Electronic device and method for spectrum management device, and storage medium

Abstract

The present disclosure provides an electronic device, method and computer-readable storage medium for a spectrum management device, the electronic device comprising: a processing circuit configured to: obtain a primary user behavior model for representing the behavior of a primary user using spectrum resources through federated learning; and store the primary user behavior model on a model blockchain.

Description

Electronic device and method for spectrum management device, and storage medium

[0001] This application claims priority to the Chinese patent application filed with the China Patent Office on August 10, 2022, with application number 202210954611.4 and invention name “Electronic device, method and storage medium for spectrum management device”, the entire contents of which are incorporated by reference into this application. Technical Field

[0002] Embodiments of the present disclosure generally relate to the field of wireless communications, and more particularly to spectrum management technology, and more particularly to an electronic device and method for a spectrum management device, and a computer-readable storage medium. Background Art

[0003] With the adoption of 5G technology, network base station deployment has become more dense. Simultaneously, the number of various terminal types, including mobile phones, IoT devices, and wearables, has increased dramatically, driving an increasing demand for spectrum resources, particularly in the sub-6 GHz frequency bands suitable for mobile communications. Due to the heterogeneous and dynamic nature of services, traditional static spectrum management solutions suffer from low spectrum utilization and waste precious spectrum resources. Therefore, dynamic spectrum sharing technology has been proposed. Secondary users, through spectrum sensing or databases, discover idle spectrum for primary users and use it for transmission, improving spectrum utilization.

[0004] The U.S. Federal Communications Commission first proposed introducing blockchain technology into spectrum management at the 2018 Mobile World Congress to make the design more secure and flexible.

[0005] ETSI EN 303 145 defines a reference model for a TV white space system, including a geolocation database (GLDB) and a spectrum coordinator (SC). The SC is used to manage coexistence issues between multiple cognitive radio systems (CRS) and can provide priority-based channel allocation. At the same time, the SC provides the GLDB with necessary protection standard information to ensure priority-based channel allocation when other white space devices (WSDs) that do not use the SC coexist. The GLDB records the actual usage of the white spectrum. However, this technology relies on a centralized database, has poor scalability, and only records historical information on spectrum usage, failing to fully utilize spectrum data.

[0006] Summary of the Invention

[0007] A brief overview of the present disclosure is provided below to provide a basic understanding of certain aspects of the present disclosure. It should be understood that this overview is not an exhaustive overview of the present disclosure. It is not intended to identify key or important aspects of the present disclosure, nor is it intended to limit the scope of the present disclosure. Its purpose is simply to present certain concepts in a simplified form as a prelude to the more detailed description discussed later.

[0008] According to one aspect of the present application, an electronic device for a spectrum management device is provided, including: a processing circuit configured to: obtain a primary user behavior model for representing the behavior of a primary user using spectrum resources through federated learning; and store the primary user behavior model on a model blockchain.

[0009] According to another aspect of the present application, a method for a spectrum management device is provided, including: obtaining a primary user behavior model for representing the behavior of a primary user using spectrum resources through federated learning; and storing the primary user behavior model on a model blockchain.

[0010] The electronic device and method according to the above aspects of the present application use federated learning to obtain a primary user behavior model and use blockchain technology to store the obtained primary user behavior model, thereby enabling analysis and prediction of a decentralized spectrum environment and optimizing spectrum resource allocation for secondary users.

[0011] According to one aspect of the present application, an electronic device for a spectrum management device is provided, including: a processing circuit, configured to: obtain a primary user behavior model obtained through federated learning to represent the behavior of a primary user in using spectrum resources; and determine a spectrum allocation scheme for secondary users to use the primary user's spectrum resources based on the primary user behavior model and relevant information of the secondary user.

[0012] According to another aspect of the present application, a method for a spectrum management device is provided, including: obtaining a primary user behavior model obtained through federated learning to represent the behavior of a primary user in using spectrum resources; and determining a spectrum allocation scheme for secondary users to use the primary user's spectrum resources based on the primary user behavior model and relevant information of the secondary user.

[0013] The electronic devices and methods according to the above aspects of the present application utilize the primary user behavior model obtained through federated learning when allocating spectrum resources to secondary users, that is, spectrum resources are allocated based on the analysis and prediction of the decentralized spectrum environment, thereby obtaining an optimized spectrum allocation plan for secondary users.

[0014] According to one aspect of the present application, an electronic device for a spectrum management device is provided, comprising: a processing circuit configured to: determine a spectrum allocation scheme for a secondary user to use the spectrum resources of a primary user based at least on a primary user behavior model for representing the behavior of the primary user in using the spectrum resources; obtain a performance analysis report on the secondary user's use of the spectrum resources of the primary user allocated to it from the secondary user, evaluate the accuracy of the determined available spectrum resources of the secondary user based on the performance analysis report and obtain an analysis score; and perform a model score on the accuracy of the primary user behavior model based on the analysis score.

[0015] According to another aspect of the present application, a method for a spectrum management device is provided, comprising: determining a spectrum allocation scheme for a secondary user to use the spectrum resources of a primary user based at least on a primary user behavior model for representing the behavior of a primary user in using spectrum resources; obtaining a performance analysis report on the secondary user's use of the spectrum resources of the primary user allocated to it from the secondary user, evaluating the accuracy of the determined available spectrum resources of the secondary user based on the performance analysis report and obtaining an analysis score; and performing a model score on the accuracy of the primary user behavior model based on the analysis score.

[0016] The electronic device and method according to the above aspects of the present application evaluate the accuracy of spectrum resources allocated to secondary users by utilizing performance analysis reports of secondary users, thereby providing an evaluation of the accuracy of the primary user behavior model.

[0017] According to other aspects of the present disclosure, a computer program code and a computer program product for implementing the above-mentioned method for network management, as well as a computer-readable storage medium having the computer program code for implementing the above-mentioned method for network management recorded thereon, are also provided.

[0018] These and other advantages of the present disclosure will become more apparent through the following detailed description of the preferred embodiments of the present disclosure in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to further illustrate the above and other advantages and features of the present disclosure, the following is a further detailed description of the specific embodiments of the present disclosure in conjunction with the accompanying drawings. The drawings, together with the detailed description below, are included in this specification and form a part of this specification. Elements with the same function and structure are represented by the same reference numerals. It should be understood that these drawings only depict typical examples of the present disclosure and should not be regarded as limiting the scope of the present disclosure. In the drawings:

[0020] FIG1 shows a functional module block diagram of an electronic device for a spectrum management apparatus according to an embodiment of the present application;

[0021] FIG2 shows a schematic diagram of a system scenario for dynamic spectrum access;

[0022] FIG3 shows a schematic example of federated learning participation determination;

[0023] FIG4 shows a schematic information flow diagram of a federated learning process of a primary user behavior model based on a model blockchain;

[0024] FIG5 shows a schematic information flow diagram of a federated learning process of a primary user behavior model not based on a model blockchain;

[0025] FIG6 shows a functional module block diagram of an electronic device used in a spectrum management apparatus according to an embodiment of the present application;

[0026] FIG7 shows an example of information flow related to spectrum allocation;

[0027] Figure 8 shows a schematic diagram of a multi-agent deep reinforcement learning algorithm;

[0028] FIG9 shows a schematic diagram of spectrum resource allocation based on a multi-agent deep reinforcement learning algorithm;

[0029] Figure 10 shows an example of the information flow associated with scoring;

[0030] Figure 11 shows an example of a system reference model;

[0031] FIG12 shows a functional module block diagram of an electronic device for a spectrum management apparatus according to another embodiment of the present application;

[0032] FIG13 shows a functional module block diagram of an electronic device for a spectrum management apparatus according to another embodiment of the present application;

[0033] FIG14 shows a functional module block diagram of an electronic device for a spectrum management apparatus according to another embodiment of the present application;

[0034] FIG15 shows a flowchart of a method for a spectrum management apparatus according to an embodiment of the present application;

[0035] FIG16 shows a flowchart of a method for a spectrum management apparatus according to another embodiment of the present application;

[0036] FIG17 shows a flowchart of a method for a spectrum management apparatus according to another embodiment of the present application;

[0037] FIG18 is a block diagram showing an example of a schematic configuration of a server; and

[0038] FIG19 is a block diagram of an exemplary structure of a general-purpose personal computer in which the method and / or apparatus and / or system according to the embodiments of the present disclosure may be implemented. DETAILED DESCRIPTION

[0039] Exemplary embodiments of the present disclosure are described below with reference to the accompanying drawings. For the sake of clarity and conciseness, not all features of an actual implementation are described in this specification. However, it should be understood that in the process of developing any such actual implementation, many implementation-specific decisions must be made in order to achieve the developer's specific goals, such as compliance with system and business-related constraints, which may vary from implementation to implementation. In addition, it should be understood that although the development work may be very complex and time-consuming, it is a routine task for those skilled in the art who benefit from the contents of this disclosure.

