multi-domain combination according to the application scenario of the network node at the preset time, and use the The preset time activates the multi-domain combination of the network node; wherein, the multi-domain combination and the application scenario of the network node are in a corresponding relationship, and the above method predicts the scene where the network node is located, and according to the network node
  • the multi-domain combination to be activated is determined according to the scene, so that the network node can adapt to the change of the application scene, maximize the network capacity and throughput, and meet the needs of users.
  • FIG. 1 is a schematic diagram of an application scenario provided by an embodiment of the present application.
  • FIG. 2 is a schematic flowchart of a wireless intelligent decision-making communication method provided by an embodiment of the present application
  • FIG. 3 is a schematic flowchart of another wireless intelligent decision-making communication method provided by an embodiment of the present application.
  • FIG. 4(a) is a schematic structural diagram of a unified scheduling management protocol frame provided by an embodiment of the present application.
  • FIG. 4(b) is a schematic structural diagram of a frame control field provided by an embodiment of the present application.
  • FIG. 4(c) is a schematic structural diagram of another unified scheduling management protocol frame provided by an embodiment of the present application.
  • 5 is a schematic structural diagram of a protocol stack provided by an embodiment of the present application.
  • FIG. 6 is a schematic flowchart of another wireless intelligent decision-making communication method provided by an embodiment of the present application.
  • FIG. 7 is a schematic structural diagram of a wireless intelligent decision-making communication device provided by an embodiment of the present application.
  • FIG. 8 is a wireless intelligent decision-making communication system provided by an embodiment of the present application.
  • FIG. 1 is a schematic diagram of an application scenario provided by an embodiment of the present application.
  • the method can be applied to a network node, a network autonomous decision-making node, or a cloud scheduling component and a network autonomous decision-making node.
  • the network node may be a base station, an access point, a user terminal, an ad hoc network node, a gateway, and the like.
  • the application scenarios of the network nodes at a certain moment in the future can be predicted, and the resources of the five domains of the network can be automatically switched according to the application scenarios. domain, air domain, code domain and delay Doppler domain.
  • the application scenario of the network node can be predicted according to the data information, and the corresponding multi-domain combination can be activated according to the application scenario.
  • the network resources cannot be switched intelligently in time, and the original single network protocol will be maintained for communication, resulting in a small network capacity and throughput, which cannot meet the needs of users. usage requirements.
  • the embodiments of the present application can automatically perceive the application scenario according to the acquired data information, and then determine the multi-domain combination to be activated according to the application scenario, and activate the multi-domain combination at a preset time.
  • the modified multi-domain combination can To meet the communication requirements of network nodes, the network capacity and throughput are increased, thereby meeting the needs of users.
  • FIG. 2 is a schematic flowchart of a wireless intelligent decision-making communication method provided by an embodiment of the present application. As shown in FIG. 2 , the method includes:
  • the data information includes scheduling request information, resource reservation indication information, network state information, wireless channel information, and communication capability information of the network node.
  • the main body that obtains the data information may be a network node, a network autonomous decision-making node, or a network autonomous decision-making node and a cloud scheduling component.
  • a communication method of a centerless ad hoc network is formed.
  • each network node is an autonomous decision-making node of the network. Determine the outcome of the decision.
  • a distributed network communication method is formed.
  • the network autonomous decision-making node can obtain the data information of all network nodes in the area and determine the decision-making result.
  • the main body of data information is the network autonomous decision node and cloud scheduling component
  • a centralized network communication method is formed, and the second level scheduling is realized through the cloud scheduling component and the network autonomous decision node, and the decision result is finally determined.
  • the acquired data information of the network node is multi-dimensional information.
  • the data information can be obtained including scheduling request information, resource reservation indication information, network status information and wireless channel information.
  • the obtained network status information includes but is not limited to the number of network nodes, the occupancy of air interface resources, and the proportion of different types of services. , communication rate, packet loss rate, distance from other network nodes, and signal strength.
  • the acquired wireless channel information includes the transmission channel from the sender to the receiver during communication between nodes.
  • the obtained scheduling request information may be the request sending data amount and the receiving buffer size, etc.
  • the resource reservation indication information may be indication information of resource reservation occupation.
  • acquiring the data information of the network node may be acquiring data information by sending request data to other nodes, or may be receiving data information actively sent by other nodes according to the agreement.
  • the communication capability information of the network node is related to the hardware accelerator and radio frequency resources of the network node itself, and different network nodes have different communication capability information. Only when the network node supports a certain domain resource, the corresponding resource can be activated.
  • a network node can acquire continuous frame data. If the frame data contains 100 time slots, the first 10 time slots represent control scheduling information. Any node can acquire the frame data. After the frame data is acquired, the frame data can be analyzed to obtain the data of the first 10 time slots, and the occupancy of time-frequency resources can be obtained through the data of the first 10 time slots. For example, when the time-frequency resource is obtained through the frame data When resource A is occupied, the node cannot perform data transmission through time-frequency resource A.
  • the communication capability information includes but is not limited to network node identifiers, network node types, whether relaying is supported, whether it can be moved, and the moving range, supported protocol types, supported frequency bands, MIMO (Multi Input Multi Output, Multiple Input Multiple Output) capability, multi-carrier capability and carrier aggregation capability, etc.
  • the communication capability information here is information of all communication capabilities supported by the node.
  • the types of network nodes may be base stations, access points, user terminals, gateways, and relay nodes.
  • the data information may also be saved.
  • the application scenario at the preset time can be predicted according to the acquired data information. Specifically, prediction can be made according to the acquired data information. For example, when it is detected that there will be interference in some areas in the future, it can be determined that the network node is in an application scenario of interference. Specifically, the application scenario of the network node is determined by the acquired scheduling request information, resource reservation indication information, network state information and wireless channel information.
  • the multi-domain combination to be activated can be determined according to the application scenario of the network node, so that the activated multi-domain combination can adapt to the network The application scenario of the node at the preset time. For example, when the application scenario of the network node at the preset time is scenario A, and there is no need to activate five-domain resources at the same time in scenario A, only the multi-domain combination resources corresponding to the scenario can be activated.
  • the wireless communication protocol of frequency division multiple access When the wireless communication protocol of frequency division multiple access is used, only the corresponding time domain resources and frequency domain resources can be activated, so that when the application scenario of the network node changes, the network capacity and throughput can be maximized, and power consumption can be reduced at the same time.
  • time domain resources, frequency domain resources, air domain resources, code domain resources and delay Doppler domain resources will correspond to different software and hardware accelerators respectively.
  • the hardware accelerator is in the working state, that is, the corresponding software is in the loading and running state, and the hardware accelerator is in the power-on state.
  • the programs corresponding to the five-domain resources are essentially different mathematical transformation methods, and the program in the running state can perform corresponding transformation processing on the data to be transmitted, so as to realize the data transmission through the corresponding multi-domain combination method.
  • the hardware accelerator is not a necessary device, and the performance of data processing can be improved by using the hardware accelerator.
  • the network node or the cloud scheduling component or the network autonomous decision-making node can obtain the data information of the network node, predict the application scenario of the network node at the preset time according to the obtained data information, and deactivate the network according to the application scenario
  • the multi-domain combination of nodes can autonomously switch and manage the network's time domain resources, frequency domain resources, airspace resources, code domain resources and delay Doppler domain resources to adapt to changes in application scenarios, maximize network capacity and throughput.
  • predicting the application scenario of the network node at a preset time according to the data information including:
  • the deterministic change information of the network node is predicted according to the data information; the application scenario of the network node is determined according to the deterministic change information.
  • the deterministic change information of the network node can be determined first according to the data information.
  • a convolutional long short-term memory hybrid neural network can be used for prediction. Taking the scheduling request information, resource preemption indication information, network status information and wireless channel information of network nodes in the local or global area as input data, the convolutional long short-term memory hybrid neural network processes the input data, and outputs deterministically changing data. Models of cycles, cyclic patterns, trends and random changes, etc.
  • the number of user terminals and the number of network nodes can be obtained according to the network status information and wireless channel information, and the occupancy of air interface resources can be obtained according to the scheduling request information and the resource reservation indication information,
  • the neural network model it can be predicted whether the number of user terminals and the number of network nodes will exceed the preset value in the future, and whether the occupancy rate of the air interface resources of the network node will exceed the preset value.
  • the predicted information is deterministic change information.
  • the neural network uses a convolutional long-term and short-term memory hybrid neural network.
  • the convolutional long-term and short-term memory hybrid neural network is a combination of the convolutional layer and the pooling layer of the convolutional neural network and the input layer of the long-term and short-term memory neural network.
  • the local features are extracted through the convolution layer and the pooling layer, and then the time series-related features in the local features are obtained through the long short-term memory neural network, which can realize the memory and prediction of short-term or long-term data.
  • the advantages of the convolutional neural network and the long short-term memory network can be used at the same time, and the prediction results with high accuracy can be obtained.
  • the application scenario of the network node After acquiring the deterministic change information, the application scenario of the network node can be obtained according to the deterministic change information, and the application scenario of the network node can be a static application scenario, a mobile application scenario, and the like.
  • the application scenario of the network node can be determined according to the characteristics in the deterministic change information of the network node. For example, when the deterministic change information is that fixed interference or random interference occurs one hour in the future, it can be determined that the application scenario of the network node is an application scenario with interference.
  • the application scenario of the network node can be determined according to the data information of the network node, and the deterministic change information of the network node is predicted first, and then the application scenario is predicted, which can make the determined application scenario more accurate.
  • the use of convolutional long short-term memory hybrid neural network can process the data information with time series relationship very well.
  • determine the decision result including the multi-domain combination including:
  • the application scenario of the network node at the preset time and the communication capability information of the network node are input into the reinforcement learning model, and the output result of the reinforcement learning model is determined as the decision result.
  • the input of reinforcement learning includes application scenarios and communication capability information of network nodes, as well as scheduling strategies.
  • a Q table can be obtained through reinforcement learning training. Each row represents the input application scenario and the communication capability information of network nodes, and each column represents Enter the decision result, the value of each cell of the Q table represents the reward expectation value in the execution of the corresponding decision result.
  • the training can be stopped, and the Q table after training can be obtained.
  • the reinforcement learning model the expected reward value corresponding to each decision result can be obtained, and the decision result with the largest corresponding reward expected value is selected when the decision result is determined.
  • the decision result is to activate the multi-domain combination in the five-domain resource.
  • the above method predicts the decision result through the reinforcement learning model, and can accurately determine the decision result corresponding to the application scenario of the current network node.
  • determine the decision result including the multi-domain combination including:
  • the multi-domain combination includes time domain resources and frequency domain resources of the network node; or ,
  • the decision result is to activate the time domain resources and frequency domain resources of the network node.
  • resources and airspace resources or,
  • the decision result is when the network node is activated. domain resources, frequency domain resources, spatial domain resources, and code domain resources; or,
  • the The decision result is to activate the time domain resources, frequency domain resources, space domain resources, code domain resources and delay Doppler domain resources of the network node.
  • the application scenario in which the network node is stationary refers to the application scenario in which the network node is stationary or relatively stationary.
  • the communication capability information of the network node includes time domain resources and frequency domain resources , then only the combination of time-domain resources and frequency-domain resources can be activated, and there is no need to activate spatial-domain resources, code-domain resources, and delay-Doppler-domain resources.
  • Activating time domain resources refers to carrying data in different time periods for data transmission.
  • Activating frequency domain resources refers to carrying data in different frequency bands for data transmission.
  • Activating time domain resources and frequency domain resources is to carry data in a certain time and frequency segment for data transmission.
  • the new The capacity can be expanded by adding carriers or carrier aggregation.
  • the network can be switched to an orthogonal frequency division multiple access network to accommodate more users. Terminals and network nodes. Therefore, when the network node is in a static application scenario, by activating the time domain resource and the frequency domain resource, the wireless communication network is switched between the time domain resource and the frequency domain resource.
  • the application scenario in which the network node is moving refers to the scenario where the network node is in a moving state, or the spatial span of multiple network nodes is large, or the electromagnetic environment between each network node is different, etc.
  • the communication capability information of the network node includes time domain resources, frequency domain resources and air domain resources, only the combination of time domain resources, frequency domain resources and air domain resources can be activated, and there is no need to activate code domain resources. and time-delay Doppler domain resources.
  • Activating airspace resources refers to carrying data on different antenna ports for data transmission.
  • the network node when the network node is in a mobile state, if the network node uses the broadband multi-antenna space division multiplexing method to send and receive data, and it is predicted that the distance between network nodes will exceed the set value in the future, you can activate time domain resources, frequency Domain resources and airspace resources, realize the switching of wireless communication networks in time domain resources, frequency domain resources and airspace resources.
  • the network can be switched to a narrowband single-transmit-multiple-receive communication protocol, which can improve the power spectral density and reduce the reception signal. Noise ratio requirements.
  • the application scenario in which the network node is in interference refers to that the network node is in interference, or in a complex electromagnetic environment, or in a scenario that requires high transmission bit error rate.
  • the capability information includes time domain resources, frequency domain resources, space domain resources and code domain resources, you can only activate the combination of time domain resources, frequency domain resources, air domain resources and code domain resources without activating the delay Doppler again.
  • domain resources. Activation code domain resources refer to different encoding of data for data transmission.
  • a combination of time domain resources, frequency domain resources, air domain resources and code domain resources can be activated, so that the wireless communication network can be switched to a preset wireless communication network, such as code division
  • a preset wireless communication network such as code division
  • the multiple access communication protocol and the code division multiple access communication protocol have the characteristics of anti-interference, and can adjust the frequency band to the frequency band without interference or the frequency band with little interference, so as to solve the situation of packet loss or inability to communicate due to interference.
  • the application scenario in which the network node is in multiple obstacles refers to the scenario where the network node is in a multi-obstacle scenario, a complex diffraction environment, or a supersonic moving scenario.
  • the communication capability of the network node is If the information includes time domain resources, frequency domain resources, air domain resources, code domain resources and delay Doppler domain resources, the above five network resources can be activated, so that the wireless communication network can be switched to the preset communication network.
  • Alternating time-frequency space communication protocol and orthogonal time-frequency space communication protocol can carry the transmitted or received information in the delay Doppler domain to avoid packet loss or inability to communicate when moving at supersonic speed and in the presence of obstacles.
  • Activating the delay-Doppler domain resources refers to carrying data at different delays and Doppler frequency offsets for data transmission.
  • the application scenario in which the network node is located also includes that the network node moves to an area without signal coverage, or the central control access point may fail in the future.
  • the communication capability information of the network node includes the time domain resources, frequency domain resources and airspace resources, time domain resources, frequency domain resources and airspace resources can be activated, so that the wireless communication network can be switched to the preset wireless communication network, such as satellite relay communication protocol, satellite relay communication protocol
