A method and device for sensing-assisted multi-user beam management

Abstract

An embodiment of the present invention provides a perception-assisted multi-user beam management method and device, including: a base station perceives and identifies UEs to obtain motion information of each UE, determines, for each UE, a target transmit beam set for the UE from a complete beam set based on its motion information, traverses all directions of the target transmit beam set to send a main synchronization signal to the UE, receives and determines a first alignment direction with the UE based on a random access preamble sent by the UE, establishes an initial connection with the UE, obtains initial physical feature information of the UE, establishes an initial association relationship between GUTI information and the initial physical feature information, sends a target signal to the UE through the first alignment direction, and receives an echo signal reflected by the UE, determines a long-term association relationship between each GUTI information and the physical feature information of each UE based on each echo signal and the initial association relationship, thereby reducing the initial access delay and improving the alignment accuracy between the base station and the UE.

Description

Technical Field

[0001] The present invention relates to the field of communication technology, specifically to the fields of millimeter waves, beam management, visual perception, radar perception, communication perception integration, and more particularly to a perception-assisted multi-user beam management method and device. Background Art

[0002] The mmWave (millimeter wave) frequency band is considered a driving force behind the performance requirements of 5G (fifth generation mobile communication technology). Compared to sub-6 GHz (electromagnetic waves with frequencies below 6 GHz) bands, mmWave bands offer greater bandwidth and higher data rates. Signals propagating in the mmWave band are blocked by many common materials, such as bricks and mortar, resulting in signal strength degradation and severe path loss. Therefore, the user equipment (UE) and gNB (next generation Node B) in next-generation cellular networks require highly directional transmission links to reduce path loss. To this end, high-dimensional phased arrays are often used in the mmWave band to benefit from the resulting beamforming gain and maintain acceptable communication quality. Establishing directional links requires a set of operations between the transmitter and receiver called beam management. Beam management is fundamental to performing various control communication tasks, including initial access and beam tracking. Among them, initial access drives the UE to initialize the physical link connection with the gNB, and the purpose of beam tracking is to achieve alignment between the gNB and the mobile UE by adjusting the antenna array direction between the gNB and the UE.

[0003] In related technologies, initial access is achieved through an exhaustive search method, and beam tracking is achieved through a time correlation prediction method based on communication feedback. Specifically, the gNB and UE traverse all their respective beam directions and select the beam direction with the highest communication SNR (Signal to Noise Ratio) as the alignment direction, and perform initial access between the gNB and the UE using this alignment direction. Furthermore, when the gNB sends a communication signal to the UE, a pilot signal is sent to the UE via a downlink. After receiving the pilot signal, the UE determines the direction corresponding to the highest SNR and feeds back data to the gNB via the uplink in this direction. The gNB calculates the time correlation between each data frame fed back by the UE. Based on each data frame fed back by the UE and the time correlation between each data frame, the Kalman filter algorithm is used to predict the UE's mobility to update the beam direction on the gNB side.

[0004] However, the exhaustive search method requires searching all beam directions between the gNB and UE, which takes a long time and results in a long initial access delay. Furthermore, the time-correlation prediction method based on communication feedback requires using data (such as channel state) fed back by the UE in the previous period to estimate the current channel state. However, the channel state estimate from the previous period may be outdated by the current period, making it possible for the gNB and UE to misalign their beams due to time-varying channel conditions. Furthermore, in high-mobility communication scenarios, the gNB and UE may lose a brief opportunity for beam alignment, resulting in poor gNB-UE alignment accuracy. Summary of the Invention

[0005] The purpose of the embodiments of the present invention is to provide a method and apparatus for sensing-assisted multi-user beam management to reduce the initial access latency and improve the alignment accuracy between the NB and UE during beam tracking. The specific technical solution is as follows:

[0006] In the first aspect, an embodiment of the present invention provides a perception-assisted multi-user beam management method, which is applied to the base station side. The method includes: perceiving and identifying the user terminal device UE through a first target sensor device to obtain the motion information of each UE; for each UE, based on the motion information of the UE, determining the target transmission beam direction set for the UE from a first preset complete beam set; traversing all directions in the target transmission beam direction set for the UE, and sending a main synchronization signal to the UE; receiving the random access preamble code sent by the UE from each direction in the target transmission beam direction set for the UE, and based on the received random access preamble code, The random access preamble determines the first alignment direction with the UE, and the random access preamble contains the UE's globally unique temporary identity GUTI information; based on the first alignment direction, an initial connection is established with the UE, the initial physical feature information of the UE is obtained, and an initial association relationship between the UE's GUTI information and the initial physical feature information is established; a target signal is sent to the UE through the first alignment direction, and an echo signal reflected by the UE is received; based on the echo signal of each UE and the initial association relationship between the GUTI information of each UE and the initial physical feature information, a long-term association relationship between the GUTI information of each UE and its respective physical feature information is determined.

[0007] Optionally, the motion information includes location information; for each UE, based on the motion information of the UE, a target transmission beam direction set for the UE is determined from a first preset complete beam set, including: for each UE, based on the location information of the UE, determining the relative angle between the UE and the base station NB; based on the relative angle, determining the target transmission beam direction set for the UE from the first preset complete beam set.

[0008] Optionally, traversing all directions in a target transmit beam direction set for the UE and sending a primary synchronization signal to the UE includes: traversing each direction in the target transmit beam direction set for the UE in turn, and sending a primary synchronization signal to the UE once through the direction at a preset time interval;

[0009] Receiving a random access preamble sent by the UE from each direction in a target transmit beam direction set for the UE, and determining a first alignment direction with the UE based on the received random access preamble, including: receiving a random access preamble sent by the UE from each direction in a target transmit beam direction set for the UE, and determining a direction corresponding to a highest signal-to-noise ratio (SNR) of the received random access preamble as the first alignment direction with the UE.

[0010] Optionally, an initial connection is established with the UE based on the first alignment direction, the initial physical feature information of the UE is obtained, and an initial association relationship between the GUTI information of the UE and the initial physical feature information is established, including: sending an access signal to the UE through the first alignment direction, and receiving a response signal fed back by the UE; based on the response signal, determining the current physical feature information of the UE; and establishing an initial association relationship between the GUTI information of the UE and the current physical feature information.

[0011] Optionally, based on the echo signal of each UE and the initial association relationship between the GUTI information of each UE and the initial physical feature information, the long-term association relationship between the GUTI information of each UE and its respective physical feature information is determined, including: estimating the motion state of each UE based on the echo signal of each UE, and determining the similarity between the states of every two UEs in adjacent time slots based on the estimation results of each state; determining the similarity between every two UEs in adjacent time slots based on the similarity between the states of every two UEs in adjacent time slots; tracking and matching the physical feature information of each UE in different time slots based on the similarity between every two UEs in adjacent time slots to obtain the association relationship between the physical feature information of each UE in adjacent time slots; determining the long-term association relationship between the GUTI information of each UE and its respective physical feature information based on the association relationship between the physical feature information of each UE in adjacent time slots and the initial association relationship between the GUTI information of each UE and the initial physical feature information.

[0012] Optionally, the motion state of each UE is estimated based on the echo signal of each UE, and the similarity between the states of every two UEs in adjacent time slots is determined according to the state estimation results, including: for each UE, based on the echo signal of the UE, determining the state information of the UE at the current moment, the state information includes: position information, speed information and angle information; based on the state information of the UE at the current moment, using the extended Kalman filter algorithm to track the state of the UE to obtain the state estimation information of the UE at the next moment; based on the state information of each UE at the current moment and the state estimation information of the next moment, calculating the similarity between the states of every two UEs in adjacent time slots.

[0013] Optionally, based on the similarity between every two UEs in adjacent time slots, the physical feature information of each UE in different time slots is tracked and matched to obtain the correlation between the physical feature information of each UE in adjacent time slots, including: based on the state information of each UE at the current moment, the state estimation information at the next moment, and the similarity between every two UEs in adjacent time slots, a weighted bipartite graph matching model is established to track and match the physical feature information of each UE in different time slots to obtain the correlation between the physical feature information of each UE in adjacent time slots.

[0014] In the second aspect, an embodiment of the present invention provides a perception-assisted multi-user beam management method, which is applied to the terminal side, and the method includes: perceiving and identifying the base station NB through a second target sensor device to obtain the location information of the NB; based on the location information of the NB, determining a target receiving beam direction set from a second preset complete beam set; traversing all directions in the target receiving beam direction set to receive the main synchronization signal sent by the NB; based on the received main synchronization signal, determining a second alignment direction with the NB, and sending a random access preamble code to the NB in ​​the second alignment direction, the random access preamble code containing the globally unique temporary identity GUTI information of the user terminal device UE; receiving the target signal sent by the NB through the second alignment direction, and reflecting the echo signal to the NB, so that the NB determines the long-term association relationship between the UE's GUTI information and its physical feature information based on the echo signal.

[0015] In a third aspect, an embodiment of the present invention provides a perception-assisted multi-user beam management device, which is applied to the base station side, and the device includes: a first perception module, which is used to perceive and identify the user terminal device UE through a first target sensor device to obtain the motion information of each UE; a first beam direction determination module, which is used to determine, for each UE, a target transmission beam direction set for the UE from a first preset complete beam set based on the motion information of the UE; a first signal sending module, which is used to traverse all directions in the target transmission beam direction set for the UE and send a main synchronization signal to the UE; a first alignment direction determination module, which is used to receive the random access preamble code sent by the UE from each direction in the target transmission beam direction set for the UE, and based on the A first alignment direction between the UE and the random access preamble is determined based on the received random access preamble, where the random access preamble contains the UE's globally unique temporary identity GUTI information; a connection establishment module is used to establish an initial connection with the UE based on the first alignment direction, obtain the initial physical feature information of the UE, and establish an initial association relationship between the UE's GUTI information and the initial physical feature information; a second signal sending module is used to send a target signal to the UE through the first alignment direction and receive an echo signal reflected by the UE; a beam tracking module is used to determine the long-term association relationship between the GUTI information of each UE and its respective physical feature information based on the echo signal of each UE and the initial association relationship between the GUTI information of each UE and the initial physical feature information.

