Interactive aware multi-user joint spatio-temporal beam management

Abstract

The techniques described herein relate to joint space / time domain beam management, including spatial domain beam management over an observation window and time domain beam management over a prediction window. Incorporating spatial and temporal dependencies increases the prediction interval and makes the prediction more accurate because spatial beam management is performed over the observation interval. Interaction aware multi-user equipment (UE) beam management techniques are also described that model a plurality of user interactions in a beam management procedure to predict future trajectory data and to predict an optimal future beam and / or a future optimal subset of beams per UE. For predicted future trajectory data and / or beam prediction data, a single user category or multiple user categories (e.g., vehicle category users, pedestrian category users) may be considered.

Description

Related applications

[0001] This application claims priority to U.S. nonprovisional patent application No. 18 / 350,859, filed July 12, 2023, entitled “INTERACTION-AWARE MULTIPLE-USER JOINT SPATIAL-TEMPORAL BEAM MANAGEMENT,” the entire contents of which are incorporated herein by reference. Background Technology

[0002] To facilitate accurate beam alignment between user equipment (UE) and serving base stations in New Radio (NR) networks, the 3GPP (3rd Generation Partnership Project) standard provides reasonable flexibility in configuring beam measurement, beam reporting, and beam indication. These configurations incur signaling overhead, leading to trade-offs between network performance and the amount of signaling overhead implemented. For example, to ensure continuous connectivity between UE and base stations, frequent beam measurements and measurement reports are required, especially in high-mobility scenarios. This frequent beam measurement and measurement reporting increases signaling overhead, which often degrades overall network performance. Attached Figure Description

[0003] The techniques described herein are illustrated by way of example and are not limited to the accompanying drawings, in which the same reference numerals indicate similar elements, and wherein:

[0004] Figure 1 This is an example block diagram representation of a system / architecture based on various aspects and implementations of this disclosure, in which interactively aware multi-user joint spatiotemporal beam management can be incorporated.

[0005] Figure 2 and Figure 3 This includes example sequences and data flow diagrams for joint spatial and temporal beam management according to various aspects and embodiments of this disclosure.

[0006] Figure 4 Example machine learning model structures (e.g., social long short-term memory (LSTM)) for beam management / beam data prediction with respect to user equipment are shown according to various aspects and embodiments of this disclosure.

[0007] Figure 5 This is an example representation of a social LSTM for beam management / beam data prediction with respect to user equipment, based on various aspects and embodiments of this disclosure.

[0008] Figure 6This is a flowchart illustrating example operations related to performing the spatial beam management observation phase and the joint temporal beam management prediction phase according to various aspects and embodiments of this disclosure.

[0009] Figure 7 This is a flowchart illustrating example operations related to determining trajectory data via joint spatiotemporal beam management, according to various aspects and embodiments of this disclosure, which is used in conjunction with trajectory cross-dependency data to predict future trajectory data and future optimal beam data for a user equipment.

[0010] Figure 8 This is a flowchart illustrating example operations related to determining future trajectory data and future optimal beam data of a user equipment based on joint spatiotemporal and cross-dependency data of other user equipment, according to various aspects and embodiments of this disclosure.

[0011] Figure 9 This is a block diagram representing an example computing environment in which various aspects of the topics described in this paper can be incorporated.

[0012] Figure 10 Example schematic block diagrams depict a computing environment according to various aspects and embodiments of this disclosure, the disclosed subjects being at least partially interacting with / implemented in such computing environment. Detailed Implementation

[0013] The various aspects of the techniques described in this paper generally involve joint spatial and temporal beam management, where spatial-temporal beam management operations increase the accuracy of prediction intervals and skip the temporal beam management observation phase. As will be understood, the techniques described in this paper can predict the optimal beam or optimal... K A beam set can be used without performing full / sparse beam scanning during the prediction interval, which reduces signaling overhead. As described in this paper, beam prediction can be considered as a sequence generation task, where the network is interested in predicting the optimal beam for the user equipment based on the history of the selected beam and the reported Reference Signal Received Power (RSRP) of its corresponding scanned beam.

[0014] Furthermore, the techniques described in this paper facilitate beam management for perceiving interactions between multiple user equipments (UEs). This interaction-aware multi-UE beam prediction is highly correlated with UE mobility patterns and their corresponding trajectories. For example, users in vehicles and pedestrians typically follow paths defined by roads and sidewalks, respectively; these users do not move randomly and often follow social conventions, such as adjusting their movement based on the movement of other users nearby in crowded scenarios. To capture such deviations in trajectories, the techniques described in this paper facilitate observing social interactions rather than observing users in isolation.

[0015] Observations of multiple users are used to build a model that can take into account the behavior of other users in the neighborhood while predicting the optimal beam for each user. One such model described in this paper learns beam space, environment, and multi-user mobility pattern data, and predicts the optimal beam for each UE.

[0016] Throughout this specification, references to "an embodiment," "an embodiment," "an implementation," "implementation," etc., mean that a particular feature, structure, or characteristic described in connection with an embodiment / implementation is included in at least one embodiment / implementation. Therefore, the appearance of such phrases "in one embodiment," "in an implementation," etc., in different places throughout the specification does not necessarily refer to the same embodiment / implementation. Furthermore, in one or more embodiments / implementations, a particular feature, structure, or characteristic can be combined in any suitable manner. It should also be noted that terms used herein, such as "optimize," "optimization," "optimal," etc., only indicate a goal of moving towards a better state, and do not necessarily mean achieving an ideal result. For example, "optimal" may mean the highest-performing entity available (e.g., the highest-rated beam in a finite set of available beams), and does not necessarily mean achieving a completely optimal result. Similarly, "maximize" means moving towards a maximum state (e.g., up to a certain threshold limit, if present), and does not necessarily mean achieving such a state.

[0017] Various aspects of this disclosure will now be described more fully below with reference to the accompanying drawings, in which example components, diagrams, and / or operations are illustrated. In the following description, numerous specific details are set forth for purposes of explanation in order to provide a thorough understanding of various embodiments. However, this disclosure may be embodied in many different forms and should not be construed as limited to the instances set forth herein.

[0018] Figure 1 An example system / architecture 100 is shown, comprising various O-RAN (Open Radio Access Network) layers coupled to a radio unit (RU) 102. In the example system / architecture 100, a service management and orchestration framework 104 includes a non-real-time RIC layer 106 configured to run a cloud-based application 108 (referred to as rApp in O-RAN). rApp 108 is coupled to an O1 service 110 via an R1 interface; in turn, the O1 service 110 is coupled to other layers via O1 interfaces to deliver various data.