[0040] It is also necessary to explain here that, in order to avoid obscuring the present disclosure due to unnecessary details, the accompanying drawings only show the device structure and / or processing steps that are closely related to the solution according to the present disclosure, while other details that are not closely related to the present disclosure are omitted.

[0041] <First embodiment>

[0042] As previously mentioned, dynamic spectrum sharing technology is proposed to improve spectrum utilization. Dynamic spectrum sharing requires consideration of the coexistence of different wireless systems using the same or adjacent frequencies, as well as the protection of the primary system (hereinafter referred to as the primary user). For example, TDD-LTE and 5G NR systems may operate simultaneously in the Citizens Broadband Radio Service (CBRS) band. This embodiment provides a decentralized distributed spectrum management technology to fully utilize spectrum resources.

[0043] Figure 1 shows a functional module block diagram of an electronic device 100 for a spectrum management device according to an embodiment of the present application. As shown in Figure 1, the electronic device 100 includes: a model acquisition unit 101, configured to obtain a primary user behavior model for representing the behavior of a primary user (PU) in using spectrum resources through federated learning; and a blockchain unit 102, configured to store the primary user behavior model on a model blockchain.

[0044] The model acquisition unit 101 and the blockchain unit 102 can be implemented by one or more processing circuits, such as chips or processors. Furthermore, it should be understood that the various functional units in the electronic device shown in FIG1 are merely logical modules divided according to the specific functions they implement, and are not intended to limit specific implementation methods.

[0045] Electronic device 100 may, for example, be located on the side of a spectrum management device or communicatively connected to the spectrum management device. The spectrum management device may be located on the side of a base station, such as implemented by a macro base station or a small base station, or located in a core network, such as implemented by the Evolved Packet Core (EPC) or 5G Core (5GC) under the LTE protocol. It may also be implemented independently of the wireless network by a third-party entity and connected to the wireless network through the wireless network's network operation and management (OAM) unit.

[0046] It should also be noted that electronic device 100 can be implemented at the chip level or at the device level. For example, electronic device 100 can function as a spectrum management device itself and may also include external devices such as a memory and a transceiver (not shown). The memory can be used to store programs and related data required for the electronic device to implement various functions. The transceiver may include one or more communication interfaces to support communication with different devices (e.g., other spectrum management devices, base stations, etc.). The implementation form of the transceiver is not specifically limited here.

[0047] In dynamic spectrum sharing, the use of spectrum resources by secondary users (SUs) must be conducted under the premise of meeting the protection requirements for primary users. Therefore, the historical and / or predicted behavior of the primary user in using spectrum resources has a significant impact on the allocation of spectrum resources to secondary users. This embodiment proposes a primary user behavior model for representing the primary user's behavior in using spectrum resources. The primary user behavior model can be a statistical model of historical behavior, a predictive model of future behavior, or a combination of the two, which is not restrictive.

[0048] Figure 2 shows a schematic diagram of a system scenario for dynamic spectrum access. In this system, multiple spectrum management devices, such as SASs, exist. Each spectrum management device allocates spectrum resources to its registered secondary users, enabling dynamic access. For example, the primary user database of a spectrum management device can store protection requirements for primary users within a management area, such as information about protected or exclusive zones, as well as spectrum usage information. Furthermore, the spectrum management device can manage a certain number of spectrum sensing devices, which periodically provide feedback on spectrum environment sensing results to the spectrum management device. The electronic device 100 of this embodiment is implemented, for example, on the spectrum management device shown in Figure 2.

[0049] Model acquisition unit 101 obtains a primary user behavior model through federated learning (FL). Federated learning here can be understood as the process of multiple spectrum management devices jointly training the primary user behavior model. In FL, one of the multiple spectrum management devices acts as a FL server, and the others act as FL clients.

[0050] For example, the spectrum management device corresponding to the electronic device 100 serves as a federated learning server, and the model acquisition unit 101 is configured to determine one or more other spectrum management devices as federated learning clients participating in the federated learning, and aggregate the local primary user behavior models trained by each federated learning client to obtain a global model as the primary user behavior model.

[0051] The model acquisition unit 101 may initiate federated learning in response to one or more of the following: a primary user status update, a model accuracy requirement update, or a periodic trigger. For example, when at least one of these three events occurs, the model acquisition unit 101 determines that federated learning of the corresponding primary user's behavior model is required.

[0052] Specifically, the model acquisition unit 101 can send a federated learning request to other spectrum management devices within a predetermined area, receive federated learning responses from other spectrum management devices that agree to participate in federated learning, and determine the spectrum management device that serves as the federated learning client among the other spectrum management devices from which the federated learning responses are received.

[0053] For example, a federated learning request may include one or more of the following: a federated learning instruction, a primary user ID, the range of the primary user's prohibited area or protected area, the primary user's location, and an analysis identifier for identifying the primary user behavior model training. The federated learning instruction indicates to other spectrum management devices that federated learning is to be performed; the primary user ID indicates the primary user for which federated learning is to be performed; and the analysis identifier indicates the type of training analysis to be performed, such as analysis of network performance or user equipment-related performance. Here, the identifier is set to indicate the primary user behavior model training.

[0054] In addition, the range of the prohibited area or protection area of ​​the primary user and the location of the primary user can be used by other spectrum management devices to determine whether to participate in the federated learning. For example, the spectrum management device that receives the federated learning request determines whether the coverage of the secondary user devices and spectrum sensing devices it manages overlaps with the range of the prohibited area or protection area of ​​the primary user based on the federated learning request. If there is no overlap, it will not participate in the federated learning. If there is an overlap and it has the function of federated learning, it will participate in the federated learning and send a response agreeing to participate in the federated learning. It should be understood that when this spectrum management device receives a federated learning request from other spectrum management devices that act as federated learning servers with model aggregation capabilities, it also performs this judgment and response accordingly.

[0055] Figure 3 illustrates an example of determining participation in federated learning. In Figure 3, the first spectrum management device sends a federated learning request to the second through fourth spectrum management devices. The fourth spectrum management device finds that the coverage of the secondary user devices and spectrum sensing devices it manages does not overlap with the protection zone of the primary user and, therefore, decides not to participate in the federated learning request. The second and third spectrum management devices, however, decide to participate based on the overlapping coverage areas.

[0056] Furthermore, the model obtaining unit 101 may determine the predetermined area based on one or more of the following: the location of the spectrum management device, the location of the primary user, and the extent of the primary user's prohibited area or protected area. The model obtaining unit 101 may obtain various information about the primary user from a primary user database. The model obtaining unit 101 sends a federated learning request to the spectrum management devices within the determined predetermined area. Alternatively, the model obtaining unit 101 may send a federated learning request to all spectrum management devices capable of communicating with the spectrum management device.