[0016] In a fourth aspect, an embodiment of the present invention provides a perception-assisted multi-user beam management device, which is applied to the terminal side, and the device includes: a second perception module, which is used to perceive and identify the base station NB through a second target sensor device to obtain the location information of the NB; a second beam direction determination module, which is used to determine the target receiving beam direction set from the second preset complete beam set based on the location information of the NB; a first signal receiving module, which is used to traverse all directions in the target receiving beam direction set and receive the main synchronization signal sent by the NB; a second alignment direction determination module, which is used to determine the second alignment direction between the NB and the NB based on the received main synchronization signal, and send a random access preamble code to the NB in ​​the second alignment direction, wherein the random access preamble code contains the globally unique temporary identity GUTI information of the user terminal device UE; a second signal receiving module, which is used to receive the target signal sent by the NB through the second alignment direction, and reflect the echo signal to the NB, so that the NB can determine the long-term association relationship between the GUTI information of the UE and its physical feature information based on the echo signal.

[0017] In a fifth aspect, an embodiment of the present invention provides an electronic device comprising a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other through the communication bus; the memory is used to store computer programs; and the processor is used to implement any of the above-described method steps when executing the program stored in the memory.

[0018] In a sixth aspect, an embodiment of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, any of the method steps described above is implemented.

[0019] Beneficial effects of the embodiments of the present invention: The embodiments of the present invention provide a perception-assisted multi-user beam management method and device, in which the base station side perceives and identifies the UE to obtain the motion information of each UE, and then determines the target transmission beam direction set from the corresponding complete beam set based on the motion information of each UE, and the terminal side perceives and identifies the NB to obtain the position information of the NB, and then determines the target receiving beam direction set from the preset complete beam set based on the position information of the NB, which narrows the range of beam traversal during the initial access process, reduces the overhead and delay during the initial access, and facilitates rapid access between the base station and the UE. During the beam tracking process, the NB can use the echo signal of each UE to track the physical characteristic information of each UE for predictive beamforming in the next time slot, eliminating the downlink pilot and uplink feedback in the beam tracking process; further, by establishing a physical characteristic association relationship between adjacent time slots of each UE, based on the initial association relationship between the UE's GUTI information and the initial physical characteristic information established at the time of initial access, the long-term association relationship between the GUTI information and the physical characteristic information of each UE can be obtained, thereby ensuring the alignment accuracy between the NB and the UE during the beam tracking process.

[0020] Of course, it is not necessary to achieve all of the advantages described above simultaneously in order to implement any product or method of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other embodiments can also be obtained based on these drawings.

[0022] Figure 1 Schematic diagram of beam pointing error in a multi-user scenario during beam tracking in related technologies;

[0023] Figure 2A schematic diagram of a flow chart of a perception-assisted multi-user beam management method provided by an embodiment of the present invention;

[0024] Figure 3 A schematic diagram of binocular stereo vision geometric imaging provided by an embodiment of the present invention;

[0025] Figure 4 A schematic diagram of a weighted bipartite graph matching model provided by an embodiment of the present invention;

[0026] Figure 5 A schematic diagram of a physical feature association relationship provided by an embodiment of the present invention;

[0027] Figure 6 A schematic flow chart of another sensing-assisted multi-user beam management method provided by an embodiment of the present invention;

[0028] Figure 7 A schematic diagram of perception-assisted multi-user beam management provided by an embodiment of the present invention;

[0029] Figure 8 A schematic diagram of beam alignment comparison provided by an embodiment of the present invention;

[0030] Figure 9 A schematic diagram comparing the accuracy of association between a terminal and a physical feature provided by an embodiment of the present invention;

[0031] Figure 10 A schematic diagram of the weight distribution of location characteristics and speed characteristics of terminals in different time slots provided by an embodiment of the present invention;

[0032] Figure 11a A schematic diagram showing a comparison of the first angle estimation method provided by an embodiment of the present invention;

[0033] Figure 11b A schematic diagram showing a comparison of the second angle estimation method provided by an embodiment of the present invention;

[0034] Figure 11c A schematic diagram showing a comparison of the third angle estimation method provided by an embodiment of the present invention;

[0035] Figure 11d A schematic diagram showing a fourth angle estimation comparison method provided by an embodiment of the present invention;

[0036] Figure 12a A schematic diagram showing a comparison of the first angle estimation error provided by an embodiment of the present invention;

[0037] Figure 12b A schematic diagram showing a comparison of the second angle estimation error provided by an embodiment of the present invention;

[0038] Figure 12cA schematic diagram showing a comparison of the third angle estimation error provided by an embodiment of the present invention;

[0039] Figure 12d A schematic diagram showing a comparison of the fourth angle estimation error provided by an embodiment of the present invention;

[0040] Figure 13a A schematic diagram showing a comparison of beam average reachability performance provided by an embodiment of the present invention;

[0041] Figure 13b A schematic diagram showing another comparison of beam average reachability performance provided by an embodiment of the present invention;

[0042] Figure 14 A schematic structural diagram of a perception-assisted multi-user beam management device provided by an embodiment of the present invention;

[0043] Figure 15 A schematic structural diagram of another perception-assisted multi-user beam management device provided by an embodiment of the present invention;

[0044] Figure 16 A schematic structural diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0045] The following will be combined with the accompanying drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field based on this application are within the scope of protection of the present invention.

[0046] In the related art, initial access is achieved through an exhaustive search method, and beam tracking is achieved through a time correlation prediction method based on communication feedback. The exhaustive search method needs to traverse and search all beam directions of the gNB and UE, which takes a long time and results in a large initial access delay. The time correlation prediction method based on communication feedback needs to use the data fed back by the UE in the previous cycle (such as the channel state) to estimate the channel state of the current period, but the channel state estimate of the previous cycle may be outdated in the current period, so that the beam between the gNB and the UE may not be aligned due to the time-varying channel state, and in high-mobility communication scenarios, the gNB and the UE may lose a short opportunity for beam alignment, resulting in poor alignment accuracy between the gNB and the UE. In addition, in the beam tracking process of multiple users, the similarity of features between different UEs is ignored, and UE state association errors are prone to occur, which in turn leads to the problem that a specific beam points to an unintended receiver (UE). For example, as Figure 1 As shown, the UEs are UAV 1, UAV 2 and UAV 3, and at the n-1th moment ( Figure 1 On the left side, the gNB forms directional beams for three UEs respectively. Each beam carries specific information about the intended UE. Due to the mobility of the UE, the positions of UAV 2 and UAV 3 at time n ( Figure 1 However, when updating the state estimation equation using the time correlation prediction method based on communication feedback, the states of UAV 2 and UAV 3 are easily confused, resulting in the beam at time n pointing to an unintended receiver.

[0047] To solve at least one of the above problems, an embodiment of the present invention provides a perception-assisted multi-user beam management method, in which the base station side perceives and identifies the user terminal device UE through a first target sensor device to obtain the motion information of each UE; for each UE, based on the motion information of the UE, a target transmission beam direction set for the UE is determined from a first preset complete beam set; all directions in the target transmission beam direction set for the UE are traversed, and a main synchronization signal is sent to the UE; a random access preamble code sent by the UE is received from each direction in the target transmission beam direction set for the UE, and a primary synchronization signal is sent to the UE based on the received random access preamble code. The guide code determines the first alignment direction with the UE, and the random access preamble code contains the UE's globally unique temporary identity GUTI information; based on the first alignment direction, an initial connection is established with the UE, the initial physical feature information of the UE is obtained, and an initial association relationship between the UE's GUTI information and the initial physical feature information is established; a target signal is sent to the UE through the first alignment direction, and an echo signal reflected by the UE is received; based on the echo signal of each UE and the initial association relationship between the GUTI information of each UE and the initial physical feature information, a long-term association relationship between the GUTI information of each UE and its respective physical feature information is determined.

[0048] The terminal side senses and identifies the base station NB through the second target sensor device to obtain the location information of the NB; based on the location information of the NB, a target receiving beam direction set is determined from the second preset complete beam set; all directions in the target receiving beam direction set are traversed to receive the main synchronization signal sent by the NB; based on the received main synchronization signal, a second alignment direction with the NB is determined, and a random access preamble code is sent to the NB in ​​the second alignment direction; the target signal sent by the NB is received through the second alignment direction, and an echo signal is reflected to the NB, so that the NB can determine the long-term association between the UE's GUTI information and its physical feature information based on the echo signal.

[0049] An embodiment of the present invention provides a sensing-assisted multi-user beam management method. The base station side senses and identifies the UE through sensing and identification to obtain the motion information of each UE, and then determines the target transmission beam direction set from the corresponding complete beam set based on the motion information of each UE. The terminal side senses and identifies the NB through sensing and identification to obtain the location information of the NB, and then determines the target reception beam direction set from the preset complete beam set based on the location information of the NB. This reduces the range of beam traversal during the initial access process, reduces the overhead and delay during the initial access, and facilitates rapid access between the base station and the UE. During the beam tracking process, the NB can use the echo signal of each UE to track the physical feature information of each UE for predictive beamforming in the next time slot, eliminating the downlink pilot and uplink feedback during the beam tracking process. Furthermore, by establishing a physical feature association relationship between adjacent time slots of each UE, based on the initial association relationship between the UE's GUTI information and the initial physical feature information established at the time of initial access, a long-term association relationship between the GUTI information and the physical feature information of each UE can be obtained, ensuring the alignment accuracy between the NB and the UE during the beam tracking process.