[0019] The near real-time RIC layer 112 is coupled to the non-real-time RIC layer 106 via the A1 interface. The near real-time RIC layer 112 is configured to run edge-based applications (e.g., xApp 114 (extended applications) in O-RAN). One or more xApps 114 can be coupled to or incorporated into the ML inference host 116.

[0020] like Figure 1 As shown, typically, the E2 node 118, coupled to the near real-time RIC layer via the E2 interface, includes centralized unit (control plane) CU-CP 120 and centralized unit (user plane) CU-UP 122 components. The E2 node also includes distributed units or a pool of distributed units 124, which may include the real-time RIC layer 126. The real-time RIC layer 126 runs dApps 128, one or more of which may be coupled to or otherwise incorporated into beam management (programs) 130, including joint spatiotemporal beam management as described herein, possibly based on information obtained from other layers 106 and 112.

[0021] Figure 1 The diagram also shows an external application server 132. The external application server 132 can participate in various data collection operations.

[0022] Regarding beam scanning and beam management, a three-stage or procedure, P1, P2, and P3, is typically used. In P1, the base station performs a wide-beam scan covering its entire cell by transmitting a set of reference signals, where each reference signal corresponds to a specific beam. A finite number of wide beams can be used for wide-angle beam scanning based on an initial coarse estimate of the optimal beam orientation. For example, the reference signal could be a synchronization signal block (SSB) transmitted during initial access or a periodic channel state information reference signal (CSI-RS) configured for beam management. The UE uses the wide receive beam to measure the power of the received reference signals from all transmitted beams and reports the beam with the highest received power to the base station.

[0023] In P2, the base station performs a beam refinement procedure by transmitting aperiodic CSI-RS using a narrower beam in the angular direction around the optimal beam reported by the UE in P1. The UE measures the received power of the CSI-RS in each transmitted beam using a wide receive beam and reports the narrower beam with the highest received power to the base station. In P3, the base station repeatedly transmits the CSI-RS in the optimal narrow beam reported by the UE in P2, and the UE finds its optimal receive beam by measuring the power of the received CSI-RS in each receive beam. For data transmission, the base station uses the optimal transmitted beam found during P2, and the UE uses the optimal receive beam found during P3. It should be noted that what is considered "optimal" or best can be considered a beam that achieves at least a threshold of high performance, for example, relative to the beam from the previous... K The highest performance of other candidate beams in each subgroup; similarly, the previous K Each subgroup can be considered as a beam set that achieves at least some threshold high performance.

[0024] Moving towards joint spatial and temporal beam management, this paper describes the simultaneous acquisition of spatially and temporally correlated data; (as opposed to spatial and temporal beam management, which acquires spatial and temporal data separately). More specifically, the aspects of the techniques described herein involve joint spatial / temporal beam management, which leverages the advantages of both spatial and temporal beam management. As will be understood, joint spatial / temporal beam management includes spatial beam management on the observation window, which results in additional pre- K Several RSRP measurements were used as input data. Additionally, the previous time instances of the prediction window were considered. K A single RSRP measurement can also be considered as input to both the model training and inference phases. The joint case can lead to the removal of the observation period and the acquisition of refined feedback measurement data from the prediction period.

[0025] Figure 2 and Figure 3 This includes an example of a joint spatial-temporal beam management data flow / sequence graph for network-side AI / ML beam management procedures. In other words, in this example, model training and inference are performed at the base station.

[0026] During a training phase, the network (e.g., one or more E2 nodes 118) scans all potential beams (sets of beams). A 240, so that UE 242 can determine and report the optimal beam ID, as shown by arrows one (1) to three (3). The reference signal received power data (RSRP) of a small set of sparse beams (set B) along with the corresponding optimal beam ID obtained from the full set of scanned beams are fed back by UE 242 (arrow three (3)) to facilitate the collection and use of training input data, as shown by block four (4).

[0027] These training input data (RSRPs of a small set of beams and their corresponding optimal beam IDs) (which can come from multiple UEs within the coverage area) are therefore collected by the network. For the model training phase, the output is (or can be used to determine) a sparse set. B 244, the sparse set is a subset of the full beam set that is determined by the model during inference given the current conditions to be sufficient to scan the coverage area.

[0028] Therefore, during further inference, the network (E2 node 118) only needs to scan from a small set of sparse beams (set) B 244) beam. This is caused by Figure 2 Arrows five (5) and six (6) in the diagram indicate the process. Figure 3 At block seven (7), spatial beam inference based on AL / ML continues.

[0029] like Figure 2 and Figure 3 As shown, spatial beam management is performed during the observation period / phase (block 246). (Previous) K One CSI-RS beam scan (block 350) was used to determine the target beams using the example steps shown. N The optimal CSI-RS beam for each time interval (block 352) is shown as arrows five (5) and six (6), block seven, and arrows eight (8) through eleven (11), and these steps are repeated as shown as arrow twelve (12). It should be noted that although blocks 350 and 352 are respectively example static representations used to illustrate beam subsets and optimal beams, during this observation phase 246, the preceding... K One CSI-RS beam and / or the optimal CSI-RS beam can be used N The time interval can vary, for example, especially for mobile UEs / dynamic environments.

[0030] from Figure 3 The lower part shows that temporal beam management is performed during the prediction cycle / stage (block 354). (Previous) K One CSI-RS beam scan (block 356) was used to determine the target beams by the example steps shown. M The optimal CSI-RS beam for each time interval (block 358) is shown in blocks thirteen (13) and arrows fourteen (14) through seventeen (17), and these steps are repeated as shown in arrow eighteen (18). Again, it should be noted that although blocks 356 and 358 are illustrative static representations for purposes of illustration, during the prediction phase 354, the preceding... K One CSI-RS beam and / or the optimal CSI-RS beam can be used M The time interval can vary, for example, especially for mobile UEs / dynamic environments.

[0031] Understandably, since the beam considered optimal may change over time, this information can be used to determine the UE's trajectory data. That is, the time interval is known, and changes from one beam to another, etc., can be used as input to the model that determines the UE's trajectory and predicts the beam.

[0032] Beam management based on periodic measurement reports imposes significant overhead on the network, which does not scale with the number of users and beams. This paper describes incorporating AI / ML into beam management to predict the next optimal serving beam for the UE and switch beams based on measurement and indication history. In one implementation, a Long Short-Term Memory (LSTM) network is able to predict the optimal serving beam by learning user mobility patterns. However, as described in this paper, users typically do not move randomly but follow social customs. To capture this rule in the trajectory, instead of observing a group of users in isolation, a social LSTM network can be used to model multiple user interactions in the beam management process.