[0057] As an example, multiple spectrum management devices form a blockchain network. In other words, the spectrum management devices have model blockchain functionality. In this example, the blockchain unit 102 can store the primary user behavior model on the model blockchain for synchronization and sharing among the multiple spectrum management devices. The model acquisition unit 101 is also configured to send a notification to the spectrum management device identified as a federated learning client to initiate the training process. In response to this notification, the federated learning client trains the local primary user behavior model and stores it on the model blockchain.

[0058] For ease of understanding, Figure 4 shows a schematic information flow diagram of the federated learning process of the primary user behavior model based on the model blockchain. Note that in Figure 4, for clarity, the primary user (PU) model training function and the model blockchain function of the spectrum management device (SMD) are shown as two separate logical modules. The SMD here is, for example, a SAS. The PU model training function basically corresponds to the function of the model acquisition unit 101, and the model blockchain function basically corresponds to the function of the blockchain unit 102. In the following description, no special distinction will be made.

[0059] First, as described above, the PU model training function of SMD-1 determines that SMD-1 starts federated learning as the federated learning server, sends a federated learning request to SMD-2 to SMD-N, and receives the corresponding federated learning responses from SMD-2 to SMD-N (specifically, their PU model training functions). Subsequently, SMD-1's PU model training selects a federated learning client based on the received federated learning response and sends a notification to the SMD selected as the federated learning client to start the training process.

[0060] Next, in each round of FL training, the PU model training function of the SMD serving as the federated learning client performs local training to obtain a local master user behavior model and records the resulting local master user behavior model on the corresponding local model blockchain. The blockchain synchronizes the model blockchain across all SMD nodes through consensus authentication and block propagation. Then, the PU model training function of SMD-1 retrieves the local master user behavior models trained by other SMDs from its local model blockchain and aggregates these local master user behavior models to obtain a global model. SMD-1's PU model training function records the global model on its local model blockchain. Similarly, the blockchain synchronizes the model blockchain across all SMD nodes through consensus authentication and block propagation. SMD clients participating in federated learning can retrieve this global model from their local model blockchain and proceed to the next round of training based on this global model.

[0061] When the predetermined conditions are met, federated learning stops and the final global model is used as the master user behavior model. The final master user behavior model is shared among multiple SMDs through the model blockchain.

[0062] Accordingly, the model acquisition unit 101 is configured to retrieve the local main user behavior model trained by each federated learning client from the model blockchain in each round of federated learning training for aggregation and store the aggregated global model on the model blockchain. Each federated learning client verifies the global model using the consensus mechanism of the blockchain, obtains the global model from the model blockchain and conducts the next round of training based on the global model; and stops federated learning until a predetermined condition is met, and uses the final obtained global model as the main user behavior model.

[0063] For example, the predetermined condition may include that the obtained global model reaches a predetermined accuracy or has been trained for a predetermined number of rounds.

[0064] In addition, the blockchain unit 102 is further configured to store the primary user behavior model in association with the analysis identifier in the model blockchain. In the case where training is performed for multiple primary users, the blockchain unit 102 is further configured to store the primary user behavior model in association with the ID of the corresponding primary user.

[0065] Note that in the above-mentioned federated learning process, in addition to model aggregation, the spectrum management device as a federated learning server can also perform local training of the primary user behavior model, which is not restrictive.

[0066] For example, the local training of the primary user behavior model in the aforementioned information flow can be performed by the model acquisition unit 101 as follows: based on the protection requirements of the primary user and the measurement and perception information of the secondary user, a machine learning algorithm is used to train the local primary user behavior model. The machine learning algorithm can be, for example, a long short-term memory (LSTM) algorithm.

[0067] Furthermore, the model acquisition unit 101 can be configured to monitor the status of each federated learning client and delete the client if its local primary user behavior model is not updated in a timely manner or if other anomalies exist. Furthermore, if an existing federated learning client is deleted, the model acquisition unit 101 can be configured to send a federated learning request to other spectrum management devices not participating in federated learning to identify a new federated learning client. This ensures the effectiveness and accuracy of federated learning.

[0068] In another example, the spectrum management device may not have a model blockchain function. In this case, the model acquisition unit 101 can also perform federated learning to obtain a primary user behavior model. Figure 5 shows a schematic information flow diagram of the federated learning process of the primary user behavior model that is not based on a model blockchain. It can be seen that the process of selecting and notifying the FL client is the same as the corresponding process in Figure 4. The difference between Figure 5 and Figure 4 is that during the federated learning training process, the local user behavior model trained by each FL client is directly transmitted to the FL server via an interface function or a communication function, and the global model obtained by the FL server through model aggregation is directly transmitted to the FL client via an interface function or a communication function. Other details can be referred to the description of Figure 4 and will not be repeated here.

[0069] FIG6 shows another functional module block diagram of electronic device 100 according to an embodiment of the present application. Compared to FIG1 , electronic device 100 in FIG6 further includes a spectrum allocation unit 103, which is configured to determine a spectrum allocation scheme for secondary users to use the primary user's spectrum resources based on the primary user behavior model and relevant information about the secondary user. Spectrum allocation unit 103 can implement spectrum allocation functionality.

[0070] The spectrum allocation unit 103 may also provide the spectrum usage information in the spectrum allocation plan to the corresponding secondary user and store the spectrum usage information confirmed by the secondary user in the spectrum blockchain. In other words, the spectrum usage information of the secondary user may be stored in the spectrum blockchain to record the spectrum usage history of the secondary user.

[0071] For example, the relevant information of the secondary user may include one or more of the following: the secondary user's registration information, the secondary user's spectrum usage request, the spectrum sensing result obtained by the spectrum sensing device, and the current secondary user's spectrum usage information. For example, the spectrum allocation unit 103 may obtain the primary user behavior model from the model blockchain, obtain the secondary user's registration information from the secondary user database, obtain the current secondary user's spectrum usage information from the spectrum blockchain, and obtain the secondary user's spectrum usage request and the spectrum sensing result obtained by the spectrum sensing device from the coexistence management function.

[0072] Figure 7 illustrates an example of the information flow related to the aforementioned spectrum allocation. In Figure 7, the spectrum allocation unit's functions of acquiring information and determining a spectrum allocation plan are illustrated as two functional modules: the spectrum coordinator and the secondary user analysis function. It should be understood that this is not intended to be limiting. As shown in Figure 7, the spectrum coordinator receives spectrum usage requests from secondary users and spectrum sensing results from spectrum sensing devices, and then sends an analysis request to the secondary user analysis function. In response to the analysis request, the secondary user analysis function retrieves data for analysis from the model blockchain, secondary user database, spectrum blockchain, and other sources, including primary user behavior models and information related to secondary users. The secondary user analysis function then performs secondary user analysis based on this data to determine a spectrum allocation plan, such as the available spectrum resources for the secondary user, and sends this available spectrum information to the spectrum coordinator. The spectrum coordinator sends the corresponding operating parameters to the secondary user, who confirms them and sends back a spectrum usage response. The spectrum coordinator records the spectrum usage information confirmed by the secondary user in the spectrum blockchain. Through blockchain consensus authentication and dissemination, each spectrum management device can obtain the spectrum usage information of the secondary user.

[0073] Exemplarily, the spectrum allocation unit 103 may use an artificial intelligence method such as a Multi-Agent Deep Reinforcement Learning (MADRL) algorithm to determine an optimized spectrum allocation scheme by minimizing mutual interference between secondary users and protecting primary users.

