Electronic device, communication method, and storage medium in a wireless communication system

Abstract

The present disclosure relates to electronic devices, communication methods, and storage media in a wireless communication system. An electronic device for a low-frequency base station includes processing circuitry configured to: obtain a channel state information (CSI) matrix based on a reference signal received from a user equipment via a low-frequency communication; determine, from a plurality of high-frequency base stations, a candidate high-frequency base station suitable for a high-frequency communication with the user equipment based on the CSI matrix; determine access assistance information associated with the candidate high-frequency base station; and transmit the access assistance information to the user equipment.

Description

Technical Field

[0001] This disclosure relates to electronic devices, communication methods, and storage media in wireless communication systems, and more specifically, to electronic devices, communication methods, and storage media that utilize low-frequency communication to assist millimeter-wave communication access in a high-low frequency hybrid network architecture. Background Technology

[0002] With the continuous increase in wireless communication users and the growing demands and service volumes, mobile communication data volume is experiencing explosive growth. Current mobile communication, limited by the bandwidth of microwave frequency bands, can no longer meet user service needs. Therefore, developing broadband mobile communication utilizing higher frequency bands has become a hot trend. Millimeter waves can greatly enrich available spectrum resources, meaning wider bandwidth and faster transmission rates. Furthermore, according to antenna theory, the antenna size used in millimeter-wave communication is also on the order of millimeters, allowing hundreds or even thousands of millimeter-wave antennas to be placed in a small space, which is beneficial for the application of Massive MIMO technology in practical systems. Therefore, millimeter-wave technology has become one of the key technologies in 5G wireless communication systems.

[0003] On the other hand, millimeter-wave communication suffers from drawbacks such as limited coverage and excessive power consumption. Beamforming technology can be used to create directional spatial beams, concentrating energy in a specific spatial direction to combat channel path fading and thus expand coverage. Furthermore, a hybrid high- and low-frequency deployment topology can be utilized. Low-frequency base stations (e.g., traditional LTE base stations) can act as anchor points to achieve stable, wide-area coverage and provide low-rate transmission services to user equipment (UEs), while switching to high-frequency base stations (e.g., millimeter-wave base stations) is considered only when the UE requires high-rate data transmission. This hybrid high- and low-frequency deployment combines the advantages of both low-frequency and high-frequency communication.

[0004] However, when deciding which millimeter-wave base station to connect to, the UE cannot determine which base station is best suited for connection because the radio channel quality between the UE and each base station is unknown. Searching for each base station sequentially would result in inefficient access. Even in the worst-case scenario, if the channel quality of any millimeter-wave base station is insufficient for communication requirements, prematurely activating the millimeter-wave communication module would lead to wasted power.

[0005] Therefore, in wireless communication systems using applications such as millimeter waves, there is a need for high-frequency communication access methods. Summary of the Invention

[0006] In response to the problems mentioned above and others, various aspects of this disclosure provide solutions suitable for efficient access to high-frequency base stations, such as millimeter-wave base stations, in high- and low-frequency hybrid network architectures.

[0007] A brief overview of this disclosure is given below to provide a basic understanding of some aspects of it. However, it should be understood that this overview is not an exhaustive summary of this disclosure. It is not intended to identify key or essential parts of this disclosure, nor is it intended to limit the scope of this disclosure. Its purpose is merely to present certain concepts of this disclosure in a simplified form as a prelude to the more detailed description that follows.

[0008] According to one aspect of this disclosure, an electronic device for a low-frequency base station is provided, including processing circuitry configured to: acquire a channel state information (CSI) matrix based on a reference signal received from a user equipment via low-frequency communication; determine, using a neural network, candidate high-frequency base stations suitable for high-frequency communication with the user equipment from a plurality of high-frequency base stations based on the CSI matrix; determine access assistance information associated with the candidate high-frequency base stations; and transmit the access assistance information to the user equipment.

[0009] According to one aspect of this disclosure, an electronic device for a user equipment is provided, including processing circuitry configured to: transmit a reference signal to a low-frequency base station via a low-frequency link for the low-frequency base station to acquire a channel state information (CSI) matrix; receive access assistance information associated with a candidate high-frequency base station determined by the low-frequency base station, wherein the candidate high-frequency base station is a high-frequency base station suitable for high-frequency communication with the user equipment, determined by the low-frequency base station based on the CSI matrix using a neural network; and access the candidate high-frequency base station using the access assistance information.

[0010] According to one aspect of this disclosure, an electronic device for a high-frequency base station is provided, including processing circuitry configured to: receive from a low-frequency base station an identification code of a user equipment and information about a beam of the high-frequency base station that can be used by the user equipment, wherein the beam is determined by the low-frequency base station by inputting a CSI matrix obtained based on a reference signal transmitted by the user equipment via a low-frequency link into a neural network; and establish high-frequency communication with the user equipment using the beam.

[0011] According to one aspect of this disclosure, a method for training a neural network is provided, comprising: receiving a reference signal transmitted via low-frequency communication from a user equipment; obtaining a channel state information (CSI) matrix based on the reference signal; receiving identification information of a high-frequency base station with which the user equipment is communicating at high frequency; and performing deep learning to determine parameters of the neural network by using the CSI matrix as input and the identification information of the high-frequency base station as output.

[0012] According to one aspect of this disclosure, a communication method for a low-frequency base station is provided: obtaining a channel state information (CSI) matrix based on a reference signal received from a user equipment via low-frequency communication; determining, using a neural network, candidate high-frequency base stations suitable for high-frequency communication with the user equipment from a plurality of high-frequency base stations based on the CSI matrix; determining access assistance information associated with the candidate high-frequency base stations; and sending the access assistance information to the user equipment.

[0013] According to one aspect of this disclosure, a communication method for a user equipment is provided: transmitting a reference signal to a low-frequency base station via low-frequency communication for the low-frequency base station to acquire a channel state information (CSI) matrix; receiving access assistance information associated with a candidate high-frequency base station determined by the low-frequency base station, wherein the candidate high-frequency base station is a high-frequency base station suitable for high-frequency communication with the user equipment, determined by the low-frequency base station using a neural network based on the CSI matrix; and accessing the candidate high-frequency base station using the access assistance information.

[0014] According to one aspect of this disclosure, a communication method for a high-frequency base station is provided: receiving from a low-frequency base station an identification code of a user equipment and information about a beam of the high-frequency base station that can be used by the user equipment, wherein the beam is determined by the low-frequency base station by inputting a CSI matrix obtained based on a reference signal transmitted by the user equipment via a low-frequency link into a neural network; and establishing high-frequency communication with the user equipment using the beam.

[0015] According to another aspect of this disclosure, a non-transitory computer-readable storage medium is provided storing executable instructions that, when executed, implement any of the methods described above. Attached Figure Description

[0016] This disclosure can be better understood by referring to the detailed description given below in conjunction with the accompanying drawings, in which the same or similar reference numerals are used throughout the drawings to denote the same or similar elements. All the drawings, together with the following detailed description, are incorporated in and form a part of this specification, and are used to further illustrate embodiments of this disclosure and explain the principles and advantages of this disclosure. Wherein:

[0017] Figure 1 This is a simplified diagram illustrating the architecture of an NR communication system;

[0018] Figure 2A and 2B The NR radio protocol architectures for the user plane and control plane are illustrated separately.

[0019] Figure 3 This is a schematic diagram illustrating a high- and low-frequency hybrid network architecture;

[0020] Figure 4 This illustrates the frame structure used in 5G NR;

[0021] Figure 5 This is a schematic diagram illustrating the relationship between the beam and the synchronization signal;

[0022] Figure 6 The time-frequency structure of the synchronization signal block (SSB) in 5G NR is illustrated;

[0023] Figure 7 This illustrates the relationship between the amplitude and angle of the angle-domain CSI;

[0024] Figure 8 The relationship between path loss and distance between the UE and the base station is illustrated.

[0025] Figure 9 An example of a convolutional neural network (CNN) according to this disclosure is illustrated;

[0026] Figure 10 An example of a convolution operation using a convolution kernel is shown;

[0027] Figure 11 An example of pooling processing is shown;

[0028] Figure 12 Another example of a convolutional neural network (CNN) according to this disclosure is illustrated;

[0029] Figure 13 An example of a communication process for high-frequency communication access according to this disclosure is illustrated;

[0030] Figure 14 The communication flowchart for initial access is illustrated;

[0031] Figures 15A-15B Another example of a communication process for high-frequency communication access according to this disclosure is illustrated;

[0032] Figure 16 Another example of a communication process for high-frequency communication access according to this disclosure is illustrated;

[0033] Figure 17An example of a communication flow for collecting training data according to this disclosure is illustrated;

[0034] Figure 18 and Figure 19 Examples of CNNs according to this disclosure are illustrated;

[0035] Figures 20A-20C A simulation diagram illustrating the effect of high-frequency communication access according to this disclosure is shown;

[0036] Figures 21A-21B An electronic device for a low-frequency base station and a communication method thereof according to the present disclosure are illustrated;

[0037] Figures 22A-22B An electronic device for a UE and a communication method thereof according to the present disclosure are illustrated;

[0038] Figures 23A-23B An electronic device for a high-frequency base station and a communication method thereof according to the present disclosure are illustrated;

[0039] Figure 24 A first example of an illustrative configuration of a base station according to this disclosure is shown;

[0040] Figure 25 A second example of an illustrative configuration of a base station according to this disclosure is shown;

[0041] Figure 26 An illustrative configuration example of a smartphone according to this disclosure is shown;

[0042] Figure 27 An illustrative configuration example of a car navigation device according to this disclosure is shown.

[0043] The features and aspects of this disclosure will become clear from the following detailed description taken in conjunction with the accompanying drawings. Detailed Implementation

[0044] Various exemplary embodiments of this disclosure will be described in detail below with reference to the accompanying drawings. For clarity and brevity, not all features of the embodiments are described in this specification. However, it should be noted that many implementation-specific settings can be made in implementing embodiments of this disclosure to meet, for example, those constraints related to the device and services, and these constraints may vary depending on the implementation. Furthermore, it should be understood that while development work may be complex and time-consuming, such development is merely a routine task for those skilled in the art who benefit from this disclosure.

[0045] Furthermore, it should be noted that, in order to avoid obscuring this disclosure with unnecessary details, some figures only show processing steps and / or device structures that are closely related to at least the technical content of this disclosure, while in other figures, existing processing steps and / or device structures are additionally shown for better understanding of this disclosure.

[0046] Exemplary embodiments and application examples according to this disclosure will be described in detail with reference to the accompanying drawings. The following description of exemplary embodiments is merely illustrative and is not intended to limit this disclosure or its applications in any way.

[0047] For ease of explanation, various aspects of this disclosure will be described below in the context of 5G NR. However, it should be noted that this is not a limitation on the scope of this disclosure; one or more aspects of this disclosure can also be applied to existing wireless communication systems such as 4G LTE / LTE-A or various future wireless communication systems. The architectures, entities, functions, processes, etc., mentioned in the following description can be found in NR or other communication standards.

[0048] [Overview]

[0049] Figure 1 This is a simplified diagram illustrating the architecture of a 5G NR communication system. (For example...) Figure 1 As shown, on the network control side, the NR communication system's radio access network (NG-RAN) nodes include gNBs and ng-eNBs. The gNB is a newly defined node in the 5G NR communication standard, connected to the 5G core network (5GC) via the NG interface, and provides NR user plane and control plane protocols for termination with terminal equipment (also known as "user equipment," hereinafter referred to as "UE"). The ng-eNB is a node defined for compatibility with 4G LTE communication systems. It can be an upgrade of the LTE radio access network's evolved Node B (eNB), connected to the 5G core network via the NG interface, and provides Evolved Universal Terrestrial Radio Access (E-UTRA) user plane and control plane protocols for termination with the UE. An Xn interface exists between NG-RAN nodes (e.g., gNBs and ng-eNBs) to facilitate inter-node communication. Hereinafter, gNBs and ng-eNBs are collectively referred to as "base stations."

[0050] However, it should be noted that the term "base station" as used in this disclosure is not limited to the two types of nodes mentioned above, but rather refers to an example of a control device in a wireless communication system, encompassing the full breadth of its usual meaning. For example, in addition to the gNB and ng-eNB specified in the 5G communication standard, depending on the scenario in which the technical solutions of this disclosure are applied, a "base station" can also be, for example, an eNB in ​​an LTE or LTE-A communication system, a remote radio head unit (RRH), a wireless access point, a relay node, a drone control tower, a control node in an automated factory, or a communication device or element thereof performing similar control functions. Application examples of base stations will be described in detail in later sections.

[0051] Furthermore, the term "UE" as used in this disclosure has the full breadth of its usual meaning, encompassing various terminal devices or in-vehicle devices that communicate with a base station. As examples, a UE can be a terminal device or component thereof, such as a mobile phone, laptop, tablet, in-vehicle communication device, drone, sensor and actuator in an automated factory, etc. Application examples of UEs will be described in detail in later sections.

[0052] Next, combine Figure 2A and 2B To describe for Figure 1 The NR radio protocol architecture of base stations and UEs. Figure 2A The radio protocol stack for the user plane of the UE and base station is shown. Figure 2B The radio protocol stack for the control plane of the UE and base station is shown.

[0053] Layer 1 (L1) of the radio protocol stack is the lowest layer, also known as the physical layer. L1 implements various physical layer signal processing functions to provide transparent signal transmission. For example, when transmitting data, L1 performs a series of physical layer processing on user data from the MAC layer, such as Cyclic Redundancy Check (CRC), channel coding, rate matching, scrambling, modulation, precoding, and resource mapping, to map it to the transmission channel. Conversely, when receiving data, it can perform a series of inverse processing.