  • the generated radio waves have a larger coverage area and a longer communication distance to avoid packet loss or inability to communicate due to no signal coverage.
  • the application scenario where the network node is located also includes that the proportion of multicast services between network nodes exceeds the preset value.
  • the communication capability information of the network node includes time domain resources, frequency domain resources and air domain resources, it can be Activate time domain resources, frequency domain resources and air domain resources, so that the wireless communication network can be switched to the preset wireless communication network, such as the blocking relay network forwarding protocol, which enables nodes to realize relay forwarding, and then Improve multicast communication speed and reliability.
  • the method is applied to a network node, and the network node predicts the application scenario where the network node is located by using a convolutional long short-term memory hybrid neural network model; the network node determines the multi-domain combination to be activated by using a reinforcement learning model. .
  • the method can be applied to network nodes, that is, each network node is an autonomous decision-making node of the network, and a communication method of an ad hoc network without a center is realized.
  • the network nodes use the convolutional long short-term memory hybrid neural network to predict the application scenarios where the network nodes are located.
  • the reinforcement learning model is used to determine the multi-domain combination to be activated. In the above process, the process of acquiring the data information and processing the data information is performed by the network node.
  • Adopting the structure of a centerless ad hoc network can facilitate the control and management of itself, and has the advantage of high reliability.
  • the method can also form a communication method of a distributed network.
  • the method includes:
  • S303 Determine the decision result according to the application scenario and the communication capability information of the network node through the network autonomous decision node; and send the decision result to the network node through a unified scheduling management protocol frame, so that the network node is in the Preset times activate multi-domain combinations in decision results.
  • wireless network resources can also be scheduled through a network autonomous decision node.
  • the network node will acquire data information that changes in real time, such as network status information, wireless channel information, scheduling request information, and resource reservation instructions.
  • the network autonomous decision node can predict the application scenario of the network node, such as predicting the application scenario through the convolutional long short-term memory hybrid neural network model.
  • the decision result is obtained according to the application scenario and the communication capability information of the network node through the reinforcement learning model, and the decision result is sent to the network node in the form of a unified scheduling management protocol frame.
  • the autonomous decision-making node of the network is used to receive the data reported by the network node, and to determine the decision-making result after processing.
  • FIG. 4 is a schematic structural diagram of a unified scheduling management protocol frame provided by an embodiment of the present application. As shown in FIG. 4( a ), the unified scheduling management protocol frame includes a frame control field, an address format field, and a frame body field.
  • the frame control field is used to store the current wireless network type of the network node;
  • the address format field is used to store the network autonomous decision node address, relay node address and destination network node address; the frame body field to store the decision result.
  • the frame control field and the address format field may be of a fixed number of bytes, and the frame body field may be of a variable length of bytes. In this embodiment, there is no specific limitation on the number of bytes of each field.
  • the frame control field can store the relevant information of the current wireless network of the network node, as shown in Figure 4(b), the content stored in the frame control field includes: the protocol type, protocol version, address format of the current network node, and reserved other frame control Function.
  • the address format field can be divided into several address fields, and the address information required during data transmission is stored in the corresponding address fields respectively.
  • the frame body field is used to store the decision result so that the network node can execute the decision result.
  • the address format field can be divided into network autonomous decision-making node address, receiving address, sending address and extension address, wherein, the network autonomous decision-making node address is the address of the network autonomous decision-making node that sends the decision result;
  • the address is the address of the network node currently receiving the decision result, which may be the address of the relay node;
  • the sending address is the address of the network node currently sending the decision result, which may be the address of the relay node.
  • the extended address can store other addresses, such as the destination address, for the address of the network node that finally receives the decision result.
  • the unified scheduling management frame format may further include a sequence control field, a service quality control field, a frame body field, and a checksum field.
  • the sequence control field is used to store the serial number identifier of the frame
  • the quality of service field is used to store the service category of the frame and the priority of the business
  • the checksum field is used to store the checksum value, which can be used for correctness detection.
  • Using the above unified scheduling management frame format can ensure that the decision result is sent to the preset network node, and the network node can also obtain the decision result according to the received protocol frame.
  • the process of obtaining the decision result through the reinforcement learning model is the same as the process of predicting the application scenario and obtaining the decision result by the network node in the above embodiment, here No longer.
  • the decision result is predicted by the network autonomous decision node, and the decision result is sent to the network node, which can facilitate the management of the network node.
  • the method can also be applied to cloud scheduling components and network autonomous decision-making nodes.
  • the method includes:
  • S501 Obtain data information of a network node through a cloud scheduling component and a network autonomous decision node respectively; wherein the data information obtained by the cloud scheduling component is first data information, and the data information obtained by the network autonomous decision node is second data information; wherein, the first data information is long-period data information; the second data information is short-period data information;
  • S505. Determine a third decision result according to the first decision result and the second decision result by the network autonomous decision node, and send the third decision result to the network node through a unified scheduling management protocol frame, so that the network node The multi-domain combination in the third decision result is activated at the preset time.
  • the cloud scheduling component may include one active cloud scheduling component and multiple backup cloud scheduling components.
  • the network autonomous decision-making nodes in an area it also includes multiple network autonomous decision-making nodes.
  • the main cloud scheduling component or network autonomous decision-making node is abnormal, it can be replaced by the backup cloud scheduling component or other available network autonomous decision-making nodes.
  • the scheduling request information, resource reservation indication information, network state information and wireless channel information acquired by the network node may be short-period data information or long-period data information.
  • the short-period data information can be millisecond-level data information, which has strong real-time performance.
  • the acquired scheduling request information, resource reservation indication information, network status information and wireless channel information are information that changes every 10 milliseconds; long-period data information may be minute-level or longer data information.
  • the acquired scheduling request information, resource reservation indication information, network state information and wireless channel information are information that changes every minute or even longer.
  • the specific cycle durations of the long cycle and the short cycle may be set according to actual needs, as long as the cycle duration corresponding to the long cycle is greater than the cycle duration corresponding to the short cycle.
  • the cloud scheduling component and the network autonomous decision node can separately process the acquired data information. After processing, the cloud scheduling component can obtain the first decision result, and the network autonomous decision node can obtain the second decision result. The cloud scheduling component can send the first decision result to the network autonomous decision node, and the network autonomous decision node generates the third decision result and sends it to the network node.
  • the first decision result generated by the cloud scheduling component is: activate time domain resources and frequency domain resources in the third scheduling cycle in the future
  • the second decision result generated by the network autonomous decision node is: in the next scheduling cycle
  • the network autonomous decision-making node will send the indication information to the network node: activate the time domain resources, frequency domain resources and airspace resources in the next scheduling cycle, and in the third scheduling cycle Activate time domain resources and frequency domain resources.
  • the network node After receiving the indication information, the network node will activate the preset wireless network resources in the preset scheduling period.
  • the above method can process the information in the long cycle time and the information in the short cycle time respectively through two-level scheduling, that is, the cloud scheduling component and the network autonomous decision-making node, and can ensure that the network node can successfully activate the preset time at the preset time.
  • Wireless network resources that is, the cloud scheduling component and the network autonomous decision-making node
  • FIG. 6 is a schematic structural diagram of a protocol stack provided by an embodiment of the present application, where the protocol stack is applied in a network autonomous decision node and a cloud scheduling component.
  • both the cloud scheduling component and the network autonomous decision-making node include a unified scheduling management module and at least one protocol conversion module; the data information of the network node is obtained through the cloud scheduling component and the network autonomous decision-making node, including:
  • the cloud scheduling component or the network autonomous decision node switches the network type to be consistent with the network node through the protocol conversion module the type of network;
  • predicting the first application scenario of the network node at the preset time by the cloud scheduling component includes:
  • predicting the second application scenario of the network node at the preset time through the network autonomous decision node includes:
  • the second decision result is determined according to the second application scenario by the unified scheduling management module in the autonomous decision-making node of the network.
  • both the cloud scheduling component and the network autonomous decision-making node include a unified scheduling management module and at least one protocol conversion module, and the protocol conversion module can implement wireless network switching between the cloud scheduling component and the network autonomous decision-making node, for example :
  • the network autonomous decision node is receiving the information sent by the network node
  • the wireless network adopted by the network node is a time division multiple access wireless network
  • the wireless network adopted by the network autonomous decision node is an orthogonal frequency division multiple access wireless network
  • the network autonomous The decision-making node can convert the wireless network adopted by itself into a time-division multiple access wireless network through the protocol conversion module, so as to realize the information transmission with the network node.
  • the unified scheduling management module is used to determine the decision result according to the obtained application scenario. And send the determined decision result to the network node.
  • the protocol stack includes a unified scheduling management module 601 and several protocol conversion modules, such as multiple protocol conversion modules 602 .
  • the unified scheduling management module 601 is respectively connected with the service access points of each protocol conversion module 602 through service access points, wherein the service access points are logical interfaces, which are interfaces for communication between upper and lower layers.
  • the protocol stack also includes a data link layer L2 module 603 and a physical layer L1 module 604.
  • the physical layer L1 module is mainly used for operations such as encoding and decoding data, so that the transmitted data is not free in the channel. Error transmission; the data link layer L2 module is mainly used to realize the control of accessing network resources to the air interface of the corresponding network node.
  • the protocol conversion module 602 can perform data conversion and transmission with several data link layer L2 modules 603 through the L2 service access point, and can perform data conversion with several physical layer L1 modules 604 through the L1 service access point. Conversion and transmission, for example, the protocol conversion module 602 can be connected to the service access points of the data link layer L21 module 603 and the data link layer L22 module 603 respectively through the L2 service access point to realize data conversion and transmission.
  • the L1 service access point can be connected to the service access points of the physical layer L11 module 604 and the physical layer L12 module 604 respectively to realize data conversion and transmission.
  • the data transmitted by the physical layer L1 module 603 and the data link layer L2 module 604 to the protocol conversion module 602 can ensure that the protocol conversion module 602 switches the network to a preset wireless communication module.
  • the decision result can be determined by setting the unified scheduling management module.
  • the protocol conversion module can switch the network type to ensure that the information sent by the network node can be received. , and generate decision results based on the received information.
  • the cloud scheduling component and the network autonomous decision node can also store the acquired data information.
  • both the cloud scheduling component and the network autonomous decision node include a data storage module, and the method further includes:
  • the cloud scheduling component and the network autonomous decision node store the communication capability information of the network node in the blockchain
  • the network state information, wireless channel information, scheduling request information and resource pre-emption indication information reported by the network node are stored in chronological order through the data storage module.
  • the communication capability information of the network node is data information with low frequency changes, wherein the low frequency changes indicate that the data information changes slowly.
  • the communication capability information of the network node refers to all the communication capability information supported by the network node.
  • the blockchain is developed based on the hyperledger platform, and is used to store correct network node information and network node communication capability information, and to solve the problem of possible destruction or tampering of data information through a Byzantine fault-tolerant algorithm.
  • the stored information has strong anti-tampering modification, so as to ensure that the decision result obtained according to the communication capability information of the network node is suitable for the current network state information of the network node. , wireless channel information, network scheduling request information and the decision result under the resource reservation indication information.
  • the cloud scheduling component and the network autonomous decision node will store the received real-time changing data information in the data storage module, such as a time series database, in the order of time.
  • the time series database can store the acquired real-time changing data as an independent table, realize the continuous storage of data, and reduce random read operations.
  • the data storage module stores the network status information, wireless channel information, scheduling request information and resource reservation indication information, it can be stored in sequence according to the time of receiving data, thereby realizing the earlier network status information, wireless Channel information, scheduling request information and resource reservation indication information are processed to prevent network nodes from being unable to activate preset wireless network resources for a long time.
  • FIG. 7 is a schematic structural diagram of a wireless intelligent decision-making communication device provided by an embodiment of the present application, and the device includes:
  • an acquisition module 701, configured to acquire data information of each network node; the data information includes scheduling request information, resource reservation indication information, network status information, wireless channel information, and communication capability information of the network node;
  • a prediction module 702 for a network node, configured to predict the scene where the network node is located at a preset time according to the data information;
  • An activation module 703, configured to determine a decision result including a multi-domain combination according to the application scenario of the network node at a preset time and the communication capability information of the network node, and activate the multi-domain of the network node at the preset time combination; wherein, the multi-domain combination is in a corresponding relationship with the application scenario of the network node.
  • the wireless intelligent decision-making communication device provided in the embodiment of the present application can implement the wireless intelligent decision-making communication method of the embodiment shown in FIG.
  • the embodiment of the present application also provides a wireless intelligent decision-making communication system, the system includes: a network autonomous decision-making node and a network node;
  • the network node is configured to send data information to an autonomous decision-making node of the network;
  • the data information includes scheduling request information, resource reservation indication information, network state information, wireless channel information and communication capability information of the network node;
  • the network autonomous decision node is used to obtain data information of the network node, predict the application scenario of the network node at a preset time, and determine the decision result according to the application scenario and the communication capability information of the network node;
  • the network autonomous decision node is further configured to send the decision result to the network node through a unified scheduling management protocol frame, so that the network node activates the multi-domain combination in the decision result at the preset time.
  • the wireless intelligent decision-making communication system provided by the embodiment of the present application can implement the wireless intelligent decision-making communication method of the embodiment shown in FIG.
  • FIG. 8 is a wireless intelligent decision-making communication system provided by an embodiment of the present application. As shown in FIG. 8 , the system includes: a network autonomous decision-making node 801 , a cloud scheduling component 802 , and a network node 803 .
  • the cloud scheduling component 802 is configured to determine a first application scenario according to the acquired first data information, and determine a first decision result according to the first application scenario;
  • the network autonomous decision node 801 is configured to determine a second application scenario according to the acquired second data information, and determine a second decision result according to the second application scenario;
  • the network autonomous decision node 801 determines a third decision result according to the first decision result and the second decision result, and sends the third decision result to the network node through a unified scheduling management protocol frame, so that the network node Activate the multi-domain combination mode in the third decision result at the preset time;
  • the first data information is long-period data information
  • the second data information is short-period data information
  • the wireless intelligent decision-making communication system provided by the embodiment of the present application can implement the wireless intelligent decision-making communication method of the embodiment shown in FIG.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