[0050] The perception-assisted multi-user beam management method provided in embodiments of the present invention can be applied to any multi-user beam management scenario. In one scenario, the Node B (NB) can be a gNB, and the UE can be a drone or other terminal device capable of perception and recognition. Both the gNB and UE operate in the mmWave frequency band and are equipped with massive multiple-input-multiple-output (mMIMO) UPA (Uniform Planar Array) technology. They also have perception capabilities using computer vision and integrated sensing and communication (ISAC) technology. The gNB is equipped with independent transmit and receive antennas, enabling it to receive echo signals for perception while maintaining uninterrupted downlink communication. Before initial beam engagement, since there is no need for radio transmission, computer vision is used to provide perception capabilities. During beam tracking, ISAC technology provides both communication and perception functions. ISAC technology achieves unified design of communication and perception functions through joint design of air interfaces and protocols, reuse of time-frequency-space resources, and sharing of hardware devices. This enables wireless networks to achieve high-precision and refined perception functions while performing high-quality communication interactions, thereby improving overall network performance and business capabilities.

[0051] Compared to existing systems where the gNB and UE passively "hear" each other's presence only in the radio domain, the two are more like a blind person with fully functional hearing. This embodiment of the present invention fully utilizes sensory information from the "visual domain" to improve beam management efficiency. The gNB applies radar perception and visual (camera) perception to beam management, employing both ISAC and computer vision technologies in the downlink. Video frames generated using computer vision can be used to perceive UE location information during initial access, while ISAC technology receives echo signals reflected from the UE for beam tracking. The gNB transmits ISAC signals and receives echo signals reflected from the UE, eliminating downlink pilots and uplink feedback during beam tracking, achieving significant matched filter gain. During beam tracking, an extended Kalman filter is used to accurately estimate user mobility. In multi-user beam tracking, a similarity definition method based on relative entropy and a feature weighting scheme based on popularity are employed. A weighted bipartite graph matching model is established to correlate multi-user state measurements and predictions, enabling accurate alignment of specific beams with intended UEs.

[0052] The following describes in detail the perception-assisted multi-user beam management method provided by the present invention through specific embodiments.

[0053] See also Figure 2 , Figure 2 An embodiment of the present invention provides a perception-assisted multi-user beam management method, which is applied to a base station side and includes:

[0054] S201: UEs are sensed and identified by a first target sensor device to obtain motion information of each UE.

[0055] In this embodiment of the present invention, the NB may specifically be a gNB. The gNB and UE each utilize computer vision technology to perceive and identify each other. The perception and identification process includes target detection and state measurement. The UE motion information includes the UE's location and speed information. The first target sensor device may be a sensor device such as a radar or a camera.

[0056] For example, during target detection, the gNB uses computer vision technology to capture video frames and then uses deep learning models such as YOLO (You Only Look Once) to detect the presence of a UE in the captured video frames to confirm the gNB's perception of the UE. Furthermore, during state measurement, the gNB uses a binocular camera to measure the relative distance and speed between the perceived UE and the gNB, as well as the perceived position and velocity of the UE.

[0057] For example, the two cameras of a binocular camera are used to obtain different information about the same target from two perspectives, thereby obtaining depth information from the depth of field of the image. Figure 3 As shown in the figure, taking the gNB binocular camera detecting a UAV (UE) as an example, the specific ranging principle is as follows. According to the similarity relationship B′ / d′=B / d, we can get Therefore, the relative distance between the gNB binocular camera and the drone is d = Bf l / (X L -X R ), where B represents the optical center distance between the two cameras (binocular cameras), f l represents the focal length of the camera, d′ represents the difference between the relative distance d and the focal length of the camera, L represents the maximum length of the single lens observation base, X L 、X R Respectively represent the endpoints of binocular imaging to P L ′、P R ′ length, P L ′、P R ′ respectively represent the UAV’s position at the binocular camera imaging point P L 、P R The mirror image position of P L ′ and P R ′, the distance between L -X R Describes the position difference of the spatial point mapped to the left image plane and the right image plane of the binocular camera. According to the above distance measurement method d = Bf l / (X L -X R ) can calculate the relative distance between the gNB binocular camera and the drone at different time slots. By recording these relative distances and combining them with the binocular camera's rotation angle and position, the relative position of the target (drone) and the gNB at each time slot can be calculated. Furthermore, based on the binocular camera's position and the relative position of the drone and gNB, the drone's true position can be calculated. Finally, based on the drone's true position at two adjacent moments (time slots), the drone's velocity can be calculated.

[0058] S202: For each UE, determine a target transmit beam direction set for the UE from a first preset complete beam set based on motion information of the UE.

[0059] In one example, after sensing and identifying the motion information of each UE, the relative angle between the UE and the base station is estimated based on the location information included in the UE's motion information. A preset number of directions adjacent to and to the left of the direction corresponding to the estimated relative angle in a first preset complete beam set are determined as a set of target transmit beam directions for the UE. The preset number can be set based on actual needs, and the specific value is not limited in this embodiment of the present invention.

[0060] S203: traverse all directions in the target transmission beam direction set for the UE and send a primary synchronization signal to the UE.

[0061] S204: Receive a random access preamble code sent by the UE from each direction in the target transmit beam direction set for the UE, and determine a first alignment direction with the UE based on the received random access preamble code.

[0062] In one example, the base station sequentially receives a random access preamble code sent by the UE from each direction in the target transmission beam direction set for the UE, where the random access preamble code includes the UE's GUTI (Global Unique Temporary Identity) information. The base station calculates whether the SNR in each direction of the random access preamble code received from the UE exceeds a threshold. If so, the direction in which the SNR exceeds the threshold is determined as the first alignment direction with the UE. The threshold value can be set according to actual needs, and the specific value of the threshold value is not limited in this embodiment of the present invention.

[0063] S205: Establish an initial connection with the UE based on the first alignment direction, obtain initial physical feature information of the UE, and establish an initial association relationship between the GUTI information of the UE and the initial physical feature information.

[0064] In one example, after determining a first alignment direction with a UE, an access request is sent to the UE via the first alignment direction, response information fed back by the UE is received, an initial connection is established with the UE, and initial physical characteristic information of the UE is perceived and identified via the received response information. This initial physical characteristic information may include location information, speed information, etc. of the UE. Furthermore, an initial association relationship is established between the UE's GUTI information and the initial physical characteristic information. This initial association relationship may be a mapping table, etc. The process of perceiving and identifying the initial physical characteristic information of lamp E via the received response information can refer to the aforementioned process of perceiving and identifying UE motion information, and will not be further described in detail in this embodiment of the present invention.

[0065] S206: Send a target signal to the UE in a first alignment direction, and receive an echo signal reflected by the UE.

[0066] The target signal may be a communication signal or an ISAC signal. The base station sends the target signal to the UE in a first alignment direction and receives an echo signal reflected by the UE through a radar.

[0067] For example, after the target signal sent by the base station is reflected by the UE, the signal received by the base station receiver can be expressed as:

[0068]

[0069] The echo signal reflected by the kth UE received by the base station is expressed as:

[0070]

[0071] c n (t) represents the echo signal at the nth moment, t represents the time variable of the signal waveform, K represents the number of UEs, represents the antenna gain factor of the gNB, β k，n 、μ k，n and τ k，n Respectively represent the number of receiving antennas, reflection coefficient, Doppler frequency and delay of the gNB. The subscripts k and n represent the variables corresponding to the kth UE at the nth time. t represents the number of transmit antennas of gNB, p k，n represents the transmit power of gNB, u k，n represents the steering vector of the gNB’s receive array, a k，n represents the transmission steering vector of the kth UE at the nth time, represents the signal of the kth UE at the nth time sent by the base station transmitting antenna, where s k，n (t) represents the non-beam-formed signal that the base station will send to the k-th UE at time n, z c (t) represents the variance The noise, c k，n (t) represents the echo signal of the kth UE at the nth time, f k，n Denotes the beamforming matrix F n The kth column vector of (used to beamform the signal of the kth UE), s k，n (t-τ k，n ) means s k，n (t) after τ k，n The delayed signal, z k，n (t) represents the variance σ 2 noise.

[0072] The SNR of the above echo signal transmission is expressed as The above μk，n and τ k，n It can be estimated by matched filtering. Therefore, the position vector p of the kth UE is k，n and velocity vector v k，n It can be obtained by the following expression:

[0073]

[0074] Among them, p k，n =[p k，n (1), P k，n (2), p k，n (3)] T , v k，n =[v k，n (1), v k，n (2), v k，n (3)] T Represent the position and velocity components of the kth UE on the x-axis, y-axis and z-axis respectively, Indicates v k，n The transpose of p b =[p b (1), p b (2), p b (3)] T represents the geographical location of the base station, f c represents the carrier frequency, c represents the speed of light, and μ k，n and τ k，n An estimated value of z f and z τ The variances are and The noise of and c k，n (t) is compensated and based on p k，n After normalizing with the matched filter gain, we can get Indicates c k，n (t) compensation, Represents the normalized noise. The likelihood function can be further defined as And the maximum likelihood estimator can be used Get the UE's azimuth angle φ k，n and pitch angle θ k，n Angle estimate of and Φ and Θ respectively represent the inclusion and The set of all possible values ​​of φ. k，n and θ k，n The variance of the measurement noise is expressed as and

[0075] S207 : Determine a long-term association relationship between the GUTI information of each UE and its physical feature information based on the echo signal of each UE and the initial association relationship between the GUTI information of each UE and the initial physical feature information.