[0074] Deep reinforcement learning (DRL) is a machine learning algorithm that combines deep neural networks (DNNs) with reinforcement learning (RL). In DRL, an intelligent agent learns to develop an optimal action strategy through interaction with its environment. Leveraging the strengths of DNNs, DRL achieves better convergence and accuracy than traditional RL, making it suitable for solving complex control problems. MADRL, on the other hand, is suitable for distributed decision-making problems involving multiple intelligent agents, such as the blockchain-based distributed spectrum management scenario in this embodiment.

[0075] Figure 8 shows a schematic diagram of MADRL. Figure 9 shows a schematic diagram of spectrum resource allocation based on MADRL. In Figure 9, one agent corresponds to, for example, a secondary user. The agent determines the selected channel and RF power based on the predicted PU state (e.g., obtained based on the primary user behavior model) and the actions of other agents. Through the implementation of MADRL, spectrum resource allocation among secondary users can be optimized.

[0076] In addition, the spectrum allocation unit 103 can also be configured to obtain a performance analysis report of the secondary user's use of spectrum resources from the secondary user, and evaluate the accuracy of the determined available spectrum resources for the secondary user based on the performance analysis report and obtain an analysis score. That is, the secondary user uses the spectrum resources allocated to it by the spectrum allocation unit 103 to communicate and generates a corresponding performance analysis report, which reflects the performance of the secondary user when using the allocated available spectrum resources for communication. Therefore, based on the report, the spectrum allocation unit 103 can evaluate the accuracy of the available spectrum resources allocated to the secondary user and obtain an analysis score. Furthermore, the spectrum allocation unit 103 can also be configured to perform a model score on the accuracy of the primary user behavior model based on the analysis score, and record the model score in the model blockchain in association with the primary user behavior model after verification by the consensus mechanism.

[0077] Exemplarily, the consensus mechanism can be configured as follows. The blockchain unit 102 is configured to initiate a score credibility request to other spectrum management devices in the network and receive score credibility responses from the other spectrum management devices. The score credibility response is a response from the other spectrum management devices, based on predetermined rules, regarding whether the model score is credible in response to the score credibility request. For example, the score credibility request may include the model score, and the other spectrum management devices may determine whether the model score is credible based on the difference between the model score and the locally stored scores of related models. For example, if the difference is greater than a predetermined degree, the other spectrum management devices will deem the score unreliable and return a "no" response; otherwise, they will return a "yes" response.

[0078] The blockchain unit 102 is further configured to record the model score in the model blockchain if more than a predetermined percentage of responses indicate that the model score is trustworthy. Otherwise, the model score is not recorded. The predetermined percentage is, for example, half. Furthermore, the blockchain unit 102 indicates whether the model score has been recorded to the spectrum allocation unit 103 that generated the model score.

[0079] The model obtaining unit 101 may also be configured to re-initiate federated learning for the primary user behavior model when the model score for the primary user behavior model is lower than a predetermined level, thereby further ensuring the accuracy of the primary user behavior model.

[0080] Figure 10 illustrates an example of the information flow related to the aforementioned scoring. Similarly, Figure 10 illustrates the spectrum allocation unit's information acquisition and model scoring functions as two functional modules: the spectrum coordinator and the secondary user analysis function. It should be understood that this is not intended to be limiting. As shown in Figure 10, a secondary user sends a performance analysis report to the spectrum coordinator. Based on the performance analysis report, the spectrum coordinator evaluates the accuracy of the available spectrum resources allocated to the secondary user to obtain an analysis score. The analysis score is then sent to the secondary user analysis function, which then performs a model score on the primary user's behavior model based on the analysis score. The model score is then verified through a consensus mechanism and recorded on the model blockchain. Through blockchain consensus authentication and dissemination, each spectrum management device can obtain the secondary user's model score for the primary user's behavior model. As described above, within the consensus mechanism, the model blockchain function can also indicate to the secondary user analysis function whether the model score has been recorded. The model training function can retrieve the model score from the model blockchain and, if the model score falls below a predetermined threshold, trigger federated learning of the primary user's behavior model. When multiple secondary users have model scores, the model score can be assessed to determine whether it falls below the predetermined threshold, for example, by averaging.

[0081] Furthermore, to facilitate understanding of the distribution of the logical functions of the various units described in this embodiment within the spectrum management system, Figure 11 illustrates an example of a system reference model. As shown in Figure 11, the system includes multiple logical functions, which are described in detail below. The spectrum coordination function, equivalent to the spectrum coordinator shown in Figures 7 and 10, performs a portion of the functions of the spectrum allocation unit 103. Based on the spectrum allocation plan obtained from the SU analysis function, it notifies the SU of the permitted spectrum and communication parameters, and controls the sensing and measurement functions of the spectrum sensing device to obtain sensing results. The spectrum coordination function's service access point (SAP) is the SC-SAP. The PU training function performs the functions of the model acquisition unit 101, and its service access point is the PAI-SAP. The SU analysis function, i.e., the SU analysis function shown in Figures 7 and 10, performs a portion of the functions of the spectrum allocation unit 103. Based on the relevant information, it calculates an optimized spectrum allocation plan using an artificial intelligence method (e.g., the aforementioned MADRL algorithm). The SU analysis function's service access point is the SAI-SAP. The federated learning function performs part of the functions of model acquisition unit 101 and, if the spectrum management device has blockchain functionality, part of the functions of blockchain unit 102. The access point for the federated learning function is FL-SAP. The model blockchain function performs the functions of blockchain unit 102, recording the primary user behavior model and the corresponding model score. During the federated learning process, it also synchronously records the results of each iteration of the local model and the aggregated model. Its service access point is MBC-SAP. The spectrum blockchain function records spectrum usage information of secondary users for use by the secondary user analysis function. Its service access point is SBC-SAP. The PU database function stores PU information, PU protection requirements, etc. Its service access point is PDB-SAP. The SU database function stores SU base station information, communication parameters, coverage information, and other information provided by the regulatory agency for calculating available spectrum resources. Its service access point is SDB-SAP. The communication function exchanges various information. Its service access point is Com-SAP. The interface function is an abstract function that can be implemented through the combination of the above functional blocks to ensure the integrity of the physical entity. Its service access point can be all the above SAPs.

[0082] According to an embodiment of the present application, electronic device 100 uses federated learning to train a primary user behavior model, thereby obtaining a more accurate primary user behavior model. Furthermore, electronic device 100 uses blockchain technology to store the obtained primary user behavior model, enabling decentralized spectrum environment analysis and prediction, thereby optimizing spectrum resource allocation for secondary users. Furthermore, electronic device 100 scores the primary user behavior model based on secondary user performance analysis reports, allowing for timely updates based on the scoring. This further improves the accuracy of the primary user behavior model and optimizes spectrum resource allocation for secondary users.

[0083] <Second embodiment>

[0084] Figure 12 shows a functional module block diagram of an electronic device 200 according to another embodiment of the present application. As shown in Figure 12, the electronic device 200 includes: a communication unit 201, configured to obtain a primary user behavior model obtained through federated learning for representing the behavior of the primary user in using spectrum resources; and a determination unit 202, configured to determine a spectrum allocation scheme for the secondary user to use the spectrum resources of the primary user based on the primary user behavior model and relevant information of the secondary user.

[0085] The communication unit 201 and the determination unit 202 may be implemented by one or more processing circuits, which may be implemented as chips or processors, for example. Furthermore, it should be understood that the various functional units in the electronic device shown in FIG12 are merely logical modules divided according to the specific functions they implement, and are not intended to limit specific implementation methods.

[0086] The electronic device 200 may be, for example, disposed on a spectrum management device side or communicatively connected to the spectrum management device. The spectrum management device may be located on a base station side, such as implemented by a macro base station or a small base station, or located in a core network, such as implemented by an Evolved Packet Core (EPC) under the LTE protocol.