[0033] Considering multiple UE scenarios necessitates environmental modeling and capturing spatial and temporal correlations across UEs. These cross-UE spatial and temporal correlations help to more accurately define the serving beam or optimal beam. K Multiple beams. In other words, the multi-UE beam management strategy described in this paper helps predict the future movement of each UE, taking into account the surrounding environment (e.g., to avoid users colliding with each other).

[0034] Predicting UE trajectories in advance leads to a significant reduction in measurement procedural overhead for prediction intervals / dwell times and improves the accuracy of prediction aids using UE position and velocity. Capturing this cross-UE correlation in the beam management procedure takes into account interference conditions and dynamic shadowing effects, thereby improving accuracy. Aspects of the techniques described in this paper can be based on environmental modeling, capturing the correlations (both spatial and temporal) of multiple UEs, resulting in implicit or explicit trajectory predictions. Therefore, the type of user considered for such a model significantly affects the results.

[0035] One model considers a single class of users with the same mobility pattern. Another model considers multiple classes of users with different mobility patterns, such as high-speed vehicle users, medium-speed bicycle users, low-speed pedestrian users, etc. Considering multiple user classes leads to multimodal trajectory predictions that can be hidden in the learned beam management model or visible in the model output.

[0036] Furthermore, techniques for simulating cross-UE correlations (spatial / temporal) in beam management procedures learn environmental and channel variations accordingly. Learning channel variations in the system improves accuracy and reduces the cardinality of the measurement set data (i.e., correspondingly reduces overhead).

[0037] To model the environment for capturing individual UE trajectories, recurrent neural network (RNN) models (e.g., LSTM) can be used. Furthermore, this paper describes modeling the environment for predicting multi-UE trajectories, which requires a more advanced RNN model capable of capturing cross-UE correlations in both single-UE and multi-UE scenarios. Social LSTM allows neighboring UEs to share state data. In a social LSTM model, hidden state information from (up to) all neighboring states can be aggregated to capture cross-correlation. Hidden states include each UE's location and velocity, as well as the UE's measurement data history. The social LSTM architecture is used to learn characteristics of beamspace, environment, and associated multi-user mobility patterns.

[0038] Furthermore, the set of neighboring UEs can be defined based on a metric, such as distance. This metric can be adjusted based on UE profile data (e.g., speed, measured signal-to-noise ratio (SINR)) and other environmental characteristics.

[0039] For reference Figure 1 The beam management 130, which may include multiple ML models, can be incorporated into or coupled to a dApp instance of dApp 128 (where an application deployed on the Real-Time RAN Intelligent Controller (RT-RIC) 126 is referred to as a dApp, which implements real-time control in the RAN node). The RT-RIC 126 is co-located with the Distributed Unit (DU) 124 (which may be aggregated) and is responsible for real-time RAN control actions.

[0040] Figure 4 The example ML model structure 460 for beam management dApp is shown below. Input data 462 includes, but is not limited to, sets. B Beam RSRP measurement data history and best K The model includes a history of beam RSRP (or CSI-RS) measurement data, beam indication data (selected optimal beam), and rich information such as UE position and velocity (and possible orientation) data. Another input dataset 464 includes hidden state data capturing cross-UE trajectory dependency data. The model output includes UE trajectory prediction data 466 and the optimal beam / optimal beam. K 468 beam prediction data.

[0041] Regarding interaction-aware single-class multi-user beam management, as described in this paper, beam prediction can be viewed as a sequence generation task used to predict the optimal beam based on the history of selected beams. Optimal beam prediction is highly correlated with UE mobility patterns and their corresponding trajectories. It is well known that users in congested environments adjust their movement based on the movement of other users in their vicinity. Therefore, using a single LSTM model for each user cannot capture the interactions of users in the neighborhood.

[0042] Therefore, observing the behavior of multiple users can take into account the behavior of other users within a large neighborhood (e.g., a cell or its distance-based portion) while predicting the optimal beam for each user. Thus, a modified pooling-based LSTM model (Social LSTM) is used to jointly predict the optimal service beam for all users in the neighborhood (e.g., a cell), as described in this paper.

[0043] Social LSTM connects adjacent LSTMs to capture user interactions. In the social LSTM approach, a separate LSTM network is used for each user in the cell, and a social pooling (S-pooling) layer connects the LSTMs to each other. The pooling operation is as follows: Figure 5 As shown, spatially neighboring LSTMs share information with each other through this pooling layer. The hidden states of the LSTMs in the neighborhood are pooled together and used as input for the next time step.

[0044] In the beam management procedure, features and utility metrics of the hidden states are defined to capture the states of neighboring users. LSTM in time t Hidden state Capture the first in the cell i The potential representation of a user. For example... Figure 4 The representation includes a set. B RSRP measurement, optimal K Individual beam RSRP measurements and available UE trajectory information. Utilizing the social hidden state tensor. This representation is shared with neighboring users; the social hidden state tensor is defined as follows: , in It corresponds to in t -1 hour j Hidden states of an LSTM for each user It is used to check users j Is it in the user i The indicator function in the defined neighborhood, and It corresponds to the user i The neighbors' group. It should be noted that... It is a user-related metric (a function in the UE profile) that defines the proximity radius. When UE location information is available, Is t -1 hour i The location of each user, otherwise, in t -1 is the index of the service beam.

[0045] Using pooled social hidden state tensors This leads to the following recursion: , in W These are LSTM weights, and It is a set B Predicted RSRP measurements and for the first i individual users at time t The best time K Each beam. In addition, when UE location information is available, the output also includes UE trajectory prediction.

[0046] Multi-user beam management considers agent classification, such as cars, pedestrians, or bicycles, and predicts their different behaviors. Interactions between different agent classes can also play a crucial role in real-world business scenarios, and thus, in beam management. For multi-user beam management scenarios, interaction-aware models capture inter-class / intra-class interactions and dependencies.

[0047] For this scenario, compared to the single-class multi-user beam management case mentioned above, a social LSTM method with modified hidden state features is used. In time... t Hidden state of LSTM The first in the capture unit i The potential representation of each user and in time t Time user i The category. This representation includes sets. B RSRP measurement, optimal K Individual beam RSRP measurements, user category, and available UE trajectory information.