[0076] After the base station receives the echo signal of each UE, it tracks the physical characteristic information (position information and speed information) of each UE according to the echo signal, determines the correlation between the physical characteristic information of each UE in adjacent time slots, and uses it for the predicted beamforming of the next time slot. Then, based on the correlation between the physical characteristic information of each UE in adjacent time slots, it traces back and combines the initial correlation between the GUTI information of each UE and the initial physical characteristic information to establish a long-term correlation between the GUTI information of each UE and its respective physical characteristic information. Then, through this long-term correlation, the target signal corresponding to each UE is accurately sent to the corresponding UE.

[0077] An embodiment of the present invention provides a perception-assisted multi-user beam management method, in which the base station side perceives and identifies the UE through perception and identification to obtain the motion information of each UE, and then determines the target transmission beam direction set from the corresponding complete beam set based on the motion information of each UE, thereby reducing the range of beam traversal during the initial access process, reducing the overhead and delay during the initial access, and facilitating rapid access between the base station and the UE. During the beam tracking process, the NB can use the echo signal of each UE to track the physical characteristic information of each UE for predictive beamforming of the next time slot, eliminating the downlink pilot and uplink feedback during the beam tracking process; further, by establishing a physical characteristic association relationship between adjacent time slots of each UE, based on the initial association relationship between the UE's GUTI information and the initial physical characteristic information established at the time of initial access, a long-term association relationship between the GUTI information and the physical characteristic information of each UE can be obtained, thereby ensuring the alignment accuracy between the NB and the UE during the beam tracking process.

[0078] In one possible implementation, the motion information includes location information. Accordingly, the implementation of step S202 for each UE, based on the motion information of the UE, determining the target transmission beam direction set for the UE from the first preset complete beam set, may include: for each UE, based on the location information of the UE, determining the relative angle between the UE and the base station NB; based on the relative angle, determining the target transmission beam direction set for the UE from the first preset complete beam set.

[0079] In one example, for each UE, the gNB calculates the relative angle between the UE and the gNB based on the UE's location information and the gNB's location information (the gNB knows its own location), determines the direction corresponding to the relative angle as an optimal direction, and selects a preset number of directions adjacent to the optimal direction from a first preset complete beam set as the target transmit beam direction set for the UE. For example, the first preset complete beam set can be expressed as:

[0080]

[0081] Among them, S B represents the first preset complete beam set, is the transmit beamforming vector on the gNB side, expressed as:

[0082]

[0083] Assuming that the UPA in gNB and UE have half-wavelength antenna spacing, and They represent the elements in the azimuth and elevation angle sets in the first preset complete beam set, m and m′ represent the indexes of the elements in the azimuth and elevation angle sets, respectively. Q B The number of bits representing the gNB's beam phase, n a and n b Respectively represent the index of the X-axis and Y-axis antenna array of UPA, N a and N b Respectively represent the number of X-axis and Y-axis antenna arrays of UPA, N t =N a ×N b denotes the total number of transmit antennas, φ∈[-π,π] and θ∈(0,π / 2] denote the azimuth and elevation components of the transmission angle from the gNB to the UE, respectively.

[0084] For example, for each UE, based on the location information of the UE and the location information of the gNB, the relative angle between the UE and the gNB is calculated, the direction corresponding to the relative angle is determined as the optimal direction, and then the sequence number in the first preset complete beam set corresponding to the optimal direction is selected in descending order. directions, and then select the best direction in ascending order directions, and obtain the target transmit beam direction set for the UE. Where floor() is a rounding function. represents the target transmission beam direction set, Represents the target transmission beam direction set The cardinality size of the set Cardinality The size depends on the number of bits QB of the gNB beam phase, that is, Determines the actual search domain size of the gNB transmit beam direction,

[0085] In this embodiment of the present invention, since the perceived UE position information may have certain errors, the relative angle calculated based on the perceived UE position information and the gNB position information may not be the accurate direction of alignment between the UE and the gNB. Therefore, based on the calculated relative angle, a target transmit beam direction set for the UE is determined from the first preset complete beam set, so as to further traverse each direction in the target transmit beam direction set and determine a more accurate alignment direction between the UE and the gNB.

[0086] In one possible implementation, the above-mentioned step S203 traverses all directions in the target transmission beam direction set for the UE and sends a main synchronization signal to the UE, which may include: traversing each direction in the target transmission beam direction set for the UE in turn, and sending a main synchronization signal to the UE through the direction once at a preset time interval.

[0087] For example, the preset duration can be set according to actual needs, for example, the preset duration can be 3 milliseconds, 5 milliseconds, or 10 milliseconds, etc.

[0088] For example, the transmitted signal of the gNB after beamforming at time n can be expressed as in, Indicates the primary synchronization signal to be sent to K UEs, s k，n (t) represents the primary synchronization signal sent by the gNB to the kth UE at time n, F n represents the beamforming matrix.

[0089] In a possible implementation, the above-mentioned step S204 receives the random access preamble code sent by the UE from each direction in the target transmission beam direction set for the UE, and determines the first alignment direction between the UE and the UE based on the received random access preamble code. The implementation method may include: receiving the random access preamble code sent by the UE from each direction in the target transmission beam direction set for the UE, and determining the direction corresponding to the highest signal-to-noise ratio SNR of the received random access preamble code as the first alignment direction between the UE.

[0090] While the gNB traverses all directions in the target transmit beam direction set for the UE and sends the primary synchronization signal to the UE, the UE traverses all directions in the target receive beam direction set it has determined, receives the primary synchronization signal from the gNB, and feeds back the random access preamble to the gNB. The gNB sequentially receives the random access preamble sent by the UE from each direction in the target transmit beam direction set for each UE and determines the direction with the highest received random access preamble signal-to-noise ratio (SNR) as the first alignment direction with the corresponding UE.

[0091] In one possible implementation, the above-mentioned step S205 establishes an initial connection with the UE based on the first alignment direction, obtains the initial physical feature information of the UE, and establishes an initial association relationship between the GUTI information of the UE and the initial physical feature information. The implementation method may include: sending an access signal to the UE through the first alignment direction, receiving a response signal fed back by the UE; determining the current physical feature information of the UE based on the response signal; and establishing an initial association relationship between the GUTI information of the UE and the current physical feature information.

[0092] For each UE, the gNB transmits an access signal to the UE via a first alignment direction. The access signal may be an ISAC signal, an access request signal, or the like. The gNB receives a response from the UE and establishes an initial connection with the UE. The gNB senses and identifies the UE's initial physical characteristic information based on the received response information. The initial physical characteristic information may include the UE's location information, velocity information, and other physical characteristic information representing the UE's physical characteristics at the time the gNB establishes the connection with the UE. Furthermore, the gNB establishes an initial association between the UE's GUTI information and the initial physical characteristic information. The initial association may be a mapping table, for example. The process of sensing and identifying the UE's initial physical characteristic information based on the received response information can be referenced to the process of sensing and identifying UE motion information described above and will not be further described in detail in this embodiment of the present invention.

[0093] In a possible implementation, the implementation of determining the long-term association relationship between the GUTI information of each UE and its respective physical feature information based on the echo signal of each UE and the initial association relationship between the GUTI information of each UE and the initial physical feature information in step S207 may include:

[0094] Step A: Estimate the motion state of each UE based on the echo signal of each UE, and determine the similarity between the states of every two UEs in adjacent time slots according to the estimation results of each state.

[0095] Step B: Based on the similarity between the states of every two UEs in adjacent time slots, determine the similarity between every two UEs in adjacent time slots.

[0096] Step C: Based on the similarity between every two UEs in adjacent time slots, the physical feature information of each UE in different time slots is tracked and matched to obtain the correlation relationship between the physical feature information of each UE in adjacent time slots.

[0097] Step D: Based on the association relationship between the physical feature information of each UE in adjacent time slots and the initial association relationship between the GUTI information of each UE and the initial physical feature information, determine the long-term association relationship between the GUTI information of each UE and its respective physical feature information.

[0098] In one example, the gNB receives radar-based echo signals from each UE. Using the echo signals from each UE at the current moment, it estimates the state of each UE's echo signals at the next moment. Furthermore, based on the estimated state results for each UE's echo signals at the current moment and the next moment, it calculates the similarity between the states of each two UEs in adjacent time slots. This state may include the UE's position, velocity, and angle. Based on the similarity between the states of each two UEs in adjacent time slots, it then calculates the similarity between each two UEs in adjacent time slots. Using this similarity, a matching algorithm is used to track and match the physical characteristic information of each UE in different time slots to determine the correlation between the physical characteristic information of each UE in adjacent time slots. This correlation is then traced back based on the initial correlation between each UE's GUTI information and the initial physical characteristic information to establish a long-term correlation between each UE's GUTI information and its physical characteristic information.

[0099] In a possible implementation, the above-mentioned step A estimates the motion state of each UE based on the echo signal of each UE, and determines the similarity between the states of every two UEs in adjacent time slots according to each state estimation result, which may include:

[0100] Step A1: For each UE, determine the current state information of the UE based on the echo signal of the UE, where the state information includes: position information, speed information, and angle information;

[0101] Step A2: Based on the current state information of the UE, the state of the UE is tracked using an extended Kalman filter algorithm to obtain state estimation information of the UE at the next moment;

[0102] Step A3: Based on the current state information of each UE and the estimated state information of the next time, the similarity between the states of every two UEs in adjacent time slots is calculated.

[0103] The implementation process of determining the UE's current state information based on the UE's echo signal can refer to the process of the gNB perceiving and identifying the UE's motion information described above.