本申请实施例提供一种无线智能决策通信方法、装置及***，所述方法包括：获取网络节点的数据信息；所述数据信息包括调度请求信息、资源预占指示信息、网络状态信息、无线信道信息和网络节点的通信能力信息；根据所述数据信息预测所述网络节点在预设时间的应用场景；根据所述网络节点在预设时间的应用场景，确定包含多域组合的决策结果，并在所述预设时间激活所述网络节点的多域组合；其中，所述多域组合与所述网络节点的应用场景为对应关系，上述方法通过对网络节点所处的场景进行预测，并根据网络节点所处场景确定所要激活的多域组合，使得所述网络节点能够适应应用场景的变化，最大化网络容量和吞吐量，满足用户的需求。

Description

无线智能决策通信方法、装置和***
本公开要求于2021年01月22日提交中国专利局、申请号为202110084903.2、申请名称为“无线智能决策通信方法、装置和***”的中国专利申请的优先权，其全部内容通过引用结合在本公开中。
技术领域
本申请涉及无线通信技术领域，尤其涉及一种无线智能决策通信方法、装置和***。
背景技术
随着通信技术的发展，对无线电磁波信号和信息进行预设的变换处理和调度的方法逐渐增多，形成了多种无线通信协议。
现有技术中，各种无线通信网络资源各自适用于某一特定的应用场景，例如，基于正交频分复用技术的3GPP LTE协议适用于陆地低速移动通信，基于时分复用技术的齐转发自组网协议适用于广播组播业务占比高的场景。然而，当网络节点移动或网络状态及环境发生变化时，原有的单一网络协议将会呈现出劣势。因此，需要提出一种无线网络资源的决策通信方法，以解决当网络节点的应用场景发生变化时无法对资源进行智能切换，导致网络容量和吞吐量较小，无法满足用户需求的问题。
发明内容
本申请提供一种无线智能决策通信方法、装置和***，以解决当网络节点的应用场景发生变化时无法对资源进行智能切换的问题，解决网络容量和吞吐量小，无法满足用户需求的问题。
第一方面，本申请实施例提供一种无线智能决策通信方法，所述方法包括：
获取网络节点的数据信息；所述数据信息包括调度请求信息、资源预占指示信息、网络状态信息、无线信道信息和网络节点的通信能力信息；
根据所述数据信息预测所述网络节点在预设时间的应用场景；
根据所述网络节点在预设时间的应用场景和网络节点的通信能力信息，确定包含多域组合的决策结果，并在所述预设时间激活所述网络节点的多域组合；其中，所述多域组合与所述网络节点的应用场景为对应关系。
可选的，根据所述数据信息预测所述网络节点在预设时间的应用场景，包括：
根据所述数据信息预测所述网络节点的确定性变化信息；
根据所述确定性变化信息确定网络节点的应用场景。
可选的，根据所述网络节点在预设时间的应用场景和网络节点的通信能力信息，确定包含多域组合的决策结果，包括：
将所述网络节点在预设时间的应用场景和网络节点的通信能力信息输入强化学习模型，将所述强化学习模型的输出结果确定为决策结果。
可选的，根据所述网络节点在预设时间的应用场景和网络节点的通信能力信息，确定包含多域组合的决策结果，包括：
当所述网络节点处于静止的应用场景，且所述网络节点的通信能力信息包含时域资源和频域资源时，所述多域组合包括所述网络节点的时域资源和频域资源；或者，
当所述网络节点处于移动的应用场景，且所述网络节点的通信能力信息包含时域资源、频域资源和空域资源时，所述决策结果为激活所述网络节点的时域资源、频域资源和空域资源；或者，
当所述网络节点处于存在干扰的应用场景，且所述网络节点的通信能力信息包含时域资源、频域资源、空域资源和码域资源时，所述决策结果为激活所述网络节点的时域资源、频域资源、空域资源和码域资源；或者，
当所述网络节点处于多障碍物的应用场景，且所述网络节点的通信能力信息包含时域资源、频域资源时、空域资源、码域资源和时延多普勒域资源时，所述决策结果为激活所述网络节点的时域资源、频域资源、空域资源、码域资源和时延多普勒域资源。
可选的，所述方法应用于网络节点，所述网络节点通过卷积长短期记忆混合神经网络模型预测网络节点所处的应用场景；所述网络节点通过强化学习模型确定所要激活的多域组合。
可选的，获取网络节点的数据信息，包括：
通过网络自主决策节点获取网络节点的数据信息；
相应的，根据所述数据信息预测所述网络节点在预设时间的应用场景，包括：
通过网络自主决策节点预测所述网络节点在预设时间的应用场景；
相应的，根据所述网络节点在预设时间的应用场景，确定包含多域组合的决策结果，并在所述预设时间激活所述网络节点的多域组合，包括：
通过所述网络自主决策节点根据应用场景和网络节点的通信能力信息，确定决策结果；并通过统一调度管理协议帧将决策结果发送给所述网络节点，以使所述网络节点在所述预设时间激活决策结果中的多域组合。
可选的，获取网络节点的数据信息，包括：
通过云调度组件和网络自主决策节点分别获取网络节点的数据信息；其中，所述云调度组件获取的数据信息为第一数据信息，所述网络自主决策节点获取的数据信息为第二数据信息；其中，所述第一数据信息为长周期的数据信息；所述第二数据信息为短周期的数据信息；
相应的，根据所述数据信息预测所述网络节点在预设时间的应用场景，包括：
通过云调度组件预测所述网络节点在预设时间的第一应用场景；通过网络自主决策节点预测所述网络节点在预设时间的第二应用场景；
相应的，根据所述网络节点在预设时间的应用场景，确定包含多域组合的决策结果，并在所述预设时间激活所述网络节点的多域组合，包括：
通过所述云调度组件根据第一应用场景和网络节点的通信能力信息，确定第一决策结果；
通过所述网络自主决策节点根据获取的第二应用场景和和网络节点的通信能力信息，确定第二决策结果；
通过所述网络自主决策节点根据第一决策结果和第二决策结果确定第三决策结果，并通过统一调度管理协议帧将第三决策结果发送给所述网络节点，以使所述网络节点在所述预设时间激活第三决策结果中的多域组合。
可选的，所述云调度组件和网络自主决策节点均包括统一调度管理模块和至少一个协议转换模块；通过云调度组件和网络自主决策节点分别获 取网络节点的数据信息，包括：
当云调度组件或网络自主决策节点的网络类型与传输数据的网络节点采用的网络类型不一致时，云调度组件或网络自主决策节点通过所述协议转换模块将网络类型切换至与所述网络节点一致的网络类型；
相应的，通过云调度组件预测所述网络节点在预设时间的第一应用场景，包括：
通过云调度组件中的统一调度管理模块根据第一应用场景确定第一决策结果；
相应的，通过网络自主决策节点预测所述网络节点在预设时间的第二应用场景，包括：
通过网络自主决策节点中的统一调度管理模块根据第二应用场景确定第二决策结果。
可选的，所述云调度组件和网络自主决策节点均包括数据存储模块，所述方法还包括：
所述云调度组件和网络自主决策节点将所述网络节点的通信能力信息存储在区块链中；
通过数据存储模块将网络节点上报的所述网络状态信息、无线信道信息、调度请求信息调度请求信息和资源预占指示信息按照时间顺序存储。
第二方面，本申请实施例提供一种用于实现第一方面所述方法的无线智能决策通信装置，所述装置包括：
获取模块，用于获取各个网络节点的数据信息；所述数据信息包括调度请求信息、资源预占指示信息、网络状态信息、无线信道信息和网络节点的通信能力信息；
预测模块，针对一个网络节点，用于根据所述数据信息预测所述网络节点在预设时间所处的场景；
激活模块，用于根据所述网络节点在预设时间的应用场景和网络节点的通信能力信息，确定包含多域组合的决策结果，并在所述预设时间激活所述网络节点的多域组合；其中，所述多域组合与所述网络节点的应用场景为对应关系。
第三方面，本申请实施例提供一种用于实现第一方面所述方法的无线 智能决策通信***，所述***包括：网络自主决策节点和网络节点；
所述网络节点用于向网络自主决策节点发送数据信息；所述数据信息包括调度请求信息、资源预占指示信息、网络状态信息、无线信道信息和网络节点的通信能力信息；
所述网络自主决策节点用于获取网络节点的数据信息，预测所述网络节点在预设时间的应用场景，并根据应用场景和网络节点的通信能力信息，确定决策结果；
所述网络自主决策节点还用于通过统一调度管理协议帧将决策结果发送给所述网络节点，以使所述网络节点在所述预设时间激活决策结果中的多域组合。
第四方面，本申请实施例提供一种用于实现第一方面所述方法的无线智能决策通信***，所述***包括：网络自主决策节点、云调度组件和网络节点；
所述云调度组件用于根据获取的第一数据信息确定第一应用场景，根据第一应用场景确定第一决策结果；
所述网络自主决策节点用于根据获取的第二数据信息确定第二应用场景，根据第二应用场景确定第二决策结果；
所述网络自主决策节点根据所述第一决策结果和第二决策结果确定第三决策结果，并通过统一调度管理协议帧将第三决策结果发送给所述网络节点，以使所述网络节点在所述预设时间激活第三决策结果中的多域组合方式；
其中，所述第一数据信息为长周期的数据信息；所述第二数据信息为短周期的数据信息。
本申请实施例提供的无线智能决策通信方法、装置和***，该方法包括：获取网络节点的数据信息；所述数据信息包括调度请求信息、资源预占指示信息、网络状态信息、无线信道信息和网络节点的通信能力信息；根据所述数据信息预测所述网络节点在预设时间的应用场景；根据所述网络节点在预设时间的应用场景，确定包含多域组合的决策结果，并在所述预设时间激活所述网络节点的多域组合；其中，所述多域组合与所述网络节点的应用场景为对应关系，上述方法通过对网络节点所处的场景进行预 测，并根据网络节点所处场景确定所要激活的多域组合，使得所述网络节点能够适应应用场景的变化，最大化网络容量和吞吐量，满足用户的需求。
附图说明
为了更清楚地说明本申请实施例或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本申请的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动性的前提下，还可以根据这些附图获得其他的附图。
图1为本申请实施例提供的应用场景示意图；
图2为本申请实施例提供的一种无线智能决策通信方法的流程示意图；
图3为本申请实施例提供的另一种无线智能决策通信方法的流程示意图；