[0048] User categories may not be available for beam prediction models. In this case, historical trajectory information is used to identify the category. If user location and velocity are obtained via an external server (e.g., Figure 1 If the information provided by server 132 is available, then UE speed can be used to perform UE classification. If there is no information about UE speed to classify it into different categories, a feedforward neural network can be used to classify it based on the user's historical state, as described below: Where the function a It is a feedforward neural network, which is trained along with the entire network, and It is in time t Time user i The category. It's worth noting that specific users... i A user's category can change based on their speed. In other words, each user's category is not fixed.

[0049] As can be seen, the social LSTM method can share state among neighboring UEs. In the social LSTM method, hidden state information from up to all neighboring UEs can be pooled to model the user's mobility pattern dependencies. Hidden state includes the measured RSRP of the sparse scan beam over the observation period, the history of the optimal CSI-RS beam ID, and inferred UE trajectory information (i.e., location and velocity). To share the hidden state of neighboring UEs with each user, a UE-related neighbor metric, 𝒩𝑖, can be defined to specify the set of UEs whose state will be shared with the user. This metric can be a function of the UE profile (e.g., velocity). Interactively aware multi-user beam prediction can be used for temporal and joint spatiotemporal beam management scenarios.

[0050] One or more aspects can be embodied in network devices, such as Figure 6 The example operations are illustrated in the examples and may include, for example, a memory storing computer-executable components and / or operations and a processor executing the computer-executable components and / or operations stored in the memory. Example operations may include operation 602, which represents a space beam management observation phase performed over a first time interval of a first number of space observation periods, and may include example operations 604 and 606. Example operation 604 represents performing a synchronization signal block beam scan to determine a high-performance beam subgroup based on received power data returned from the user equipment corresponding to the synchronization signal block beam scan. Example operation 606 represents determining a space beam pair that can be used to achieve at least a first threshold of high performance, including performing a first channel state information reference signal beam scan using the high-performance beam subgroup to obtain a first channel state information reference signal identifier representing the space beam pair from the user equipment. Example operation 608 represents a time beam management prediction phase performed over a second time interval of a second number of time prediction periods, and may include example operations 610 and 612. Example operation 610 represents determining a time beam pair that can be used to achieve at least a second threshold of high performance, including performing a second channel state information reference signal beam scan based on the spatial beam pair to obtain a second channel state information reference signal identifier representing the time beam pair from the user equipment. Example operation 612 represents estimating the user equipment's trajectory data based on the spatial beam pair and the time beam pair.

[0051] Performing a synchronous signal block beam scan involves obtaining a sparse beam group from the model and using the sparse beam group.

[0052] Performing a second channel state information reference signal beam scan based on the spatial beam pair includes using at least some of the high-performance beam subgroups that include the spatial beam pair.

[0053] The trajectory data of the user equipment may include first trajectory data of the first user equipment, and further operations may include estimating second trajectory data of the second user equipment and predicting future optimal beam data for the first user equipment based on the first and second trajectory data.

[0054] Estimating trajectory data and predicting the future optimal beam for a first user equipment may include: inputting a first dataset into the model, the first dataset including historical beam measurement data, historical beam indication data, location data of the first user equipment, and velocity data of the first user equipment; inputting a second dataset into the model, the second dataset including cross-user equipment trajectory dependency data based on the first and second trajectory data; and obtaining an estimate of the trajectory data and a prediction of the future optimal beam data in response to inputting the first and second datasets into the model.

[0055] The future optimal beam data for the first user equipment may include first future optimal beam data, and further operations may include predicting second future optimal beam data for the second user equipment based on first trajectory data and second trajectory data.

[0056] The network device includes distributed components corresponding to a communication neighborhood. The model can be a first model of the distributed components. Determining spatial beam pairs can be performed using a second model of the distributed components that can include a spatial inference model, and determining temporal beam pairs can be performed using a third model of the distributed components that include a temporal inference model.

[0057] Historical beam measurement data may include first historical reference signal received power measurement data corresponding to a synchronization signal block beam scan and second historical reference signal received power measurement data corresponding to at least one of the following: a first channel state information reference signal beam scan or a second channel state information reference signal beam scan.

[0058] The trajectory data of a user equipment may include first trajectory data of a first user equipment, and further operations may include: estimating corresponding other trajectory data of corresponding other user equipment in a group that may include the first user equipment and corresponding other user equipment, predicting future optimal beam data for the first user equipment based on the first trajectory data and other trajectory data, and predicting corresponding other future optimal beam data for corresponding other user equipment based on the first trajectory data and corresponding other trajectory data.

[0059] A group that may include a first user device and corresponding other user devices can be classified into a category of user devices.

[0060] Further operations may include classifying the group into categories based on the first trajectory data and corresponding other trajectory data.

[0061] Further operations may include classifying groups into categories based on first historical state data of the first user device and corresponding other historical state data of other user devices.

[0062] The trajectory data of a user equipment may include first trajectory data of a first user equipment, the first user equipment and corresponding first other user equipment may be classified into a first classification group, and further operations may include: estimating corresponding first other trajectory data of the corresponding first other user equipment, estimating corresponding second other trajectory data of the corresponding second user equipment including a second classification group of the corresponding second user equipment, predicting the future optimal beam for the first user equipment based on the first trajectory data, the first other trajectory data and the second other trajectory data; and predicting corresponding other first future optimal beams for the corresponding first other user equipment based on the first trajectory data, the first other trajectory data and the second other trajectory data.

[0063] Further operations may include predicting corresponding second future optimal beams for corresponding second other user equipment based on first trajectory data, first other trajectory data, and second other trajectory data.

[0064] Figure 7 The symbol represents one or more example aspects, such as example operations corresponding to methods. Example operation 702 represents a system including a processor determining corresponding trajectory data for a group of user equipment based on corresponding joint spatial and temporal beam data obtained for corresponding user equipment in a communication neighborhood. Example operation 704 represents a system inputting first trajectory data of the first user equipment in the group and trajectory cross-dependency data based on the corresponding trajectory data into a social recurrent neural network model. Example operation 706 represents a system obtaining predicted trajectory data and predicted future optimal beam data for the first user equipment in response to the input.

[0065] The first trajectory data may be part of a dataset, which may also include historical beam measurement data and historical beam indication data, and inputting the first trajectory data may include inputting the dataset to the social recurrent neural network model.

[0066] Further operations may include the system determining the corresponding joint spatial and temporal beam data, which may include a corresponding spatial beam management observation phase to obtain the corresponding spatial beam data and a corresponding temporal beam management prediction phase based on the corresponding spatial beam data.