[0104] Exemplarily, the state transfer equation and observation equation of the extended Kalman filter algorithm are respectively expressed as:

[0105]

[0106] Among them, the state vector and observation vector are represented as x n =[p n , v n ] and y n =[p n , v n ] T , p n =[p n (1), p n (2), p n (3)] T and v n =[v n (1), v n (2), v n (3)] T Respectively represent the position and velocity components of the UE on the x-axis, y-axis and z-axis, H() represents the observation function, G=[I 3×3 , ΔT·I 3×3 ;0 3×3 , I 3×3 ] represents the state transfer matrix, δ and μ represent the observation noise and state transfer noise respectively, and the covariance matrix is and Indicates length N rb The unit vector of represents the position noise, represents velocity noise.

[0107] To linearize the measurement model, define Represents the angle-related component in the normalized echo signal. and is an estimate of φ and θ. The Jacobian matrix of the observation function is expressed as

[0108]

[0109] in,

[0110]

[0111] χ(e), b(e) and q(e) represent the e-th element in the first, second and third rows of the Jacobian matrix respectively. The e-th element starting from the third element in the second row of the Jacobian matrix is ​​an intermediate variable and has no actual physical meaning.

[0112] N tx 、N ty Respectively represent the total number of X-axis and Y-axis antennas in the transmitting array, n tx 、n ty Respectively represent the X-axis and Y-axis antenna indexes of the transmitting array, N rx 、N ry Respectively represent the total number of X-axis and Y-axis antennas in the receiving array. The partial derivative of η with respect to θ Expressed as:

[0113]

[0114] The iterative steps of state prediction and tracking are as follows:

[0115] 1) State prediction: It represents the estimated value of the state at time n based on the state at time n-1, x n-1 Indicates the state at time n-1;

[0116] 2) Linearization processing: Hn represents the observation function at the nth moment;

[0117] 3) Prediction mean square error matrix: M n|n-1 G represents the mean square error matrix at the nth moment based on the n-1 moment prediction. n-1 、M n-1 Represent the state transfer matrix and mean square error matrix at time n-1 respectively;

[0118] 4) Calculate the Kalman gain: K n represents the Kalman gain at the nth moment;

[0119] 5) Status tracking: represents the estimated value of the state at the nth moment;

[0120] 6) Update the mean square error matrix: M n =(IK n H n )M n|n-1 .

[0121] In an example, the similarity of two physical features can be measured by the JS (Jensen-Shannon) divergence of relative entropy. Specifically, for a feature F that can be represented in the form of a probability distribution, a and F b , the relative entropy R(Fa , F b ) is expressed as:

[0122]

[0123] Among them, σ a and σ b is feature F a and F b The variance of f a and f b is feature F a and F b The mean of , x represents the characteristic variable. Feature F a and F b JS divergence C(F a , F b ) can be expressed as: C(F a , F b )=R(F a ,M) / 2+R(F b , M) / 2, M represents the feature F a and F b Furthermore, the similarity between every two UE states in adjacent time slots can be calculated by the above-mentioned method of calculating the JS divergence of relative entropy.

[0124] In one possible implementation, the implementation of the above-mentioned step B determining the similarity between every two UEs in adjacent time slots based on the similarity between the states of every two UEs in adjacent time slots may include: respectively calculating the weight values ​​of the similarity between every two UE states in adjacent time slots, and determining the similarity between every two UEs in adjacent time slots based on the weight values ​​of the similarity between every two UE states in adjacent time slots and the similarity between every two UE states in adjacent time slots.

[0125] In an example, the weight wn(ξ) of the ξ-th feature of a UE can be defined as: Characterizes the characteristic pf k，n The distinguishability of (ξ), that is, pf k，n (ξ) and pf n (ξ)={pf k，n (ξ), k = 1, ..., K}, the probability that other features are different. k，n (ξ) represents the ξ-th physical characteristic of the k-th UE at time n, pf n (ξ) represents the set of the ξ-th physical characteristics of all K UEs at the n-th time. k，n (ξ) is defined as:

[0126]

[0127] Among them, C(pfk，n (ξ), pf j，n (ξ))×П q≠j，q≠k (1-C(pf k，n (ξ), pf q，n The term (ξ))) represents pf k，n (ξ) and pf j，n (ξ) but the probability that it is different from other characteristic measurements, C(pf k，n (ξ), pf j，n (ξ)) represents the characteristic pf k，n (ξ) and pf j，n (ξ), q represents the index of the qth user.

[0128] Based on this, the similarity S of the ξjth UE n (ξ, j) is defined as the harmonic mean of all physical characteristics and can be expressed as:

[0129]

[0130] Among them, w′ n (ξ) represents the normalized weight, and M represents the number of observable physical features.

[0131] In a possible implementation, the above-mentioned step C tracks and matches the physical feature information of each UE in different time slots based on the similarity between every two UEs in adjacent time slots, and obtains the association relationship between the physical feature information of each UE in adjacent time slots. The implementation method may include: based on the state information of each UE at the current moment, the state estimation information at the next moment, and the similarity between every two UEs in adjacent time slots, establishing a weighted bipartite graph matching model to track and match the physical feature information of each UE in different time slots, and obtain the association relationship between the physical feature information of each UE in adjacent time slots.

[0132] Based on the current state information of each UE, the estimated state information at the next moment, and the similarity between every two UEs in adjacent time slots, a weighted bipartite graph matching model is established and solved to track and match the physical feature information of each UE in different time slots, and obtain the correlation between the physical feature information of each UE in adjacent time slots.

[0133] In an example, you can Calculate the measurement estimation results (measurement values) of the predicted states of all UEs, represents the measurement estimation result of the predicted state of the jth UE, Represents the predicted state of the jth UE, due to the measurement value y i，n Measured value of the predicted state There will not be much difference when i=j, so we can define the characteristic difference D between the two n (i, j) is the inverse of the similarity D between the two UEs n (i, j) = S n (i, j) -1 For example, establish Figure 4 The weighted bipartite graph matching model shown in Figure 4 middle y 1，n represents the state observation value of the first UE at the nth moment, represents the estimated value of the state observation of the first UE at time n, D n (1, 1) represents y 1，n and The feature difference value of , and so on, the weight of the edge is defined as the matching cost i, j represent the i-th and j-th UE respectively. Then the physical feature matching problem is transformed into an optimization problem whose optimization goal is to minimize the overall difference and balance the individual differences, that is, min(f1+f2), where f1 and f2 represent the sub-optimization objectives of the overall cost and individual cost, respectively.

[0134] The constraints of the above optimization problem are and That is, any y i，n or Can only be used for matching once. n represents the assignment matrix, if y i，n Assigned to Then A n (i, j) = 1, otherwise A n (i, j) = 0. The above optimization problem can be solved using a standard solver, such as the Kuhn-Munkres algorithm and the vampire bat optimizer.

[0135] For example, Figure 5 As shown, after obtaining the initial association relationship between the GUTI information of each UE and the initial physical feature information, as well as the association relationship between the physical feature information of each UE in adjacent time slots, the association between the GUTI and the physical features, and the association between the physical features of different time slots are tracked to determine the long-term association relationship between the GUTI information of each UE and its respective physical feature information.

[0136] In this embodiment of the present invention, by eliminating radar signal receive beamforming at the gNB, most angular information is retained, reducing errors in the extended Kalman filter state estimation algorithm and achieving more accurate beamforming. During beam tracking, the extended Kalman filter method is used to accurately estimate user mobility. During multi-user beam tracking, a similarity definition method based on relative entropy and a feature weighting scheme are utilized. A weighted bipartite graph matching model is established to correlate multi-user measurement and prediction results, achieving accurate alignment of a specific beam with the intended UE.

[0137] See also Figure 6 , Figure 6 Another perception-assisted multi-user beam management method provided in an embodiment of the present invention is applied to a terminal side, including:

[0138] S601: Sense and identify a base station NB through a second target sensor device to obtain location information of the NB.

[0139] In this embodiment of the present invention, the terminal may be a drone or other terminal device capable of sensing and identifying, and the NB may specifically be a gNB. The terminal is the UE mentioned above, and the second target sensor device may be a sensor device such as a radar or camera. Specifically, the process by which the UE senses and identifies the gNB through the second target sensor device and obtains the gNB's location information can be similar to the process by which the gNB senses and identifies the UE through the first target sensor device and obtains the motion information of each UE, and this embodiment of the present invention will not be further described here.

[0140] S602: Determine a target receiving beam direction set from a second preset complete beam set based on the location information of the NB.

[0141] Based on the gNB location information and the UE's location information, the UE calculates the relative angle between the UE and the gNB, determines the direction corresponding to the relative angle as the target optimal direction, and selects a target number of directions adjacent to the target optimal direction on both sides from the second preset complete beam set as the target receive beam direction set. The target number can be set based on actual needs.

[0142] Exemplarily, the second preset complete beam set can be expressed as:

[0143]

[0144] in, represents the second preset complete beam set, The receive beamforming vector on the UE side is defined in a similar way to the transmit beamforming vector on the gNB side, except that the number of receive antennas on the UE is N. ru, and represents the elements in the azimuth and elevation angle sets in the second preset complete beam set, o and o′ represent the indexes of the elements in the azimuth and elevation angle sets, respectively, Q u The number of bits indicating the UE's beam phase.

[0145] For example, the UE calculates the relative angle between the UE and the gNB based on the location information of the gNB and the location information of the UE, determines the direction corresponding to the relative angle as the target optimal direction, and then selects the target optimal direction in descending order based on the sequence number in the second preset complete beam set corresponding to the target optimal direction. directions, and then select the best direction based on the target in ascending order. directions, and obtain the target receiving beam direction set. Among them, represents the target receiving beam direction set, Indicates the target receiving beam direction set The cardinality size of the set Cardinality The size depends on the number of bits Qu of the UE's beam phase, that is, Determines the actual search area size of the UE receiving beam direction,

[0146] S603: traverse all directions in the target receiving beam direction set and receive a primary synchronization signal sent by the NB.

[0147] The UE sequentially traverses each direction in the target receive beam direction set and receives the primary synchronization signal sent by the gNB in ​​each direction once at a preset interval. For example, the preset interval can be set based on actual needs, such as 3 milliseconds, 5 milliseconds, or 10 milliseconds.

[0148] For example, the primary synchronization signal transmitted by the gNB at time n is received by the kth UE through the receiving beamformer W k，n Reception. For a channel environment dominated by LoS (Line-of-sight) links, the signal received by the UE can be expressed as:

[0149]

[0150] Among them, r k，n (t) represents the received signal of the kth UE at the nth time, represents the UE's antenna gain factor, α represents the LoS channel coefficient, Indicates road loss. Indicates the power gain at unit reference distance.k，n represents the Euclidean distance between the kth UE and the gNB at time n, d k，n =||p k，n -p b ||, 2πf c d k，n / c represents the phase of the channel, b k，n represents the receiving steering vector of the kth UE at the nth time, z r (t) indicates zero mean and variance is Gaussian noise, γ k，n Indicates the Doppler frequency of the UE.

[0151] S604: Determine a second alignment direction with the NB based on the received primary synchronization signal, and send a random access preamble to the NB in ​​the second alignment direction.

[0152] The random access preamble includes the UE's GUTI information. The UE calculates the SNR of the primary synchronization signal received in each direction in the target receive beam direction set, determines the direction with the highest primary synchronization signal SNR as the second alignment direction with the gNB, and sends the random access preamble to the gNB using the second alignment direction. This allows the gNB to receive the random access preamble in all directions in the target transmit beam direction set for the UE, thereby determining the first alignment direction with the UE. Furthermore, the gNB sends an access signal to the UE using the first alignment direction. The UE receives the access signal using the second alignment direction and sends a response message back to the gNB, establishing the initial connection between the UE and the gNB.

[0153] S605 , receiving a target signal sent by the NB through a second alignment direction, and reflecting an echo signal to the NB, so that the NB determines a long-term association relationship between the GUTI information of the UE and its physical characteristic information based on the echo signal.

[0154] After the connection between the UE and the gNB is established, the gNB sends a target signal to the UE via a first alignment direction. The UE receives the target signal sent by the gNB via a second alignment direction and reflects an echo signal to the gNB, so that the gNB determines the long-term association between the UE's GUTI information and its physical characteristic information based on the echo signal.

[0155] An embodiment of the present invention provides a sensing-assisted multi-user beam management method. The terminal side senses and identifies the NB through sensing and identification to obtain the NB's location information. Then, based on the NB's location information, a target receiving beam direction set is determined from a preset complete beam set, thereby reducing the range of beam traversal during the initial access process, reducing the overhead and delay during the initial access, and facilitating rapid access between the base station and the UE. During the beam tracking process, the UE reflects an echo signal to the NB through a second alignment direction, so that the NB determines the long-term association between the UE's GUTI information and its physical feature information based on the echo signal. Because the NB can use the echo signal of each UE to track the physical feature information of each UE for predictive beamforming in the next time slot, the downlink pilot and uplink feedback in the beam tracking process are eliminated. Furthermore, by establishing a physical feature association relationship between adjacent time slots of each UE, based on the initial association relationship between the UE's GUTI information and the initial physical feature information established during initial access, a long-term association relationship between the GUTI information and the physical feature information of each UE can be obtained, thereby ensuring the alignment accuracy between the NB and the UE during the beam tracking process.

[0156] For example, Figure 7 As shown, beam management includes initial access and beam tracking. During the initial access phase, the gNB uses visual perception technology to sense and identify the UE's state (position and velocity) and obtain each UE's state measurement value (position and velocity). The UE also uses visual perception technology to sense and identify the gNB's state (position) and obtain the gNB's state measurement value (position). For each UE, the gNB determines a target transmit beam direction set for that UE from a first preset complete beam set based on the UE's state measurement value. The UE determines a target receive beam direction set from a second preset complete beam set based on the gNB's state measurement value. For each UE, the gNB iterates through all directions in the target transmit beam direction set for that UE and sends a primary synchronization signal to the UE. The UE iterates through all directions in the target receive beam direction set, receives the primary synchronization signal sent by the gNB, and based on the received primary synchronization signal, determines a secondary alignment direction with the gNB and sends a random access preamble to the gNB in ​​the second alignment direction. For each UE, the gNB receives a random access preamble sent by the UE from each direction in the set of target transmit beam directions for the UE, determines a first alignment direction with the UE based on the received random access preamble, and establishes an initial connection with the UE based on the first alignment direction.

[0157] After the gNB and UE complete initial access, the gNB obtains the initial physical feature information of the UE and establishes an initial association between the UE's GUTI information and its corresponding initial physical feature information. Then, in the beam tracking phase, the gNB sends a target signal to the UE in the first alignment direction and receives the echo signal reflected by the UE through radar sensing. The gNB predicts and tracks the UE's status based on the echo signal. Specifically, for each UE, the gNB determines the UE's current status information based on the UE's echo signal. Based on the UE's current status information ( Figure 7 The UE is tracked using the extended Kalman filter algorithm to obtain the state estimation information of the UE at the next moment ( Figure 7 Furthermore, the physical characteristics of the states of each UE in different time slots are processed, and the similarity between each two UE states in adjacent time slots is calculated by calculating the JS divergence of relative entropy, and then the weight value of the similarity between each two UE states in adjacent time slots is calculated ( Figure 7 Dynamic weight calculation in the middle), based on the weight value of the similarity between each two UE states in adjacent time slots and the similarity between each two UE states in adjacent time slots, determine the similarity between each two UEs in adjacent time slots ( Figure 7 Then, the UE and its physical features are matched with each other. Specifically, based on the current state information of each UE, the estimated state information of the next moment, and the similarity between each two UEs in adjacent time slots, a weighted bipartite graph matching model is established to track and match the physical feature information of each UE in different time slots, and the correlation relationship between the physical feature information of each UE in adjacent time slots is obtained ( Figure 7 Finally, based on the association relationship between the physical feature information of each UE in adjacent time slots and the initial association relationship between the GUTI information of each UE and the initial physical feature information, the long-term association relationship between the GUTI information of each UE and its respective physical feature information is determined (i.e. Figure 7 Long-term association between physical features and GUTI in the gNB is used to achieve beam alignment between gNB and UE.

[0158] For example, Figure 8 As shown, Figure 8 The demonstration compares beam tracking using beam training, beam prediction, and the solution of the present invention. In the beam training approach, the gNB selects a beam direction and sends a pilot signal to the UE. The UE estimates the gNB's direction based on the received pilot signal and feeds back channel state information to the gNB, enabling beam tracking. The beam prediction approach is the time-correlation prediction method based on communication feedback described above. Figure 8The beam training phase includes the GUTI identity verification process, while the beam tracking phase narrows the beam search space. However, both beam training-based and beam prediction-based approaches require downlink pilots and uplink feedback, which still incur significant overhead and time consumption. Furthermore, the angle information obtained during training at the previous moment can become outdated, leading to beam degradation / alignment failure due to time-varying channel conditions. However, the present invention's solution allows the entire downlink to function as both a radar sensing signal and a carrier of communication data symbols, eliminating the need for dedicated downlink pilots and uplink feedback. Furthermore, after obtaining GUTI information during the initial access phase, the gNB tracks the UE and, by correlating the physical characteristics of adjacent time slots, maps the GUTIs of multiple users to their physical characteristics at any given moment. This eliminates the dedicated resources reserved for GUTI feedback while ensuring beam mismatch prevention.

[0159] For example, Figure 9 As shown, Figure 9 This paper compares the accuracy of matching gNBs and UEs during beam tracking using position, velocity, and the proposed design. The velocity-based solution, which uses only velocity and direction to distinguish UEs, performs the worst because UEs can have similar velocities during movement. Furthermore, as the upper speed limit decreases and the number of UEs increases, more UEs will have similar velocities, resulting in more matching errors. Because UEs occasionally move close to each other, the position-based matching solution, which uses only location information to distinguish UEs, can also experience matching errors. However, this situation is less common than when UEs have similar velocities, resulting in slightly better performance than the velocity-based matching solution. The proposed design, however, achieves nearly 100% matching accuracy. This is because the similarity of physical features depends not only on the observations but also on the observation errors, and each physical feature is dynamically weighted based on its prevalence. In other words, when velocities are similar, location information plays a more important role in the matching matrix, and vice versa.

[0160] For example, Figure 10 As shown, Figure 10 The figure shows the weight distribution of the location and speed characteristics of terminals in different time slots. Figure 10Because the initial positions of all UEs are randomly generated, more distinguishable location features are given greater weight during the 0-1800ms (milliseconds) period. Even if some UEs have similar speeds during this period, they can still be distinguished by the more heavily weighted location features. However, during the 1800ms-3200ms period, UEs converge on the adjacent gNB from different directions, even with intersecting trajectories. This makes similar location features too common to distinguish multiple UEs. In this case, the speed feature is given a greater weight and plays a more important role in the matching process. The matching accuracy of the design of the present invention is 18.67% and 9.24% higher than that of the position-based and speed-based matching schemes, respectively.

[0161] For example, Figure 11a 、 Figure 11b 、 Figure 11c 、 Figure 11d and Figure 12a 、 12b , 12c, 12d, Figure 11a-Figure 11d The demonstration compares the angle estimation between the gNB and UE during beam tracking using the actual angle between the gNB and UE, a feedback-based solution (i.e., the time correlation prediction method based on communication feedback described above), and the design of the present invention. Figures 12a-12d The comparison of the angle estimation error between gNB and UE in the beam tracking process between the feedback-based scheme and the design scheme of the present invention is shown. Figure 10 Compared to the proposed design, the UE approaches the gNB between 1800ms and 3200ms, resulting in a higher rate of angle change. The feedback-based approach suffers from reduced prediction accuracy due to: 1) only one pilot being used for tracking, and 2) the need for a receive beamformer to combine the received signals, which inevitably results in a loss of detailed angle information. Compared to the proposed design, the feedback-based approach estimates angles with larger errors, and its performance deteriorates with 64 antennas, as the increased number of antennas results in narrower beams and a higher probability of misalignment. The proposed design utilizes the entire echo signal block for sensing, boasts a matched filter gain several times greater than that of the feedback-based approach, and does not perform receive beamforming on the echo, preserving most angle information. This results in a better RMSE (Root Mean Squared Error) than the feedback-based approach, and even a smaller RMSE when equipped with 64 antennas due to the higher array gain.

[0162] For example, Figure 13a and Figure 13b As shown, Figure 13a and Figure 13bThe performance comparison of the average beam reachability obtained by tracking using a location-based solution, a feedback-based solution, the design of the present invention, and a static weight-based solution during beam tracking is shown. The static weight-based solution assigns a static weight to each UE feature during beam tracking to match the gNB and UE. Perfect alignment means that the predicted angle and the actual angle are completely matched. It should be noted that if the physical characteristics of the UE are not correctly associated with its GUTI, that is, the beam is mistakenly associated with an unintended receiver, the reachability rate of the link will be zero. Figure 13a and Figure 13b As can be seen, with 32 antennas, the proposed design maintains the highest rate due to its near-zero matching errors and low RMSE. The feedback-based scheme notifies the gNB of the UE's GUTI in real time, eliminating the issue of beam failure due to matching errors. During the relatively slow angular tracking period (0-1800 ms), the achievable rate performance is comparable to that of the proposed design. Subsequently, as the UE approaches the gNB, the feedback-based scheme's rate decreases for the same reason as the angle tracking performance degradation. With 64 antennas, the achievable rate significantly decreases due to the higher probability of misalignment caused by the narrower beam. The location-based scheme utilizes only location information to correlate the UE's physical characteristics with the GUTI. However, due to the matching errors it suffers in each trial, its performance gradually deviates from the achievable rate of perfect alignment, with achievable rates decreasing by 36.12% and 22.59% for the 32 and 64 antennas, respectively. The static weight-based scheme gives equal weight to the matching matrix for position and velocity information. As a result, similar velocities at the beginning and positions near the gNB play an equal role, despite their poor distinguishability. Statistically, the average rate ultimately drops to 6.16% and 8.72% of the design of the present invention with 32 and 64 antennas, respectively.

[0163] For example, as shown in Table 1 and Table 2 below, Table 1 shows different Q B Table 2 shows the initial access delay performance comparison of different Q UComparison of initial access delay performance. Tables 1 and 2 compare the initial access delays of five initial access schemes, in milliseconds. The scheme based on gNB prior information utilizes prior distribution information about the gNB location to determine the UE's location, thereby narrowing the gNB's beam search space. The scheme based on UE prior information utilizes prior distribution information about the UE's location to determine the UE's location, thereby narrowing the gNB's beam search space. An exhaustive search traverses all possible beam directions to find the one with the highest SNR. An iterative search uses a hierarchical traversal of beam directions to find the one with the highest SNR. The design of the present invention leverages the visual perception capabilities of both the gNB and UE to narrow the range of beam traversal during initial access.

[0164] Table 1 Different Q B Initial access delay performance comparison

[0165] <![CDATA[Q B ]]> 2 4 6 8 Based on gNB prior information 20 80 320 1280 Based on UE's prior information 80 80 80 80 Exhaustive search 340 1360 5440 21760 Iterative Search 320 320 320 320 Design of the present invention 20 20 20 20

[0166] Table 2 Different Q U Initial access delay performance comparison

[0167] <![CDATA[Q U ]]> 2 4 6 8 Based on gNB prior information 80 80 80 80 Based on UE's prior information 20 80 320 1280 Exhaustive search 400 1360 5200 20560 Iterative Search 80 320 1280 5120 Design of the present invention 20 20 20 20

[0168] The delay of the exhaustive solution increases with Q U and Q B This is because it needs to increase The delay of the iterative search scheme will not be affected by the Q B The gNB has acquired the UE's prior information, which means that only the UE side performs iterative search. However, when the number of beam phase bits on the UE side increases, it still suffers from a large delay, which is close to 1 / 4 of the exhaustive method. This is because 16 exhaustive searches are replaced by 4 hierarchical searches. For the two methods based on prior information, due to the introduction of gNB (UE) location information, the UE (gNB) side no longer needs to search. Therefore, the schemes based on gNB / UE prior information need to be and Compared with these methods, the design of the present invention fully utilizes the sensor information of gNB and UE, accelerates beam training with the help of camera sensing, and searches in two potential beam sets, avoiding the time-consuming initial access process. The scanning space on both sides of gNB and UE is reduced to and This means that the worst case only requires Initial access can be completed with one search, because Therefore, it has the best delay performance and is U and Q B The advantages are more obvious when it gets bigger.

[0169] Corresponding to the above Figure 2 The embodiment provides a method for multi-user beam management assisted by perception, and the embodiment of the present invention also provides a multi-user beam management device assisted by perception, such as Figure 14 As shown, applied to the base station side, the device includes: a first perception module 1401, used to perceive and identify the user terminal device UE through the first target sensor device to obtain the motion information of each UE; a first beam direction determination module 1402, used to determine, for each UE, based on the motion information of the UE, a target transmission beam direction set for the UE from a first preset complete beam set; a first signal sending module 1403, used to traverse all directions in the target transmission beam direction set for the UE and send a main synchronization signal to the UE; a first alignment direction determination module 1404, used to receive the random access preamble code sent by the UE from each direction in the target transmission beam direction set for the UE, and based on the received random access preamble code Determine the first alignment direction with the UE, and the random access preamble code includes the UE's globally unique temporary identity GUTI information; a connection establishment module 1405 is used to establish an initial connection with the UE based on the first alignment direction, obtain the initial physical feature information of the UE, and establish an initial association relationship between the UE's GUTI information and the initial physical feature information; a second signal sending module 1406 is used to send a target signal to the UE through the first alignment direction, and receive an echo signal reflected by the UE; a beam tracking module 1407 is used to determine the long-term association relationship between the GUTI information of each UE and its respective physical feature information based on the echo signal of each UE and the initial association relationship between the GUTI information of each UE and the initial physical feature information.

[0170] In one possible embodiment, the above-mentioned motion information includes location information; the above-mentioned first beam direction determination module 1402 is specifically used to: for each UE, based on the location information of the UE, determine the relative angle between the UE and the base station NB; based on the relative angle, determine the target transmission beam direction set for the UE from the first preset complete beam set.

[0171] In a possible implementation, the first signal sending module 1403 is specifically configured to sequentially traverse each direction in the target transmission beam direction set for the UE, and send a main synchronization signal to the UE through the direction once at a preset interval; the first alignment direction determination module 1404 is specifically configured to receive the random access preamble code sent by the UE from each direction in the target transmission beam direction set for the UE, and determine the direction corresponding to the highest signal-to-noise ratio (SNR) of the received random access preamble code as the first alignment direction with the UE.

[0172] In one possible implementation, the connection establishment module 1405 is specifically used to: send an access signal to the UE through a first alignment direction, and receive a response signal fed back by the UE; determine the current physical feature information of the UE based on the response signal; and establish an initial association relationship between the GUTI information of the UE and the current physical feature information.

[0173] In one possible embodiment, the beam tracking module 1407 includes: a first similarity determination submodule, which is used to estimate the motion state of each UE based on the echo signal of each UE, and determine the similarity between the states of every two UEs in adjacent time slots based on the estimation results of each state; a second similarity determination submodule, which is used to determine the similarity between every two UEs in adjacent time slots based on the similarity between the states of every two UEs in adjacent time slots; a first association relationship establishment submodule, which is used to track and match the physical feature information of each UE in different time slots based on the similarity between every two UEs in adjacent time slots, and obtain the association relationship between the physical feature information of each UE in adjacent time slots; a second association relationship establishment submodule, which is used to determine the long-term association relationship between the GUTI information of each UE and its respective physical feature information based on the association relationship between the physical feature information of each UE in adjacent time slots and the initial association relationship between the GUTI information of each UE and the initial physical feature information.

[0174] In one possible implementation, the above-mentioned first similarity determination submodule is specifically used to: for each UE, determine the current state information of the UE based on the echo signal of the UE, where the state information includes: position information, speed information and angle information; based on the current state information of the UE, use the extended Kalman filter algorithm to track the state of the UE to obtain the state estimation information of the UE at the next moment; based on the current state information of each UE and the state estimation information of the next moment, calculate the similarity between the states of every two UEs in adjacent time slots.

[0175] In one possible implementation, the above-mentioned first association relationship establishment submodule is specifically used to: establish a weighted bipartite graph matching model based on the current state information of each UE, the state estimation information of the next moment, and the similarity between every two UEs in adjacent time slots to track and match the physical feature information of each UE in different time slots, and obtain the association relationship between the physical feature information of each UE in adjacent time slots.

[0176] Corresponding to the above Figure 6 The embodiment provides a method for multi-user beam management assisted by perception, and the embodiment of the present invention also provides a multi-user beam management device assisted by perception, such as Figure 15As shown, applied to the terminal side, the device includes: a second perception module 1501, which is used to perceive and identify the base station NB through a second target sensor device to obtain the location information of the NB; a second beam direction determination module 1502, which is used to determine the target receiving beam direction set from the second preset complete beam set based on the location information of the NB; a first signal receiving module 1503, which is used to traverse all directions in the target receiving beam direction set and receive the main synchronization signal sent by the NB; a second alignment direction determination module 1504, which is used to determine the second alignment direction between the NB and the NB based on the received main synchronization signal, and send a random access preamble code to the NB in ​​the second alignment direction, and the random access preamble code contains the globally unique temporary identity GUTI information of the user terminal device UE; a second signal receiving module 1505, which is used to receive the target signal sent by the NB through the second alignment direction, and reflect the echo signal to the NB, so that the NB determines the long-term association relationship between the GUTI information of the UE and its physical feature information based on the echo signal.

[0177] The embodiment of the present invention further provides an electronic device, such as Figure 16 As shown, it includes a processor 1601, a communication interface 1602, a memory 1603 and a communication bus 1604, wherein the processor 1601, the communication interface 1602, and the memory 1603 communicate with each other through the communication bus 1604, and the memory 1603 is used to store computer programs; the processor 1601 is used to implement the steps of any of the above-mentioned perception-assisted multi-user beam management methods when executing the program stored in the memory 1603 to achieve the same technical effect.

[0178] The communication bus mentioned in the above-mentioned electronic device may be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The communication bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus. The communication interface is used for communication between the above-mentioned electronic device and other devices. The memory may include a random access memory (RAM) or a non-volatile memory (NVM), such as at least one disk storage. Optionally, the memory may also be at least one storage device located away from the aforementioned processor. The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, and discrete hardware components.

[0179] In another embodiment of the present invention, a computer-readable storage medium is provided, storing a computer program. When executed by a processor, the computer program implements the steps of any of the aforementioned sensing-assisted multi-user beam management methods, achieving the same technical effect. In another embodiment of the present invention, a computer program product comprising instructions is provided. When executed on a computer, the computer executes the steps of any of the sensing-assisted multi-user beam management methods described in the aforementioned embodiments, achieving the same technical effect.

[0180] In the above embodiments, all or part of the embodiments can be implemented by software, hardware, firmware, or any combination thereof. When implemented using software, all or part of the embodiments can be implemented in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via a wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) method. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that includes one or more available media. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state drive (SSD)).

[0181] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply the existence of any such actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or device comprising the element.

[0182] Each embodiment in this specification is described in a related manner. Similar portions between the various embodiments can be referenced to each other. Each embodiment focuses on the differences from other embodiments. In particular, the device / electronic device embodiments are generally similar to the method embodiments, so their description is relatively simple. For related portions, reference can be made to the description of the method embodiments.

[0183] The above description is only a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention are included in the scope of protection of the present invention.

Claims

1. A perception-assisted multi-user beam management method, characterized in that: Applied to the base station side, the method includes: The user terminal device UE is sensed and identified by the first target sensor device to obtain motion information of each UE; For each UE, based on the motion information of the UE, determine a target transmit beam direction set for the UE from the first preset complete beam set; Traversing all directions in the target transmit beam direction set for the UE and sending a primary synchronization signal to the UE; receiving a random access preamble sent by the UE in each direction of the target transmit beam direction set for the UE, and determining a first alignment direction with the UE based on the received random access preamble, wherein the random access preamble includes GUTI information of the UE; Establishing an initial connection with the UE based on the first alignment direction, obtaining initial physical feature information of the UE, and establishing an initial association relationship between the GUTI information of the UE and the initial physical feature information; sending a target signal to the UE in the first alignment direction, and receiving an echo signal reflected by the UE; Based on the echo signal of each UE and the initial association relationship between the GUTI information of each UE and the initial physical feature information, a long-term association relationship between the GUTI information of each UE and the respective physical feature information is determined.

2. The method according to claim 1, characterized in that The motion information includes position information; The determining, for each UE, a target transmit beam direction set for the UE from the first preset complete beam set based on the motion information of the UE includes: For each UE, determine the relative angle between the UE and the base station NB based on the location information of the UE; Based on the relative angle, a target transmit beam direction set for the UE is determined from a first preset complete beam set.

3. The method according to claim 1, characterized in that The traversing all directions in the target transmit beam direction set for the UE and sending the primary synchronization signal to the UE includes: traversing each direction in the target transmit beam direction set for the UE in turn, and sending the primary synchronization signal to the UE once through the direction at a preset time interval; The receiving a random access preamble sent by the UE from each direction in the target transmit beam direction set for the UE, and determining a first alignment direction with the UE based on the received random access preamble, includes: receiving a random access preamble sent by the UE from each direction in the target transmit beam direction set for the UE, and determining a direction corresponding to a highest signal-to-noise ratio (SNR) of the received random access preamble as the first alignment direction with the UE.

4. The method according to claim 1, wherein The establishing an initial connection with the UE based on the first alignment direction, obtaining initial physical feature information of the UE, and establishing an initial association relationship between the GUTI information of the UE and the initial physical feature information includes: Sending an access signal to the UE in the first alignment direction, and receiving a response signal fed back by the UE; Determining current physical characteristic information of the UE based on the response signal; An initial association relationship is established between the GUTI information of the UE and the current physical feature information.

5. The method according to claim 1, wherein The determining, based on the echo signal of each UE and the initial association relationship between the GUTI information of each UE and the initial physical feature information, a long-term association relationship between the GUTI information of each UE and the respective physical feature information includes: Estimate the motion state of each UE based on the echo signal of each UE, and determine the similarity between the states of every two UEs in adjacent time slots according to the state estimation results; Determining the similarity between every two UEs in adjacent time slots based on the similarity between every two UE states in adjacent time slots; Based on the similarity between every two UEs in adjacent time slots, the physical feature information of each UE in different time slots is tracked and matched to obtain the correlation relationship between the physical feature information of each UE in adjacent time slots; Based on the association relationship between the physical feature information of each UE in adjacent time slots and the initial association relationship between the GUTI information of each UE and the initial physical feature information, a long-term association relationship between the GUTI information of each UE and its respective physical feature information is determined.

6. The method according to claim 5, characterized in that The estimating the motion state of each UE based on the echo signal of each UE, and determining the similarity between the states of every two UEs in adjacent time slots according to the state estimation results, includes: For each UE, determine the current state information of the UE based on the echo signal of the UE, the state information including: position information, speed information and angle information; Based on the current state information of the UE, the state of the UE is tracked using the extended Kalman filter algorithm to obtain the state estimation information of the UE at the next moment; Based on the current state information of each UE and the estimated state information of the next time, the similarity between the states of every two UEs in adjacent time slots is calculated.

7. The method according to claim 5, characterized in that The tracking and matching of the physical feature information of each UE in different time slots based on the similarity between every two UEs in adjacent time slots to obtain the correlation relationship between the physical feature information of each UE in adjacent time slots includes: Based on the current state information of each UE, the estimated state information of the next moment, and the similarity between every two UEs in adjacent time slots, a weighted bipartite graph matching model is established to track and match the physical feature information of each UE in different time slots, and obtain the correlation between the physical feature information of each UE in adjacent time slots.

8. A perception-assisted multi-user beam management method, characterized in that: Applied to the terminal side, the method includes: Sensing and identifying the base station NB by a second target sensor device to obtain location information of the NB; Determining a target receive beam direction set from a second preset complete beam set based on the location information of the NB; Traversing all directions in the target receive beam direction set, and receiving a primary synchronization signal sent by the NB; Determine a second alignment direction with the NB based on the received primary synchronization signal, and send a random access preamble to the NB in ​​the second alignment direction, where the random access preamble includes GUTI information of a user terminal device UE; The target signal sent by the NB is received through the second alignment direction, and an echo signal is reflected to the NB, so that the NB determines the long-term association relationship between the GUTI information of the UE and its physical feature information based on the echo signal.

9. A sensing-assisted multi-user beam management device, characterized in that: Applied to a base station side, the device includes: A first perception module is configured to perceive and identify user terminal devices UE through a first target sensor device to obtain motion information of each UE; a first beam direction determination module, configured to determine, for each UE, a target transmit beam direction set for the UE from a first preset complete beam set based on motion information of the UE; A first signal sending module, configured to traverse all directions in a target transmit beam direction set for the UE and send a primary synchronization signal to the UE; A first alignment direction determination module is configured to receive a random access preamble sent by the UE from each direction in the target transmit beam direction set for the UE, and determine a first alignment direction with the UE based on the received random access preamble, where the random access preamble includes the globally unique temporary identity GUTI information of the UE; a connection establishing module, configured to establish an initial connection with the UE based on the first alignment direction, obtain initial physical feature information of the UE, and establish an initial association relationship between the GUTI information of the UE and the initial physical feature information; a second signal sending module, configured to send a target signal to the UE through the first alignment direction, and receive an echo signal reflected by the UE; The beam tracking module is used to determine the long-term association relationship between the GUTI information of each UE and its respective physical feature information based on the echo signal of each UE and the initial association relationship between the GUTI information of each UE and the initial physical feature information.

10. A sensing-assisted multi-user beam management device, characterized in that: Applied to the terminal side, the device includes: A second sensing module is configured to sense and identify the base station NB through a second target sensor device to obtain location information of the NB; a second beam direction determination module, configured to determine a target receive beam direction set from a second preset complete beam set based on the location information of the NB; A first signal receiving module is configured to traverse all directions in the target receiving beam direction set and receive a primary synchronization signal sent by the NB; A second alignment direction determining module is configured to determine a second alignment direction with the NB based on the received primary synchronization signal, and send a random access preamble to the NB in ​​the second alignment direction, where the random access preamble includes GUTI information of a user terminal device UE; The second signal receiving module is used to receive the target signal sent by the NB through the second alignment direction and reflect the echo signal to the NB, so that the NB determines the long-term association relationship between the GUTI information of the UE and its physical feature information based on the echo signal.

Images