图4(a)为本申请实施例提供的统一调度管理协议帧的结构示意图；
图4(b)为本申请实施例提供的帧控制字段的结构示意图；
图4(c)为本申请实施例提供的另一种统一调度管理协议帧的结构示意图；
图5为本申请实施例提供的协议栈结构示意图；
图6为本申请实施例提供的又一种无线智能决策通信方法的流程示意图；
图7为本申请实施例提供的无线智能决策通信装置的结构示意图；
图8为本申请实施例提供的无线智能决策通信***。
具体实施方式
下面将结合本申请实施例中的附图，对本申请实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本申请一部分实施例，而不是全部的实施例。基于本申请中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都属于本申请保护的范围。
本申请的说明书和权利要求书及上述附图中的术语“第一”、“第二”、 “第三”“第四”等(如果存在)是用于区别类似的对象，而不必用于描述特定的顺序或先后次序。应该理解这样使用的数据在适当情况下可以互换，以便这里描述的本申请的实施例例如能够以除了在这里图示或描述的那些以外的顺序实施。此外，术语“包括”和“具有”以及他们的任何变形，意图在于覆盖不排他的包含，例如，包含了一系列步骤或单元的过程、方法、***、产品或设备不必限于清楚地列出的那些步骤或单元，而是可包括没有清楚地列出的或对于这些过程、方法、产品或设备固有的其它步骤或单元。
图1为本申请实施例提供的应用场景示意图。如图1所示，所述方法可以应用在网络节点中，也可以应用在网络自主决策节点中，或者还可以应用在云调度组件和网络自主决策节点中。其中，网络节点可以为基站，接入点，用户终端，自组网节点，网关等。通过获取网络节点的数据信息可以预测网络节点在未来某一时刻的应用场景，并根据应用场景对网络的五域的资源进行自主切换管理，其中，此处的五域指的是时域、频域、空域、码域及时延多普勒域。对于一个网路节点来说，在获取数据信息后，可以根据数据信息来预测网络节点的应用场景，并根据应用场景来激活相应的多域组合。
在一些技术中，当网络节点的网络状态及环境发生改变时，无法及时对网络资源进行智能切换，还会保持原有的单一网络协议进行通信，导致网络容量和吞吐量较小，无法满足用户的使用需求。
基于上述问题，本申请实施例能够根据获取的数据信息来自动感知应用场景，进而根据应用场景来确定所要激活的多域组合，并在预设时间激活多域组合，修改后的多域组合能够满足网络节点的通信需求，使得网络容量和吞吐量增大，进而满足用户的使用需求。
下面以具体地实施例对本申请的技术方案进行详细说明。下面这几个具体的实施例可以相互结合，对于相同或相似的概念或过程可能在某些实施例不再赘述。
图2为本申请实施例提供的一种无线智能决策通信方法的流程示意图，如图2所示，所述方法包括：
S201、获取网络节点的数据信息；所述数据信息包括调度请求信息、资 源预占指示信息、网络状态信息、无线信道信息和网络节点的通信能力信息。
在本实施例中，获取数据信息的主体可以为网络节点，还可以为网络自主决策节点，还可以是网络自主决策节点和云调度组件。当获取数据信息的主体为网络节点时，形成的是无中心的自组网的通信方法，在该方法中每个网络节点都是网络自主决策节点，通过各个网络自主决策节点之间的协同机制确定决策结果。当获取数据信息的主体为网络自主决策节点时，形成的是分布式网络的通信方法，网络自主决策节点可以获取区域内所有网络节点的数据信息并确定决策结果。当获取数据信息的主体为网络自主决策节点和云调度组件时，形成的是集中式网络的通信方法，通过云调度组件和网络自主决策节点实现二级调度，最终确定决策结果。
其中，获取的网络节点的数据信息为多维信息。数据信息可以获取包括调度请求信息、资源预占指示信息、网络状态信息和无线信道信息，获取的网络状态信息包括但不限于网络节点数量信息、空口资源的占用情况、不同类别业务所占的比例、通信速率、丢包率、与其他网络节点之间的距离以及信号强度等。获取的无线信道信息包括节点之间通信时的发送端至接收端的传输通道。获取的调度请求信息可以为请求发送数据量和接收缓存大小等。资源预占指示信息可以为资源预留占用的指示信息。
其中，获取网络节点的数据信息可以是通过向其他节点发送请求数据的方式来获取数据信息，还可以是按协议约定接收其他节点主动发送的数据信息。
其中，网络节点的通信能力信息与网络节点自身的硬件加速器和射频资源有关，不同的网络节点具有不同的通信能力信息。只有当网络节点支持某一域资源时，才可以激活对应的资源。
下面以获取空口资源为例进行说明，网络节点可以获取连续的帧数据，若帧数据包含100个时隙，其中，前10个时隙表示控制调度信息，任一节点可以获取该帧数据，在获取该帧数据后，可以对该帧数据进行解析，得到前10个时隙的数据，通过该前10个时隙的数据可以获取时频资源的占用情况，例如，当通过帧数据获知时频资源A被占用时，则该节点不能通过时频资源A进行数据传输。
此外，还可以获取自身的通信能力信息，其中，通信能力信息包括但 不限于网络节点标识、网络节点类型、是否支持中继、是否可以移动以及移动范围、支持的协议类型、支持的频段、MIMO(Multi Input Multi Output，多输入多输出)能力、多载波能力以及载波聚合能力等。此处的通信能力信息为该节点所支持的全部通信能力的信息。网络节点类型可以为基站、接入点、用户终端、网关和中继节点等。
在获取到上述多维数据信息后，还可以将上述数据信息进行保存。
S202、根据所述数据信息预测所述网络节点在预设时间的应用场景。
在本实施例中，可以根据获取的数据信息来预测在预设时间的应用场景。具体的，可以根据获取的数据信息来进行预测，例如，当检测到某些区域未来会出现干扰时，则可以确定网络节点处于干扰的应用场景。具体的，通过获取的调度请求信息、资源预占指示信息、网络状态信息和无线信道信息来确定网路节点的应用场景。
S203、根据所述网络节点在预设时间的应用场景和网络节点的通信能力信息，确定包含多域组合的决策结果，并在所述预设时间激活所述网络节点的多域组合；其中，所述多域组合与所述网络节点的应用场景为对应关系。
在本实施例中，在确定网络节点在预设时间的应用场景和网络节点的通信能力信息后，可以根据网络节点的应用场景确定要激活的多域组合，使得激活的多域组合能够适应网络节点在预设时间的应用场景。例如，当网络节点在预设时间的应用场景为场景A，而在场景A下无需同时激活五域资源，此时可以仅激活与该场景对应的多域组合的资源，例如当需要采用正交频分多址的无线通信协议时，可以仅激活对应的时域资源和频域资源，从而实现当网络节点的应用场景变化时，能够最大化网络容量和吞吐量，同时可以降低功耗。
其中，时域资源、频域资源、空域资源、码域资源及时延多普勒域资源分别会对应不同的软件及硬件加速器，激活多域组合的资源是指将与多域组合对应的软件及硬件加速器处于工作状态，即将对应的软件处于加载运行状态，将硬件加速器处于上电状态。其中，五域资源对应的程序本质为不同的数学变换方法，处于运行状态的程序可以对待传输的数据进行相应的变换处理，以实现通过对应的多域组合方式进行数据传输。其中，硬件加速器并非为必要器件，通过采用硬件加速器可以提高数据处理的性能。
在本实施例中，网络节点或云调度组件或网络自主决策节点，可以获取网络节点的数据信息，并根据获取的数据信息预测网络节点在预设时间的应用场景，并根据应用场景去激活网络节点的多域组合，可以对网络的时域资源、频域资源、空域资源、码域资源和时延多普勒域资源进行自主切换管理，以适应应用场景的变化，能够最大化网络容量和吞吐量。
可选的，根据所述数据信息预测所述网络节点在预设时间的应用场景，包括：
根据所述数据信息预测所述网络节点的确定性变化信息；根据所述确定性变化信息确定网络节点的应用场景。
在本实施例中，当根据数据信息确定网络节点的应用场景时，可以先根据数据信息确定网络节点的确定性变化信息，具体的，可以采用卷积长短期记忆混合神经网络来进行预测。将局部或全局区域的网络节点的调度请求信息、资源预占指示信息、网络状态信息和无线信道信息作为输入数据，卷积长短期记忆混合神经网络对输入的数据进行处理，输出确定性变化的周期、循环规律、趋势和随机变化的模型等等。
例如，当一个网络节点采用的网络为单载波网络时，根据网络状态信息和无线信道信息可以得到用户终端数量和网络节点数量，根据调度请求信息和资源预占指示信息可以得到空口资源占用情况，通过神经网络模型可以预测到用户终端数量和网络节点数量在未来时间是否会超过预设数值，以及网络节点空口资源的占用率是否会超过预设数值，预测的信息即为确定性变化信息。
神经网络采用的是卷积长短期记忆混合神经网络，其中，卷积长短期记忆混合神经网络是由卷积神经网络的卷积层和池化层以及长短期记忆神经网络的输入层相结合，通过卷积层和池化层提取局部特征，再通过长短期记忆神经网络获取局部特征中与时间序列相关的特征，可以实现对短期或长期的数据进行记忆和预测。
其中，在使用卷积长短期记忆混合神经网络来预测网络节点的确定性变化信息之前，需要通过历史数据对卷积长短期记忆混合神经网络进行训练，其中，历史数据为历史调度请求信息、历史资源预占指示信息、历史网络状态信息、历史无线信道信息以及在该历史调度请求信息、历史资源 预占指示信息、历史网络状态信息和历史无线信道信息下的历史网络变化信息，通过将上述历史数据输入至预设网络模型，经过多次训练得到确定的神经网络模型。其中，在进行训练时需要输入大量的历史数据。
通过将两种神经网络进行结合可以同时利用卷积神经网络和长短期记忆网络的优点，得到准确度高的预测结果。
在获取确定性变化信息后，可以根据确定性变化信息得到网络节点的应用场景，网络节点的应用场景可以为静止应用场景、移动应用场景等等。可以根据网络节点的确定性变化信息中的特征确定网络节点的应用场景。例如，当确定性变化信息为在未来一小时的时刻，出现固定干扰或出现随机干扰，则可以确定网络节点的应用场景为存在干扰的应用场景。
通过上述方法可以根据网络节点的数据信息确定网络节点的应用场景，先对网络节点的确定性变化信息进行预测，再对应用场景进行预测，可以使得确定的应用场景更加准确。采用卷积长短期记忆混合神经网络可以很好的处理存在时序关系的数据信息。