[0067] Figure 8This summary includes various example operations corresponding to machine-readable media, including executable instructions that facilitate the execution of the operations when executed by a processor. Example operation 802 represents determining corresponding joint spatial and temporal beam data, including performing a corresponding spatial beam management observation phase to obtain the corresponding spatial beam data and a corresponding temporal beam management prediction phase based on the corresponding spatial beam data. Example operation 804 represents determining corresponding trajectory data for a corresponding user equipment based on the corresponding joint spatial and temporal beam data. Example operation 806 represents obtaining cross-dependency trajectory data based on the corresponding trajectory data. Example operation 808 represents inputting cross-dependency trajectory data and first trajectory data of a first user equipment to a model. Example operation 810 represents obtaining first predicted trajectory data for a first user equipment and first predicted future optimal beam data for the first user equipment in response to the input.

[0068] Further operations may include inputting cross-dependent trajectory data and second trajectory data of a second user equipment to the model, and obtaining second predicted trajectory data for the second user equipment and second predicted future optimal beam data for the first user equipment in response to the input.

[0069] The corresponding spatial beam management observation phase may include: performing a corresponding synchronization block beam scan to determine a corresponding high-performance beam subgroup based on the corresponding received power data corresponding to the synchronization block beam scan, and determining the corresponding optimal spatial beam data, including using the corresponding high-performance beam subgroup to perform a corresponding first channel state information reference signal beam scan to obtain a corresponding first channel state information reference signal identifier representing the corresponding optimal spatial beam data from the corresponding user equipment; the corresponding temporal beam management prediction phase may include determining the corresponding optimal temporal beam data, which may include performing a corresponding second channel state information reference signal beam scan based on the corresponding spatial beam data to obtain a corresponding second channel state information reference signal identifier representing the corresponding temporal beam data from the corresponding user equipment.

[0070] As can be seen, the techniques described in this paper promote improved reliability and throughput in communication networks, as well as reduced latency and signaling overhead. In joint spatial / temporal beam management, the observation-only phase of the temporal beam is skipped, with spatial beam management performed on the observation period. Performing spatial beam management on the observation period provides optimal CSI-RS ID, which improves the optimality on the prediction interval. K The accuracy of beam selection is improved. Furthermore, spatial beam management during the observation period enhances spectral efficiency over that period.

[0071] Furthermore, observing multiple user facilitations allows for the consideration of the behavior of other users within the neighborhood while simultaneously predicting the optimal beam and / or future trajectory data for each user. Social LSTMs connect adjacent LSTMs to capture user interactions. For prediction data, a single user category or multiple user categories can be considered.

[0072] Figure 9 This is a schematic block diagram of a computing environment 900 with which the disclosed subject can interact. System 900 includes one or more remote components 910. The remote components 910 may be hardware and / or software (e.g., threads, processes, computing devices). In some embodiments, the remote components 910 may be a distributed computer system, connected via a communication framework 940 to local autoscaling components and / or programs that use the resources of the distributed computer system. The communication framework 940 may include wired network devices, wireless network devices, mobile devices, wearable devices, radio access network devices, gateway devices, femtocell devices, servers, etc.

[0073] System 900 also includes one or more local components 920. The local components 920 may be hardware and / or software (e.g., threads, processes, computing devices). In some embodiments, the local components 920 may include autoscaling components and / or programs that communicate / use remote resources 910, etc., and are connected to a remote distributed computing system via a communication framework 940.

[0074] One possible form of communication between (multiple) remote components 910 and (multiple) local components 920 may be the form of data packets transmitted between two or more computer processes. Another possible form of communication between (multiple) remote components 910 and (multiple) local components 920 may be the form of circuit-switched data transmitted between two or more computer processes in a radio time slot. System 900 includes a communication framework 940 that can be used to facilitate communication between (multiple) remote components 910 and (multiple) local components 920, and may include an air interface, such as the Uu interface of a UMTS network, via a Long Term Evolution (LTE) network, etc. (Multiple) remote components 910 may be operatively connected to one or more remote data repositories 950, such as hard disk drives, solid-state drives, SIM cards, device memory, etc., that can be used to store information on the (multiple) remote component 910 side of the communication framework 940. Similarly, the local components 920 can be operatively connected to one or more local data repositories 930, which can be used to store information on the local component 920 side of the communication framework 940.

[0075] To provide additional context for the various embodiments described herein, Figure 10 The following discussion is intended to provide a brief overview of a suitable computing environment 1000 in which various embodiments of the embodiments described herein may be implemented. While the embodiments have been described above in the general context of computer-executable instructions that can run on one or more computers, those skilled in the art will recognize that the embodiments may also be implemented in combination with other program modules and / or as a combination of hardware and software.

[0076] Typically, program modules include routines, programs, components, data structures, etc., that perform specific tasks or implement specific abstract data types. Furthermore, those skilled in the art will appreciate that the method can be practiced with other computer system configurations, including single-processor or multi-processor computer systems, minicomputers, mainframes, Internet of Things (IoT) devices, distributed computing systems, and personal computers, handheld computing devices, microprocessor-based or programmable consumer electronics, each of which can be operatively coupled to one or more associated devices.

[0077] The embodiments described herein can also be practiced in a distributed computing environment, where certain tasks are performed by remote processing devices linked via a communication network. In a distributed computing environment, program modules can reside on both local and remote memory storage devices.

[0078] Computing devices typically include various media, which may include computer-readable storage media, machine-readable storage media, and / or communication media, these two terms being used interchangeably herein as follows. A computer-readable storage medium or a machine-readable storage medium can be any available storage medium accessible by a computer, and includes both volatile and non-volatile media, removable and non-removable media. By way of example and not limitation, a computer-readable storage medium or a machine-readable storage medium can be implemented in conjunction with any method or technique for storing information such as computer-readable or machine-readable instructions, program modules, structured data, or unstructured data.

[0079] Computer-readable storage media may include, but are not limited to, random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other storage technologies, optical disc read-only memory (CD-ROM), digital versatile disc (DVD), Blu-ray disc (BD) or other optical disc storage devices, magnetic tape cassettes, magnetic tape, disk storage devices or other magnetic storage devices, solid-state drives or other solid-state storage devices, or other tangible and / or non-transitory media that can be used to store desired information. In this regard, when applied herein to storage devices, memories, or computer-readable media, the terms “tangible” or “non-transitory” should be understood to exclude only the propagation of transient signals themselves as a modifier, and do not waive the rights to all standard storage devices, memories, or computer-readable media that do not only propagate transient signals themselves.