[0054] Layer 2 (L2) sits above the physical layer and is responsible for managing the radio link between the UE and the base station. In the user plane, L2 includes the Medium Access Control (MAC) sublayer, the Radio Link Control (RLC) sublayer, the Packet Data Convergence Protocol (PDCP) sublayer, and the Service Data Adaptation Protocol (SDAP) sublayer. Additionally, in the control plane, L2 includes the MAC sublayer, the RLC sublayer, and the PDCP sublayer. The relationships between these sublayers are as follows: the physical layer provides transport channels for the MAC sublayer, such as the Physical Uplink Shared Channel (PUSCH), Physical Uplink Control Channel (PUCCH), Physical Random Access Channel (PRACH), Physical Downlink Shared Channel (PDSCH), Physical Downlink Control Channel (PDCCH), and Physical Broadcast Channel (PBCH); the MAC sublayer provides logical channels for the RLC sublayer; the RLC sublayer provides RLC channels for the PDCP sublayer; and the PDCP sublayer provides radio bearers for the SDAP sublayer.

[0055] In the control plane, the UE and base station also include a Layer 3 (L3) Radio Resource Control (RRC) sublayer. The RRC sublayer is responsible for acquiring radio resources (i.e., radio bearers) and configuring the lower layers using RRC signaling. Additionally, the Non-Access Stratum (NAS) control protocol in the UE performs functions such as authentication, mobility management, and security control.

[0056] In 5G NR, both downlink and uplink transmissions are organized into frames. Figure 4 A diagram illustrating the frame structure in a 5G communication system is shown. As a fixed architecture compatible with LTE / LTE-A, NR frames also have a length of 10ms, consisting of two 5ms half-frames, and further comprising ten equal-sized subframes, each 1ms long. Unlike LTE / LTE-A, the frame structure in NR has a flexible architecture that depends on the subcarrier spacing. Each subframe has configurable... There are several time slots, such as 1, 2, 4, 8, and 16. Each time slot also has configurable... Each time slot contains 14 consecutive OFDM symbols for a normal cyclic prefix and 12 consecutive OFDM symbols for an extended cyclic prefix. In the frequency domain, each time slot comprises several resource blocks, and each resource block contains 12 consecutive subcarriers in the frequency domain. Therefore, resource elements (REs) in a time slot can be represented using a resource grid, such as... Figure 4 As shown in the image.

[0057] 5G NR communication systems consider three major application scenarios: enhanced mobile broadband (eMBB), massive machine-type communications (mMTC), and ultra-reliable low-latency communications (URLLC). These feature wider bandwidth (e.g., greater than 1Gbps), more user access (1 million connections per square kilometer), and lower latency (less than 1 millisecond), all of which rely on abundant spectrum resources. The spectrum available for NR can be divided into two parts: FR1, approximately 450MHz to 6GHz, also known as the sub-6GHz band; and FR2, approximately 24GHz to 52GHz. The electromagnetic wave wavelengths in FR2 are essentially at the millimeter level, belonging to the so-called millimeter wave category.

[0058] In NR communication systems, base stations and UEs have numerous antennas supporting massive MIMO, such as dozens, hundreds, or even thousands. By adjusting the antenna parameters, constructive interference of radio signals at certain angles and destructive interference at others can be created, forming a narrow beam to provide strong power coverage in a specific direction. This process is also known as beamforming. Base stations and UEs can use multiple beams in different directions to achieve cell coverage. Utilizing massive MIMO and beamforming techniques can overcome the excessive path fading inherent in millimeter waves.

[0059] Furthermore, hybrid low-frequency and high-frequency networking is a feasible option for overcoming some of the shortcomings of millimeter waves in practical communication systems. As used in this disclosure, "low frequency" refers to communication frequency bands lower than millimeter wave bands, such as those used in LTE or LTE-A communication systems, and the FR1 band (sub-6GHz band) used in NR communication systems. The advantages of low-frequency bands are their low frequency, strong diffraction capability, and good coverage, thus they can serve as basic coverage bands to enable rapid deployment of 5G networks. "High frequency" refers to communication frequency bands in or near the millimeter wave band, such as the FR2 band used in NR communication systems. The advantages of high-frequency bands are their ultra-large bandwidth, clean spectrum, and less interference, thus they can serve as capacity supplementation bands to support high-speed applications.

[0060] Figure 3 A schematic diagram of a high- and low-frequency hybrid network architecture is shown. (For example...) Figure 3 As shown, the communication network comprises a hybrid network of two or more layers, including macrocells and microcells / pecicells. Macrocells are deployed using low-frequency bands, providing stable coverage over a wide area through low-frequency base stations. Within the coverage area of ​​the low-frequency base stations, multiple (e.g., Figure 3 There are 5 (but not limited to) high-frequency base stations, each of which is deployed using high-frequency bands to achieve small-scale hotspot enhancement.

[0061] Under this network architecture, UEs can support dual connectivity. That is, in most cases, control signals and low-rate data transmission services are completed through low-frequency links with low-frequency base stations. Only when the UE has a need for high-rate data transmission will it be considered to switch to a connection with a high-frequency base station (e.g., a millimeter-wave base station). This is because it is difficult for UEs to communicate using millimeter waves due to power consumption and stability limitations, especially when the UE is a resource-constrained device.

[0062] When the UE decides whether to enable the high-frequency communication module, the channel conditions of each high-frequency base station are unknown to the UE. The UE does not know which high-frequency base station is most suitable for providing services, nor does it know whether the channel quality of the high-frequency base station meets the transmission requirements. The UE will sequentially detect the possible frequency points of each high-frequency base station's cell across all frequency ranges; that is, the UE "blindly searches" for available cells and attempts to access them.

[0063] Compared to the broadcast-based mechanism of 4G LTE, 5G NR uses a beam management mechanism for initial access. Specifically, such as... Figure 5 As shown, the gNB broadcasts synchronization signal bursts (SS Bursts) at regular periods T. Each SS Burst consists of one or more synchronization signal / physical broadcast channel blocks (SSBs), and each SSB corresponds to a beam with a different direction. Thus, the gNB can periodically use SS Bursts to scan all predetermined directions with the beam. Typically, an SS Burst can be transmitted within a 5ms time window (half a frame) and repeated at, for example, a 20ms period.

[0064] Figure 6 The time-frequency structure of the SSB in an NR communication system is shown. The SSB consists of the primary synchronization signal (PSS), the secondary synchronization signal (SSS), and the PBCH. Figure 6 As shown, in the time domain, each SSB occupies 4 consecutive OFDM symbols, and in the frequency domain, each SSB contains 240 consecutive subcarriers. The PSS and SSS occupy 1 OFDM symbol and 127 subcarriers, respectively, and the PBCH spans 3 OFDM symbols and 240 subcarriers. However, within one OFDM symbol ( Figure 6 The OFDM symbol 2) has a portion in the middle for embedding the SSS. Each SSB within an SS burst has a corresponding index (SSB_index), therefore, the SSB_index can also be used to identify the corresponding beam.

[0065] For example Figure 3 For multiple high-frequency base stations, each high-frequency base station can transmit SSBs with different frequency positions. REF This is indicated by the corresponding Global Synchronization Channel Number (GSCN). SS for all frequency ranges.REF The GSCN is shown in Table 1 below:

[0066] Table 1: SS REF Relationship with GSCN

[0067]

[0068] In the traditional 5G NR initial access mechanism, the UE blindly detects all possible SSB frequency locations within the frequency range used by its Public Land Mobile Network (PLMN). When the signal quality of the SSB (e.g., Reference Signal Received Power (RSRP)) meets the requirements, the UE attempts to access the cell of the high-frequency base station transmitting that SSB. However, the base station the UE accesses may not be the one that can provide it with the best high-frequency communication. Detecting all SSBs transmitted by high-frequency base stations would incur significant overhead and is inefficient.

[0069] This disclosure considers using low-frequency communication to maintain network connectivity to provide information that facilitates high-frequency communication access, thereby improving the efficiency and quality of initial access to high-frequency base stations.

[0070] In such Figure 3 In the high-low frequency hybrid network shown, the low-frequency base station can act as an anchor point to achieve a stable connection between the UE and the wireless communication network. The low-frequency base station can be, for example, an eNB in ​​a 4G communication system or an ng-eNB in ​​a 5G communication system. For the low-frequency link between the low-frequency base station and the UE, low-frequency reference signals can be used to assess the channel conditions. For example, the low-frequency base station can transmit downlink reference signals such as Channel State Information Reference Signal (CSI-RS). The UE obtains the channel state information of the downlink low-frequency channel by measuring the reference signal and feeds it back to the low-frequency base station. Alternatively, the UE can transmit uplink reference signals such as Sounding Reference Signal (SRS) or CSI-RS. The low-frequency base station obtains the channel state information of the uplink low-frequency channel by measuring the reference signal. In particular, when using time-division duplexing, the channel state information of both the uplink and downlink channels can be obtained from measurements in a single direction.

[0071] The CSI obtained by measuring the reference signal can be a three-dimensional complex matrix associated with the number of antennas of the base station, the number of antennas of the UE, and the number of subcarriers; it can also be called spatial domain CSI. The elements in the CSI matrix describe the conditions of the wireless transmission path from each antenna of the UE to each antenna of the low-frequency base station, such as signal scattering, fading, and other information.

[0072] The inventors of this disclosure have noted that, in addition to information about channel conditions, the CSI matrix actually implicitly contains the UE's location information. To illustrate this more clearly, Figure 7This diagram schematically illustrates the relationship between the amplitude and angle of the CSI for three UEs when the spatial domain CSI is transformed into the angular domain CSI, where the horizontal axis represents the angular range of 0 to π, and the vertical axis represents the amplitude value of the CSI. Figure 7 It is evident that CSI exhibits a significant peak in a certain direction, which is often the LOS direction in environments with a line-of-sight (LOS) path. In MIMO, the antenna response vector is a function of the channel angle of arrival (AoA) and the channel angle of departure (AoD), therefore CSI includes angular information such as AoA and AoD. Furthermore, Figure 8 A scatter plot is shown, with path loss on the vertical axis and the distance between the UE and the base station on the horizontal axis. From Figure 8 It can be seen that the amplitude attenuation of CSI is related to the distance between the UE and the base station. The greater the distance, the smaller the amplitude of CSI tends to be, which indicates that CSI contains distance information.

[0073] Considering that angle and distance can represent position in polar coordinates, low-frequency CSI contains the UE's position information. However, the UE position information hidden in the CSI matrix is ​​difficult to obtain using traditional methods because the limited number of low-frequency antennas restricts the performance of traditional angle estimation methods; simultaneously, from Figure 8 It can also be seen that the relationship between low-frequency CSI and millimeter-wave path attenuation is difficult to express directly with an explicit expression.

[0074] According to embodiments of this disclosure, a deep learning-based prediction model is introduced to extract the complex hidden relationship between CSI and UE location, and further predict out-of-band information to assist UE decision-making regarding high-frequency communication access. As an example, a convolutional neural network (CNN) is a widely used deep learning model that includes convolutional computation and a deep structure, possessing extremely strong non-linear fitting capabilities. Leveraging the non-linear fitting capabilities of, for example, CNNs, when a CSI matrix obtained via low-frequency communication is input into a CNN, information about the UE location hidden within the low-frequency CSI matrix can be extracted. CNNs have self-learning capabilities, allowing the parameters of the neural network to be determined through training on real data without the need for complex manual parameter design.

[0075] The out-of-band information obtained from the prediction model includes information about the most suitable high-frequency base station for providing services to the UE, information about the channel conditions between the UE and the high-frequency base station, information about the optimal beam used for communication between the high-frequency base station and the UE, etc. This information is beneficial to the UE's high-frequency communication access and is also referred to as "access assistance information" in this disclosure.

[0076] This disclosure provides various examples of deep learning-based prediction models. The following description primarily uses convolutional neural networks as an example; however, it is conceivable that this disclosure is not limited to convolutional neural networks, but can use other types of neural networks or any suitable deep learning model, as long as it can produce the desired output based on the same input.

[0077] [The First Example of a Convolutional Neural Network]

[0078] Figure 9 A schematic diagram illustrating the construction of a convolutional neural network based on the first example is shown. Typically, a CNN consists of five parts: convolutional layers, activation functions, fully connected layers, pooling layers, and batch normalization layers.

[0079] According to the first example, the input data of the convolutional neural network consists of the low-frequency CSI matrix obtained by the low-frequency base station based on the uplink reference signal and the location data of all candidate high-frequency base stations. The CSI matrix is ​​a three-dimensional complex matrix. The location data of the high-frequency base stations can be represented as a two-dimensional matrix. Generally, the location of the high-frequency base stations is fixed, so the location data of the high-frequency base stations can be predetermined and stored in the memory of the low-frequency base stations. When a high-frequency base station is added or removed, the location data stored in the low-frequency base station can be updated accordingly. There are various ways to represent the location data of the high-frequency base stations, such as relative position with reference to the low-frequency base station, absolute geographic coordinates, etc. The different representations of location data essentially only involve linear transformations and have no fundamental difference.

[0080] Batch Normalization (BN) Layers: Batch normalization layers normalize the input data matrix to a standard distribution with a mean of 0 and a variance of 1, thereby accelerating the convergence of the neural network. A CNN can include one or more batch normalization layers. For example, batch normalization layers can be used to batch normalize the input data fed into the CNN before it is fed into the convolutional layers, or they can be used to batch normalize the intermediate data of the neural network.

[0081] Convolutional layers: such as Figure 9 As shown, in a convolutional layer, a self-learning filter (i.e., a convolutional kernel) is convolved with a data matrix to extract hidden features from the input data. Since the size of the convolutional kernel is often much smaller than the data matrix, the kernel moves across the data matrix to traverse it; the distance moved is called the stride. Furthermore, to match the movement of the convolutional kernel, the data matrix may undergo edge expansion (i.e., edge padding). Convolutional kernels with different parameters are used to extract different features from the data matrix, and their corresponding convolutional outputs are called feature channels. To extract richer features, the number of feature channels gradually increases with the depth of the network layers. Figure 10 The diagram illustrates convolving the same data matrix with different convolution kernels, where the stride of the convolution is 2.

[0082] Activation Functions: The output of a convolutional layer is typically passed through an activation function before being input to the next layer. Activation functions are usually non-linear functions, thus introducing non-linear fitting capabilities into CNNs. Deep learning exhibits high performance precisely because it achieves high non-linearity through repeated non-linear transformations using multi-layered structures. Without activation functions to handle non-linear transformations and the network only including linear transformations, there would only be equivalent single-layer linear transformations regardless of the number of layers, and multiple layers would be useless. Clearly, as the number of layers increases, deep learning exhibits stronger non-linearity and higher performance.

[0083] Pooling layers: Pooling layers downsample the input matrix to reduce the amount of data and computation in the neural network. Pooling operations include max pooling and average pooling. Figure 11 Schematic diagrams of two different pooling processes are shown. For example... Figure 11 As shown, max pooling retains the maximum value of the data matrix, while average pooling retains the average value. Feature vectors obtained from different pooling layers can be combined into a single feature vector to facilitate subsequent network structure prediction of the output.

[0084] Fully connected layer: In a fully connected layer, the input feature vector is linearly fitted to obtain the output. Fully connected layers can control the output size, therefore they are commonly used to implement size transformations from extracted data features to output.

[0085] Convolutional neural networks can output the most suitable high-frequency base station (hereinafter referred to as "candidate high-frequency base station") for providing high-frequency communication services to the UE. This means that at the UE's current location, the high-frequency base station is likely to provide the high-frequency link connection with the best channel conditions. The output of the candidate high-frequency base station can be its identification information, such as an identifier used to uniquely identify the high-frequency base station—the high-frequency base station ID, including but not limited to, eNodeB ID, gNBID (which, together with the Physical Layer Cell ID (PCI), constitutes the NR Cell ID (NCI) that uniquely identifies the cell), or a simpler base station number, as long as the low-frequency base station can associate the high-frequency base station ID with the corresponding high-frequency base station one by one.

[0086] Considering that the predicted candidate high-frequency base station is selected from a limited number of alternative high-frequency base stations, therefore Figure 9 The classifier in the convolutional neural network corresponds to the candidate high-frequency base station.

[0087] To achieve multi-class prediction, features extracted from the low-frequency CSI matrix and high-frequency base station locations are transformed into an output of size equal to the number of candidate high-frequency base stations through a fully connected layer, and then activated by a function such as Softmax.

[0088]

[0089] The Softmax activation function normalizes the output of the fully connected layer into probabilities. Under hard decision, the convolutional neural network selects the classifier with the highest probability as the predicted output. This output representation provides a ranking of the predicted high-frequency base stations; the higher the probability, the more likely the user is to connect to that high-frequency base station. The convolutional neural network can output the probability of only one high-frequency base station, or multiple high-frequency base stations ranked by probability.

[0090] For the convolutional neural network in the first example, the prediction criterion for candidate high-frequency base stations can be the path loss of the high-frequency link between the UE and the high-frequency base station. In other words, the desired outcome is that the candidate high-frequency base station output by the convolutional neural network has the minimum path loss with the UE's wireless channel. In fact, the prediction criterion of the convolutional neural network depends on the principles for acquiring the training data; different training data may lead to different candidate high-frequency base stations in the output.

[0091] like Figure 9 As shown, convolutional neural networks can also output other access assistance information.

[0092] For example, a branch of the convolutional neural network (branch 1) can output the path loss value corresponding to the candidate high-frequency base station. Predicting the path loss value helps estimate whether the channel conditions of the high-frequency communication link between the UE and the candidate high-frequency base station meet the connection requirements. If the path loss value indicates that the radio channel does not meet the requirements, the UE can temporarily suspend the high-frequency communication module and wait for a better opportunity.

[0093] Since path loss is a continuous value, predicting path loss can be modeled as a regression problem. That is, the fully connected layer transforms the extracted features into a scalar, which is then directly output as the path loss value. To reduce the dynamic range of the path loss value, the path loss value predicted by the neural network can be expressed in dB.

[0094] For example, a branch of the convolutional neural network (branch 2) can output the optimal beam for a candidate high-frequency base station to provide high-frequency communication services to the UE. Generally, for the UE's current location, the optimal beam's AOD and AOA are closest to the channel direction, such as the LOS direction. The low-frequency base station can notify the candidate high-frequency base station of the predicted optimal beam, so that the high-frequency base station can use this beam to establish high-frequency communication with the UE, or further refine the beam used for the UE.

[0095] Since the beamforming codebook size of high-frequency base stations is limited, the optimal beam prediction by a convolutional neural network essentially involves selecting one from the finite number of beams available at the high-frequency base station. Therefore, this type of problem can also be modeled as a multi-classification problem, where the number of classifiers is equal to the number of candidate beams, and the classifier types are equal to the finite number of beams available at the high-frequency base station. The beams can be indexed using associated reference signals (e.g., SSBs).

[0096] Furthermore, a loss function is used to measure the difference between the predicted result and the target output. To reduce prediction error (i.e., loss), the gradient backpropagation algorithm is used to update the parameters of the convolutional neural network. As an example, the cross-entropy loss function can be used for multi-class prediction of candidate high-frequency base stations or optimal beams, and the Smooth-L1 loss function can be used for regression prediction of path loss values. When the neural network makes multiple predictions simultaneously, a linear weighting can be used to apply to the overall loss function, Loss, i.e.:

[0097] Loss=μ BS loss BS +μ beam loss beam +μ path loss path

[0098] Where loss BS loss beam loss path Let μ represent the loss function for the three prediction targets: candidate high-frequency base station, optimal beam, and path loss value. BS μ beam μ path These represent the corresponding linear weighting coefficients.

[0099] [A Second Example of a Convolutional Neural Network]

[0100] In the first example described above, the convolutional neural network, used as the prediction model, is used to predict the high-frequency base station with the minimum path loss as a candidate high-frequency base station suitable for providing high-frequency communication services to the UE. However, in reality, the transmit power of various high-frequency base stations often varies, and the high-frequency base station with the minimum path loss may not necessarily have the maximum signal received power at the UE. Since the UE usually selects the base station to access by comparing the received signal power, this may lead to a deviation between the predicted candidate high-frequency base station and the actual candidate high-frequency base station.

[0101] Therefore, the convolutional neural network in the second example takes into account this difference in transmit power, and the prediction criterion for candidate high-frequency base stations can be the received power (RSRP) of the synchronization signal at the UE. In other words, the desired outcome is that the synchronization signal transmitted by the candidate high-frequency base station output by the prediction model has the best signal quality when received by the UE.

[0102] Figure 12 A schematic diagram illustrating the construction of a convolutional neural network as an example of a prediction model according to the second embodiment is shown. The convolutional neural network comprises five parts: convolutional layers, activation functions, fully connected layers, pooling layers, and batch normalization layers. The following focuses on describing... Figure 9 The differences between the convolutional neural networks shown are as follows.

[0103] like Figure 12 As shown, in addition to the low-frequency CSI matrix and the location data of the high-frequency base stations, the input to the convolutional neural network also includes the transmit power of each high-frequency base station. Here, the transmit power can be the power of the SSB signal broadcast by the high-frequency base station. The location data of each high-frequency base station can be input into the neural network along with the transmit power as high-frequency base station information. Generally, the transmit power of high-frequency base stations does not change frequently, so it can be pre-stored at the low-frequency base stations, just like the location data.

[0104] Figure 12 The convolutional neural network outputs the candidate high-frequency base station ID in the sense of RSRP. Optionally, the convolutional neural network also includes a branch network (branch 1) that outputs the corresponding RSRP value and a branch network (branch 2) that outputs the optimal beam of the candidate high-frequency base station.

[0105] Predicting the RSRP value helps estimate whether the channel conditions of the high-frequency communication link between the UE and the candidate high-frequency base station meet the connection requirements. If the RSRP value indicates that the radio channel does not meet the requirements, the UE can temporarily suspend the high-frequency communication module and wait for a better opportunity.

[0106] Similar to the path loss value in the first embodiment, considering that RSRP is a continuous value, the prediction of RSRP can also be modeled as a regression problem. That is, the fully connected layer transforms the extracted features into a scalar, which is directly output as the RSRP value.

[0107] For the loss function, the cross-entropy loss function can be used for multi-class prediction of candidate high-frequency base stations or optimal beams, while the Smooth-L1 loss function can be used for regression prediction of RSRP values. When the neural network performs multiple predictions simultaneously, linear weighting can be used to affect the overall loss function, i.e.:

[0108] Loss=μ BS loss BS +μ beam loss beam +μ RSRP loss RSRP

[0109] Where loss BS lossbeam loss RSRP Let μ represent the loss function for the three prediction targets: candidate high-frequency base station, optimal beam, and RSRP value. BS μ beam μ RSRP These represent the corresponding linear weighting coefficients.

[0110] [Initial Access Procedure for High-Frequency Communication]

[0111] The following description, in conjunction with the accompanying diagram, describes the process of using prediction results from a deep learning-based prediction model to assist in the initial access of high-frequency communication.

[0112] Figure 13 This is an example illustrating the communication process for initial access to high-frequency communication. Figure 13 The communication process shown can begin with the UE sending an uplink reference signal (S1) to the low-frequency base station. The uplink reference signal can be, for example, SRS, CSI-RS, etc., and is sent to the low-frequency base station via the low-frequency link.

[0113] The UE's act of sending an uplink reference signal can occur when the UE has a high-speed data transmission requirement. When the UE needs to connect to a high-frequency base station, the UE can send a high-frequency link access request to a low-frequency base station. Figure 13 (Not shown in the image) to request information from a low-frequency base station that facilitates access to a high-frequency link. As an example, the UE can use pre-configured communication resources (e.g., time-frequency resource blocks) to send the uplink reference signal and the high-frequency link access request separately or together. As another example, the UE can first send a high-frequency link access request to a low-frequency base station. In response to this request, the low-frequency base station can allocate communication resources to the UE (e.g., via DCI), thereby allowing the UE to use the allocated communication resources to send the uplink reference signal.

[0114] In response to receiving an uplink reference signal, the low-frequency base station can acquire the CSI matrix (S2). The low-frequency base station can measure the reference signal received through its multiple antennas and estimate the CSI matrix based on the measurements. The CSI matrix can be obtained using conventional estimation methods, which will not be detailed here. The estimated CSI matrix can be in the form of a three-dimensional matrix as described above.

[0115] Then, the low-frequency base station uses the acquired CSI matrix to predict the optimal high-frequency base station for the UE (S3). Specifically, the low-frequency base station inputs the CSI matrix as input data into a pre-trained prediction model, for example... Figure 9 Or the convolutional neural network shown in Figure 12. For example, Figure 9The first example of the convolutional neural network shown in the diagram also includes location data of all available high-frequency base stations as input data, which can be pre-stored at low-frequency base stations. And for... Figure 12 The second example of the convolutional neural network shown in the diagram also includes the location data and transmit power of all available high-frequency base stations as input data, which can be pre-stored at low-frequency base stations. The convolutional neural network can determine candidate high-frequency base stations suitable for providing high-frequency communication services to the UE as output. For example, depending on the prediction model utilized by the low-frequency base stations, the candidate high-frequency base stations can be high-frequency base stations with minimum path loss for the high-frequency communication link with the UE, or high-frequency base stations with maximum received signal power at the UE.

[0116] The low-frequency base station can determine access assistance information associated with the candidate high-frequency base station (S4). While directly informing the UE of the candidate high-frequency base station's ID is a viable approach, preferably, the low-frequency base station can determine information that is more convenient for the UE to use when accessing the candidate high-frequency base station, which is beneficial for compatibility with traditional NR initial access procedures. In one example, the access assistance information determined by the low-frequency base station includes information that helps the UE identify the synchronization signal of the candidate high-frequency base station, such as the frequency position SS of the SSB broadcast by the candidate high-frequency base station. REF SSB frequency position SS REF It can also be indicated by GSCN.

[0117] The low-frequency base station can send determined access assistance information (e.g., GSCN) to the UE (S5). The UE can then use this access assistance information to access the candidate high-frequency base station (S6). In particular, when the UE's high-frequency communication module (e.g., millimeter-wave communication module) is normally turned off, the UE can turn on the high-frequency communication module upon receiving access assistance information associated with the candidate high-frequency base station.

[0118] The following is combined Figure 14 This describes the initial access process between the UE and the base station. When the UE powers on or needs to switch to a specific base station, it first needs to perform a cell search. Cell search is the process by which the UE obtains time and frequency synchronization with the cell and detects the physical layer cell ID of the cell.

[0119] At position 101, the UE performs cell search by receiving SSBs. According to a first embodiment of this disclosure, the UE can directly search for synchronization signals at frequency locations indicated in the access assistance information. At these frequency locations, the UE can receive SS bursts periodically transmitted by candidate high-frequency base stations, including one or more SSBs corresponding to different beams. Compared to traditional frequency-domain blind search, directly locating the SSBs of candidate high-frequency base stations is obviously more efficient.

[0120] When a candidate high-frequency base station broadcasts an SS burst containing two or more SSBs, the UE can detect whether the signal quality (e.g., RSRP) of each SSB meets a threshold requirement, or detect the SSB with the best signal quality. The UE can decode the SSBs that meet the threshold requirement for synchronization to the downlink timing, for example, through the following steps:

[0121] 1) Detect and decode the PSS in the SSB to obtain the transmission group.

[0122] 2) Based on the relative time domain position between SSS and PSS, detect and decode SSS to obtain the transmission group. And according to Obtain the Physical Layer Cell ID (PCI) of the corresponding cell, where

[0123] 3) In addition, by decoding PSS / SSS, the synchronization of symbols can be obtained, thereby indirectly obtaining the subcarrier spacing (SCS) and absolute frequency (SSB) of SSB.

[0124] 4) After obtaining the PCI, the location of the DMRS of the PBCH can be determined. For example, the location offset of the DMRS is PCI mod 4.

[0125] 5) By demodulating the DMRS and payload of the PBCH, the SSB index (i) can be obtained. SSB ) and half-frame information (n hf The UE can obtain 10ms frame synchronization.

[0126] After obtaining downlink cell synchronization, the UE can receive cell system information, such as the Master Information Block (MIB) and various System Information Blocks (SIBs), at appropriate locations in the downlink frame. The system information can be periodically broadcast by the base station through broadcast channels (such as the Broadcast Channel PBCH, the Shared Channel PDSCH, etc.) and can include information necessary for the UE to access the base station, such as random access related information.

[0127] Subsequently, in order to achieve uplink cell synchronization, the UE needs to undergo a random access procedure. For example... Figure 14As shown, at point 102, the UE can notify the base station of its access behavior by sending a random access preamble (e.g., included in MSG-1) to the candidate high-frequency base station. Sending the random access preamble enables the base station to estimate the uplink timing advance of the terminal device. At point 103, the base station can notify the UE of the aforementioned timing advance by sending a random access response (e.g., included in MSG-2). The UE can use this timing advance to achieve uplink cell synchronization. The random access response may also include uplink resource information, which the UE can use in operation 104. For contention-based random access procedures, at point 104, the UE can send its UE identifier and any other possible information (e.g., included in MSG-3) using the scheduled uplink resources. The base station can determine the contention resolution result using the UE identifier. At point 105, the base station can inform the UE of the contention resolution result (e.g., included in MSG-4). If the contention is successful, the UE successfully accesses the base station, and the random access procedure ends; otherwise, the UE needs to repeat the random access procedure from 102 to 105.

[0128] Once the random access procedure between the UE and the candidate high-frequency base station is completed, the UE establishes high-frequency communication with the candidate high-frequency base station and can conduct subsequent communication.

[0129] When in Figure 13 In step S3, if the neural network outputs more than one candidate high-frequency base station, the low-frequency base station can determine the access assistance information associated with each high-frequency base station (S4) and send it to the UE (S5). The UE can search for the corresponding SS bursts sequentially according to the priority of these candidate high-frequency base stations. If none of the SSBs in the SS burst of the candidate high-priority (i.e., the most probable) candidate high-frequency base station meet the threshold requirement, the UE can search for the SS burst of the candidate high-priority candidate high-frequency base station, and so on. If no SSB that meets the requirements is found, the initial access fails.

[0130] The following is combined Figure 15A and 15B To describe another example of the communication process of high-frequency communication access according to this disclosure. The main description will be related to... Figure 13 The differences in the communication process will be explained, and the other identical parts will not be described again.

[0131] like Figure 15AOr as shown in 15B, in S31, the low-frequency base station, using a convolutional neural network, can predict not only candidate high-frequency base stations but also the path loss value or RSRP value corresponding to the candidate high-frequency base stations, depending on whether the low-frequency base station uses the first or second example of the convolutional neural network. The predicted path loss value or RSRP value can be used to determine whether the predicted candidate high-frequency base station is worth accessing. The convolutional neural network used in S31 may include a branch network for predicting the path loss value or RSRP value (…). Figure 9 Or branch 1 in 12).

[0132] In one example, such as Figure 15A As shown, the low-frequency base station can compare the path loss value (or RSRP value) output by the convolutional neural network with a predetermined threshold (S40). If the path loss value is lower than the predetermined threshold (or the RSRP value exceeds the predetermined threshold), the low-frequency base station can assume that the predicted preferred high-frequency base station can provide a qualified high-frequency communication link, and determine the associated access assistance information in S4 to facilitate the initial access of the UE. The UE, in response to receiving the access assistance information, activates the high-frequency communication module and begins the initial access process. Alternatively, the low-frequency base station can generate an activation indication and send it to the UE along with the access assistance information (e.g., ...). Figure 15A (as shown in parentheses in S5) to instruct the UE to enable high-frequency communication mode and begin the initial access procedure.

[0133] Conversely, if the path loss value exceeds a predetermined threshold (or the RSRP value is lower than a predetermined threshold), the low-frequency base station may assume that the predicted candidate high-frequency base station may not be able to provide a suitable high-frequency communication link and will not proceed with the subsequent steps. In this case, the low-frequency base station can instead send an indication to the UE that there is currently no suitable high-frequency base station for access.

[0134] In another example, such as Figure 15B As shown, the low-frequency base station can send the determined access assistance information along with the path loss value (or RSRP value) to the UE (S51). The UE can compare the received path loss value (or RSRP value) with a predetermined threshold. If the path loss value is lower than the predetermined threshold (or the RSRP value exceeds the predetermined threshold), the UE can decide to activate the high-frequency communication mode and connect to the candidate high-frequency base station through the initial access procedure. Conversely, if the path loss value exceeds the predetermined threshold (or the RSRP value is lower than the predetermined threshold), the UE can decide that there is currently no suitable high-frequency base station for access and attempt access at a later time, such as by repeating the process. Figure 15B Steps S1 to S6 in the process.

[0135] It should be understood that the predetermined threshold used for comparison with path loss values ​​mentioned above can be an adjustable parameter. The predetermined threshold used by the low-frequency base station or UE can depend on various factors, such as: the UE's current battery level (the more battery the UE has remaining, the higher the predetermined threshold can be, thus increasing the chance of high-frequency access); the UE's connection preferences (e.g., the UE or its user can be set to access high-frequency base stations with better radio channel conditions (i.e., lower path loss); the transmission success rate of high-frequency communication (e.g., dynamically adjusting the predetermined threshold based on the result of the previous high-frequency access or the data transmission success rate of high-frequency communication); the urgency of the service (e.g., if the service requires immediate high-frequency communication, the UE can temporarily increase the predetermined threshold); and so on. Similarly, the predetermined threshold used for comparison with RSRP values ​​can also be an adjustable parameter, which the low-frequency base station or UE can adjust based on factors such as the UE's current battery level, connection preferences, the transmission success rate of high-frequency communication, and the urgency of the service. It should be understood that the predetermined threshold for path loss and the predetermined threshold for RSRP are different thresholds.

[0136] The following is combined Figure 16 To describe another example of the communication process of high-frequency communication access according to this disclosure. The main description will be related to... Figure 13 The differences in the communication process will be explained, and the other identical parts will not be described again.

[0137] like Figure 16 As shown, at S32, the low-frequency base station can use a convolutional neural network to predict the optimal beam for the UE from the candidate high-frequency base station. The optimal beam may correspond to the SSB in the SS burst transmitted by the candidate high-frequency base station. At this time, the access assistance information determined by the low-frequency base station in S4 may include information about the time-frequency resources of the SSB, such as the frequency location of the SSB (SS). REF ) and index (SSB_index).

[0138] As mentioned above, convolutional neural networks can output more than one (n>1) optimal beam based on the predicted probabilities. Therefore, in one example, according to the order of the probabilities of each beam from high to low, the low-frequency base station can determine the time-frequency resource information containing the corresponding n SSBs and send it to the UE. That is, these n SSBs have different priorities.

[0139] When a UE receives such access assistance information, it can directly detect the corresponding SSB and estimate its signal quality (e.g., RSRP). If the SSB meets the threshold requirements, it can continue to decode the SSB to attempt access to the candidate high-frequency base station, as described above. Figure 13If the SSB does not meet the threshold requirement, the UE can continue to detect other SSBs with lower priority at the same frequency location. If no SSB that meets the requirements is found, the search scope is expanded to other candidate high-frequency base stations or other possible frequency locations. Through this priority-based search method, the UE's search scope can be further narrowed in the time domain, improving the efficiency of initial access.

[0140] Optionally, the low-frequency base station can also notify the corresponding candidate high-frequency base station of the predicted beam and the UE's identification code through an interface between base stations (e.g., the Xn interface) (S53). As used in this disclosure, the UE's identification code refers to unique identification information of the UE, such as the International Mobile Subscriber Identity (IMSI). When the UE accesses the low-frequency base station, the low-frequency base station can obtain the user's identification code. Thus, when the UE accesses the candidate high-frequency base station, the high-frequency base station can use the predicted optimal beam to communicate with the UE. For example, in... Figure 14 In the uplink random access process shown, the high-frequency base station can use the optimal beam to transmit MSG-1 and / or MSG-4, etc.

[0141] Furthermore, according to embodiments of this disclosure, after obtaining the UE's identification code and the optimal beam, the high-frequency base station can further refine the beam through beam training. Specifically, the high-frequency base station prioritizes scanning multiple narrow beams contained within the optimal beam, receives signal quality (e.g., RSRP) reports from the UE for each narrow beam, and determines the narrow beam with the highest quality as the beam for subsequent data transmission. If the received signal quality of these narrow beams does not meet the requirements, it continues to try beam directions adjacent to the optimal beam until the signal quality meets the requirements or all beam directions have been traversed. Therefore, by utilizing the optimal beam predicted by a neural network, the range of the scanning beam can be narrowed, effectively reducing the overhead of beam training.

[0142] When the convolutional neural network outputs more than one (n>1) optimal beam, the low-frequency base station can send the beam IDs (e.g., the identifiers of the corresponding reference signals) sorted by predicted probability from high to low to the candidate high-frequency base station. The high-frequency base station can then scan these beams sequentially to improve the efficiency of beam training.

[0143] Although the above describes separately the cases of low-frequency base stations using convolutional neural networks to predict candidate high-frequency base stations + path loss (or RSRP) and candidate high-frequency base stations + optimal beam, it should be understood that low-frequency base stations can simultaneously predict candidate high-frequency base stations, path loss (or RSRP), and optimal beam. In this case, the initial access procedure for high-frequency communication can be combined... Figure 15A Or 15B and Figure 16 This will be understood in more detail, and will not be described again here.

[0144] The following describes the learning scheme based on the prediction model of this disclosure.

[0145] Predictive models such as convolutional neural networks are self-learning deep learning models. Low-frequency base stations can collect a large amount of training data from UEs and high-frequency base stations to determine or update the parameters of convolutional neural networks.

[0146] Figure 17 A communication flowchart for collecting training data for a convolutional neural network is shown. Figure 17 As shown, a UE accessing a high-frequency communication link can periodically report the identification information of the high-frequency base station it communicates with, which serves as the model output for training a convolutional neural network by the low-frequency base station. For example... Figure 9 The first example of the convolutional neural network shown aims to minimize the path loss of the high-frequency base station reported by the UE in order to ensure the accuracy of the convolutional neural network's predictions. However, for... Figure 12 The second example of the convolutional neural network shown expects the high-frequency base station reported by the UE to have the maximum RSRP.

[0147] The UE also periodically transmits low-frequency reference signals to a low-frequency base station, which then obtains a low-frequency CSI matrix based on the reference signals, which is used as input to train the convolutional neural network model.

[0148] Optionally, the UE can also periodically estimate the path loss (or RSRP) on the current high-frequency communication link and report it to the low-frequency base station as a branch network for training the convolutional neural network. Figure 9 The model output of branch 1) in 12.

[0149] Optionally, the high-frequency base station communicating with the UE can periodically notify the low-frequency base station of the UE's identification code and the beam used for that UE through an inter-base station interface (e.g., the Xn interface), as a branch network for training the convolutional neural network. Figure 9 The model output of branch 2 in 12.

[0150] The low-frequency base station pairs training data collected from the UE and the high-frequency base station based on the UE's identification code. This data includes, for example, the CSI matrix associated with the same UE, the high-frequency base station ID, path loss value (or RSRP value), and the beam used by the high-frequency base station. Furthermore, the location data and transmit power of all candidate high-frequency base stations can also be used as model input data for the corresponding convolutional neural network; this data can be pre-acquired and stored.

[0151] Real-world data collected by low-frequency base stations is used to learn the parameters of neural networks, such as the kernels of each convolutional layer, enabling the convolutional neural network to acquire predictive capabilities. Typically, the convolutional neural network is trained on a large amount of data and then deployed to low-frequency base stations for practical prediction. Furthermore, online learning strategies can also be applied to the updating of convolutional neural networks; that is, after deployment, the low-frequency base station continues to collect training data in real time and train the prediction model. Online learning strategies allow the prediction model to adapt to changes in the communication environment.

[0152] For example Figure 9 Or, for the convolutional neural network including branch networks shown in Figure 12, the training scheme may include the following steps:

[0153] 1. Using the low-frequency CSI matrix and the location data of high-frequency base stations as model inputs and the high-frequency base station IDs as model outputs, the parameters of the main network structure of the pre-trained convolutional neural network are used.

[0154] 2. Use the corresponding model output to train the parameters of each branch network. For example, use path loss values ​​for training. Figure 9 The parameters of branch 1 are trained using RSRP values. Figure 12 The parameters of branch 1 are used to train the parameters of branch 2 using the beams of the high-frequency base station.

[0155] Since accurately predicting the candidate high-frequency base station ID is fundamental to predicting the corresponding path loss (or RSRP) and the optimal beam of the high-frequency base station, during the pre-training phase, the neural network is trained with the goal of accurately predicting the high-frequency base station ID. This training optimizes the network structure parameters shared by all outputs to fully extract the essential features commonly needed by all outputs. After pre-training, each output further learns its branch network based on the extracted essential features.

[0156] In practical deployment of the convolutional neural network (CNN) described above, the parameters of the CNN can be determined first using training data collected from one or more low-frequency base stations. The pre-trained CNN can then be applied to more low-frequency base stations. In other words, the CNN applied to all low-frequency base stations can be initialized in advance using training data collected from a subset of these base stations. Then, each low-frequency base station can use the location data and / or transmit power of high-frequency base stations within its coverage area, along with the acquired low-frequency CSI matrix, as model input to continuously optimize the CNN parameters through online learning. Online learning primarily studies environmental information such as obstructions and reflectors around the current low-frequency base station, thereby improving the adaptability of the parameters to the surrounding communication environment.

[0157] Next, an example of the convolutional neural network according to this disclosure and the simulation effect of high-frequency communication access assisted by the prediction results of the convolutional neural network will be described.

[0158] The simulation described below is based on, for example, Figure 3 The low-frequency and millimeter-wave hybrid network shown is illustrated in Table 2. Specific settings are shown in Table 2.

[0159] Table 2: Specific Simulation Settings

[0160] Coverage radius of low-frequency base stations (meters) 500 Number of low-frequency base station antennas n 64 Low-frequency UE antenna count m 4 Low-frequency UE subcarrier number k 288 Number of millimeter-wave base stations b 5 Beamforming codebook size of millimeter-wave base stations 4 SNR / dB 5-25, randomly generated

[0161] The millimeter-wave path loss model is as follows:

[0162] PL[dB] = 32.4 + 20log 10 d+20log 10 f c +σ s

[0163] Where d represents the distance between the UE and the millimeter-wave base station, f c f represents the center frequency of millimeter waves. c =28GHz, σ s The shadow fading represents a Gaussian distribution with a mean of 0 and a variance of 4. The simulation settings described above are used to generate the input dataset for the convolutional neural network.

[0164] Figure 18 An example of a convolutional neural network, as shown in the first example, is presented. Figure 19 An example of a convolutional neural network is shown in the second example.

[0165] Figure 18 The input to the convolutional neural network in the example is: is the low-frequency CSI matrix, k is the number of low-frequency UE subcarriers, m is the number of low-frequency UE antennas, and n is the number of low-frequency base station antennas. B represents the candidate millimeter-wave base station locations. The millimeter-wave base station locations, represented in rectangular coordinates, are sequentially concatenated to form a vector input of length 2b, where b is the number of candidate base stations. This vector input is considered as a two-dimensional matrix input with a dimension of 1 (i.e., ).

[0166] Figure 19 The input to the convolutional neural network in the example is: is the low-frequency CSI matrix, k is the number of low-frequency UE subcarriers, m is the number of low-frequency UE antennas, and n is the number of low-frequency base station antennas. B' represents the candidate millimeter-wave base station locations and transmit power, constructed with dimensions of... The matrix, where b ij ,i=1,2,3 represent the two-dimensional horizontal coordinate, two-dimensional vertical coordinate, and transmission power of the j-th millimeter-wave base station, respectively.

[0167] Figure 18 and 19The meanings of the symbols or numbers in the table are shown in Table 3 below:

[0168] Table 3: Meaning of symbols or numbers in convolutional neural networks

[0169]

[0170]

[0171] Figure 18 In addition to outputting the millimeter-wave base station ID with minimum path loss, the convolutional neural network in the system also outputs the corresponding path loss value and the beam of the millimeter-wave base station.

[0172] Figure 19 In addition to outputting the millimeter-wave base station ID with the largest RSRP, the convolutional neural network in the model also outputs the corresponding RSRP value and the beam of the millimeter-wave base station.

[0173] Figures 20A-20C It shows the use of Figure 18 A schematic diagram illustrating the effect of the prediction results of the convolutional neural network in assisting high-frequency communication access.

[0174] Figure 20A This figure shows the proportion of millimeter-wave base stations with the lowest actual path loss that rank among the top X predicted probabilities. As can be seen from the figure, in 67.1% of the samples, the millimeter-wave base station with the lowest actual path loss was accurately predicted as a candidate high-frequency base station. Furthermore, in over 90% of the samples, the millimeter-wave base station with the lowest actual path loss was among the two high-frequency base stations with the highest predicted probabilities.

[0175] Figure 20B The cumulative distribution function of the absolute error between the predicted millimeter-wave path loss and the actual path loss is shown. As can be seen from the figure, over 70% of the prediction absolute errors are within 5 dB, and approximately 92.5% are within 10 dB. Meanwhile, the mean absolute error is 5.1 dB. This indicates that predicting the minimum path loss between the user and surrounding millimeter-wave base stations based on low-frequency CSI is feasible.

[0176] Figure 20C The results show the proportion of the actual optimal beam that ranks among the top X positions with the highest predicted probability in the prediction results. In 44.5% of the samples, the actual beam used can be accurately predicted as the optimal beam for the high-frequency base station; at the same time, the actual beam used ranks among the top two positions with the highest predicted probability in approximately 75% of the prediction results. Simulation results show that prioritizing the predicted optimal beam can effectively reduce the beam search overhead.

[0177] Electronic Equipment and Communication Methods

[0178] The electronic devices and communication methods according to this disclosure are described below.

[0179] Figure 21A This is a block diagram illustrating an electronic device 100 according to the present disclosure. The electronic device 100 may be a low-frequency base station or a component thereof.

[0180] like Figure 21A As shown, the electronic device 100 includes a processing circuit 101. The processing circuit 101 includes at least a CSI matrix acquisition unit 102, a candidate high-frequency base station determination unit 103, an access assistance information determination unit 104, and an access assistance information transmission unit 105. The processing circuit 101 can be configured to perform... Figure 21B The communication method shown. The processing circuit 101 can refer to various implementations of digital circuit systems, analog circuit systems, or mixed-signal (combination of analog and digital signals) circuit systems that perform functions in a low-frequency base station.

[0181] The CSI matrix acquisition unit 102 of the processing circuit 101 is configured to acquire a low-frequency CSI matrix based on a reference signal received from the UE via low-frequency communication, i.e., to perform... Figure 21B Step S101. When the UE needs to access a high-frequency base station, it can send a reference signal, such as CSI-RS or SRS, to the low-frequency base station.

[0182] The candidate high-frequency base station determination unit 103 is configured to determine, based on the CSI matrix obtained by the CSI matrix acquisition unit 102, a candidate high-frequency base station suitable for high-frequency communication with the UE from multiple high-frequency base stations using a neural network, i.e., to perform... Figure 21B Step S102 in the process. The neural network can be as follows: Figure 9 or Figure 12 The convolutional neural network shown in the figure. The determined candidate high-frequency base station can be a high-frequency base station predicted to have the minimum path loss for high-frequency communication with the UE, or a high-frequency base station predicted to have the maximum received power for high-frequency communication with the UE at the UE.

[0183] The access assistance information determination unit 104 is configured to determine the access assistance information associated with the candidate high-frequency base station, that is, to perform... Figure 21B Step S103. Access assistance information may include the frequency location of the SSB broadcast by the candidate high-frequency base station or the GSCN indicating the frequency location of the SSB. Furthermore, access assistance information may also include the path loss value or RSRP value corresponding to the candidate high-frequency base station predicted using a branch network of a neural network. If the optimal beam to be used by the candidate high-frequency base station can be predicted using a branch network of a neural network, the access assistance information may also include the frequency location (or GSCN indicating the frequency location) of the SSB corresponding to the optimal beam and its index.

[0184] Access assistance information sending unit 105 is configured to send the access assistance information determined by access assistance information determining unit 104 to the UE, i.e., to perform... Figure 21B Step S104. This access assistance information will help the UE decide whether to access the high-frequency base station and can narrow down the range of SSBs the UE searches for, thereby improving the efficiency and quality of initial access to high-frequency communication.

[0185] Electronic device 100 may also include communication unit 106. Communication unit 106 may be configured to communicate with UE under the control of processing circuitry 101. In one example, communication unit 106 may be implemented as a transceiver, including communication components such as an antenna array and / or a radio frequency link. Communication unit 106 is drawn with dashed lines because it may also be located outside electronic device 100.

[0186] The electronic device 100 may also include a memory 107. The memory 107 can store various data and instructions, such as programs and data for the operation of the electronic device 100, various data generated by the processing circuit 101, and various control signals or service data sent or received by the communication unit 106. The memory 107 is drawn with dashed lines because it may be located inside the processing circuit 101 or outside the electronic device 100.

[0187] Figure 22A This is a block diagram illustrating an electronic device 200 according to the present disclosure. The electronic device 200 may be a UE or a component thereof.

[0188] like Figure 22A As shown, the electronic device 200 includes a processing circuit 201. The processing circuit 201 includes at least a reference signal transmitting unit 202, an access assistance information receiving unit 203, and an access unit 204. The processing circuit 201 can be configured to perform... Figure 22B The communication method shown. Processing circuit 201 can refer to various implementations of digital circuit systems, analog circuit systems, or mixed-signal (combination of analog and digital signals) circuit systems that perform functions in the UE.

[0189] Reference signal transmission unit 202 can be configured to transmit a reference signal to a low-frequency base station via a low-frequency link so that the low-frequency base station can obtain the CSI matrix, i.e., perform... Figure 22B Step S201. The reference signal transmitted by the reference signal transmitting unit 202 can be, for example, a low-frequency reference signal of CSI-RS or SRS.

[0190] The access assistance information receiving unit 203 can be configured to receive access assistance information associated with a candidate high-frequency base station, determined by the low-frequency base station, i.e., to perform... Figure 22BStep S202 in the process. The candidate high-frequency base station is the low-frequency base station that inputs the acquired CSI matrix into, for example... Figure 9 The candidate high-frequency base station is determined by the convolutional neural network shown in Figure 12. The candidate high-frequency base station is predicted to be a suitable high-frequency base station for high-frequency communication with the UE, for example, having minimum path loss for high-frequency communication with the UE, or having maximum received power for high-frequency communication with the UE at the UE. The access assistance information received by the access assistance information receiving unit 203 may include the frequency location and / or index of the SSB broadcast by the candidate high-frequency base station.

[0191] Access unit 204 can be configured to access candidate high-frequency base stations based on the received access assistance information, i.e., to perform... Figure 22B Step S203. Access unit 204 can search for one or more SSBs transmitted by candidate high-frequency base stations in the time-frequency resources indicated by the access assistance information, and obtain downlink synchronization and uplink synchronization by decoding the SSBs. Compared with the UE blindly searching at all synchronization channel frequency positions and all time domain positions of its PLMN, the initial access using access assistance information is more efficient.

[0192] Electronic device 200 may also include a communication unit 206. Communication unit 206 may be configured to communicate with a base station under the control of processing circuitry 201. In one example, communication unit 206 may be implemented as a transmitter or transceiver, including communication components such as an antenna array and / or a radio frequency link. Communication unit 206 is drawn with dashed lines because it may also be located outside electronic device 200.

[0193] The electronic device 200 may also include a memory 207. The memory 207 may store various data and instructions, programs and data for the operation of the electronic device 200, various data generated by the processing circuit 201, data to be transmitted by the communication unit 207, etc. The memory 207 is drawn with dashed lines because it may be located inside the processing circuit 201 or outside the electronic device 200.

[0194] Figure 23A This is a block diagram illustrating an electronic device 300 according to the present disclosure. The electronic device 300 may be a high-frequency base station or a component thereof.

[0195] like Figure 23A As shown, the electronic device 300 includes a processing circuit 301. The processing circuit 301 includes at least a receiving unit 302 and a communication establishment unit 303. The processing circuit 301 can be configured to perform... Figure 23B The communication method shown. The processing circuit 301 can refer to various implementations of digital circuit systems, analog circuit systems, or mixed-signal (a combination of analog and digital signals) circuit systems that perform functions in the base station equipment.

[0196] The receiving unit 302 can be configured to receive the UE's identification code and information about the beams available to the UE from the low-frequency base station, i.e., to perform... Figure 23B Step S301 in the above steps. The beam is determined by the low-frequency base station by inputting a CSI matrix obtained based on reference signals transmitted by the UE via the low-frequency link into a neural network. The neural network can be as follows: Figure 9 Or the convolutional neural network shown in 12, which includes a branch network for predicting the optimal beam for a high-frequency base station.

[0197] The communication establishment unit 303 can be configured to establish high-frequency communication with the UE using the beam, i.e., to perform... Figure 23B Step S302. For example, when the UE initially accesses the high-frequency base station, the communication establishment unit 303 of the high-frequency base station can send signaling to the UE using the beam recommended by the low-frequency base station. In addition, the communication establishment unit 303 can also refine the beam through beam training.

[0198] Electronic device 300 may also include a communication unit 306. Communication unit 306 may be configured to communicate with a base station under the control of processing circuitry 301. In one example, communication unit 306 may be implemented as a transmitter or transceiver, including communication components such as an antenna array and / or a radio frequency link. Communication unit 306 is drawn with dashed lines because it may also be located outside electronic device 300.

[0199] The electronic device 300 may also include a memory 307. The memory 307 may store various data and instructions, programs and data for the operation of the electronic device 300, various data generated by the processing circuit 301, data to be transmitted by the communication unit 307, etc. The memory 307 is drawn with dashed lines because it may be located inside the processing circuit 301 or outside the electronic device 300.

[0200] Various aspects of the embodiments of this disclosure have been described in detail above. However, it should be noted that the above description of the structure, arrangement, type, number, etc. of the illustrated antenna arrays, ports, reference signals, communication devices, communication methods, etc., is not intended to limit the aspects of this disclosure to these specific examples.

[0201] It should be understood that the various units of the electronic devices 100, 200, and 300 described in the above embodiments are merely logical modules divided according to their specific functions, and are not intended to limit the specific implementation methods. In actual implementation, the above units can be implemented as independent physical entities, or they can be implemented by a single entity (e.g., a processor (CPU or DSP, etc.), integrated circuit, etc.).

[0202] It should be understood that the processing circuits 101, 201, or 301 described in the above embodiments may include, for example, circuits such as integrated circuits (ICs), application-specific integrated circuits (ASICs), portions or circuits of a single processor core, an entire processor core, a single processor, programmable hardware devices such as field-programmable gate arrays (FPGAs), and / or systems including multiple processors. Memory 107, 207, or 307 may be volatile memory and / or non-volatile memory. For example, memory 107, 207, or 307 may include, but is not limited to, random access memory (RAM), dynamic random access memory (DRAM), static random access memory (SRAM), read-only memory (ROM), and flash memory.

[0203] It should be understood that the various units of the electronic devices 100, 200, or 300 described in the above embodiments are merely logical modules divided according to their specific functions, and are not intended to limit the specific implementation method. In actual implementation, the above units can be implemented as independent physical entities, or they can be implemented by a single entity (e.g., a processor (CPU or DSP, etc.), integrated circuit, etc.).

[0204] [Exemplary Implementation of this Disclosure]

[0205] According to embodiments of this disclosure, various implementations of the concepts of this disclosure are conceivable, including but not limited to: 1) an electronic device for a low-frequency base station, comprising: a processing circuit configured to: acquire a channel state information (CSI) matrix based on a reference signal received from a user equipment via low-frequency communication; determine, using a neural network, candidate high-frequency base stations suitable for high-frequency communication with the user equipment from a plurality of high-frequency base stations based on the CSI matrix; determine access assistance information associated with the candidate high-frequency base stations; and send the access assistance information to the user equipment.

[0206] 2) The electronic device as described in 1), wherein the processing circuit is further configured to: determine the high-frequency base station with the minimum path loss for high-frequency communication with the user equipment as a candidate high-frequency base station by inputting the CSI matrix and the location data of the plurality of high-frequency base stations into the neural network.

[0207] 3) The electronic device as described in 2), wherein the processing circuit is further configured to send the path loss value corresponding to the candidate high-frequency base station to the user equipment.

[0208] 4) The electronic device as described in 2), wherein the processing circuit is further configured to: instruct the user equipment to activate the high-frequency communication module when the path loss value corresponding to the candidate high-frequency base station is lower than a predetermined threshold.

[0209] 5) The electronic device as described in 1), wherein the processing circuit is further configured to: determine the high-frequency base station with the maximum received power for high-frequency communication at the user equipment as a candidate high-frequency base station by inputting the CSI matrix and the location data and transmit power of the plurality of high-frequency base stations into the neural network.

[0210] 6) The electronic device as described in 5), wherein the processing circuit is further configured to send the received power value corresponding to the candidate high-frequency base station to the user equipment.

[0211] 7) The electronic device as described in 5), wherein the processing circuit is further configured to: instruct the user equipment to activate the high-frequency communication module when the received power value corresponding to the candidate high-frequency base station exceeds a predetermined threshold.

[0212] 8) The electronic device as described in 1), wherein the processing circuit is further configured to: use the neural network to determine the beam of the candidate high-frequency base station for the user equipment; and notify the candidate high-frequency base station of the identification code of the user equipment and the information of the beam.

[0213] 9) An electronic device as described in 8), wherein the information about the beam is an index of the synchronization signal / physical broadcast channel block (SSB) corresponding to the beam.

[0214] 10) The electronic device as described in 1), wherein the access assistance information includes the frequency location of the synchronization signal / physical broadcast channel block (SSB) of the candidate high-frequency base station.

[0215] 11) The electronic device as described in 8), wherein the access assistance information includes the frequency location and index of the synchronization signal / physical broadcast channel block (SSB) of the candidate high-frequency base station.

[0216] 12) The electronic device as described in 1), wherein the processing circuit is further configured to: receive a reference signal transmitted via low-frequency communication and identification information of the high-frequency base station from a user equipment that communicates with any of the plurality of high-frequency base stations at high frequency; and update the parameters of the neural network by using a CSI matrix obtained based on the reference signal and the location data of the high-frequency base station as input and the identification information of the high-frequency base station as output.

[0217] 13) The electronic device as described in 2), wherein the processing circuit is further configured to: receive a path loss value of high-frequency communication estimated by the user equipment from a user equipment that is communicating at high frequency with any one of the plurality of high-frequency base stations; and update the parameters of the branch network of the neural network by using the path loss value as an output.

[0218] 14) The electronic device as described in 2), wherein the processing circuit is further configured to: receive, from one of the plurality of high-frequency base stations that communicates with the user equipment at high frequency, an identification code of the user equipment and information about the beam used for the high-frequency communication; and update the parameters of the branch network of the neural network by using the information about the beam used for the high-frequency communication as output.

[0219] 15) The electronic device as described in 9), wherein the low-frequency communication operates in the LTE band, the LTE-A band or the sub-6GHz band, and wherein the high-frequency communication operates in the millimeter-wave band.

[0220] 16) An electronic device for a user equipment, comprising: a processing circuit configured to: transmit a reference signal to a low-frequency base station via low-frequency communication for the low-frequency base station to acquire a channel state information (CSI) matrix; receive access assistance information associated with a candidate high-frequency base station determined by the low-frequency base station, wherein the candidate high-frequency base station is a high-frequency base station suitable for high-frequency communication with the user equipment, determined by the low-frequency base station based on the CSI matrix using a neural network; and access the candidate high-frequency base station using the access assistance information.

[0221] 17) The electronic device as described in 16), wherein the candidate high-frequency base station is the high-frequency base station with the minimum path loss for high-frequency communication with the user equipment predicted by the neural network.

[0222] 18) An electronic device as described in 16), wherein the candidate high-frequency base station is the high-frequency base station with the highest received power at the user equipment for high-frequency communication with the user equipment predicted by the neural network.

[0223] 19) The electronic device as described in 17), wherein the processing circuit is further configured to: receive a path loss value corresponding to a candidate high-frequency base station from a low-frequency base station; and, if the path loss value is lower than a predetermined threshold, activate the high-frequency communication module to access the candidate high-frequency base station.

[0224] 20) The electronic device as described in 18), wherein the processing circuit is further configured to: receive a received power value corresponding to a candidate high-frequency base station from a low-frequency base station; and, if the received power value exceeds a predetermined threshold, activate the high-frequency communication module to access the candidate high-frequency base station.

[0225] 21) The electronic device as described in 16), wherein the access assistance information includes the frequency location of the synchronization signal / physical broadcast channel block (SSB) of the candidate high-frequency base station.

[0226] 22) The electronic device as described in 21), wherein the access assistance information further includes an index of the SSB of the candidate high-frequency base station.

[0227] 23) An electronic device as described in 19) or 20), wherein the processing circuit is further configured to adjust the predetermined threshold based on the current battery level of the user device, connection preferences, and the transmission success rate of high-frequency communication.

[0228] 24) An electronic device for a high-frequency base station, comprising: a processing circuit configured to: receive from a low-frequency base station an identification code of a user equipment and information about a beam of the high-frequency base station that can be used by the user equipment, wherein the beam is determined by the low-frequency base station by inputting a CSI matrix obtained based on a reference signal transmitted by the user equipment via a low-frequency link into a neural network; and establish high-frequency communication with the user equipment using the beam.

[0229] 25) The electronic device as described in 24), wherein the processing circuitry is further configured to: determine a narrower beam for high-frequency communication with the user equipment by beam scanning based on the beam.

[0230] 26) The electronic device as described in 25), wherein the beam comprises a plurality of beams with different priorities, and the processing circuit is further configured to scan the beam based on the priorities of the plurality of beams.

[0231] 27) A method for training a neural network, comprising: receiving a reference signal transmitted via low-frequency communication from a user equipment; obtaining a channel state information (CSI) matrix based on the reference signal; receiving identification information of a high-frequency base station with which the user equipment conducts high-frequency communication; and performing deep learning by using the CSI matrix as input and the identification information of the high-frequency base station as output to determine the parameters of the neural network.

[0232] 28) The method as described in 27) further includes: receiving from a user equipment a path loss value of high-frequency communication estimated by the user equipment; and determining parameters of a branch network of the neural network by using the path loss value as an output.

[0233] 29) The method as described in 27) further includes: receiving a received power value for high-frequency communication measured by the user equipment; and determining parameters of the branch network of the neural network by using the received power value as an output.

[0234] 30) The method as described in 27) further includes: receiving, from a high-frequency base station communicating with the user equipment at high frequency, an identification code of the user equipment and information about the beam used for the high-frequency communication; and determining parameters of the branch network of the neural network by using the information about the beam used for the high-frequency communication as output.

[0235] 31) A communication method, comprising: acquiring a channel state information (CSI) matrix based on a reference signal received from a user equipment via low-frequency communication; determining, using a neural network, candidate high-frequency base stations suitable for high-frequency communication with the user equipment from a plurality of high-frequency base stations based on the CSI matrix; determining access assistance information associated with the candidate high-frequency base stations; and sending the access assistance information to the user equipment.

[0236] 32) A communication method, comprising: transmitting a reference signal to a low-frequency base station via low-frequency communication for the low-frequency base station to acquire a channel state information (CSI) matrix; receiving access assistance information associated with a candidate high-frequency base station determined by the low-frequency base station, wherein the candidate high-frequency base station is a high-frequency base station suitable for high-frequency communication with the user equipment, determined by the low-frequency base station based on the CSI matrix using a neural network; and accessing the candidate high-frequency base station using the access assistance information.

[0237] 33) A communication method, comprising: receiving from a low-frequency base station an identification code of a user equipment and information about a beam of the high-frequency base station that can be used by the user equipment, wherein the beam is determined by the low-frequency base station by inputting a CSI matrix obtained based on a reference signal transmitted by the user equipment via a low-frequency link into a neural network; and establishing high-frequency communication with the user equipment using the beam.

[0238] 34) A non-transitory computer-readable storage medium storing executable instructions that, when executed, implement the method as described in any one of 27) to 33).

[0239] [Application Examples of this Disclosure]

[0240] The technology described in this disclosure can be applied to a variety of products.

[0241] For example, electronic devices 100 and 300 according to embodiments of the present disclosure can be implemented as various base stations or installed in base stations, and electronic device 200 can be implemented as various user equipment or installed in various user equipment.

[0242] The communication methods according to embodiments of this disclosure can be implemented by various base stations or user equipment; the methods and operations according to embodiments of this disclosure can be embodied as computer-executable instructions, stored in a non-transitory computer-readable storage medium, and can be executed by various base stations or user equipment to achieve one or more of the functions described above.

[0243] The techniques according to embodiments of this disclosure can be used to create various computer program products that can be used in various base stations or user equipment to achieve one or more of the functions described above.

[0244] The base station described in this disclosure can be implemented as any type of base station, preferably such as macro gNB and ng-eNB as defined in the 3GPP 5G NR standard. A gNB can be a gNB covering a cell smaller than a macro cell, such as a pico gNB, micro gNB, and femtocell gNB. Alternatively, the base station can be implemented as any other type of base station, such as a NodeB, eNodeB, and Base Transceiver Station (BTS). The base station may also include: a main body configured to control wireless communication and one or more remote radio heads (RRHs), wireless relay stations, drone towers, control nodes in automated factories, etc., located at locations different from the main body.

[0245] User equipment can be implemented as a mobile terminal (such as a smartphone, tablet PC, laptop PC, portable gaming terminal, portable / dongle-type mobile router, and digital camera device) or an in-vehicle terminal (such as a car navigation device). User equipment can also be implemented as a terminal performing machine-to-machine (M2M) communication (also known as a machine-type communication (MTC) terminal), a drone, a sensor and actuator in an automated factory, etc. Furthermore, user equipment can be a wireless communication module (such as an integrated circuit module comprising a single chip) installed on each of the aforementioned terminals.

[0246] First application example of base stations

[0247] Figure 24 This is a block diagram illustrating a first example of a schematic configuration of a base station to which the technologies of this disclosure can be applied. Figure 24 In this implementation, the base station can be a gNB 1400. The gNB 1400 includes multiple antennas 1410 and a base station device 1420. The base station device 1420 and each antenna 1410 can be connected to each other via RF cables. In one implementation, the gNB 1400 (or base station device 1420) here can correspond to the aforementioned electronic devices 100 or 300.

[0248] Antenna 1410 includes multiple antenna elements. Antenna 1410 can be arranged, for example, as an antenna array matrix and used by base station equipment 1420 to transmit and receive wireless signals. For example, multiple antennas 1410 can be compatible with multiple frequency bands used by gNB 1400.

[0249] The base station equipment 1420 includes a controller 1421, a memory 1422, a network interface 1423, and a wireless communication interface 1425.

[0250] The controller 1421 may be, for example, a CPU or a DSP, and operates various higher-level functions of the base station device 1420. For example, the controller 1421 may include the processing circuitry 101 or 301 described above to perform... Figure 21B The communication method described in 23B, or the various components controlling base station equipment 1420, may be used. For example, controller 1421 generates data packets based on data in signals processed by wireless communication interface 1425 and transmits the generated packets via network interface 1423. Controller 1421 may bundle data from multiple baseband processors to generate bundled packets and transmit the generated bundled packets. Controller 1421 may have logical functions to perform controls such as radio resource control, radio bearer control, mobility management, admission control, and scheduling. This control may be performed in conjunction with nearby gNBs or core network nodes. Memory 1422 includes RAM and ROM and stores programs executed by controller 1421 and various types of control data (such as terminal lists, transmission power data, and scheduling data).

[0251] Network interface 1423 is a communication interface for connecting base station equipment 1420 to core network 1424 (e.g., a 5G core network). Controller 1421 can communicate with core network nodes or other gNBs via network interface 1423. In this case, gNB 1400 and core network nodes or other gNBs can be connected to each other via logical interfaces (such as NG and Xn interfaces). Network interface 1423 can also be a wired communication interface or a wireless communication interface for wireless backhaul. If network interface 1423 is a wireless communication interface, it can use a higher frequency band for wireless communication compared to the frequency band used by wireless communication interface 1425.

[0252] Wireless communication interface 1425 supports any cellular communication scheme (such as 5G NR) and provides wireless connectivity to terminals located in the cell of gNB 1400 via antenna 1410. Wireless communication interface 1425 typically includes, for example, a baseband (BB) processor 1426 and RF circuitry 1427. BB processor 1426 can perform, for example, encoding / decoding, modulation / demodulation, and multiplexing / demultiplexing, and performs various types of signal processing at each layer (e.g., physical layer, MAC layer, RLC layer, PDCP layer, SDAP layer). Instead of controller 1421, BB processor 1426 may have some or all of the above-described logical functions. BB processor 1426 may be a memory storing communication control programs, or a module including a processor and associated circuitry configured to execute programs. Update programs can change the functionality of BB processor 1426. The module may be a card or blade inserted into a slot in base station equipment 1420. Alternatively, the module may be a chip mounted on a card or blade. Meanwhile, the RF circuit 1427 may include, for example, a mixer, a filter, and an amplifier, and transmits and receives wireless signals via the antenna 1410. Although Figure 24An example of an RF circuit 1427 connected to an antenna 1410 is shown, but this disclosure is not limited to the illustration, and an RF circuit 1427 can be connected to multiple antennas 1410 simultaneously.

[0253] like Figure 24 As shown, the wireless communication interface 1425 may include multiple BB processors 1426. For example, the multiple BB processors 1426 may be compatible with multiple frequency bands used by the gNB 1400. Figure 24 As shown, the wireless communication interface 1425 may include multiple RF circuits 1427. For example, the multiple RF circuits 1427 may be compatible with multiple antenna elements. Although Figure 24 An example is shown in which the wireless communication interface 1425 includes multiple BB processors 1426 and multiple RF circuits 1427, but the wireless communication interface 1425 may also include a single BB processor 1426 or a single RF circuit 1427.

[0254] exist Figure 24 In the gNB 1400 shown, refer to Figure 21A The processing circuit 101 described includes one or more units (e.g., access auxiliary information transmission unit 105) or references Figure 23A One or more units (e.g., receiving unit 302) included in the described processing circuitry 301 may be implemented in the wireless communication interface 825. Alternatively, at least a portion of these components may be implemented in the controller 821. For example, the gNB 1400 may include a portion (e.g., BB processor 1426) or the entirety of the wireless communication interface 1425, and / or a module including the controller 1421, and one or more components may be implemented in the module. In this case, the module may store a program for allowing the processor to function as one or more components (in other words, a program for allowing the processor to perform the operation of one or more components), and may execute the program. As another example, a program for allowing the processor to function as one or more components may be installed in the gNB 1400, and the wireless communication interface 1425 (e.g., BB processor 1426) and / or the controller 1421 may execute the program. As described above, the gNB 1400, the base station device 1420, or the module may be provided as an apparatus including one or more components, and a program for allowing the processor to function as one or more components may be provided. Additionally, a readable medium in which the program is recorded may be provided.

[0255] Second application example of base stations

[0256] Figure 25 This is a block diagram illustrating a second example of a schematic configuration of a base station to which the techniques of this disclosure can be applied. Figure 25In the diagram, the base station is shown as gNB 1530. gNB 1530 includes multiple antennas 1540, base station equipment 1550, and RRH 1560. RRH 1560 and each antenna 1540 can be connected to each other via RF cables. Base station equipment 1550 and RRH 1560 can be connected to each other via high-speed lines such as fiber optic cables. In one implementation, gNB 1530 (or base station equipment 1550) here may correspond to the aforementioned electronic equipment 100 or 300.

[0257] Antenna 1540 includes multiple antenna elements. Antenna 1540 can be arranged, for example, as an antenna array matrix and used by base station equipment 1550 to transmit and receive wireless signals. For example, multiple antennas 1540 can be compatible with multiple frequency bands used by gNB 1530.

[0258] Base station equipment 1550 includes a controller 1551, a memory 1552, a network interface 1553, a wireless communication interface 1555, and a connection interface 1557. The controller 1551, memory 1552, and network interface 1553 are related to a reference... Figure 24 The controller 1421, memory 1422 and network interface 1423 described are the same.

[0259] The wireless communication interface 1555 supports any cellular communication scheme (such as 5G NR) and provides wireless communication to terminals located in the sector corresponding to RRH 1560 via RRH 1560 and antenna 1540. The wireless communication interface 1555 may typically include, for example, a BB processor 1556. In addition to the BB processor 1556 being connected to the RF circuitry 1564 of RRH 1560 via connection interface 1557, the BB processor 1556 is connected to the reference... Figure 24 The BB processor 1426 is described as identical. Figure 25 As shown, the wireless communication interface 1555 may include multiple BB processors 1556. For example, the multiple BB processors 1556 may be compatible with multiple frequency bands used by the gNB 1530. Although Figure 25 An example is shown in which the wireless communication interface 1555 includes multiple BB processors 1556, but the wireless communication interface 1555 may also include a single BB processor 1556.

[0260] Connection interface 1557 is an interface for connecting base station device 1550 (wireless communication interface 1555) to RRH 1560. Connection interface 1557 may also be a communication module for communication in the aforementioned high-speed line connecting base station device 1550 (wireless communication interface 1555) to RRH 1560.

[0261] The RRH 1560 includes a connectivity interface 1561 and a wireless communication interface 1563.

[0262] Connection interface 1561 is an interface for connecting RRH 1560 (wireless communication interface 1563) to base station equipment 1550. Connection interface 1561 can also be a communication module for communication in the aforementioned high-speed line.

[0263] Wireless communication interface 1563 transmits and receives wireless signals via antenna 1540. Wireless communication interface 1563 typically includes, for example, RF circuitry 1564. RF circuitry 1564 may include, for example, a mixer, filter, and amplifier, and transmits and receives wireless signals via antenna 1540. Although Figure 25 An example of an RF circuit 1564 connected to an antenna 1540 is shown, but this disclosure is not limited to the illustration, and an RF circuit 1564 can be connected to multiple antennas 1540 simultaneously.

[0264] like Figure 25 As shown, the wireless communication interface 1563 may include multiple RF circuits 1564. For example, the multiple RF circuits 1564 may support multiple antenna elements. Although Figure 25 An example is shown in which the wireless communication interface 1563 includes multiple RF circuits 1564, but the wireless communication interface 1563 may also include a single RF circuit 1564.

[0265] exist Figure 25 In the gNB 1500 shown, refer to Figure 21A The processing circuit 101 described includes one or more units (e.g., access auxiliary information transmission unit 105) or references Figure 23AOne or more units (e.g., receiving unit 302) included in the described processing circuitry 301 may be implemented in the wireless communication interface 1525. Alternatively, at least a portion of these components may be implemented in the controller 1521. For example, the gNB 1500 may include a portion (e.g., BB processor 1526) or the entirety of the wireless communication interface 1525, and / or a module including the controller 1521, and one or more components may be implemented in the module. In this case, the module may store a program for allowing the processor to function as one or more components (in other words, a program for allowing the processor to perform the operation of one or more components), and may execute the program. As another example, a program for allowing the processor to function as one or more components may be installed in the gNB 1500, and the wireless communication interface 1525 (e.g., BB processor 1526) and / or the controller 1521 may execute the program. As described above, the gNB 1500, the base station device 1520, or the module may be provided as an apparatus including one or more components, and a program for allowing the processor to function as one or more components may be provided. Additionally, a readable medium in which the program is recorded may be provided.

[0266] First application example of user equipment

[0267] Figure 26 This is a block diagram illustrating an example of a schematic configuration of a smartphone 1600 to which the technologies of this disclosure can be applied. In one example, the smartphone 1600 may be implemented as the electronic device 200 described in this disclosure.

[0268] The smartphone 1600 includes a processor 1601, a memory 1602, a storage device 1603, an external connection interface 1604, a camera device 1606, a sensor 1607, a microphone 1608, an input device 1609, a display device 1610, a speaker 1611, a wireless communication interface 1612, one or more antenna switches 1615, one or more antennas 1616, a bus 1617, a battery 1618, and an auxiliary controller 1619.

[0269] Processor 1601 may be, for example, a CPU or a system-on-a-chip (SoC), and controls the application layer and other functions of smartphone 1600. Processor 1601 may include or act as a reference. Figure 22A The described processing circuit 201. The memory 1602 includes RAM and ROM, and stores data and programs executed by the processor 1601 to implement the referenced... Figure 22BThe communication method is described above. Storage device 1603 may include storage media such as semiconductor memory and hard disk. External connection interface 1604 is an interface for connecting external devices (such as memory cards and Universal Serial Bus (USB) devices) to smartphone 1600.

[0270] The camera device 1606 includes an image sensor (such as a charge-coupled device (CCD) and complementary metal-oxide-semiconductor (CMOS)) and generates captured images. The sensor 1607 may include a set of sensors, such as a measurement sensor, a gyroscope sensor, a magnetometer sensor, and an accelerometer sensor. The microphone 1608 converts sound input to the smartphone 1600 into an audio signal. The input device 1609 includes, for example, a touch sensor, keypad, keyboard, buttons, or switches configured to detect touches on the screen of the display device 1610 and receive operations or information input from the user. The display device 1610 includes a screen (such as a liquid crystal display (LCD) and an organic light-emitting diode (OLED) display) and displays the output image of the smartphone 1600. The speaker 1611 converts the audio signal output from the smartphone 1600 into sound.

[0271] The wireless communication interface 1612 supports any cellular communication scheme (such as 4G LTE or 5G NR, etc.) and performs wireless communication. The wireless communication interface 1612 typically includes, for example, a BB processor 1613 and RF circuitry 1614. The BB processor 1613 can perform, for example, encoding / decoding, modulation / demodulation, and multiplexing / demultiplexing, and performs various types of signal processing for wireless communication. Meanwhile, the RF circuitry 1614 can include, for example, a mixer, filters, and amplifiers, and transmits and receives wireless signals via antenna 1616. The wireless communication interface 1612 can be a single chip module on which the BB processor 1613 and RF circuitry 1614 are integrated. Figure 26 As shown, the wireless communication interface 1612 may include multiple BB processors 1613 and multiple RF circuits 1614. Although Figure 26 An example is shown in which the wireless communication interface 1612 includes multiple BB processors 1613 and multiple RF circuits 1614, but the wireless communication interface 1612 may also include a single BB processor 1613 or a single RF circuit 1614.

[0272] In addition to cellular communication schemes, wireless communication interface 1612 can support other types of wireless communication schemes, such as short-range wireless communication schemes, near-field communication schemes, and wireless local area network (LAN) schemes. In this case, wireless communication interface 1612 may include a BB processor 1613 and RF circuitry 1614 for each wireless communication scheme.

[0273] Each of the antenna switches 1615 switches the connection destination of the antenna 1616 among multiple circuits (e.g., circuits for different wireless communication schemes) included in the wireless communication interface 1612.

[0274] Antenna 1616 includes multiple antenna elements. Antenna 1616 may be arranged, for example, as an antenna array matrix, and used for transmitting and receiving wireless signals through wireless communication interface 1612. Smartphone 1600 may include one or more antenna panels (not shown).

[0275] Furthermore, the smartphone 1600 may include an antenna 1616 for each wireless communication scheme. In this case, the antenna switch 1615 can be omitted from the configuration of the smartphone 1600.

[0276] Bus 1617 connects processor 1601, memory 1602, storage device 1603, external connection interface 1604, camera device 1606, sensor 1607, microphone 1608, input device 1609, display device 1610, speaker 1611, wireless communication interface 1612, and auxiliary controller 1619 to each other. Battery 1618 supplies power to... Figure 26 The various blocks of the smartphone 1600 shown are powered, and the feeders are partially shown as dashed lines in the figure. The auxiliary controller 1619 operates the minimum necessary functions of the smartphone 1600, for example, in sleep mode.

[0277] exist Figure 26 Among the smartphones 1600 shown, refer to Figure 22AOne or more components (e.g., reference signal transmitting unit 202, access auxiliary information receiving unit 203) included in the described processing circuitry 201 may be implemented in the wireless communication interface 1612. Alternatively, at least a portion of these components may be implemented in the processor 1601 or the auxiliary controller 1619. As an example, the smartphone 1600 includes a portion (e.g., BB processor 1613) or the entirety of the wireless communication interface 1612, and / or includes a module comprising the processor 1601 and / or the auxiliary controller 1619, and one or more components may be implemented in the module. In this case, the module may store and execute a program that allows the processor to function as one or more components (in other words, a program that allows the processor to perform the operation of one or more components). As another example, a program that allows the processor to function as one or more components may be installed in the smartphone 1600, and the wireless communication interface 1612 (e.g., BB processor 1613), the processor 1601, and / or the auxiliary controller 1619 may execute the program. As described above, a smartphone 1600 or module may be provided as an apparatus comprising one or more components, and a program for allowing the processor to function as one or more components may be provided. Additionally, a readable medium in which the program is recorded may be provided.

[0278] Second application example of user equipment

[0279] Figure 27 This is a block diagram illustrating an example of a schematic configuration of a car navigation device 1720 to which the techniques of this disclosure can be applied. The car navigation device 1720 can be implemented as described above. Figure 22A The described electronic device 200. The car navigation device 1720 includes a processor 1721, a memory 1722, a Global Positioning System (GPS) module 1724, a sensor 1725, a data interface 1726, a content player 1727, a storage medium interface 1728, an input device 1729, a display device 1730, a speaker 1731, a wireless communication interface 1733, one or more antenna switches 1736, one or more antennas 1737, and a battery 1738. In one example, the car navigation device 1720 can be implemented as the UE described in this disclosure.

[0280] The processor 1721 can be, for example, a CPU or a SoC, and controls the navigation functions and other functions of the car navigation device 1720. The memory 1722 includes RAM and ROM, and stores data and programs executed by the processor 1721.

[0281] GPS module 1724 uses GPS signals received from GPS satellites to measure the location (such as latitude, longitude, and altitude) of car navigation device 1720. Sensor 1725 may include a set of sensors, such as a gyroscope sensor, a geomagnetic sensor, and an air pressure sensor. Data interface 1726 is connected to, for example, an in-vehicle network 1741 via a terminal not shown, and acquires data generated by the vehicle (such as vehicle speed data).

[0282] Content player 1727 reproduces content stored on storage media (such as CDs and DVDs), which is inserted into storage media interface 1728. Input device 1729 includes, for example, a touch sensor, button, or switch configured to detect touch on the screen of display device 1730, and receives operations or information input from the user. Display device 1730 includes a screen such as an LCD or OLED display and displays images or reproduced content for navigation functions. Speaker 1731 outputs sound for navigation functions or reproduced content.

[0283] The wireless communication interface 1733 supports any cellular communication scheme (such as 4G LTE or 5G NR) and performs wireless communication. The wireless communication interface 1733 typically includes, for example, a BB processor 1734 and RF circuitry 1735. The BB processor 1734 can perform, for example, encoding / decoding, modulation / demodulation, and multiplexing / demultiplexing, and performs various types of signal processing for wireless communication. Meanwhile, the RF circuitry 1735 can include, for example, a mixer, filter, and amplifier, and transmits and receives wireless signals via antenna 1737. The wireless communication interface 1733 can also be a chip module on which the BB processor 1734 and RF circuitry 1735 are integrated. Figure 27 As shown, the wireless communication interface 1733 may include multiple BB processors 1734 and multiple RF circuits 1735. Although Figure 27 An example is shown in which the wireless communication interface 1733 includes multiple BB processors 1734 and multiple RF circuits 1735, but the wireless communication interface 1733 may also include a single BB processor 1734 or a single RF circuit 1735.

[0284] In addition to cellular communication schemes, the wireless communication interface 1733 can support other types of wireless communication schemes, such as short-range wireless communication schemes, near-field communication schemes, and wireless LAN schemes. In this case, for each wireless communication scheme, the wireless communication interface 1733 may include a BB processor 1734 and an RF circuit 1735.

[0285] Each of the antenna switches 1736 switches the connection destination of the antenna 1737 among multiple circuits (such as circuits for different wireless communication schemes) included in the wireless communication interface 1733.

[0286] Antenna 1737 includes multiple antenna elements. Antenna 1737 may be arranged, for example, as an antenna array matrix, and used by wireless communication interface 1733 to transmit and receive wireless signals.

[0287] Furthermore, the car navigation device 1720 may include an antenna 1737 for each wireless communication scheme. In this case, the antenna switch 1736 can be omitted from the configuration of the car navigation device 1720.

[0288] Battery 1738 via feeder to Figure 27 The various blocks of the car navigation device 1720 shown are powered, and the feeders are partially shown as dashed lines in the figure. Battery 1738 accumulates the power supplied from the vehicle.

[0289] exist Figure 27 In the car navigation device 1720 shown in the figure, refer to Figure 22A One or more components (e.g., reference signal transmitting unit 202, access assistance information receiving unit 203) included in the described processing circuitry 201 may be implemented in the wireless communication interface 1733. Alternatively, at least a portion of these components may be implemented in the processor 1721. As an example, the car navigation device 1720 includes a portion (e.g., BB processor 1734) or the entirety of the wireless communication interface 1733, and / or a module including the processor 1721, and one or more components may be implemented in the module. In this case, the module may store a program that allows the processor to function as one or more components (in other words, a program that allows the processor to perform the operation of one or more components), and may execute the program. As another example, a program that allows the processor to function as one or more components may be installed in the car navigation device 1720, and the wireless communication interface 1733 (e.g., BB processor 1734) and / or the processor 1721 may execute the program. As described above, the car navigation device 1720 or the module may be provided as a device including one or more components, and a program that allows the processor to function as one or more components may be provided. Additionally, a readable medium in which the program is recorded can be provided.

[0290] The technology disclosed herein can also be implemented as an in-vehicle system (or vehicle) 1740 including one or more blocks of an automotive navigation device 1720, an in-vehicle network 1741, and a vehicle module 1742. The vehicle module 1742 generates vehicle data (such as vehicle speed, engine speed, and fault information) and outputs the generated data to the in-vehicle network 1741.

[0291] Exemplary embodiments of the present disclosure have been described above with reference to the accompanying drawings; however, the present disclosure is by no means limited to the examples described above. Various changes and modifications can be made by those skilled in the art within the scope of the appended claims, and it should be understood that such changes and modifications naturally fall within the technical scope of the present disclosure.

[0292] For example, the multiple functions included in one unit in the above embodiments can be implemented by separate devices. Alternatively, the multiple functions implemented by multiple units in the above embodiments can be implemented by separate devices respectively. In addition, one of the above functions can be implemented by multiple units. Needless to say, such a configuration is included within the scope of the present disclosure.

[0293] In this specification, the steps described in the flowchart include not only processes executed sequentially in the stated order, but also processes executed in parallel or individually, rather than necessarily sequentially. Furthermore, even within the steps of sequential processing, needless to say, the order can be appropriately altered.

[0294] While this disclosure and its advantages have been described in detail, it should be understood that various changes, substitutions, and modifications can be made without departing from the spirit and scope of this disclosure as defined by the appended claims. Furthermore, the terms "comprising," "including," or any other variations thereof used in embodiments of this disclosure are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

Claims

1. An electronic device for a low-frequency base station, comprising: The processing circuit is configured as follows: The channel state information (CSI) matrix is ​​obtained based on the reference signal received from the user equipment via low-frequency communication; By inputting the CSI matrix and the location data of multiple high-frequency base stations into a neural network, the neural network is used to determine candidate high-frequency base stations suitable for high-frequency communication with the user equipment from the multiple high-frequency base stations; Determine the access assistance information associated with the candidate high-frequency base station; as well as The access assistance information is sent to the user equipment.

2. The electronic device of claim 1, wherein the processing circuit is further configured to: The high-frequency base station with the minimum path loss for high-frequency communication with the user equipment is identified as the candidate high-frequency base station.

3. The electronic device of claim 2, wherein the processing circuit is further configured to: The path loss value corresponding to the candidate high-frequency base station is sent to the user equipment.

4. The electronic device of claim 2, wherein the processing circuit is further configured to: If the path loss value corresponding to the candidate high-frequency base station is lower than a predetermined threshold, the user equipment is instructed to activate the high-frequency communication module.

5. The electronic device of claim 1, wherein the processing circuit is further configured to: By inputting the CSI matrix along with the location data and transmit power of the plurality of high-frequency base stations into the neural network, the high-frequency base station with the maximum receive power at the user equipment is determined as the candidate high-frequency base station.

6. The electronic device of claim 5, wherein the processing circuit is further configured to: The received power value corresponding to the candidate high-frequency base station is sent to the user equipment.

7. The electronic device of claim 5, wherein the processing circuit is further configured to: If the received power value corresponding to the candidate high-frequency base station exceeds a predetermined threshold, the user equipment is instructed to activate the high-frequency communication module.

8. The electronic device of claim 1, wherein the processing circuit is further configured to: Using the neural network, the beams of the candidate high-frequency base stations for the user equipment are determined; The user equipment identification code and the beam information are notified to the candidate high-frequency base station.

9. The electronic device of claim 8, wherein the information about the beam is an index of the synchronization signal / physical broadcast channel block (SSB) corresponding to the beam.

10. The electronic device of claim 1, wherein the access assistance information includes the frequency location of the synchronization signal / physical broadcast channel block (SSB) of the candidate high-frequency base station.

11. The electronic device of claim 8, wherein the access assistance information includes the frequency location and index of the synchronization signal / physical broadcast channel block (SSB) of the candidate high-frequency base station.

12. The electronic device of claim 1, wherein the processing circuit is further configured to: The user equipment that communicates with any of the plurality of high-frequency base stations at high frequency receives a reference signal transmitted via low-frequency communication and identification information of any of the high-frequency base stations. The parameters of the neural network are updated by using the CSI matrix obtained based on the reference signal and the location data of any high-frequency base station as input and the identification information of any high-frequency base station as output.

13. The electronic device of claim 2, wherein the processing circuit is further configured to: The user equipment that communicates with any of the plurality of high-frequency base stations receives the path loss value of the high-frequency communication estimated by the user equipment. The parameters of the branch network of the neural network are updated by using the path loss value as the output.

14. The electronic device of claim 2, wherein the processing circuit is further configured to: The high-frequency base station that communicates with the user equipment from the plurality of high-frequency base stations receives the identification code of the user equipment and information about the beam used for the high-frequency communication. The parameters of the branch network of the neural network are updated by using the information about the beam used for the high-frequency communication as output.

15. The electronic device of claim 1, wherein the low-frequency communication operates in the LTE band, the LTE-A band, or the sub-6GHz band, and wherein, The high-frequency communication operates in the millimeter-wave band.

16. An electronic device for a user equipment, comprising: The processing circuit is configured as follows: Reference signals are sent to low-frequency base stations via low-frequency communication so that the low-frequency base stations can obtain the channel state information (CSI) matrix; Receive access assistance information associated with candidate high-frequency base stations determined by low-frequency base stations, wherein the candidate high-frequency base stations are high-frequency base stations suitable for high-frequency communication with the user equipment, determined by the low-frequency base stations by inputting the CSI matrix and the location data of multiple high-frequency base stations into a neural network. as well as The candidate high-frequency base station is accessed using the access assistance information.

17. The electronic device of claim 16, wherein the candidate high-frequency base station is the high-frequency base station with the minimum path loss for high-frequency communication with the user equipment predicted by the neural network.

18. The electronic device of claim 16, wherein the candidate high-frequency base station is the high-frequency base station with the highest received power at the user equipment predicted by the neural network for high-frequency communication with the user equipment.

19. The electronic device of claim 17, wherein the processing circuitry is further configured to: Receive the path loss value corresponding to the candidate high-frequency base station from the low-frequency base station; and If the path loss value is lower than a predetermined threshold, the high-frequency communication module is activated to access the candidate high-frequency base station.

20. The electronic device of claim 18, wherein the processing circuitry is further configured to: Receive the received power value corresponding to the candidate high-frequency base station from the low-frequency base station; and If the received power value exceeds a predetermined threshold, the high-frequency communication module is activated to access the candidate high-frequency base station.

21. The electronic device of claim 16, wherein the access assistance information includes the frequency location of the synchronization signal / physical broadcast channel block (SSB) of the candidate high-frequency base station.

22. The electronic device of claim 21, wherein the access assistance information further includes an index of the SSB of the candidate high-frequency base station.

23. The electronic device of claim 19 or 20, wherein the processing circuitry is further configured to: The predetermined threshold is adjusted based on the user device's current battery level, connection preferences, and the transmission success rate of high-frequency communication.

24. An electronic device for a high-frequency base station, comprising: The processing circuit is configured as follows: The user equipment receives an identification code from a low-frequency base station and information about a beam available to the user equipment from a high-frequency base station, wherein the beam is determined by the low-frequency base station by inputting a CSI matrix obtained based on a reference signal transmitted by the user equipment via a low-frequency link and location data from multiple high-frequency base stations into a neural network. The beam is used to establish high-frequency communication with the user equipment.

25. The electronic device of claim 24, wherein the processing circuitry is further configured to: Based on the beam, a narrower beam for high-frequency communication with the user equipment is determined by beam scanning.

26. The electronic device of claim 25, wherein the beam comprises a plurality of beams with different priorities, and the processing circuitry is further configured to: The beams are scanned based on their priority.

27. A method for training a neural network, comprising: Receive reference signals transmitted via low-frequency communication from user equipment; The channel state information (CSI) matrix is ​​obtained based on the reference signal; Receive identification information of the high-frequency base station with which the user equipment communicates at high frequencies; The parameters of the neural network are determined by using the CSI matrix and the location data of multiple high-frequency base stations as inputs and the identification information of the high-frequency base stations as outputs through deep learning.

28. The method of claim 27, further comprising: Receive path loss values ​​for high-frequency communication estimated by the user equipment. The parameters of the branch networks of the neural network are determined by using the path loss value as the output.

29. The method of claim 27, further comprising: Receive the received power value for high-frequency communication as measured by the user equipment. The parameters of the branch networks of the neural network are determined by using the received power value as the output.

30. The method of claim 27, further comprising: The system receives the identification code of the user equipment and information about the beam used for the high-frequency communication from the high-frequency base station that communicates with the user equipment at high frequencies. The parameters of the branch networks of the neural network are determined by using information about the beam used for the high-frequency communication as output.

31. A communication method, comprising: The channel state information (CSI) matrix is ​​obtained based on the reference signal received from the user equipment via low-frequency communication; By inputting the CSI matrix and the location data of multiple high-frequency base stations into a neural network, the neural network is used to determine candidate high-frequency base stations suitable for high-frequency communication with the user equipment from the multiple high-frequency base stations; Determine the access assistance information associated with the candidate high-frequency base station; as well as The access assistance information is sent to the user equipment.

32. A communication method, comprising: Reference signals are sent to low-frequency base stations via low-frequency communication so that the low-frequency base stations can obtain the channel state information (CSI) matrix; Receive access assistance information associated with candidate high-frequency base stations determined by low-frequency base stations, wherein the candidate high-frequency base stations are high-frequency base stations suitable for high-frequency communication with user equipment, determined by the low-frequency base stations by inputting the CSI matrix and the location data of multiple high-frequency base stations into a neural network. as well as The candidate high-frequency base station is accessed using the access assistance information.

33. A communication method, comprising: The user equipment receives an identification code from a low-frequency base station and information about a beam from a high-frequency base station that can be used by the user equipment, wherein the beam is determined by the low-frequency base station by inputting a CSI matrix obtained based on a reference signal transmitted by the user equipment via a low-frequency link and location data from multiple high-frequency base stations into a neural network. The beam is used to establish high-frequency communication with the user equipment.

34. A non-transitory computer-readable storage medium storing executable instructions that, when executed, implement the method as described in any one of claims 27-33.

Images