可选的，根据所述网络节点在预设时间的应用场景和网络节点的通信能力信息，确定包含多域组合的决策结果，包括：
将所述网络节点在预设时间的应用场景和网络节点的通信能力信息输入强化学习模型，将所述强化学习模型的输出结果确定为决策结果。
其中，强化学习的输入包括应用场景和网络节点的通信能力信息，还包括调度策略，通过强化学习训练可以获得一个Q表，每行表示输入的应用场景和网络节点的通信能力信息，每列表示输入的决策结果，Q表的每个单元格的值表示在执行相应决策结果的奖励期望值。其中，在达到预设训练次数时可以停止训练，得到训练后的Q表。根据强化学习模型可以获得在执行各个决策结果时对应的奖励期望值，在确定决策结果时选取对应奖励期望值最大的决策结果。其中，决策结果为激活五域资源中的多域组合。
上述方法通过强化学习模型来对决策结果进行预测，能够准确确定与当前网络节点的应用场景对应的决策结果。
可选的，根据所述网络节点在预设时间的应用场景和网络节点的通信能力信息，确定包含多域组合的决策结果，包括：
当所述网络节点处于静止的应用场景，且所述网络节点的通信能力信息 包含时域资源和频域资源时，所述多域组合包括所述网络节点的时域资源和频域资源；或者，
当所述网络节点处于移动的应用场景，且所述网络节点的通信能力信息包含时域资源、频域资源和空域资源时，所述决策结果为激活所述网络节点的时域资源、频域资源和空域资源；或者，
当所述网络节点处于存在干扰的应用场景，且所述网络节点的通信能力信息包含时域资源、频域资源、空域资源和码域资源时，所述决策结果为激活所述网络节点的时域资源、频域资源、空域资源和码域资源；或者，
当所述网络节点处于多障碍物的应用场景，且所述网络节点的通信能力信息包含时域资源、频域资源时、空域资源、码域资源和时延多普勒域资源时，所述决策结果为激活所述网络节点的时域资源、频域资源、空域资源、码域资源和时延多普勒域资源。
在本实施例中，网络节点处于静止的应用场景指的是网络节点处于静止或相对静止的应用场景下，在该应用场景下，若网络节点的通信能力信息包含时域资源和频域资源时，则仅激活时域资源和频域资源的组合即可，而无需再激活空域资源、码域资源和时延多普勒域资源。激活时域资源是指将数据承载在不同的时间段以进行数据传输。激活频域资源是指将数据承载在不同的频率段以进行数据传输。激活时域资源和频域资源则是将数据承载在某一时间及频率段来进行数据传输。
例如，当网络节点为静止状态时，若网络节点采用的网络为单载波网络，且网络容量发生变化时，即单载波网络下的用户终端数量和网络节点数量超过设定数值时，可以通过新增载波或载波聚合的方法进行扩容。或者，若当前网络为时分多址网络，一些网络节点下的用户终端数量和网络节点数量超过设定数值时，可以通过切换网络为正交频分多址的网络，以实现容纳更多的用户终端和网络节点。因此，当网络节点处于静止的应用场景时，通过激活时域资源和频域资源，从而在时域资源和频域资源中进行无线通信网络的切换。
在本实施例中，网络节点处于移动的应用场景指的是网络节点处于移动状态，或者，多个网络节点的空间跨度范围较大，或者，各个网络节点之间的电磁环境不同等场景，在该应用场景下，若网络节点的通信能力信息包 含时域资源、频域资源和空域资源时，则仅激活时域资源、频域资源和空域资源的组合即可，而无需再激活码域资源和时延多普勒域资源。激活空域资源是指将数据承载在不同的天线端口以进行数据传输。
例如，当网络节点为移动状态时，若网络节点使用宽带多天线空分复用方式收发数据，且预测到网络节点之间的距离在未来会超过设定数值，可以通过激活时域资源、频域资源和空域资源，实现在时域资源、频域资源和空域资源中进行无线通信网络的切换，具体的，可以切换网络为窄带单发多收通信协议，能够提升功率谱密度，降低接收信噪比要求。
在本实施例中，网络节点处于存在干扰的应用场景指的是网络节点处于干扰，或者复杂电磁环境，或者对传输误码率要求较高的场景，在该应用场景下，若网络节点的通信能力信息包含时域资源、频域资源、空域资源和码域资源，则可以仅激活时域资源、频域资源、空域资源和码域资源的组合即可，而无需再激活时延多普勒域资源。激活码域资源是指对数据进行不同的编码以进行数据传输。
例如，当网络节点处于存在干扰的应用场景时，可以激活时域资源、频域资源、空域资源和码域资源的组合，从而能够实现将无线通信网络切换为预设无线通信网络，如码分多址通信协议，码分多址通信协议具有抗干扰的特性，可以调节频段为无干扰频段或干扰小的频段，解决由于干扰而出现的丢包或无法通信的情况。
在本实施例中，网络节点处于多障碍物的应用场景指的是网络节点处于多障碍物场景，或者复杂绕射环境，或者超音速移动场景，在该应用场景下，若网络节点的通信能力信息包含时域资源、频域资源、空域资源、码域资源和时延多普勒域资源，则可以激活上述五种网络资源，从而能够实现将无线通信网络切换为预设通信网络，如正交时频空通信协议，正交时频空通信协议能够将发送或接收的信息承载在时延多普勒域，避免在超音速移动及存在障碍物时出现丢包或无法通信的情况。激活时延多普勒域资源是指将数据承载在不同的时延及多普勒频偏以进行数据传输。
此外，所述网络节点所处的应用场景还包括网络节点移动到无信号覆盖的区域，或者中心控制接入点未来会出现故障，在该应用场景下，若网络节点的通信能力信息包含时域资源、频域资源和空域资源，则可以激活时域资 源、频域资源和空域资源，从而能够实现将无线通信网络切换为预设无线通信网络，如卫星中继通信协议，卫星中继通信协议使得产生的无线电波具有较大的覆盖区域和较远的通信距离，避免由于无信号覆盖而出现丢包或无法通信的情况。
网络节点所处的应用场景还包括网络节点间的多播业务占比超过预设值，在该应用场景下，若网络节点的通信能力信息包含时域资源、频域资源和空域资源，则可以激活时域资源、频域资源和空域资源，从而能够实现将无线通信网络切换为预设无线通信网络，如阻继网络齐转发协议，阻继网络齐转发协议使得节点可以实现中继转发，进而提升多播通信速率和可靠性。
可选的，所述方法应用于网络节点，所述网络节点通过卷积长短期记忆混合神经网络模型预测网络节点所处的应用场景；所述网络节点通过强化学习模型确定所要激活的多域组合。
在本实施例中，所述方法可以应用于网络节点，即每个网络节点都是网络自主决策节点，实现无中心的自组网的通信方法。在此种方法中，网络节点通过卷积长短期记忆混合神经网络来预测网络节点所处的应用场景，在确定应用场景后，通过强化学习模型确定所要激活的多域组合。在上述过程中，获取数据信息以及对数据信息进行处理的过程均通过网络节点执行。
采用无中心的自组网的结构可以便于对自身进行控制及管理，具有可靠性高的优点。
如图3所示，所述方法还可以形成分布式网络的通信方法，可选的，所述方法包括：
S301、通过网络自主决策节点获取网络节点的数据信息；
S302、通过网络自主决策节点预测所述网络节点在预设时间的应用场景；
S303、通过所述网络自主决策节点根据应用场景和网络节点的通信能力信息，确定决策结果；并通过统一调度管理协议帧将决策结果发送给所述网络节点，以使所述网络节点在所述预设时间激活决策结果中的多域组合。
在本实施例中，还可以通过网络自主决策节点对无线网络资源进行调度，具体的，网络节点会获取实时变化的数据信息，如网络状态信息、无线 信道信息、调度请求信息和资源预占指示信息，并向网络自主决策节点上报数据信息，网络自主决策节点在获取到数据信息后，可以预测网络节点的应用场景，如通过卷积长短期记忆混合神经网络模型预测应用场景。在得到应用场景后，通过强化学习模型根据应用场景和网络节点的通信能力信息得到决策结果，并将决策结果通过统一调度管理协议帧的形式发送给网络节点。在形成分布式网络的通信方法中，网络自主决策节点用于接收网络节点上报的数据，进行处理后确定决策结果。
图4为本申请实施例提供的统一调度管理协议帧的结构示意图，如图4(a)所示，所述统一调度管理协议帧包括帧控制字段、地址格式字段和帧体字段。
所述帧控制字段，用于存储所述网络节点的当前无线网络类型；所述地址格式字段，用于存储所述网络自主决策节点地址、中继节点地址和目的网络节点地址；所述帧体字段，用于存储所述决策结果。
如图4所示，所述帧控制字段和地址格式字段可以为固定字节数，帧体字段为可变长度的字节数。本实施例中，对于各个字段的字节个数均不作具体的限制。
帧控制字段可以存储网络节点的当前无线网络的相关信息，如图4(b)所示，帧控制字段存储的内容包括：当前网络节点的协议类型、协议版本、地址格式以及预留其他帧控制功能。
地址格式字段可以划分为若干个地址字段，将数据传输时所需的地址信息分别存储至对应的地址字段。帧体字段用于存储决策结果，以使网络节点执行所述决策结果。
如图4(c)所示，地址格式字段可以划分为网络自主决策节点地址、接收地址、发送地址以及扩展地址，其中，网络自主决策节点地址为发送决策结果的网络自主决策节点的地址；接收地址为当前接收决策结果的网络节点的地址，可以为中继节点的地址；发送地址为当前发送该决策结果的网络节点的地址，可以为中继节点的地址。扩展地址可以存储其他地址，比如目的地址，用于最终接收该决策结果的网络节点的地址。
如图4(c)所示，统一调度管理帧格式还可以包括序列控制字段、服务质量控制字段、帧体字段以及校验和字段。其中，序列控制字段用于存储该 条帧的序列号标识；服务质量字段用于存储该帧的服务类别和业务的优先级；校验和字段用于存储校验和数值，可以用于正确性检测。
通过采用上述的统一调度管理帧格式可以保证将决策结果发送至预设网络节点，网络节点也可以根据接收的协议帧获取决策结果。
对于网络自主决策节点通过卷积长短期记忆混合神经网络模型预测应用场景，以及通过强化学习模型得到决策结果的过程和上述实施例中的网络节点预测应用场景及得到决策结果的过程相同，此处不再赘述。
通过网络自主决策节点对决策结果进行预测，并将决策结果发送给网络节点，可以便于对网络节点进行管理。
如图5所示，所述方法还可以应用于云调度组件和网络自主决策节点，可选的，所述方法包括：