[0080] Computer-readable storage media can be accessed by one or more local or remote computing devices, for example via access requests, queries or other data retrieval protocols, for various operations concerning the information stored on the media.

[0081] Communication media typically embody computer-readable instructions, data structures, program modules, or other structured or unstructured data in data signals such as modulated data signals, carrier waves, or other transmission mechanisms, and include any medium for delivering or transmitting information. The term "modulated data signal" or signal refers to a signal having one or more of its characteristics, which are set or altered in a manner that encodes information in one or more signals. By way of example and not limitation, communication media include: wired media, such as wired networks or straight-line connections, and wireless media, such as acoustic, RF, infrared, and other wireless media.

[0082] Refer again Figure 10 An example environment 1000 for implementing various embodiments of the aspects described herein includes a computer 1002, which includes a processing unit 1004, a system memory 1006, and a system bus 1008. The system bus 1008 couples system components, including but not limited to the system memory 1006, to the processing unit 1004. The processing unit 1004 can be any of a variety of commercial processors. Dual microprocessors and other multiprocessor architectures can also be used as the processing unit 1004.

[0083] System bus 1008 can be any of several types of bus architectures, which can be further interconnected to memory buses (with or without memory controllers), peripheral buses, and local buses using any of the various commercial bus architectures. System memory 1006 includes ROM 1010 and RAM 1012. The Basic Input / Output System (BIOS) can be stored in non-volatile memory such as ROM, erasable programmable read-only memory (EPROM), or EEPROM, containing basic routines such as those that help transfer information between components within computer 1002 during startup. RAM 1012 may also include high-speed RAM, such as static RAM for caching data.

[0084] Computer 1002 also includes an internal hard disk drive (HDD) 1014 (e.g., EIDE, SATA) and may include one or more external storage devices 1016 (e.g., floppy disk drive (FDD) 1016, memory stick or flash drive reader, memory card reader, etc.). Although the internal HDD 1014 is shown as being located within computer 1002, the internal HDD 1014 may also be configured for external use within a suitable chassis (not shown). Additionally, although not shown in environment 1000, a solid-state drive (SSD) may be used in addition to, or in place of, HDD 1014.

[0085] Other internal or external storage devices may include at least one other storage device 1020 having storage medium 1022 (e.g., solid-state storage device, non-volatile storage device, and / or optical disc drive that can be read from or written to removable media such as CD-ROM, DVD, BD, etc.). External storage device 1016 may be facilitated by a network virtual machine. HDD 1014, (multiple) external storage devices 1016, and storage devices (e.g., drives) 1020 may be connected to system bus 1008 via HDD interface 1024, external storage interface 1026, and drive interface 1028, respectively.

[0086] The drive and its associated computer-readable storage medium provide non-volatile storage of data, data structures, computer-executable instructions, etc. For computer 1002, the drive and storage medium accommodate storage of any data in a suitable digital format. Although the above description of computer-readable storage media refers to various types of storage devices, those skilled in the art will understand that other types of computer-readable storage media, whether currently existing or developed in the future, may also be used in the example operating environment, and further, any such storage medium may contain computer-executable instructions for performing the methods described herein.

[0087] Numerous program modules can be stored in the drive and RAM 1012, including the operating system 1030, one or more application programs 1032, other program modules 1034, and program data 1036. All or part of the operating system, applications, modules, and / or data can also be cached in RAM 1012. The systems and methods described herein can be implemented using various commercial operating systems or combinations of operating systems.

[0088] Computer 1002 may optionally include emulation technology. For example, a hypervisor (not shown) or other intermediary may emulate the hardware environment for operating system 1030, and the emulated hardware may optionally differ from... Figure 10 The hardware is shown. In this embodiment, the operating system 1030 may include one of a plurality of virtual machines (VMs) hosted at the computer 1002. Furthermore, the operating system 1030 may provide a runtime environment for the application 1032, such as the Java Runtime Environment or the .NET Framework. A runtime environment is a consistent execution environment that allows the application 1032 to run on any operating system that includes a runtime environment. Similarly, the operating system 1030 may support containers, and the application 1032 may be in the form of a container, which is a lightweight, standalone, executable software package including, for example, code, runtime, system tools, system libraries, and settings specific to the application.

[0089] Furthermore, computer 1002 can enable security modules, such as a Trusted Processing Module (TPM). For example, in the case of TPM, before loading the next boot component, the boot component performs a hash operation on the next timely boot component and waits for the result to match a security value. This process can occur at any layer of the computer 1002's code execution stack, for example, it can be applied at the application execution level or the operating system (OS) kernel level, thereby achieving security at any code execution level.

[0090] Users can input commands and information to computer 1002 through one or more wired / wireless input devices, such as keyboard 1038, touchscreen 1040, and pointing devices such as mouse 1042. Other input devices (not shown) may include microphones, infrared (IR) remote controls, radio frequency (RF) remote controls or other remote controls, joysticks, virtual reality controllers and / or virtual reality headsets, gaming pads, styluses, image input devices (e.g., cameras), gesture sensor input devices, visual motion sensor input devices, emotion or face detection devices, biometric input devices (e.g., fingerprint or iris scanners), etc. These and other input devices are typically connected to processing unit 1004 via input device interface 1044, which may be coupled to system bus 1008, but may also be connected using other interfaces such as parallel ports, IEEE 1094 serial ports, gaming ports, USB ports, IR interfaces, BLUETOOTH® interfaces, etc.

[0091] Monitor 1046 or other types of display devices can also be connected to system bus 1008 via an interface such as video adapter 1048. In addition to monitor 1046, the computer typically includes other peripheral output devices (not shown), such as speakers, printers, etc.

[0092] Computer 1002 can operate in a networked environment using logical connections to one or more remote computers (such as (multiple) remote computers 1050) via wired and / or wireless communications. The (multiple) remote computers 1050 can be workstations, server computers, routers, personal computers, laptops, microprocessor-based entertainment devices, peer-to-peer devices, or other common network nodes, and typically include many or all of the elements described relative to computer 1002, although for brevity only memory / storage device 1052 is illustrated. The depicted logical connections include wired / wireless connectivity to a local area network (LAN) 1054 and / or a larger network (e.g., a wide area network (WAN) 1056). Such LAN and WAN networking environments are common in offices and companies and facilitate enterprise-wide computer networks, such as intranets, all of which can connect to global communication networks such as the Internet.