[0087] It should also be noted that electronic device 200 can be implemented at the chip level or at the device level. For example, electronic device 200 can function as a spectrum management device itself and may also include external devices such as a memory and a transceiver (not shown). The memory can be used to store programs and related data required for the electronic device to implement various functions. The transceiver may include one or more communication interfaces to support communication with different devices (e.g., other spectrum management devices, base stations, etc.). The implementation form of the transceiver is not specifically limited here.

[0088] Here, the main user behavior model can be the main user behavior model stored in the model blockchain manner as described in the first embodiment, or it can be the main user behavior model stored in the storage area of ​​the spectrum management device that is not a node of the blockchain network. This is not restrictive.

[0089] The communication unit 201 is also configured to provide the spectrum usage information in the spectrum allocation plan to the corresponding secondary user, and store the spectrum usage information confirmed by the secondary user in the spectrum blockchain.

[0090] Similar to the first embodiment, the relevant information of the secondary user may include one or more of the following: the secondary user's registration information, the secondary user's spectrum usage request, the spectrum sensing result obtained by the spectrum sensing device, and the current secondary user's spectrum usage information. For example, the communication unit 201 may obtain the primary user behavior model from the model blockchain or the storage area of ​​the spectrum management device, obtain the secondary user's registration information from the secondary user database, obtain the current secondary user's spectrum usage information from the spectrum blockchain, and obtain the secondary user's spectrum usage request and the spectrum sensing result obtained by the spectrum sensing device from the coexistence management function.

[0091] The determination unit 202 may use the MADRL algorithm, for example, to determine an optimized spectrum allocation scheme by minimizing mutual interference between secondary users and protecting primary users. The relevant detailed description has been given in the first embodiment and will not be repeated here.

[0092] In addition, the communication unit 201 is further configured to obtain a performance analysis report on the secondary user's use of spectrum resources from the secondary user. The determination unit 202 evaluates the accuracy of the available spectrum resources of the determined secondary user based on the performance analysis report and obtains an analysis score. Based on the analysis score, a model score is performed on the accuracy of the primary user behavior model, and the model score is recorded in association with the primary user behavior model. For example, when the primary user behavior model is stored in a model blockchain, the model score is recorded in association with the primary user behavior model in the model blockchain after verification by a consensus mechanism. When the primary user behavior model is stored in a storage area of ​​a spectrum management device, the model score is recorded in association with the primary user behavior model in the storage area, and the model score can be notified by the spectrum management device to other spectrum management devices for synchronization and sharing.

[0093] The determining unit 202 is further configured to re-initiate federated learning for the primary user behavior model when the model score for the primary user behavior model is lower than a predetermined level.

[0094] The relevant information flow has been described in detail with reference to FIG. 7 and FIG. 10 in the first embodiment, and will not be repeated here.

[0095] In one example, the electronic device 200 also includes the model acquisition unit 101 described with reference to Figure 1. As shown in Figure 13, the spectrum management device acts as a federated learning server, for example, and the model acquisition unit 101 can be configured to obtain the main user behavior model as follows: determine one or more other spectrum management devices as federated learning clients participating in federated learning, and aggregate the local main user behavior models trained by each federated learning client to obtain a global model as the main user behavior model.

[0096] Specifically, the model acquisition unit 101 is configured to: obtain trained local primary user behavior models from each federated learning client for aggregation in each round of federated learning training, and send the aggregated global model to each federated learning client so that each federated learning client can conduct the next round of training based on the received global model; and terminate federated learning until a predetermined condition is met, and use the final global model as the primary user behavior model. The predetermined condition may include, for example, that the global model has reached a predetermined accuracy or has undergone a predetermined number of training rounds.

[0097] When the spectrum management device does not have a blockchain function, the relevant information flow can refer to that shown in Figure 5.

[0098] Additionally, as shown in another dashed box in Figure 13 , electronic device 200 may also include a blockchain unit 102 configured to record the local primary user behavior model, the global model, and the primary user behavior model on a model blockchain for sharing among multiple spectrum management devices. Through the blockchain's consensus mechanism, this data can be synchronized across blockchain nodes. The relevant information flow in this case can be seen in Figure 4 .

[0099] The detailed description of the model obtaining unit 101 and the blockchain unit 102 has been given in the first embodiment and will not be repeated here.

[0100] To sum up, according to this embodiment, the electronic device 200 utilizes the primary user behavior model obtained through federated learning when allocating spectrum resources to secondary users, that is, spectrum resources are allocated based on the analysis and prediction of the decentralized spectrum environment, thereby obtaining an optimized spectrum allocation plan for secondary users.

[0101] <Third embodiment>

[0102] Figure 14 shows a functional module block diagram of an electronic device 300 according to another embodiment of the present application. As shown in Figure 14, the electronic device 300 includes: a determination unit 301, configured to determine a spectrum allocation scheme for a secondary user to use the spectrum resources of a primary user based at least on a primary user behavior model for representing the behavior of the primary user using spectrum resources; a communication unit 302, configured to obtain a performance analysis report on the secondary user's use of the spectrum resources of the primary user allocated to it from the secondary user; and an evaluation unit 303, configured to evaluate the accuracy of the determined available spectrum resources of this user based on the performance analysis report and obtain an analysis score, and perform a model score on the accuracy of the primary user behavior model based on the analysis score.

[0103] The determining unit 301, the communication unit 302, and the evaluating unit 303 may be implemented by one or more processing circuits, which may be implemented as chips or processors, for example. Furthermore, it should be understood that the various functional units in the electronic device shown in FIG14 are merely logical modules divided according to the specific functions they implement, and are not intended to limit specific implementations.

[0104] The electronic device 300 may be, for example, disposed on a spectrum management device side or communicatively connected to the spectrum management device. The spectrum management device may be located on a base station side, such as implemented by a macro base station or a small base station, or located in a core network, such as implemented by an Evolved Packet Core (EPC) under the LTE protocol.

[0105] It should also be noted that electronic device 300 can be implemented at the chip level or at the device level. For example, electronic device 200 can function as a spectrum management device itself and may also include external devices such as a memory and a transceiver (not shown). The memory can be used to store programs and related data required for the electronic device to implement various functions. The transceiver may include one or more communication interfaces to support communication with different devices (e.g., other spectrum management devices, base stations, etc.). The implementation form of the transceiver is not specifically limited here.

[0106] The evaluation unit 303 may also be configured to record the model score in association with the primary user behavior model.

[0107] In one example, the master user behavior model and model score can be stored on a model blockchain to be shared among multiple spectrum management devices. In another example, the master user behavior model and model score are stored in a storage area of ​​each spectrum management device via a communication function or an interface function.

[0108] The determining unit 301 is further configured to enable the spectrum management apparatus to retrain the primary user behavior model when the model score of the primary user behavior model is lower than a predetermined level. For example, the primary user behavior model can be obtained through federated learning.

[0109] In addition, a spectrum blockchain may be created, and the communication unit 302 stores the spectrum usage information confirmed by the secondary user in the spectrum blockchain. The spectrum usage information may be used by the determination unit 301 when determining the spectrum allocation plan.

[0110] The examples of obtaining the main user behavior model through federated learning, the examples of the model blockchain and the spectrum blockchain, and the example of determining the spectrum allocation scheme based on the main user behavior model have been described in detail in the first and second embodiments, which are also applicable to this embodiment and will not be repeated here.

[0111] In summary, the electronic device 300 according to this embodiment evaluates the accuracy of the spectrum resources allocated to the secondary user by utilizing the performance analysis report of the secondary user, thereby providing an evaluation of the accuracy of the primary user behavior model.

[0112] <Fourth embodiment>

[0113] While the above embodiments describe the electronic device used in a spectrum management apparatus, it is apparent that certain processes or methods are also disclosed. Below, an overview of these methods is provided without repeating some of the details discussed above. However, it should be noted that while these methods are disclosed in the description of the electronic device used in a spectrum management apparatus, these methods do not necessarily employ or are not necessarily executed by the components described. For example, the embodiments of the electronic device used in a spectrum management apparatus may be partially or completely implemented using hardware and / or firmware, while the methods used in a spectrum management apparatus discussed below may be completely implemented using computer-executable programs, although these methods may also employ the hardware and / or firmware of the electronic device used in a spectrum management apparatus.