S501、通过云调度组件和网络自主决策节点分别获取网络节点的数据信息；其中，所述云调度组件获取的数据信息为第一数据信息，所述网络自主决策节点获取的数据信息为第二数据信息；其中，所述第一数据信息为长周期的数据信息；所述第二数据信息为短周期的数据信息；
S502、通过云调度组件预测所述网络节点在预设时间的第一应用场景；通过网络自主决策节点预测所述网络节点在预设时间的第二应用场景；
S503、通过所述云调度组件根据第一应用场景和网络节点的通信能力信息，确定第一决策结果；
S504、通过所述云调度组件根据获取的第二应用场景和和网络节点的通信能力信息，确定第二决策结果；
S505、通过所述网络自主决策节点根据第一决策结果和第二决策结果确定第三决策结果，并通过统一调度管理协议帧将第三决策结果发送给所述网络节点，以使所述网络节点在所述预设时间激活第三决策结果中的多域组合。
在本实施例中，云调度组件可以包括一个主用云调度组件和多个备用云调度组件。同样的，对于一个区域内的网络自主决策节点来说，也包括多个网络自主决策节点。当主用云调度组件或网络自主决策节点发生异常时，可以通过备用云调度组件或其他可用的网络自主决策节点替补。
其中，网络节点获取的调度请求信息、资源预占指示信息、网络状态信息和无线信道信息可以是短周期的数据信息，也可以是长周期的数据信 息。短周期的数据信息可以是毫秒级的数据信息，具有较强的实时性。例如，获取的调度请求信息、资源预占指示信息、网络状态信息和无线信道信息为每10毫秒改变一次的信息；长周期的数据信息可以是分钟级或更长时间的数据信息。例如，获取的调度请求信息、资源预占指示信息、网络状态信息和无线信道信息为每1分钟甚至更长时间改变一次的信息。所述长周期和所述短周期的具体周期时长可以根据实际需要来设置，只要保证长周期对应的周期时长大于短周期对应的周期时长即可。
云调度组件和网络自主决策节点可以分别对获取的数据信息进行处理，经过处理后，云调度组件可以得到第一决策结果，网络自主决策节点可以得到第二决策结果。云调度组件可以将第一决策结果发送至网络自主决策节点，网络自主决策节点生成第三决策结果并发送至网络节点。例如，针对同一个网络节点，云调度组件产生的第一决策结果为：在未来第三个调度周期激活时域资源和频域资源，网络自主决策节点产生的第二决策结果为：在下个调度周期激活时域资源、频域资源及空域资源，则网络自主决策节点会向网络节点发送指示信息为：在下个调度周期激活时域资源、频域资源及空域资源，并在第三个调度周期激活时域资源和频域资源。网络节点在接收到指示信息后，会在预设调度周期激活预设的无线网络资源。
上述方法通过两级的调度，即云调度组件和网络自主决策节点，可以分别对长周期时间内的信息和短周期时间内的信息进行处理，可以保证网络节点在预设时间顺利激活预设的无线网络资源。
图6为本申请实施例提供的协议栈结构示意图，该协议栈应用在网络自主决策节点和云调度组件中。
可选的，所述云调度组件和网络自主决策节点均包括统一调度管理模块和至少一个协议转换模块；通过云调度组件和网络自主决策节点分别获取网络节点的数据信息，包括：
当云调度组件或网络自主决策节点的网络类型与传输数据的网络节点采用的网络类型不一致时，云调度组件或网络自主决策节点通过所述协议转换模块将网络类型切换至与所述网络节点一致的网络类型；
相应的，通过云调度组件预测所述网络节点在预设时间的第一应用场景，包括：
通过云调度组件中的统一调度管理模块根据第一应用场景确定第一决策结果；
相应的，通过网络自主决策节点预测所述网络节点在预设时间的第二应用场景，包括：
通过网络自主决策节点中的统一调度管理模块根据第二应用场景确定第二决策结果。
在本实施例中，云调度组件和网络自主决策节点中均包括统一调度管理模块和至少一个协议转换模块，通过协议转换模块可以实现对云调度组件和网络自主决策节点的无线网络的切换，例如：当网络自主决策节点在接收网络节点发送的信息时，网络节点采用的无线网络为时分多址无线网络，而网络自主决策节点采用的无线网络为正交频分多址无线网络，则网络自主决策节点可以通过协议转换模块将自身采用的无线网络转换为时分多址无线网络，实现与网络节点的信息传输。
统一调度管理模块用于根据获取的应用场景来确定决策结果。并将确定的决策结果发送给网络节点。
如图6所示，协议栈包括统一调度管理模块601和若干个协议转换模块，如多个协议转换模块602。统一调度管理模块601通过服务访问点分别与各个协议转换模块602的服务访问点连接，其中，服务访问点为逻辑接口，是上下层之间进行通信的接口。
如图6所示，所述协议栈还包括数据链路层L2模块603和物理层L1模块604，物理层L1模块主要用于对数据进行编码解码等操作，从而使得传输的数据在信道中无差错的传输；数据链路层L2模块主要用于实现将网络资源接入到对应的网络节点空口的控制。
如图6所示，所述协议转换模块602通过L2服务访问点可以与若干个数据链路层L2模块603进行数据转换和传输，通过L1服务访问点可以与若干个物理层L1模块604进行数据转换和传输，例如：所述协议转换模块602通过L2服务访问点可以分别与数据链路层L21模块603和数据链路层L22模块603的服务访问点连接，实现数据转换和传输。通过L1服务访问点可以分别与物理层L11模块604和物理层L12模块604的服务访问点连接，实现数据转换和传输。其中，物理层L1模块603和数据链路层L2模块604向协 议转换模块602传输的数据可以保证所述协议转换模块602切换网络为预设的无线通信模块。
在上述协议栈中，通过设置统一调度管理模块能够确定决策结果，当与发送数据方采用的网络类型不一致时，通过协议转换模块能够实现将网络类型进行切换，从而保证能够接收网络节点发送的信息，并根据接收的信息生成决策结果。
此外，云调度组件和网络自主决策节点还可以对获取的数据信息进行存储，下面对数据信息的存储过程进行详细说明。
可选的，所述云调度组件和网络自主决策节点均包括数据存储模块，所述方法还包括：
所述云调度组件和网络自主决策节点将所述网络节点的通信能力信息存储在区块链中；
通过数据存储模块将网络节点上报的所述网络状态信息、无线信道信息、调度请求信息和资源预占指示信息按照时间顺序存储。
在本实施例中，网络节点的通信能力信息为低频变化的数据信息，其中低频变化表示所述数据信息变化频率较慢。对于低频变换的数据信息可以将其保存在区块链中。网络节点的通信能力信息是指该网络节点支持的全部通信能力信息。具体的，所述区块链基于超级账本平台进行开发的，用于存储正确的网络节点信息以及网络节点的通信能力信息，并通过拜占庭容错算法解决数据信息可能出现的毁坏或篡改问题。
通过采用区块链来保存网络节点的通信能力信息，使得保存的信息具有较强的防篡改性，从而保证根据网络节点的通信能力信息得到的决策结果是适合所述网络节点在当前网络状态信息、无线信道信息、网络调度请求信息和资源预占指示信息下的决策结果。
云调度组件和网络自主决策节点会将接收的实时变化的数据信息按照时间的先后顺序存储在数据存储模块，例如时间序列数据库。采用时间序列数据库可以将获取的实时变化的数据作为一张独立的表进行存储，实现数据的连续存放，减少随机读取操作。
通过上述数据存储模块在对网络状态信息、无线信道信息、调度请求信息和资源预占指示信息进行存储时，能够按照接收数据时间的先后顺序依次 存储，从而实现对较早的网络状态信息、无线信道信息、调度请求信息和资源预占指示信息进行处理，避免网络节点长时间无法激活预设的无线网络资源。
图7为本申请实施例提供的无线智能决策通信装置的结构示意图，所述装置包括：
获取模块701，用于获取各个网络节点的数据信息；所述数据信息包括调度请求信息、资源预占指示信息、网络状态信息、无线信道信息和网络节点的通信能力信息；
预测模块702，针对一个网络节点，用于根据所述数据信息预测所述网络节点在预设时间所处的场景；
激活模块703，用于根据所述网络节点在预设时间的应用场景和网络节点的通信能力信息，确定包含多域组合的决策结果，并在所述预设时间激活所述网络节点的多域组合；其中，所述多域组合与所述网络节点的应用场景为对应关系。
本申请实施例提供的无线智能决策通信装置，可以实现上述如图2所示的实施例的无线智能决策通信方法，其实现原理和技术效果类似，此处不再赘述。
本申请实施例还提供一种无线智能决策通信***，所述***包括：网络自主决策节点和网络节点；
所述网络节点用于向网络自主决策节点发送数据信息；所述数据信息包括调度请求信息、资源预占指示信息、网络状态信息、无线信道信息和网络节点的通信能力信息；
所述网络自主决策节点用于获取网络节点的数据信息，预测所述网络节点在预设时间的应用场景，并根据应用场景和网络节点的通信能力信息，确定决策结果；
所述网络自主决策节点还用于通过统一调度管理协议帧将决策结果发送给所述网络节点，以使所述网络节点在所述预设时间激活决策结果中的多域组合。
本申请实施例提供的无线智能决策通信***，可以实现上述如图3所示的实施例的无线智能决策通信方法，其实现原理和技术效果类似，此处 不再赘述。
图8为本申请实施例提供的无线智能决策通信***，如图8所示，所述***包括：网络自主决策节点801、云调度组件802和网络节点803。
所述云调度组件802用于根据获取的第一数据信息确定第一应用场景，根据第一应用场景确定第一决策结果；
所述网络自主决策节点801用于根据获取的第二数据信息确定第二应用场景，根据第二应用场景确定第二决策结果；
所述网络自主决策节点801根据所述第一决策结果和第二决策结果确定第三决策结果，并通过统一调度管理协议帧将第三决策结果发送给所述网络节点，以使所述网络节点在所述预设时间激活第三决策结果中的多域组合方式；
其中，所述第一数据信息为长周期的数据信息；所述第二数据信息为短周期的数据信息。
本申请实施例提供的无线智能决策通信***，可以实现上述如图5所示的实施例的无线智能决策通信方法，其实现原理和技术效果类似，此处不再赘述。
最后应说明的是：以上各实施例仅用以说明本申请的技术方案，而非对其限制；尽管参照前述各实施例对本申请进行了详细的说明，本领域的普通技术人员应当理解：其依然可以对前述各实施例所记载的技术方案进行修改，或者对其中部分或者全部技术特征进行等同替换；而这些修改或者替换，并不使相应技术方案的本质脱离本申请各实施例技术方案的范围。