[0093] When used in a LAN networking environment, computer 1002 can connect to local area network 1054 via a wired and / or wireless communication network interface or adapter 1058. Adapter 1058 can facilitate wired or wireless communication to LAN 1054, which may also include wireless access points (APs) deployed thereon for communicating with adapter 1058 in wireless mode.

[0094] When used in a WAN networking environment, computer 1002 may include modem 1060, or may be connected to a communication server on WAN 1056 via other components (such as via the Internet) for establishing communication over WAN 1056. Modem 1060, which may be internal or external and wired or wireless, may be connected to system bus 1008 via input device interface 1044. In a networked environment, program modules described relative to computer 1002 or parts thereof may be stored in remote memory / storage device 1052. It will be understood that the network connection shown is an example, and other components may be used to establish communication links between computers.

[0095] When used in a LAN or WAN networking environment, in addition to or replacing the external storage device 1016 as described above, computer 1002 can access cloud storage systems or other network-based storage systems. Typically, the connection between computer 1002 and the cloud storage system can be established via LAN 1054 or WAN 1056, for example, via adapter 1058 or modem 1060 respectively. When computer 1002 is connected to an associated cloud storage system, external storage interface 1026 can manage the storage devices provided by the cloud storage system with the help of adapter 1058 and / or modem 1060, just as it manages other types of external storage devices. For example, external storage interface 1026 can be configured to provide access to cloud storage sources as if these sources were physically connected to computer 1002.

[0096] Computer 1002 is operable to communicate with any wireless device or entity operatively positioned in wireless communication (e.g., printer, scanner, desktop and / or portable computer, portable data assistant, communications satellite, any device or location associated with a wirelessly detectable tag (e.g., phone booth, newsstand, shelf, etc.) and telephone). This can include Wi-Fi and BLUETOOTH® wireless technologies. Therefore, communication can be a predefined structure like a traditional network, or simply self-organizing communication between at least two devices.

[0097] The above description of the embodiments illustrated in this disclosure, including the content described in the abstract, is not intended to be exhaustive or to limit the disclosed embodiments to the precise forms disclosed. While specific embodiments and examples have been described herein for illustrative purposes, those skilled in the art will recognize that various modifications are possible within the scope of such embodiments and examples.

[0098] In this regard, although the disclosed subject matter has been described in conjunction with various embodiments and corresponding drawings, it should be understood that, where applicable, other similar embodiments may be used, or modifications and additions may be made to the described embodiments to perform the same, similar, alternative, or substitute functions of the disclosed subject matter without departing from the disclosed subject matter. Therefore, the disclosed subject matter should not be limited to any single embodiment described herein, but should be interpreted in accordance with the breadth and scope of the appended claims.

[0099] As used in this subject matter specification, the term "processor" can refer to substantially any computing processing unit or device, including, but not limited to, single-core processors; single-processors with software multithreading capabilities; multi-core processors; multi-core processors with software multithreading capabilities; multi-core processors with hardware multithreading technology; parallel platforms; and parallel platforms with distributed shared memory. Additionally, a processor can refer to an integrated circuit, application-specific integrated circuit, digital signal processor, field-programmable gate array, programmable logic controller, complex programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination thereof designed to perform the functions described herein. Processors can utilize nanoscale architectures, such as, but not limited to, molecular and quantum dot-based transistors, switches, and gates, to optimize space utilization or enhance the performance of user devices. Processors can also be implemented as a combination of computing processing units.

[0100] As used herein, the terms “component,” “system,” “platform,” “layer,” “selector,” “interface,” etc., are intended to refer to a computer-related entity or an entity related to an operating device having one or more specific functionalities, wherein the entity may be hardware, a combination of hardware and software, software, or software in execution. As an example, a component may be, but is not limited to, a process running on a processor, a processor, an object, an executable program, an execution thread, a program, and / or a computer. By way of illustration and not limitation, both an application running on a server and the server itself can be components. One or more components may reside in an executing process and / or thread, and components may reside on a single computer and / or be distributed among two or more computers. Furthermore, these components may be executed from various computer-readable media on which various data structures are stored. Components may communicate via local and / or remote processes, such as according to signals having one or more data packets (e.g., data from a component that interacts with a local system, another component in a distributed system, and / or other systems via a network such as the Internet). As another example, a component can be a device having specific functionality provided by mechanical parts operated by an electrical or electronic circuit system, operated by a software or firmware application executed by a processor, wherein the processor can be internal or external to the device and executes at least a portion of the software or firmware application. As yet another example, a component can be a device that provides specific functionality without mechanical parts through an electronic component, which may include a processor therein to execute software or firmware that at least partially endows the electronic component with functionality.

[0101] Furthermore, the term "or" is intended to mean an inclusive "or," not an exclusive "or." That is, unless otherwise stated or clear from the context, "X uses A or B" is intended to mean any natural inclusive permutation. In other words, if X uses A; X uses B; or X uses both A and B, then "X uses A or B" is satisfied in any of the foregoing instances.

[0102] While the embodiments are susceptible to various modifications and alternative constructions, some of the illustrated implementations are shown in the accompanying drawings and have been described in detail above. However, it should be understood that the various embodiments are not intended to be limited to the specific forms disclosed, but rather, the invention is intended to cover all modifications, alternative constructions, and equivalents falling within its spirit and scope.

[0103] In addition to the various embodiments described herein, it should be understood that other similar embodiments may be used, or modifications and additions may be made to the described embodiments(s) to perform the same or equivalent functions of the corresponding embodiments(s) without departing from the invention. Furthermore, multiple processing chips or multiple devices may share the performance of one or more functions described herein, and similarly, storage may be implemented across multiple devices. Therefore, the various embodiments are not limited to any single embodiment, but are to be interpreted in accordance with the breadth, spirit, and scope of the appended claims.

Claims

1. A network device comprising: a processor; and a memory storing executable instructions that, when executed by the processor, facilitate performance of operations comprising: performing a spatial beam management observation phase over a first number of first time intervals of a spatial observation period, including: performing a synchronization signal block beam sweep to determine a high performance beam sub-group with respect to received power data returned from a user device corresponding to the synchronization signal block beam sweep, determining spatial beam pairs usable to achieve at least a first threshold high performance, including performing a first channel state information reference signal beam sweep using the high performance beam sub-group to obtain from the user device first channel state information reference signal identifiers representing the spatial beam pairs; performing a temporal beam management prediction phase over a second number of second time intervals of a temporal prediction period, including: determining temporal beam pairs usable to achieve at least a second threshold high performance, including performing a second channel state information reference signal beam sweep based on the spatial beam pairs to obtain from the user device second channel state information reference signal identifiers representing the temporal beam pairs; and estimating trajectory data of the user device based on the spatial beam pairs and the temporal beam pairs. obtaining a sparse beam group from a model, and using the sparse beam group.