[0114] Figure 15 shows a flowchart of a method for a spectrum management device according to one embodiment of the present application. The method includes: obtaining a primary user behavior model representing the primary user's behavior in using spectrum resources through federated learning (S11); and storing the primary user behavior model on a model blockchain (S12). This method can be performed, for example, on the spectrum management device.

[0115] For example, the spectrum management device acts as a federated learning server, and the above step S11 includes: determining one or more other spectrum management devices as federated learning clients participating in the federated learning, and aggregating the local primary user behavior models trained by each federated learning client to obtain a global model as the primary user behavior model.

[0116] For example, a federated learning request can be sent to other spectrum management devices within a predetermined area, a federated learning response can be received from other spectrum management devices that agree to participate in the federated learning, and a spectrum management device that serves as a federated learning client can be determined from the other spectrum management devices from which the federated learning response is received.

[0117] A federated learning request may include one or more of the following: a federated learning instruction, the primary user's identifier, the primary user's prohibited or protected area, the primary user's location, and an analysis identifier for training the primary user's behavior model. The primary user's behavior model and the analysis identifier may also be stored on the model blockchain in a corresponding manner.

[0118] For example, the predetermined area may be determined based on one or more of the following: the location of the spectrum management device, the location of the primary user, the range of the prohibited area or the protection area of ​​the primary user.

[0119] In step S11, a notification to initiate training can be sent to the spectrum management device identified as a federated learning client. In response, the federated learning client trains a local primary user behavior model and stores it on the model blockchain. In each round of federated learning training, the local primary user behavior models trained by each federated learning client are retrieved from the model blockchain for aggregation and the aggregated global model is stored on the model blockchain. Each federated learning client verifies the global model using the blockchain's consensus mechanism, obtains the global model from the model blockchain, and conducts the next round of training based on it. Federated learning is terminated until predetermined conditions are met, and the resulting global model is used as the primary user behavior model. For example, the local primary user behavior model can be trained using a machine learning algorithm based on the protection requirements of the primary user and the measurement and perception information of the secondary user.

[0120] The predetermined conditions include, for example, that the obtained global model reaches a predetermined accuracy or has been trained for a predetermined number of rounds.

[0121] Federated learning may be initiated in response to one or more of the following: primary user status update, model accuracy requirement update, periodic trigger.

[0122] Although not shown in the figure, the above method may also include: monitoring the status of each federated learning client, and deleting the federated learning client when the local primary user behavior model of the federated learning client is not updated in a timely manner or there are other abnormalities, and when the existing federated learning client is deleted, sending a federated learning request to other spectrum management devices that are not participating in the federated learning to determine a new federated learning client.

[0123] In addition, the above method may also include: receiving a federated learning request from other spectrum management devices that have model aggregation capabilities and serve as federated learning servers, judging based on the federated learning request whether the coverage range of the secondary user equipment and spectrum sensing equipment managed by itself overlaps with the range of the prohibited area or protection area of ​​the primary user, and sending a response agreeing to participate in the federated learning if there is an overlap.

[0124] As shown in the dashed box in Figure 15 , the method may further include step S13: determining a spectrum allocation scheme for secondary users to utilize the primary user's spectrum resources based on the primary user behavior model and relevant information about the secondary users. The relevant information about the secondary users may include, for example, one or more of the following: the secondary user's registration information, the secondary user's spectrum usage request, the spectrum sensing results of a spectrum sensing device, and current secondary user spectrum usage information. For example, a multi-agent deep reinforcement learning algorithm may be used to determine an optimized spectrum allocation scheme by minimizing mutual interference between secondary users and protecting the primary user.

[0125] As shown in another dotted box in Figure 15, the above method may also include step S14: providing the spectrum usage information in the spectrum allocation plan to the corresponding secondary user, and storing the spectrum usage information confirmed by the secondary user in the spectrum blockchain.

[0126] As shown in the other two dotted boxes in Figure 15, the above method may also include: obtaining a performance analysis report on the secondary user's use of the spectrum resources from the secondary user, evaluating the accuracy of the determined available spectrum resources of the secondary user based on the performance analysis report and obtaining an analysis score (S15); and based on the analysis score, performing a model score on the accuracy of the primary user behavior model, and recording the model score in the model blockchain in association with the primary user behavior model after verification by a consensus mechanism (S16).

[0127] In addition, when the model score for the main user behavior model is lower than a predetermined level, federated learning for the main user behavior model can be reinitiated.

[0128] The above method corresponds to the electronic device 100 in the first embodiment. The specific details can be referred to the first embodiment and will not be repeated here.

[0129] Figure 16 shows a flowchart of a method for a spectrum management device according to another embodiment of the present application. The method includes: obtaining a primary user behavior model obtained through federated learning, which represents the behavior of a primary user in using spectrum resources (S21); and determining a spectrum allocation scheme for secondary users to use the primary user's spectrum resources based on the primary user behavior model and relevant information about secondary users (S22). This method can be performed, for example, on the spectrum management device side.

[0130] The above method corresponds to the electronic device 200 in the second embodiment. The specific details can be referred to the second embodiment and will not be repeated here.

[0131] Figure 17 shows a flowchart of a method for a spectrum management device according to another embodiment of the present application. The method includes: determining a spectrum allocation scheme for secondary users to use the spectrum resources of a primary user based at least on a primary user behavior model representing the primary user's behavior in using spectrum resources (S31); obtaining a performance analysis report from the secondary user regarding the secondary user's use of the spectrum resources allocated to the primary user, evaluating the accuracy of the determined available spectrum resources for the secondary user based on the performance analysis report and obtaining an analysis score (S32); and performing a model score on the accuracy of the primary user behavior model based on the analysis score (S33). This method can be performed, for example, on the spectrum management device side.

[0132] For example, the model score can be recorded in association with the primary user behavior model. The primary user behavior model and the model score can be stored on a model blockchain for sharing among multiple spectrum management devices. If the model score of the primary user behavior model falls below a predetermined level, the spectrum management device can be instructed to retrain the primary user behavior model.

[0133] For example, the primary user behavior model is obtained through federated learning. Spectrum usage information confirmed by secondary users can also be stored in the spectrum blockchain.

[0134] The above method corresponds to the electronic device 300 in the third embodiment. For specific details, please refer to the third embodiment and will not be repeated here.

[0135] Note that the above methods can be used in combination or individually.

[0136] The technology of the present disclosure can be applied to various products. For example, the electronic device 100, 200 or 300 can be implemented as any type of server, such as a tower server, a rack server, and a blade server. The electronic device 100, 200 or 300 can be a control module installed on the server (such as an integrated circuit module including a single chip, and a card or blade inserted into a slot of a blade server).

[0137] [Application examples about servers]

[0138] 18 is a block diagram illustrating an example of a schematic configuration of a server 700 to which the technology of the present disclosure can be applied. The server 700 includes a processor 701 , a memory 702 , a storage device 703 , a network interface 704 , and a bus 706 .

[0139] The processor 701 may be, for example, a central processing unit (CPU) or a digital signal processor (DSP), and controls the functions of the server 700. The memory 702 includes a random access memory (RAM) and a read-only memory (ROM), and stores data and programs executed by the processor 701. The storage device 703 may include a storage medium such as a semiconductor memory and a hard disk.

[0140] The network interface 704 is a wired communication interface for connecting the server 700 to a wired communication network 705. The wired communication network 705 may be a core network such as an evolved packet core (EPC) or a packet data network (PDN) such as the Internet.

[0141] The bus 706 connects the processor 701, the memory 702, the storage device 703, and the network interface 704 to each other. The bus 706 may include two or more buses each having a different speed (such as a high-speed bus and a low-speed bus).

[0142] In the server 700 shown in FIG18 , the model acquisition unit 101 and blockchain unit 102 described with reference to FIG1 , the spectrum allocation unit 103 described with reference to FIG6 , the communication unit 201 and determination unit 202 described with reference to FIG12 , and the determination unit 301, communication unit 302, and evaluation unit 303 described with reference to FIG14 can be implemented by a processor 701. For example, the processor 701 can implement the acquisition and storage of the primary user behavior model on the blockchain by executing the functions of the model acquisition unit 101 and blockchain unit 102; implement the determination of a spectrum allocation scheme for a secondary user based on the primary user behavior model and the model scoring of the primary user behavior model by executing the functions of the spectrum allocation unit 103; implement the determination of a spectrum allocation scheme for a secondary user based on the primary user behavior model and the model scoring of the primary user behavior model by executing the functions of the communication unit 201 and determination unit 202; and implement the model scoring of the primary user behavior model by executing the functions of the determination unit 301, communication unit 302, and evaluation unit 303.

[0143] The basic principles of the present invention are described above in conjunction with specific embodiments. However, it should be pointed out that those skilled in the art will understand that all or any steps or components of the methods and devices of the present invention can be implemented in any computing device (including a processor, storage medium, etc.) or a network of computing devices in the form of hardware, firmware, software, or a combination thereof. This can be achieved by those skilled in the art using their basic circuit design knowledge or basic programming skills after reading the description of the present invention.

[0144] Furthermore, the present invention also provides a program product storing machine-readable instruction codes, which, when read and executed by a machine, can execute the method according to the embodiment of the present invention.

[0145] Accordingly, the storage medium for carrying the program product storing the machine-readable instruction code is also included in the disclosure of the present invention, including but not limited to a floppy disk, an optical disk, a magneto-optical disk, a memory card, a memory stick, and the like.

[0146] When the present invention is implemented through software or firmware, the programs constituting the software are installed from a storage medium or a network to a computer with a dedicated hardware structure (such as the general-purpose computer 1900 shown in Figure 19). When various programs are installed on the computer, it can perform various functions, etc.

[0147] In FIG19 , a central processing unit (CPU) 1901 executes various processes according to a program stored in a read-only memory (ROM) 1902 or a program loaded from a storage section 1908 to a random access memory (RAM) 1903. In the RAM 1903, data required when the CPU 1901 executes various processes, etc., is also stored as needed. The CPU 1901, the ROM 1902, and the RAM 1903 are connected to each other via a bus 1904. An input / output interface 1905 is also connected to the bus 1904.

[0148] The following components are connected to the input / output interface 1905: an input section 1906 (including a keyboard, a mouse, etc.), an output section 1907 (including a display such as a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and speakers, etc.), a storage section 1908 (including a hard disk, etc.), and a communication section 1909 (including a network interface card such as a LAN card, a modem, etc.). The communication section 1909 performs communication processing via a network such as the Internet. A drive 1910 may also be connected to the input / output interface 1905 as needed. Removable media 1911 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc. are installed in the drive 1910 as needed, so that computer programs read therefrom are installed in the storage section 1908 as needed.

[0149] In the case of realizing the above-described series of processing by software, a program constituting the software is installed from a network such as the Internet or a storage medium such as the removable medium 1911 .

[0150] It should be understood by those skilled in the art that such storage media is not limited to the removable medium 1911 shown in FIG. 19 , which stores the program and is distributed separately from the device to provide the program to the user. Examples of the removable medium 1911 include magnetic disks (including floppy disks (registered trademark)), optical disks (including compact disk read-only memories (CD-ROMs) and digital versatile disks (DVDs)), magneto-optical disks (including minidiscs (MDs) (registered trademark)), and semiconductor memories. Alternatively, the storage medium may be the ROM 1902, a hard disk included in the storage section 1908, or the like, in which the program is stored and distributed to the user together with the device containing the program.

[0151] It should also be noted that in the apparatus, method, and system of the present invention, each component or step can be decomposed and / or recombined. Such decomposition and / or recombination should be considered equivalent solutions of the present invention. Furthermore, the steps of performing the above series of processes can naturally be performed in chronological order according to the order described, but do not necessarily need to be performed in chronological order. Certain steps can be performed in parallel or independently of each other.

[0152] Finally, it should be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus. Furthermore, in the absence of further limitations, an element defined by the phrase "comprises a ..." does not preclude the presence of additional identical elements in the process, method, article, or apparatus that includes the element.

[0153] Although the embodiments of the present invention have been described in detail above with reference to the accompanying drawings, it should be understood that the embodiments described above are merely illustrative of the present invention and are not intended to limit the present invention. It will be apparent to those skilled in the art that various modifications and variations may be made to the embodiments described above without departing from the spirit and scope of the present invention. Therefore, the scope of the present invention is limited solely by the appended claims and their equivalents.

Claims

1. An electronic device for a spectrum management device, comprising: The processing circuit is configured to: A primary user behavior model for representing the primary user's behavior of using spectrum resources is obtained through federated learning; as well as The master user behavior model is stored on a model blockchain.

2. The electronic device according to claim 1, wherein The spectrum management device serves as a federated learning server, and the processing circuit is configured to determine one or more other spectrum management devices as federated learning clients participating in the federated learning, and aggregate the local primary user behavior models trained by each federated learning client to obtain a global model as the primary user behavior model.

3. The electronic device according to claim 2, wherein The processing circuit is configured to send a federated learning request to other spectrum management devices within a predetermined area, receive a federated learning response from other spectrum management devices that agree to participate in the federated learning, and determine a spectrum management device as the federated learning client among the other spectrum management devices from which the federated learning response is received.

4. The electronic device according to claim 3, wherein The federated learning request includes one or more of the following: a federated learning indication, an identifier of a primary user, a range of a prohibited area or a protected area of ​​the primary user, a location of the primary user, and an analysis identifier for identifying the primary user behavior model training.

5. The electronic device according to claim 4, wherein The processing circuit is further configured to store the master user behavior model corresponding to the analysis identifier on the model blockchain. The electronic device according to claim 3 , wherein: The processing circuit is further configured to determine the predetermined area based on one or more of the following: a location of the spectrum management device, a location of a primary user, and a range of a prohibited area or a protected area of ​​the primary user.

7. The electronic device according to claim 3, wherein: The processing circuit is further configured to send a notification of starting a training process to a spectrum management device determined as the federated learning client, and the federated learning client, in response to the notification, performs training of a local primary user behavior model and stores the model on the model blockchain.

8. The electronic device according to claim 7, wherein: The processing circuit is further configured to: In each round of training of the federated learning, the local primary user behavior models trained by each federated learning client are retrieved from the model blockchain for aggregation and the aggregated global model is stored on the model blockchain. Each federated learning client verifies the global model using the consensus mechanism of the blockchain, obtains the global model from the model blockchain, and conducts the next round of training based on the global model. as well as Once a predetermined condition is met, the federated learning is stopped, and the final global model is used as the main user behavior model.

9. The electronic device according to claim 8, wherein: The predetermined conditions include that the obtained global model reaches a predetermined accuracy or has been trained for a predetermined number of rounds.

10. The electronic device according to claim 2, wherein The processing circuit is configured to initiate the federated learning in response to one or more of the following: a primary user status update, a model accuracy requirement update, or a periodic trigger.

11. The electronic device according to claim 2, wherein: The processing circuit is further configured to monitor the status of each federated learning client and delete the federated learning client when the local primary user behavior model of the federated learning client is not updated in time or there are other abnormalities, and The processing circuit is further configured to, when an existing federated learning client is deleted, send a federated learning request to other spectrum management devices that have not participated in the federated learning, so as to determine a new federated learning client.

12. The electronic device according to claim 1, wherein The processing circuit is also configured to receive a federated learning request from other spectrum management devices that have model aggregation capabilities and serve as federated learning servers, and based on the federated learning request, determine whether the coverage of the secondary user devices and spectrum sensing devices managed by itself overlaps with the scope of the prohibited area or protection area of ​​the primary user, and send a response agreeing to participate in the federated learning if there is an overlap.

13. The electronic device according to claim 1, wherein The processing circuit is further configured to train the local primary user behavior model by using a machine learning algorithm based on the protection requirements of the primary user and the measurement and perception information of the secondary user.

14. The electronic device according to claim 1, wherein The processing circuit is further configured to determine a spectrum allocation scheme for secondary users to use the spectrum resources of the primary user based on the primary user behavior model and relevant information of the secondary user.

15. The electronic device according to claim 14, wherein The processing circuit is further configured to provide spectrum usage information in the spectrum allocation plan to corresponding secondary users, and store the spectrum usage information confirmed by the secondary users in a spectrum blockchain.

16. The electronic device according to claim 14, wherein The relevant information of the secondary user includes one or more of the following: registration information of the secondary user, spectrum use request of the secondary user, spectrum sensing result of the spectrum sensing device, and current spectrum use information of the secondary user.

17. The electronic device according to claim 14, the processing circuit is further configured to use a multi-agent deep reinforcement learning algorithm to determine an optimized spectrum allocation scheme by minimizing mutual interference between secondary users and protecting the primary user.

18. The electronic device according to claim 14, wherein The processing circuit is further configured to obtain a performance analysis report of the secondary user using the spectrum resources from the secondary user, and evaluate the accuracy of the determined available spectrum resources of the secondary user based on the performance analysis report to obtain an analysis score. The processing circuit is further configured to perform a model score on the accuracy of the primary user behavior model based on the analysis score, and record the model score in the model blockchain in association with the primary user behavior model after verification by a consensus mechanism.

19. The electronic device according to claim 18, wherein The processing circuit is configured to initiate a scoring credibility request to other spectrum management devices in the network and receive a scoring credibility response from other spectrum management devices. The scoring credibility response is a response made by other spectrum management devices in response to the scoring credibility request according to predetermined rules regarding whether the model score is credible.

20. The electronic device according to claim 19, wherein The score credibility request includes the model score, and the other spectrum management apparatus determines whether the model score is credible according to a difference between the model score and a locally stored related model score.

21. The electronic device according to claim 20, wherein The processing circuit is further configured to record the model score in the model blockchain if more than a predetermined proportion of the received responses indicate that the model score is credible; otherwise, not record the model score.

22. The electronic device according to claim 18, wherein The processing circuit is further configured to re-initiate federated learning for the primary user behavior model when the model score for the primary user behavior model is lower than a predetermined level.

23. An electronic device for a spectrum management device, comprising: The processing circuit is configured to: Obtaining a primary user behavior model obtained through federated learning for representing the behavior of the primary user in using spectrum resources; as well as A spectrum allocation scheme for secondary users to use the spectrum resources of the primary user is determined based on the primary user behavior model and relevant information of the secondary user.

24. The electronic device according to claim 23, wherein The processing circuit is further configured to provide spectrum usage information in the spectrum allocation plan to corresponding secondary users, and store the spectrum usage information confirmed by the secondary users in a spectrum blockchain.

25. The electronic device according to claim 23, wherein The relevant information of the secondary user includes one or more of the following: registration information of the secondary user, spectrum use request of the secondary user, spectrum sensing result of the spectrum sensing device, and current spectrum use information of the secondary user.

26. The electronic device according to claim 23, wherein The processing circuit is also configured to use a multi-agent deep reinforcement learning algorithm to determine an optimized spectrum allocation scheme by minimizing mutual interference between secondary users and protecting the primary user.

27. The electronic device according to claim 23, wherein The processing circuit is further configured to obtain a performance analysis report of the secondary user using the spectrum resources from the secondary user, evaluate the accuracy of the determined available spectrum resources of the secondary user based on the performance analysis report, and obtain an analysis score. The processing circuit is further configured to perform a model score on the accuracy of the primary user behavior model based on the analysis score, and record the model score in association with the primary user behavior model.

28. The electronic device according to claim 27, wherein The processing circuit is further configured to re-initiate federated learning for the primary user behavior model when the model score for the primary user behavior model is lower than a predetermined level.

29. The electronic device according to claim 23, wherein The spectrum management device serves as a federated learning server, and the processing circuit is configured to obtain the primary user behavior model as follows: determining one or more other spectrum management devices as federated learning clients participating in the federated learning, and aggregating the local primary user behavior models trained by each federated learning client to obtain a global model as the primary user behavior model.

30. The electronic device according to claim 29, wherein The processing circuit is configured to: In each round of training of the federated learning, the trained local primary user behavior models are obtained from each federated learning client for aggregation, and the aggregated global model is sent to each federated learning client, so that each federated learning client performs the next round of training based on the received global model; as well as Once a predetermined condition is met, the federated learning is stopped, and the final global model is used as the main user behavior model.

31. The electronic device according to claim 29, wherein The processing circuit is configured to record the local primary user behavior model, the global model, and the primary user behavior model on a model blockchain for sharing among multiple spectrum management devices.

32. An electronic device for a spectrum management device, comprising: The processing circuit is configured to: determining, based at least on a primary user behavior model representing a primary user's behavior of using spectrum resources, a spectrum allocation scheme for secondary users to use the primary user's spectrum resources; Obtaining, from a secondary user, a performance analysis report on the secondary user's use of the spectrum resources of the primary user allocated thereto, and evaluating accuracy of the determined available spectrum resources of the secondary user based on the performance analysis report to obtain an analysis score; as well as Based on the analysis score, a model score is performed on the accuracy of the main user behavior model.

33. The electronic device according to claim 32, wherein The processing circuit is further configured to record the model score in association with the master user behavior model.

34. The electronic device according to claim 33, wherein The processing circuit is further configured to enable the spectrum management apparatus to retrain the primary user behavior model when the model score of the primary user behavior model is lower than a predetermined level.

35. The electronic device according to claim 32, wherein The main user behavior model is obtained through federated learning.

36. The electronic device according to claim 32, wherein The master user behavior model and the model score are stored on a model blockchain to be shared among multiple spectrum management devices.

37. The electronic device according to claim 32, wherein The processing circuit is further configured to store the spectrum usage information confirmed by the secondary user in the spectrum blockchain.

38. A method for a spectrum management device, comprising: A primary user behavior model for representing the primary user's behavior of using spectrum resources is obtained through federated learning; as well as The master user behavior model is stored on a model blockchain.

39. A method for a spectrum management device, comprising: Obtaining a primary user behavior model obtained through federated learning for representing the behavior of the primary user in using spectrum resources; as well as A spectrum allocation scheme for secondary users to use the spectrum resources of the primary user is determined based on the primary user behavior model and relevant information of the secondary user.

40. A method for a spectrum management device, comprising: determining, based at least on a primary user behavior model representing a primary user's behavior of using spectrum resources, a spectrum allocation scheme for secondary users to use the primary user's spectrum resources; Obtaining, from a secondary user, a performance analysis report on the secondary user's use of the spectrum resources of the primary user allocated thereto, and evaluating accuracy of the determined available spectrum resources of the secondary user based on the performance analysis report to obtain an analysis score; as well as Based on the analysis score, a model score is performed on the accuracy of the main user behavior model.

41. A computer-readable storage medium having computer-executable instructions stored thereon, which, when executed by a processor, causes the processor to execute the method for a spectrum management device according to any one of claims 38 to 40.

Images