Claims (12)

  1. 一种无线智能决策通信方法，其特征在于，所述方法包括：
    获取网络节点的数据信息；所述数据信息包括调度请求信息、资源预占指示信息、网络状态信息、无线信道信息和网络节点的通信能力信息；
    根据所述数据信息预测所述网络节点在预设时间的应用场景；
    根据所述网络节点在预设时间的应用场景和网络节点的通信能力信息，确定包含多域组合的决策结果，并在所述预设时间激活所述网络节点的多域组合；其中，所述多域组合与所述网络节点的应用场景为对应关系。
  2. 根据权利要求1所述的方法，其特征在于，根据所述数据信息预测所述网络节点在预设时间的应用场景，包括：
    根据所述数据信息预测所述网络节点的确定性变化信息；
    根据所述确定性变化信息确定网络节点的应用场景。
  3. 根据权利要求1或2所述的方法，其特征在于，根据所述网络节点在预设时间的应用场景和网络节点的通信能力信息，确定包含多域组合的决策结果，包括：
    将所述网络节点在预设时间的应用场景和网络节点的通信能力信息输入强化学习模型，将所述强化学习模型的输出结果确定为决策结果。
  4. 根据权利要求1或2所述的方法，其特征在于，根据所述网络节点在预设时间的应用场景和网络节点的通信能力信息，确定包含多域组合的决策结果，包括：
    当所述网络节点处于静止的应用场景，且所述网络节点的通信能力信息包含时域资源和频域资源时，所述多域组合包括所述网络节点的时域资源和频域资源；或者，
    当所述网络节点处于移动的应用场景，且所述网络节点的通信能力信息包含时域资源、频域资源和空域资源时，所述决策结果为激活所述网络节点的时域资源、频域资源和空域资源；或者，
    当所述网络节点处于存在干扰的应用场景，且所述网络节点的通信能力信息包含时域资源、频域资源、空域资源和码域资源时，所述决策结果为激活所述网络节点的时域资源、频域资源、空域资源和码域资源；或者，
    当所述网络节点处于多障碍物的应用场景，且所述网络节点的通信能力 信息包含时域资源、频域资源时、空域资源、码域资源和时延多普勒域资源时，所述决策结果为激活所述网络节点的时域资源、频域资源、空域资源、码域资源和时延多普勒域资源。
  5. 根据权利要求1-4任一项所述的方法，其特征在于，所述方法应用于网络节点，所述网络节点通过卷积长短期记忆混合神经网络模型预测网络节点所处的应用场景；所述网络节点通过强化学习模型确定所要激活的多域组合。
  6. 根据权利要求1-5任一项所述的方法，其特征在于，获取网络节点的数据信息，包括：
    通过网络自主决策节点获取网络节点的数据信息；
    相应的，根据所述数据信息预测所述网络节点在预设时间的应用场景，包括：
    通过网络自主决策节点预测所述网络节点在预设时间的应用场景；
    相应的，根据所述网络节点在预设时间的应用场景，确定包含多域组合的决策结果，并在所述预设时间激活所述网络节点的多域组合，包括：
    通过所述网络自主决策节点根据应用场景和网络节点的通信能力信息，确定决策结果；并通过统一调度管理协议帧将决策结果发送给所述网络节点，以使所述网络节点在所述预设时间激活决策结果中的多域组合。
  7. 根据权利要求1-6任一项所述的方法，其特征在于，获取网络节点的数据信息，包括：
    通过云调度组件和网络自主决策节点分别获取网络节点的数据信息；其中，所述云调度组件获取的数据信息为第一数据信息，所述网络自主决策节点获取的数据信息为第二数据信息；其中，所述第一数据信息为长周期的数据信息；所述第二数据信息为短周期的数据信息；
    相应的，根据所述数据信息预测所述网络节点在预设时间的应用场景，包括：
    通过云调度组件预测所述网络节点在预设时间的第一应用场景；通过网络自主决策节点预测所述网络节点在预设时间的第二应用场景；
    相应的，根据所述网络节点在预设时间的应用场景，确定包含多域组合的决策结果，并在所述预设时间激活所述网络节点的多域组合，包括：
    通过所述云调度组件根据第一应用场景和网络节点的通信能力信息，确定第一决策结果；
    通过所述网络自主决策节点根据获取的第二应用场景和和网络节点的通信能力信息，确定第二决策结果；
    通过所述网络自主决策节点根据第一决策结果和第二决策结果确定第三决策结果，并通过统一调度管理协议帧将第三决策结果发送给所述网络节点，以使所述网络节点在所述预设时间激活第三决策结果中的多域组合。
  8. 根据权利要求7所述的方法，其特征在于，所述云调度组件和网络自主决策节点均包括统一调度管理模块和至少一个协议转换模块；通过云调度组件和网络自主决策节点分别获取网络节点的数据信息，包括：
    当云调度组件或网络自主决策节点的网络类型与传输数据的网络节点采用的网络类型不一致时，云调度组件或网络自主决策节点通过所述协议转换模块将网络类型切换至与所述网络节点一致的网络类型；
    相应的，通过云调度组件预测所述网络节点在预设时间的第一应用场景，包括：
    通过云调度组件中的统一调度管理模块根据第一应用场景确定第一决策结果；
    相应的，通过网络自主决策节点预测所述网络节点在预设时间的第二应用场景，包括：
    通过网络自主决策节点中的统一调度管理模块根据第二应用场景确定第二决策结果。
  9. 根据权利要求7或8所述的方法，其特征在于，所述云调度组件和网络自主决策节点均包括数据存储模块，所述通过云调度组件和网络自主决策节点分别获取网络节点的数据信息之后，所述方法还包括：
    所述云调度组件和网络自主决策节点将所述网络节点的通信能力信息存储在区块链中；
    通过数据存储模块将网络节点上报的所述网络状态信息、无线信道信息、调度请求信息和资源预占指示信息按照时间顺序存储。
  10. 一种用于实现权利要求1-9任一项所述方法的无线智能决策通信装置，其特征在于，包括：
    获取模块，用于获取各个网络节点的数据信息；所述数据信息包括调度请求信息、资源预占指示信息、网络状态信息、无线信道信息和网络节点的通信能力信息；
    预测模块，针对一个网络节点，用于根据所述数据信息预测所述网络节点在预设时间所处的场景；
    激活模块，用于根据所述网络节点在预设时间的应用场景和网络节点的通信能力信息，确定包含多域组合的决策结果，并在所述预设时间激活所述网络节点的多域组合；其中，所述多域组合与所述网络节点的应用场景为对应关系。
  11. 一种用于实现权利要求1-9任一项所述方法的无线智能决策通信***，其特征在于，所述***包括：网络自主决策节点和网络节点；
    所述网络节点用于向网络自主决策节点发送数据信息；所述数据信息包括调度请求信息、资源预占指示信息、网络状态信息、无线信道信息和网络节点的通信能力信息；
    所述网络自主决策节点用于获取网络节点的数据信息，预测所述网络节点在预设时间的应用场景，并根据应用场景和网络节点的通信能力信息，确定决策结果；
    所述网络自主决策节点还用于通过统一调度管理协议帧将决策结果发送给所述网络节点，以使所述网络节点在所述预设时间激活决策结果中的多域组合。
  12. 一种用于实现权利要求1-9任一项所述方法的无线智能决策通信***，其特征在于，所述***包括：网络自主决策节点、云调度组件和网络节点；
    所述云调度组件用于根据获取的第一数据信息确定第一应用场景，根据第一应用场景确定第一决策结果；
    所述网络自主决策节点用于根据获取的第二数据信息确定第二应用场景，根据第二应用场景确定第二决策结果；
    所述网络自主决策节点根据所述第一决策结果和第二决策结果确定第三决策结果，并通过统一调度管理协议帧将第三决策结果发送给所述网络节点，以使所述网络节点在所述预设时间激活第三决策结果中的多域组合方式；
    其中，所述第一数据信息为长周期的数据信息；所述第二数据信息为短周期的数据信息。
PCT/CN2021/086679 2021-01-22 2021-04-12 无线智能决策通信方法、装置和*** WO2022156074A1 (zh)