2. The network device of claim 1, wherein the performance of the synchronization signal block beam sweep comprises: using at least some beams of the high performance beam sub-group including the spatial beam pairs.

3. The network device of claim 1, wherein the performing of the spatial beam pair based second channel state information reference signal beam sweeping comprises: estimating second trajectory data of a second user device, and predicting future optimal beam data for the first user device based on the first trajectory data and the second trajectory data.

4. The network device of claim 1, wherein the trajectory data of the user devices comprises first trajectory data of a first user device, and wherein the operations further comprise: inputting a first data set comprising historical beam measurement data, historical beam indication data, location data of the first user device, and velocity data of the first user device to a model, inputting a second data set comprising cross-user device trajectory dependency data based on the first trajectory data and the second trajectory data to the model, and obtaining an estimate of the trajectory data and a prediction of the future optimal beam data responsive to the input of the first data set and the second data set to the model.

5. The network device of claim 4, wherein the estimate of the trajectory data and the future optimal beam for the prediction of the first user device comprises: predicting second future optimal beam data for the second user device based on the first trajectory data and the second trajectory data.

6. The network device of claim 5, wherein the future best beam data for the first user device comprises first future best beam data, and wherein the operations further comprise:

7. The network device of claim 5, wherein the network device comprises a distributed component corresponding to a communication neighborhood, wherein the model is a first model of the distributed component, wherein the determination of the spatial beam pairs is performed using a second model of the distributed component comprising a spatial reasoning model, and wherein the determination of the temporal beam pairs is performed using a third model of the distributed component comprising a temporal reasoning model. ​ 8. The network device of claim 5, wherein the historical beam measurement data comprises first historical reference signal received power measurement data corresponding to the synchronization signal block beam sweep and second historical reference signal received power measurement data corresponding to at least one of: the first channel state information reference signal beam sweep or the second channel state information reference signal beam sweep.

9. The network device of claim 1, wherein the trajectory data of the user devices comprises first trajectory data of a first user device, and wherein the operations further comprise: estimating respective other trajectory data of respective other user devices of a group comprising the first user device and the respective other user devices, predicting future optimal beam data for the first user device based on the first trajectory data and the other trajectory data, and predicting respective other future optimal beam data for the respective other user devices based on the first trajectory data and the respective other trajectory data.

10. The network device of claim 9, wherein the group comprising the first user device and the respective other user devices is classified into a category of user devices.

11. The network device of claim 10, wherein the operations further comprise: classifying the group into the category based on the first trajectory data and the respective other trajectory data.

12. The network device of claim 10, wherein the operations further comprise: classifying the group into the category based on first historical state data of the first user device and respective other historical state data of the respective other user devices.

13. The network device of claim 1, wherein the trajectory data of the user devices includes first trajectory data of a first user device, wherein the first user device and a corresponding first other user device are classified into a first classification group, and wherein the operations further comprise: estimating respective first other trajectory data of the respective first other user devices, estimating respective second other trajectory data of respective second user devices of a second classified group comprising the respective second user devices, predicting future optimal beam for the first user device based on the first trajectory data, the first other trajectory data, and the second other trajectory data, and predicting respective other first future optimal beam for the respective first other user devices based on the first trajectory data, the first other trajectory data, and the second other trajectory data.

14. The network device of claim 13, wherein the operations further comprise: predicting respective other second future optimal beam for the respective second other user devices based on the first trajectory data, the first other trajectory data, and the second other trajectory data.

15. A method comprising: determining, by a system comprising a processor, respective trajectory data for a group of respective user devices based on respective joint spatial and temporal beam data obtained for the respective user devices in a communication neighborhood; inputting, by the system, first trajectory data of a first user device of the group and trajectory cross-dependence data based on the respective trajectory data to a social recurrent neural network model; and in response to the input, obtaining, by the system, predicted trajectory data and predicted future optimal beam data for the first user device.

16. The method of claim 15, wherein the first trajectory data is part of a data set that further comprises historical beam measurement data and historical beam indication data, and wherein the input of the first trajectory data comprises inputting the data set to the social recurrent neural network model.

17. The method of claim 15, further comprising: determining the respective joint spatial and temporal beam data, including conducting a respective spatial beam management observation phase to obtain respective spatial beam data and conducting a respective temporal beam management prediction phase based on the respective spatial beam data; 18. A non-transitory machine-readable medium comprising executable instructions that, when executed by a processor, facilitate performance of operations comprising: determining respective joint spatial and temporal beam data, including conducting a respective spatial beam management observation phase to obtain respective spatial beam data and conducting a respective temporal beam management prediction phase based on the respective spatial beam data; determining respective trajectory data for respective user equipment based on the respective joint spatial and temporal beam data; obtaining cross-dependence trajectory data based on the respective trajectory data; inputting the cross-dependence trajectory data and first trajectory data for a first user equipment of the respective user equipment to a model; and in response to the input, obtaining first predicted trajectory data for the first user equipment and first predicted future optimal beam data for the first user equipment.

19. The non-transitory machine readable medium of claim 18, wherein the operations further comprise: inputting the cross-dependence trajectory data and second trajectory data for a second user equipment of the respective user equipment to the model, and in response to the input, obtaining second predicted trajectory data for the second user equipment and second predicted future optimal beam data for the first user equipment.

20. The non-transitory machine-readable medium of claim 19, wherein the conducting of the respective spatial beam management observation phase comprises: performing a respective synchronization signal block beam sweep based on respective received power data corresponding to the synchronization signal block beam sweep to determine a respective high performance beam sub-set, and determining the respective optimal spatial beam data, including performing a respective first channel state information reference signal beam sweep using the respective high performance beam sub-set to obtain, from a respective user equipment, a respective first channel state information reference signal identifier representing the respective optimal spatial beam data; and wherein the conducting of the respective temporal beam management prediction phase comprises: determining the respective optimal temporal beam data, including performing a respective second channel state information reference signal beam sweep based on the respective spatial beam data to obtain, from the respective user equipment, a respective second channel state information reference signal identifier representing the respective temporal beam data.

Images