Priority Applications (3)

Application Number Priority Date Filing Date Title
JP2021576899A JP7333839B2 (ja) 2021-01-22 2021-04-12 無線インテリジェント意思決定通信方法、装置及びシステム
EP21761959.2A EP4061048A4 (en) 2021-01-22 2021-04-12 METHOD, DEVICE AND SYSTEM FOR WIRELESS INTELLIGENT DECISION MAKING
US17/469,642 US20220240162A1 (en) 2021-01-22 2021-09-08 Method, apparatus and system for wireless intelligent decision-making communication

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202110084903.2 2021-01-22
CN202110084903.2A CN112423324B (zh) 2021-01-22 2021-01-22 无线智能决策通信方法、装置和***

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US17/469,642 Continuation US20220240162A1 (en) 2021-01-22 2021-09-08 Method, apparatus and system for wireless intelligent decision-making communication

Publications (1)

Publication Number Publication Date
WO2022156074A1 true WO2022156074A1 (zh) 2022-07-28

Family

ID=74783184

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/086679 WO2022156074A1 (zh) 2021-01-22 2021-04-12 无线智能决策通信方法、装置和***

Country Status (2)

Country Link
CN (1) CN112423324B (zh)
WO (1) WO2022156074A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112423324B (zh) * 2021-01-22 2021-04-30 深圳市科思科技股份有限公司 无线智能决策通信方法、装置和***

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20170195456A1 (en) * 2015-11-30 2017-07-06 Telefonaktiebolaget L M Ericsson (Publ) Traffic control in a communication network
CN108092740A (zh) * 2017-05-05 2018-05-29 中兴通讯股份有限公司 传输参数配置方法及装置、确定方法及装置
CN108574591A (zh) * 2017-03-10 2018-09-25 ***通信有限公司研究院 一种配置方法、网络管理设备、终端设备及网络设备
CN108900980A (zh) * 2018-09-21 2018-11-27 北京邮电大学 一种在异构网络中基于移动性预测的资源分配优化方法
CN111417194A (zh) * 2019-01-07 2020-07-14 ***通信有限公司研究院 资源映射配置方法、网络侧设备、用户设备及通信设备
CN111797867A (zh) * 2019-04-09 2020-10-20 Oppo广东移动通信有限公司 ***资源优化方法、装置、存储介质及电子设备
CN112423324A (zh) * 2021-01-22 2021-02-26 深圳市科思科技股份有限公司 无线智能决策通信方法、装置和***

Family Cites Families (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2012072869A1 (en) * 2010-11-30 2012-06-07 Nokia Corporation Method and apparatus for coordinating information request messages over an ad-hoc mesh network
CN102025680A (zh) * 2010-12-15 2011-04-20 华为技术有限公司 速度估计方法和装置
RU2628223C2 (ru) * 2011-09-14 2017-08-15 Риарден, Ллк Системы и способы эксплуатации областей когерентности в беспроводных системах
CN103516654B (zh) * 2012-06-20 2017-04-12 华为技术有限公司 频偏估计方法及***
US10604257B2 (en) * 2016-06-28 2020-03-31 International Business Machines Corporation Unmanned aerial vehicle for air sampling
CN107889251B (zh) * 2016-09-30 2020-04-14 华为技术有限公司 一种时域资源单元集合结构的确定方法、网络设备及终端
US11194809B2 (en) * 2016-12-02 2021-12-07 International Business Machines Corporation Predicting performance of database queries
CN109699037B (zh) * 2017-10-24 2021-11-19 ***通信有限公司研究院 一种波束断连判定配置方法、判定方法及装置
CN107967311B (zh) * 2017-11-20 2021-06-29 创新先进技术有限公司 一种对网络数据流进行分类的方法和装置
CN110138421B (zh) * 2018-02-08 2022-05-17 华为技术有限公司 信道探测方法及相关装置
CN109960395B (zh) * 2018-10-15 2021-06-08 华为技术有限公司 资源调度方法和计算机设备
US11828848B2 (en) * 2019-03-14 2023-11-28 Aeva, Inc. Velocity estimation using doppler per point LiDAR systems
CN111031558A (zh) * 2019-12-23 2020-04-17 安徽理工大学 一种移动流量预测方法、装置、计算机设备及存储介质
CN112952848B (zh) * 2020-01-03 2022-07-08 浙江大学台州研究院 一种基于综合识别指标的储能多场景应用切换控制方法
CN111294234B (zh) * 2020-01-17 2022-07-29 麦科思(苏州)数据科技有限公司 一种基于智能合约优化模型的并行区块链分片方法
CN111741452B (zh) * 2020-07-20 2020-11-17 武汉绿色网络信息服务有限责任公司 一种NB-IoT基站剩余容量预测的方法和装置

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20170195456A1 (en) * 2015-11-30 2017-07-06 Telefonaktiebolaget L M Ericsson (Publ) Traffic control in a communication network
CN108574591A (zh) * 2017-03-10 2018-09-25 ***通信有限公司研究院 一种配置方法、网络管理设备、终端设备及网络设备
CN108092740A (zh) * 2017-05-05 2018-05-29 中兴通讯股份有限公司 传输参数配置方法及装置、确定方法及装置
CN108900980A (zh) * 2018-09-21 2018-11-27 北京邮电大学 一种在异构网络中基于移动性预测的资源分配优化方法
CN111417194A (zh) * 2019-01-07 2020-07-14 ***通信有限公司研究院 资源映射配置方法、网络侧设备、用户设备及通信设备
CN111797867A (zh) * 2019-04-09 2020-10-20 Oppo广东移动通信有限公司 ***资源优化方法、装置、存储介质及电子设备
CN112423324A (zh) * 2021-01-22 2021-02-26 深圳市科思科技股份有限公司 无线智能决策通信方法、装置和***

Also Published As

Publication number Publication date
CN112423324B (zh) 2021-04-30
CN112423324A (zh) 2021-02-26

Similar Documents

Publication Publication Date Title
Pokhrel et al. Towards enabling critical mMTC: A review of URLLC within mMTC
Jiang et al. Low-latency networking: Where latency lurks and how to tame it
Ali et al. Fast uplink grant for machine type communications: Challenges and opportunities
Cecchini et al. LTEV2Vsim: An LTE-V2V simulator for the investigation of resource allocation for cooperative awareness
Gkotsiopoulos et al. Performance determinants in LoRa networks: A literature review
Li et al. Federated multi-agent deep reinforcement learning for resource allocation of vehicle-to-vehicle communications
KR20210117960A (ko) O-ran 기반 성능 최적화 및 구성을 위한 방법 및 장치
WO2022015221A1 (en) Managing a wireless device that is operable to connect to a communication network
Kaur et al. QoS mechanisms for MAC protocols in wireless sensor networks: a survey
US20210274449A1 (en) Method and device for controlling transmission power in wireless communication system
US20220240162A1 (en) Method, apparatus and system for wireless intelligent decision-making communication
US20230038071A1 (en) Resource configuration method and apparatus, device, and storage medium
Hoyhtya et al. Database-assisted spectrum prediction in 5G networks and beyond: A review and future challenges
CN107079463A (zh) 向通信设备分配部分重叠的资源
Mlika et al. Network slicing for vehicular communications: a multi-agent deep reinforcement learning approach
WO2022156074A1 (zh) 无线智能决策通信方法、装置和***
CN113810873A (zh) 用于旁路资源确定的方法和终端
WO2023185531A1 (zh) 资源共存方法及装置
CN113330767A (zh) 频谱管理设备、电子设备、无线通信方法和存储介质
US20230077354A1 (en) Device and method for supporting tsc in wireless communication network
US11784751B1 (en) List size reduction for polar decoding
Kongkham et al. Recurrent network based protocol design for spectrum sensing in cognitive users
WO2023184312A1 (en) Distributed machine learning model configurations
US20240040420A1 (en) Apparatus and method for reporting csi in wireless communication system
US20240089968A1 (en) Enhanced scheduling request configurations for wireless communications

Legal Events

Date Code Title Description
ENP Entry into the national phase

Ref document number: 2021761959

Country of ref document: EP

Effective date: 20210906

ENP Entry into the national phase

Ref document number: 2021576899

Country of ref document: JP

Kind code of ref document: A

121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21761959

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE