Confidence-based advanced trajectory prediction

Abstract

An apparatus comprising: at least one processor; and at least one memory including computer program code; wherein the at least one memory and the computer program code are configured to, with the at least one processor, cause the apparatus to at least: receive a sequence of past service locations of the user device as input to a machine learning model; determining a confidence vector for the predicted next location of the user device using a machine learning model; creating at least one bifurcation for at least one of the predicted next locations having a confidence that exceeds a bifurcation threshold; and iteratively determining a plurality of next locations of the user device using the at least one predicted next location having a confidence exceeding the bifurcation threshold as input.

Description

Technical Field

[0001] The exemplary and non-limiting exemplary embodiments relate generally to communications, and more particularly, to confidence-based advanced trajectory prediction. Background Art

[0002] It is known to determine the location of a user equipment in a wireless communication network. Summary of the invention

[0003] According to one aspect, the device includes: at least one processor; and at least one memory, including computer program code; wherein the at least one memory and the computer program code are configured to, together with the at least one processor, cause the device to at least: receive a sequence of past service locations of a user device as input to a machine learning model; determine a confidence vector for a predicted next location for the user device using the machine learning model; create at least one fork for at least one predicted next location among the predicted next locations having a confidence level exceeding a fork threshold; and iteratively determine multiple next locations of the user device using at least one predicted next location having a confidence level exceeding the fork threshold as input.

[0004] According to one aspect, the apparatus comprises: at least one processor; and at least one memory, including computer program code; wherein the at least one memory and the computer program code are configured to, together with the at least one processor, cause the apparatus to at least: receive a handover request from a network node; receive, through the handover request from the network node, a predicted confidence that a user equipment follows a trajectory; and determine whether to perform at least one of admission control, contention-free random access, or resource allocation based on the received prediction confidence and one or more predicted trajectory forks.

[0005] According to one aspect, an apparatus comprises: at least one processor; and at least one memory, including computer program code; wherein the at least one memory and the computer program code are configured to, together with the at least one processor, enable the apparatus to at least: form a trajectory prediction information element, the trajectory prediction information element being configured to store information for predicting a trajectory of a user device in a set of at least two candidate trajectories; wherein at least one predicted next position having a confidence exceeding a bifurcation threshold is used as an input to a machine learning model, and at least two candidate trajectories are iteratively determined by creating at least one bifurcation for at least one predicted next position having a confidence exceeding the bifurcation threshold; form a trajectory information element, the trajectory information element being configured to store information related to the trajectory of the user device; wherein the trajectory prediction information element is composed of at least one of the trajectory information elements; and form a trajectory node information element, the trajectory node information element representing at least one position in the trajectory of the user device.

[0006] According to one aspect, the method includes: receiving a sequence of past service locations of a user device as input to a machine learning model; determining a confidence vector for a predicted next location for the user device using the machine learning model; creating at least one fork for at least one of the predicted next locations having a confidence exceeding a fork threshold; and iteratively determining multiple next locations of the user device using at least one predicted next location having a confidence exceeding the fork threshold as input.

[0007] According to one aspect, the method includes: receiving a handover request from a network node; receiving, via the handover request from the network node, a predicted confidence that a user equipment follows a trajectory; and determining whether to perform at least one of admission control, contention-free random access, or resource allocation based on the received prediction confidence and one or more predicted trajectory forks.

[0008] According to one aspect, a method includes: forming a trajectory prediction information element, the trajectory prediction information element being configured to store information for predicting a trajectory of a user device in a set of at least two candidate trajectories; wherein at least one predicted next position having a confidence exceeding a bifurcation threshold is used as an input to a machine learning model, and at least two candidate trajectories are iteratively determined by creating at least one bifurcation for at least one predicted next position having a confidence exceeding the bifurcation threshold; forming a trajectory information element, the trajectory information element being configured to store information related to the trajectory of the user device; wherein the trajectory prediction information element is composed of at least one of the trajectory information elements; and forming a trajectory node information element, the trajectory node information element representing at least one position in the trajectory of the user device.

[0009] According to one aspect, the apparatus includes: a component for receiving a sequence of past service locations of a user device as input to a machine learning model; a component for determining a confidence vector for a predicted next location for the user device using the machine learning model; a component for creating at least one fork for at least one predicted next location among the predicted next locations having a confidence exceeding a fork threshold; and a component for iteratively determining multiple next locations of the user device using at least one predicted next location having a confidence exceeding the fork threshold as input.

[0010] According to one aspect, the apparatus includes: a component for receiving a handover request from a network node; a component for receiving a prediction confidence that a user equipment follows a trajectory through the handover request from the network node; and a component for determining whether to perform at least one of admission control, contention-free random access, or resource allocation based on the received prediction confidence and one or more predicted trajectory forks.

[0011] According to one aspect, an apparatus includes: a single component for forming a trajectory prediction information element, the trajectory prediction information element being configured to store information for predicting a trajectory of a user device in a set of at least two candidate trajectories; wherein at least one predicted next position having a confidence exceeding a bifurcation threshold is used as an input to a machine learning model, and at least two candidate trajectories are iteratively determined by creating at least one bifurcation for at least one predicted next position having a confidence exceeding the bifurcation threshold; a component for forming a trajectory information element, the trajectory information element being configured to store information related to the trajectory of the user device; wherein the trajectory prediction information element is composed of at least one of the trajectory information elements; and a component for forming a trajectory node information element, the trajectory node information element representing at least one position in the trajectory of the user device.

[0012] According to one aspect, a machine-readable non-volatile program storage device is described and provided, the non-volatile program storage device tangibly embodying an instruction program executable with a machine to perform operations, the operations including: receiving a sequence of past service locations of a user device as input to a machine learning model; determining a confidence vector for a predicted next location for the user device using the machine learning model; creating at least one fork for at least one predicted next location among the predicted next locations having a confidence exceeding a fork threshold; and iteratively determining multiple next locations of the user device using at least one predicted next location having a confidence exceeding the fork threshold as input.

[0013] According to one aspect, a machine-readable non-transitory program storage device is described and provided, the non-transitory program storage device tangibly embodying an instruction program executable with the machine to perform operations, the operations comprising: receiving a handover request from a network node; receiving, via the handover request from the network node, a predicted confidence that a user device follows a trajectory; and determining whether to perform at least one of admission control, contention-free random access, or resource allocation based on the received prediction confidence and one or more predicted trajectory forks.

[0014] According to one aspect, a machine-readable non-transitory program storage device is described and provided, the non-transitory program storage device tangibly embodying an instruction program executable with a machine to perform operations, the operations comprising: forming a trajectory prediction information element, the trajectory prediction information element being configured to store information for predicting a trajectory of a user device in a set of at least two candidate trajectories; wherein at least one predicted next position having a confidence exceeding a bifurcation threshold is used as an input to a machine learning model, and at least two candidate trajectories are iteratively determined by creating at least one bifurcation for at least one predicted next position having a confidence exceeding the bifurcation threshold; forming a trajectory information element, the trajectory information element being configured to store information related to the trajectory of the user device; wherein the trajectory prediction information element is composed of at least one of the trajectory information elements; and forming a trajectory node information element, the trajectory node information element representing at least one position in the trajectory of the user device. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The foregoing aspects and other features are explained in the following description in conjunction with the accompanying drawings.

[0016] Figure 1 is a block diagram of one possible and non-limiting system in which example embodiments may be practiced.

[0017] Figure 2 An example of UE trajectory prediction is shown.

[0018] Figure 3 An example beam / transmission prediction method is shown.

[0019] Figure 4 The difficulty of predicting the exact trajectory of a UE is shown.

[0020] Figure 5 The prediction accuracy is shown if the K most likely beams are predicted and one of them is correct.

[0021] Figure 6 The reliability of the prediction is shown.

[0022] Figure 7 The beam index prediction is shown.

[0023] Figure 8 Creating a bifurcated trajectory prediction for beam index prediction is shown.

[0024] Fig. 9 An example of presenting a more complex scenario with bifurcations and overall confidence is shown.

[0025] Fig.10 Table 3 shows the new parameters of the trajectory prediction function

[0026] Fig.11Information elements for bifurcation trajectory prediction are shown as a class diagram.

[0027] Fig.12 Table 4, or properties of an exemplary trajectory prediction IE, is shown.

[0028] Fig.13 Table 5, or attributes of an example trace IE, is shown.

[0029] Fig.14 A tree node representing a position in a trajectory is shown.

[0030] Fig.15 Shows part of the 3GPP TS38.300 Fig. 9 .2.3.4.2-1 Use the examples described in this article in CHO preparation.

[0031] Fig.16 is an example apparatus configured to implement the examples described herein.

[0032] Fig.17 A representation of an example of a non-volatile storage medium is shown.

[0033] Fig.18 is a method of performing the examples described herein.

[0034] Fig.19 is a method of performing the examples described herein.

[0035] Fig. 20 is a method of performing the examples described herein. DETAILED DESCRIPTION

[0036] Steering Figure 1 , which shows a block diagram of one possible and non-limiting example in which the examples may be practiced. A user equipment (UE) 110, a radio access network (RAN) node 170 and (multiple) network elements 190 are illustrated. Figure 1In the example of , a user equipment (UE) 110 is in wireless communication with a wireless network 100. A UE is a wireless device that can access the wireless network 100. The UE 110 includes one or more processors 120, one or more memories 125, and one or more transceivers 130 interconnected by one or more buses 127. Each of the one or more transceivers 130 includes a receiver (Rx) 132 and a transmitter (Tx) 133. The one or more buses 127 can be an address bus, a data bus, or a control bus, and can include any interconnection mechanism, such as a series of wires on a motherboard or integrated circuit, an optical fiber or other optical communication device, etc. The one or more transceivers 130 are connected to one or more antennas 128. The one or more memories 125 include computer program code 123. The UE 110 includes a module 140, which includes one or both of the parts 140-1 and / or 140-2 that can be implemented in a variety of ways. The module 140 can be implemented in hardware as the module 140-1, such as being implemented as part of the one or more processors 120. Module 140-1 may also be implemented as an integrated circuit or by other hardware such as a programmable gate array. In another example, module 140 may be implemented as module 140-2, which is implemented as computer program code 123 and executed by one or more processors 120. For example, one or more memories 125 and computer program code 123 may be configured to perform one or more operations as described herein with one or more processors 120 using user equipment 110. UE 110 communicates with RAN node 170 via wireless link 111.

[0037] In this example, RAN node 170 is a base station that provides access to wireless network 100 for wireless devices such as UE 110. RAN node 170 may be, for example, a base station for 5G, also known as New Radio (NR). In 5G, RAN node 170 may be a NG-RAN node, which is defined as a gNB or ng-eNB. A gNB is a node that provides NR user plane and control plane protocol terminations to a UE and is connected to a 5GC (such as, for example, (multiple) network elements 190) via an NG interface (such as connection 131). An ng-eNB is a node that provides E-UTRA user plane and control plane protocol terminations to a UE and is connected to a 5GC via an NG interface (such as connection 131). An NG-RAN node may include multiple gNBs, which may also include a central unit (CU) (g NB-CU) 196 and a distributed unit (DU) (g NB-DU), where DU 195 is shown. Note that DU 195 may include or be coupled to and control a radio unit (RU). The gNB-CU 196 is a logical node that hosts the radio resource control (RRC), SDAP, and PDCP protocols of a gNB, or the RRC and PDCP protocols of an en-gNB that controls the operation of one or more gNB-DUs. The gNB-CU 196 terminates the F1 interface connected to the gNB-DU 195. The F1 interface is shown as reference numeral 198, although reference numeral 198 also shows links between remote elements of the RAN node 170 and centralized elements of the RAN node 170, such as the link between the gNB-CU 196 and the gNB-DU 195. The gNB-DU 195 is a logical node that hosts the RLC, MAC, and PHY layers of a gNB or en-gNB, and its operation is partially controlled by the gNB-CU 196. One gNB-CU 196 supports one or more cells. One cell can be supported by one gNB-DU 195, or one cell can be supported / shared by multiple DUs under RAN sharing. The gNB-DU 195 terminates the F1 interface 198 connected to the gNB-CU 196. Note that the DU 195 is considered to include the transceiver 160, for example as part of the RU, but some examples thereof may have the transceiver 160 as part of a separate RU, for example under the control of the DU 195 and connected to the DU 195. The RAN node 170 may also be an eNB (evolved NodeB) base station for LTE (Long Term Evolution) or any other suitable base station or node.

[0038] The RAN node 170 includes one or more processors 152, one or more memories 155, one or more network interfaces ((multiple) N / WI / F) 161, and one or more transceivers 160 interconnected by one or more buses 157. Each of the one or more transceivers 160 includes a receiver (Rx) 162 and a transmitter (Tx) 163. The one or more transceivers 160 are connected to one or more antennas 158. The one or more memories 155 include computer program code 153. The CU 196 may include (multiple) processors 152, (multiple) memories 155, and network interfaces 161. Note that the DU 195 may also contain its own memory / storage and (multiple) processors and / or other hardware, but these are not shown.

[0039] The RAN node 170 includes a module 150, which includes one or both of the parts 150-1 and / or 150-2 that can be implemented in a variety of ways. The module 150 can be implemented in hardware as the module 150-1, such as being implemented as part of one or more processors 152. The module 150-1 can also be implemented as an integrated circuit or by other hardware such as a programmable gate array. In another example, the module 150 can be implemented as a module 150-2, which is implemented as a computer program code 153 and executed by one or more processors 152. For example, one or more memories 155 and the computer program code 153 are configured to, together with the one or more processors 152, cause the RAN node 170 to perform one or more of the operations described herein. Note that the functionality of the module 150 can be distributed, such as distributed between the DU 195 and the CU 196, or implemented only in the DU 195.

[0040] One or more network interfaces 161 communicate over a network, such as via links 176 and 131. Two or more gNBs 170 may communicate using, for example, link 176. Link 176 may be wired or wireless or both, and may implement, for example, an Xn interface for 5G, an X2 interface for LTE, or other suitable interfaces for other standards.

[0041] The one or more buses 157 may be an address bus, a data bus, or a control bus, and may include any interconnection mechanism, such as a series of wires on a motherboard or integrated circuit, an optical fiber or other optical communication device, a wireless channel, etc. For example, the one or more transceivers 160 may be implemented as a remote radio head (RRH) 195 for LTE or a distributed unit (DU) 195 for a gNB implementation for 5G, where other elements of the RAN node 170 may be physically located at a different location from the RRH / DU 195, and the one or more buses 157 may be partially implemented as, for example, a fiber optic cable or other suitable network connection to connect other elements of the RAN node 170 (e.g., a central unit (CU), gNB-CU 196) to the RRH / DU 195. Reference numeral 198 also indicates those suitable network links (multiple).

[0042] The RAN node / gNB may include one or more TRPs to which the methods described herein may be applied. Figure 1 The RAN node 170 is shown to include two TRPs, TRP 51 and TRP 52. The RAN node 170 may host or include Figure 1 Other TRPs not shown in FIG. TRP 51 and TRP 52 may form part of the components of transceiver 160 .

[0043] The relay nodes in NR are called integrated access and backhaul nodes. The mobile terminal part of the IAB node supports backhaul (parent link) connections. The mobile terminal part is the function that carries the UE functions. The distributed unit part of the IAB node supports the so-called access link (sub-link) connections (i.e., for access link UE in the case of multi-hop IAB, and for backhaul of other IAB nodes). The distributed unit part is responsible for certain base station functions. The IAB scenario can follow a split architecture, where the central unit hosts the higher layer protocols to the UE and terminates the control plane and user plane interfaces to the 5G core network.

[0044] Note that the description herein indicates that a "cell" performs a function, but it should be clear that the devices that form the cell can perform the function. A cell constitutes part of a base station. That is, each base station can have multiple cells. For example, there may be three cells for a single carrier frequency and associated bandwidth, each covering one-third of a 360-degree area, so that the coverage area of ​​a single base station covers an approximate ellipse or circle. In addition, each cell may correspond to a single carrier, and a base station may use multiple carriers. Therefore, if there are three 120-degree cells per carrier and there are two carriers, the base station has a total of 6 cells.

[0045] The wireless network 100 may include one or more network elements 190, which may include core network functions, and the core network functions provide connection to another network (such as a telephone network and / or a data communication network (e.g., the Internet)) via a link or link 181. Such core network functions for 5G may include location management functions ((multiple) LMFs) and / or access and mobility management functions (AMF(S)) and / or user plane functions ((multiple) UPFs) and / or (multiple) session management functions ((multiple) SMFs). Such core network functions for LTE may include MME (mobility management entity) / SGW (serving gateway) functions. Such core network functions may include SON (self-organizing / optimizing network) functions. These are merely example functions that may be supported by (multiple) network elements 190, and both 5G and LTE functions may be supported. The RAN node 170 is coupled to the network element 190 via a link 131. Link 131 may be implemented as, for example, an NG interface for 5G, or an S1 interface for LTE, or other suitable interfaces for other standards. The network element 190 includes one or more processors 175, one or more memories 171, and one or more network interfaces (N / WI / F) 180 interconnected by one or more buses 185. The one or more memories 171 include computer program code 173. The computer program code 173 may include SON and / or MRO functionality 172.

[0046] One or more network elements 190 include module 177, which may include near real-time RIC functionality. Computer program code 173 may include near real-time RIC functionality. Module 150-1 and / or module 150-2 may include near real-time RIC functionality.

[0047] The wireless network 100 may implement network virtualization, which is the process of combining hardware and software network resources and network functions into a single software-based management entity (virtual network). Network virtualization involves platform virtualization that is often combined with resource virtualization. Network virtualization is classified as external virtualization that combines many networks or parts of networks into virtual units, or internal virtualization that provides network-like functions to software containers on a single system. Note that the virtualized entities resulting from network virtualization are still implemented to some extent using hardware such as processors 152 or 175 and memories 155 and 171, and such virtualized entities also produce technical effects.

[0048] Computer readable memories 125, 155, and 171 may be of any type suitable for the local technical environment and may be implemented using any suitable data storage technology, such as semiconductor-based memory devices, flash memory, magnetic memory devices and systems, optical memory devices and systems, non-volatile memory, transient memory, fixed memory, and removable memory. Computer readable memories 125, 155, and 171 may be components for performing storage functions. As non-limiting examples, processors 120, 152, and 175 may be of any type suitable for the local technical environment and may include one or more of a general-purpose computer, a special-purpose computer, a microprocessor, a digital signal processor (DSP), and a processor based on a multi-core processor architecture. Processors 120, 152, and 175 may be components for performing functions such as controlling UE 110, RAN node 170, (multiple) network elements 190, and other functions described herein.

[0049] In general, various example embodiments of user equipment 110 may include, but are not limited to, a cellular phone, such as a smart phone, a tablet computer, a personal digital assistant (PDA) with wireless communication capabilities, a portable computer with wireless communication capabilities, an image capture device such as a digital camera with wireless communication capabilities, a gaming device with wireless communication capabilities, a music storage and playback device with wireless communication capabilities, an Internet device that allows wireless Internet access and browsing, a tablet device with wireless communication capabilities, a head-mounted display such as those that implement virtual / augmented / mixed reality, and a portable unit or terminal incorporating a combination of such functions. UE 110 may also be a vehicle such as a car, or a UE installed in a vehicle, a UAV such as, for example, a drone, or a UE installed in a UAV.

[0050] UE 110, RAN node 170 and / or network element(s) 190 (and associated memory, computer program code and modules) may be configured to implement (e.g., in part) the methods described herein, including confidence-based advanced trajectory prediction. Thus, computer program code 123 of UE 110, modules 140-1, 140-2 and Figure 1 Other elements / features shown in the figure may implement the user equipment related aspects of the methods described herein. Computer program code 153 of RAN node 170, module 150-1, module 150-2 and Figure 1 Other elements / features shown in the figure may implement the gNB / TRP related aspects of the methods described herein. Computer program code 173 of (multiple) network element 190 and Figure 1 Other elements / features shown in may be configured to implement network element related aspects of the methods described herein.

[0051] Having thus introduced a suitable but non-limiting technical context for practicing example embodiments, example embodiments are now described in more detail.

[0052] The examples described herein relate to 5G New Radio, and in particular to the Multi-RAT Mobility (MRM) concept for 3GPP Rel-18 and beyond, which enables improved mobility / cell change / CHO preparation in the RAN through trajectory prediction of the UE. The base station improves the selection of candidate beams / cells for CHO to the most likely ones through trajectory prediction of the UE. The base station determines the next / second beam / cell to which the UE is most likely to connect, and reduces the number of CHO preparation procedures to those beams / cells that are most likely to connect to the UE later.

[0053] TS38.300 Section 9.2.3.4.2 provides the basis for the CHO preparation process (e.g. Fig. 9 .2.3.4.2-1, which is used in this article as Fig.15 References). Trajectory prediction use cases with planned enhancements were discussed in 3GPP TSG-RAN WG3 meeting #116-e. The CHO preparation process (including the CHO preparation process of Rel-173 GPP) can be enhanced / improved by the methods described in this article.

[0054] In order to predict the second / next beam / cell to which the UE is most likely to connect, the source base station determines a confidence level for a set of neighboring cells and selects among those beams / cells with the highest scores (i.e., above a threshold), and performs a trajectory prediction determination for each selected cell / beam using the same input including those neighboring beams / cells with high scores to obtain a trajectory prediction for the UE including the second / next beam / cell.

[0055] UE trajectory prediction

[0056] UE trajectory prediction is discussed in 3GPP RAN3 with the intention of adding it to 3GPP TR 37.817 “Study on Enhancement for Data Collection for NR and EN-DC”. Refer to 3GPP TSG-RAN WG3 meeting #114bis-e, R3-21xxx.

[0057] The UE trajectory prediction may include the latitude, longitude, altitude and / or cell ID of the UE in the future time period. Additionally, the beam ID may be added. The beam ID may be a factor in optimizing early data forwarding, especially for CHO. In order to produce the UE trajectory prediction in the output, the NG-RAN node needs to train an ML model. Reference Figure 2 , model inference of the trajectory prediction ML algorithm may require the following information as input: a) the observed UE' trajectory T1 until the UE reaches point A, T1 may include a list of access cells (eg 210, 220, 230) on which the UE is camped in idle mode or to which the UE is connected, and b) radio measurements reported by the UE or performed by the network.

[0058] The output of model inference corresponds to Figure 2 T2p in FIG. 1 represents the trajectory-related prediction information. UE 110 is predicted to access cells 240, 250, and 260. One or more of cells 210, 220, 230, 240, 250, and 260 are hosted by gNB 170, gNB 170-2, and gNB 170-3.

[0059] L1 / L2 beam prediction and predictive L3 mobility

[0060] In 2021 China Nanjing Wireless Communications and Networking Conference (WCNC) (“Deep Learning”) Kaya, Harish Viswanathan, in "Deep Learning-Based Predictive Beam Management for 5G Millimeter Wave Systems", shows a method for accurately predicting the best serving beam and transmission point in advance as the user moves through the network, and thus eliminating the need for frequent measurement reports. The prediction method applies deep learning techniques similar to those used in natural language processing (NLP) for translation / sentence completion tasks (i.e., long short-term memory (LSTM) recurrent neural networks (RNNs)) to the problem of predicting the best serving beam and probability. The prediction strategy can be applied to both L1 beam prediction and handover prediction based on L3 RSRP. The input to the neural network is the beam that served the UE in the past at equidistant time intervals, and the output is a trajectory of beam probabilities for equidistant time intervals for a future time window Δt. Figure 3 An overview of the method is shown.

[0061] refer to Figure 3 At 302, the RSRP measurement is saved. Figure 2As shown, L1-RSRP beam measurements for different timestamps of UEs (including UE1 and UE2) are saved. At 304, the measurement entry for UE k is extracted from the memory. At 306, a past beam sequence for UE k is generated for the past Δt. The past beam index 308 is provided to the LSTM encoder 310. The LSTM encoder 310 provides an internal LSTM state 312 to the LSTM decoder 314 and provides an output elsewhere in the system (313). The LSTM decoder 314 predicts the future beam index 316 and probability. At 318, the last predicted beam index and probability are re-injected into the list 320 of future beam indices and their probabilities.

[0062] Serving beam index prediction is also a type of trajectory prediction. The predicted next beam may not belong to the current serving cell, in which case an inter-cell handover or conditional handover (CHO) may be prepared.

[0063] consider Figure 4 , where highway 214 bifurcates at fork 215 in a Y-intersection. Although a statistical prediction can be made based on the percentage of UEs traveling on highway 214 in which direction they travel, this may not be predicted for a single UE 110. Giving a wrong prediction may be worse than having no prediction at all, so predicting the most common trajectory may not be a solution for beam management, inter-cell mobility, load balancing optimization, etc.

[0064] As an example, the trajectory of the UE 110 may change over time, such as due to intersections on a highway, so relying on a single UE trajectory prediction may result in higher outage time and inefficient use of radio resources.

[0065] If it is assumed that UE 110 follows the trajectory prediction Figure 4 However, UE 110 follows the predicted trajectory T2p Figure 4 In alternative 2, a) only contention-free random access resources (CFRA) may be used to prepare cells on the path of the T2p trajectory to reduce the interruption time of the UE, and since the trajectory prediction of UE 110 is wrong, this may lead to inefficient use of limited CFRA resources at the cell, and b) cells on the path of the alternative 2 path are not informed about the UE trajectory prediction (when the prediction outputs the trajectory T2p), so they do not prepare in advance (e.g., prepare CFRA resources) for possible handover procedures of the UE (e.g., to cells 270 and 280). This may result in the UE using only CBRA resources for handover, and thus the interruption time may be increased.

[0066] Research referring to "Deep Learning" and WO2020 / 214168 indicates that the reliability of the prediction increases if more than one beam is monitored (which may belong to the same or different access points) and the network is prepared in advance for beam swapping and switching for those beams. Figure 5 The prediction accuracy is shown if the network is prepared with the most likely K beams. Curve 502 corresponds to the beams and curve 504 corresponds to the TRP.

[0067] exist Figure 6 In , the top 2 beam prediction reliability is shown for all samples with similar best beam prediction probability. Reliability is defined as the percentage of all samples where one of the top K beams is correctly predicted. Figure 6 It is shown that if the best beam prediction probability exceeds 0.5, the reliability can be improved to above the 90% target.

[0068] Therefore, although it may be difficult to predict Figure 4 10 in the image, but it is possible to predict that the UE takes either of the two trajectories, which is important information. The trajectory prediction concepts used before the methods described herein do not enable such predictions and do not present a solution on how such predictions can be made.

[0069] refer to Figure 2 and Figure 3 The beam prediction methods and trajectory prediction presented do not cover the prediction of multiple possible trajectories alone. Examples of this are given in Figure 7 Shown in. Figure 7 A Long Short-Term Memory (LSTM) Recurrent Neural Network (RNN) model is shown for predicting the next beam index using a sequence of past serving beam indices (702) as input. In this use case based on the methods in "Deep Learning" and WO2020 / 214168, trajectories are expressed and predicted as serving beam index probabilities. For simplicity, a simplified implementation of the methods in "Deep Learning" and WO2020 / 214168 is presented, omitting the splitting of the encoder and decoder. However, this does not change the underlying logic.

[0070] This scenario covers a model covering 6 beams with index AF.The input to the model 704 is a sequence 702 of past serving beam indices for a UE (eg, UE 110), and the output is a confidence vector (706, 712) that each beam in the plurality of beams will be the next serving beam.

[0071] To predict ( 708 ) multiple steps ahead, the predicted probability of the previous step ( 706 ) is fed back into the LSTM 704 as the last input element of the updated serving beam ID sequence 710 .

[0072] exist Figure 7 In the case shown, in the first prediction (706), both beams C and D have almost the same confidence in becoming the next serving beam. This may correspond to Figure 4 The method still selects the next potential beam with the highest confidence (beam D with confidence 0.49 in vector 706 and beam E with confidence 0.8 in vector 712) at (708, 714) and feeds it into LSTM 704 to predict the next step in advance. As shown, selection 714 leads to prediction 716.

[0073] Although the model can predict the subsequent steps (708, 714) with high confidence, the overall predicted future sequence still has relatively low confidence as the prediction window becomes longer because in one of the prediction steps, the model does not make a decision between two alternatives if there is a possibility of bifurcation. Existing trajectory prediction methods do not explicitly predict alternative trajectories to achieve an acceptable level of confidence and robustness of the prediction in such scenarios.

[0074] This paper introduces a method for creating several "forks" in trajectory prediction when several trajectories have similar or significantly high confidence or likelihood. Figure 4 In the case shown, predictions can be made with meaningful confidence when the most likely trajectory is included in the prediction rather than just one trajectory.

[0075] In addition, this paper describes a confidence-driven approach to predictions, where a confidence requirement can be used to configure a tradeoff between the confidence of the prediction and the number of prediction trajectories and the length of the prediction window. High confidence can be achieved with only more than one prediction trajectory. Conversely, the fewer trajectories predicted, the more powerful the prediction.

[0076] This article describes the improvements made by adding Figure 7 Instead of a single trajectory, the trajectory prediction may be a set of possible future trajectories. The trajectory is predicted as a time series of quantized positions, later referred to as "positions". The quanta may relate to, for example, a serving beam ID, a serving cell ID, or UE coordinates quantized into pixels.

[0077] A bifurcation criterion can be defined such that at each prediction step, where multiple possibilities satisfy the bifurcation criterion, e.g., the likelihood of a position being the next position compared to a confidence threshold, the prediction step is bifurcated such that each predicted next position with a confidence above the threshold is fed back into the model, thereby creating its own "fork".

[0078] The definitions in this paper are two confidence measures, trajectory confidence C T and bifurcation confidence CF .

[0079] Trajectory confidence C T is the model confidence that the UE will travel a given single predicted trajectory. A possible example implementation could be: where c i is the prediction confidence that the UE will move from the i-th position in the trajectory to the next one, and N is the prediction window length.

[0080] Bifurcation confidence C F Indicates the confidence or probability that the UE will follow any trajectory, including all (subsequent) forks. A possible example implementation could be: C F =∑ f∈F C T (f), where F is the set of all trajectories in the prediction (considering bifurcations), C T (f) is the trajectory confidence of trajectory bifurcation f. The bifurcation confidence evaluated over the entire prediction is the prediction confidence.

[0081] Additional bifurcation conditions may be defined. For example, a bifurcation may be created only when a bifurcation threshold is exceeded and the trajectory confidence of the trajectory (bifurcation) is above an additional bifurcation threshold. This may be used to prevent additional bifurcations when the confidence that the UE is about to reach a bifurcation point is not high enough. A prediction window threshold may be defined to configure the prediction window length according to the achievable confidence. In this case, future trajectories (bifurcations) are predicted only if the trajectory confidence remains above a trajectory length threshold. It may be used to ensure that only trajectory predictions with sufficient confidence are made.

[0082] Finally, the bifurcation criterion can be adjusted and the algorithm iterated until a satisfactory prediction confidence is reached. Figure 8 , Fig. 9 , Fig.10 , Fig.11 , Fig.12 , Fig.13 , Fig.14 and Fig.15 In the presented solution, forking can be done and configured entirely at inference time, and there is no need to retrain the trajectory prediction model when the forking criterion changes.

[0083] Based on the observed beam index, a fork in the prediction can be eliminated without creating a new prediction. For example, a fork can be used to select in preparation for CHO or for optimizing the contention-free random access (CFRA) procedure. The same approach can also be used when predicting the cell ID instead of the serving beam.

[0084] If predicting coordinates, the locations need to be expressed as "pixels" covering a certain area predicted to be able to utilize the confidence threshold based bifurcation method.

[0085] Figure 8 Shows the use of Figure 7 804 . 806 . 807 . 808 . 809 . 100 . 110 . 120 . 130 . 140 . 150 . 160 . 170 . 180 . 190 . 200 . 210 . 220 . 230 . 240 . 190 . 250 . 180 . 190 . 260 . 180 . 190 . 270 . 180 . 190 . 280 . 180 . 190 . 290 . 300 . 310 . 320 . 330 . 340 . 350 . 360 . 370 . 380 . 390 . 400 . 410 . 420 . 430 . 440 . 450 . 460 . 470 . 480 . 490 . 500 . 510 . 520 . 530 . 540 . 550 However, no other beam has a prediction confidence above the bifurcation threshold, so no further bifurcation is performed. Model 804 outputs a confidence vector 814 for a given serving beam ID sequence 810, and model 804 outputs a confidence vector 816 for a given serving beam ID sequence 812. This example produces the following confidence values ​​shown in Table 1. Table 1 Trajectory and bifurcation confidence in the example

[0086] Bifurcation is only done if more than one possible future position (or beam) has a sufficiently high confidence in the prediction (i.e. exceeds the bifurcation threshold). In this case, one of the predicted beams has such a high confidence (0.8) that no other beam has a confidence above the threshold, and no further bifurcation is done.

[0087] It can be noted that although the model is predicting that the UE is more likely to follow fork 2 813 (because it has a higher confidence at the bifurcation point due to D), fork 1 811 still has a higher trajectory confidence because the model has a much higher confidence in predicting the next beam in that trajectory, i.e., if the UE 110 follows the trajectory in fork 1. The fork confidence of approximately 80% indicates that forks 1 and 2 are the most relevant (the top 2 trajectories), which means a high reliability of the prediction that the UE will eventually take either of these paths.

[0088] Fig. 9The same example with an additional beam GJ is extended and by predicting two additional steps into the future after time T 903. It demonstrates the potential cyclic nature of the bifurcation. In this example, the earlier bifurcation 1 811 bifurcates into bifurcation 1.1 901 and bifurcation 1.2 902, while the original bifurcation 2 813 does not bifurcate further. The prediction confidence for the next step is shown in the lower right corner. The predicted trajectories and confidence values ​​for the bifurcations are shown in Table 2. Table 2 for Fig. 9 The confidence value of the example in

[0089] In the above example, the last step of Fork 2 may have another beam that will also be above the Fork Threshold, but since the trajectory confidence at that point is already below the Additional Fork Threshold, no new fork is created. In this way, unnecessary forks are avoided in locations where the confidence that the UE 100 will reach the predicted beam is not high enough to justify the fork. Alternatively, if a prediction window threshold is configured, Fork 2 may only predict 3 beams into the future to ensure sufficient confidence.

[0090] The idea of ​​forking is to use it when the model cannot determine which trajectory the UE is going to take. If the model can, then the model only provides one trajectory, otherwise the model gives the forking option to the consumer of the prediction. The consumer may use the machine learning model to determine a first trajectory confidence that the user device follows the first trajectory using one of the at least one second confidence vectors, use the machine learning model to determine a second trajectory confidence that the user device follows the second trajectory using another of the at least one second confidence vectors, compare the first trajectory confidence to the second trajectory confidence, and if the consumer so chooses, predict the next position of the user device associated with the larger of the first trajectory confidence and the second trajectory confidence, i.e., "flattening" the prediction again by selecting the trajectory with the highest confidence (which is the one closest to the consumer). A higher trajectory confidence does not necessarily mean that the UE is more likely to take the fork, as it may simply be because if the UE takes the fork, the next step after the fork can be predicted with more confidence for the fork, but this does not mean that the fork is more likely to be taken.

[0091] Information Elements (IE) for Bifurcation Trajectory Prediction

[0092] Fig.10 Table 3 shown in shows the configurable parameters to be added to the trajectory prediction function.

[0093] Information elements used to present trajectory predictions

[0094] Fig.11Depicts an example IE in which a trajectory prediction function (TrajectoryPredictionFunction) may be used to provide diverging trajectory predictions. Fig.11 It is shown as a class diagram 1100 using the Unified Modeling Language (UML). A trajectory prediction IE 1110 is shown, which includes attributes prediction confidence 1111, trajectory list 1112, and trajectory tree 1113. A trajectory IE 1120 includes attributes position ID list 1121 and trajectory confidence 1122. A trajectory node IE 1130 includes position ID 1131, previous 1132, next 1133, and next confidence list 1134.

[0095] like Fig.11 As further shown, the trajectory prediction information element 1110 is composed of trajectory information elements 1120. There are one or more trajectory list information elements 1112. The trajectory node information element 1130 is composed of zero to one previous trajectory node information element 1130. There are zero or more next elements 1133, and zero to one trajectory tree 1113.

[0096] Trajectory Prediction IE

[0097] Trajectory prediction IE 1110 is an IE containing a complete trajectory prediction. It may include a list of trajectory IEs 1112, which represents all possible trajectories taking into account bifurcations. In addition, it may include a tree data structure 1113 of trajectory nodes 1130 for representing a complete predicted trajectory including bifurcations. Fig.12 Table 4 shown in lists the properties of an exemplary trajectory prediction IE 1110.

[0098] Track IE

[0099] An IE representing a single trajectory, trajectory IE 1120 , is described herein. Fig.13 Table 5 shown in lists the properties of an example trace IE 1120.

[0100] Track Node IE

[0101] exist Fig.14 A tree node representing a location in a trajectory, such as international node IE 1130, is shown in FIG.

[0102] Deployment and signaling, example technical effects

[0103] Depending on where the trajectory prediction algorithm is located for inference, the following two approaches can be considered, namely 1) an approach for inference at the gNB and 2) an approach for inference at the near real-time RIC.

[0104] 1. For inference at the gNB (e.g., RAN node 170), the UE's trajectory prediction list can be used to support decision mechanisms at the source gNB and the target gNB. The source gNB can use the predicted trajectory divergence to optimize, for example, its mobility (e.g., CHO preparation), load balancing, or energy saving decisions. Fig.15 , source gNB 170 may share the predicted trajectory information including the bifurcation confidence level with target gNB 170-2 via a handover request message. Target gNB 170-2 may use this information for cell preparation. For example, target gNB 170-2 prepares a cell for UE 110 on bifurcation 1 only if the confidence level for bifurcation 1 is above a threshold (set by target gNB 170-2 itself).

[0105] 2. Reasoning at near real-time RIC (eg, module 150 - 1 , module 150 - 2 , module 177 , or computer program code 173 ), prediction can be used in many use cases in near-real-time RIC, such as energy saving, mobility optimization, or load balancing.

[0106] In the solution described herein, the forking is done at the inference stage and no specific training is required. The trajectory prediction LSTM RNN may be trained, for example, in the OAM of one or more network elements 190 .

[0107] Fig.15 An example of using predicted trajectory divergence to optimize CHO preparation is shown 1000. Compared to message 3GPP TS 38.300, new or updated steps are given as item 1502 (trajectory prediction), item 1503 (CHO decision), item 1504-1 (handover request), item 1504-2 (handover request), item 1505-1 (admission control) and item 1505-2 (admission control).

[0108] The trajectory prediction 1502 can be triggered by an existing event (e.g., A3) or a new event. The predicted trajectory can include several forks. The predicted trajectory forks can be used to determine which target cells to configure for CHO at 1503. In this example 100, cells in two target gNBs (one of 170-2, 170-3) are predicted as potential target cells and prepared for CHO. The predicted trajectory can be shared to the target gNB (170-2, 170-3) in the handover request (1504-1, 1504-2) using the trajectory prediction IE 1110. The shared trajectory can be only the trajectory related to that gNB. The target gNB (170-2, 170-3) can use the predicted trajectory to optimize its admission control (1505-1, 1505-2), contention-free random access (CFRA) process, resource allocation, etc.

[0109] therefore, Fig.15 A signaling diagram 1000 for implementing confidence-based advanced trajectory prediction based on the examples described herein is shown. The signaling diagram 1500 includes a handover preparation 1550, a handover execution 1552, and a handover completion 1554. The handover preparation 1550 includes items 1562, 1564, 1500, 1501, 1502, 1503, 1504-1, 1504-2, 1505-1, 1505-2, 1506-1, 1506-2, 1507, and 1508. The handover execution 1552 includes items 1508-a, 1566, 1568, 1570, and 1509. The handover completion 1554 includes items 1509-a, 1509-b, 1572, 1509-3, and 1574.

[0110] At 1562, user data is exchanged between the UE 110 and the source gNB 170. At 1564, user data is exchanged between the source gNB 170 and the UPF 190-2. At 1500, the AMF 190-1 provides mobility control information to the source gNB 170, the target gNB 170-2, and one or more other potential (multiple) target gNBs 170-3. At 1501, measurement control and reporting are performed by the UE 110 and the source gNB 170.

[0111] At 1502, the source gNB 170 performs trajectory prediction. At 1503, the source gNB 170 makes a CHO decision, which may be based in part on the trajectory prediction 1502. At 1504-1, the source gNB 170 sends a handover request to the target gNB 170-2. At 1504-2, the source gNB 170 sends a handover request to one or more other potential (multiple) target gNBs 170-3. At 1505-1, the target gNB 170-2 performs admission control, and at 1505-2, one or more of the other potential target gNBs 170-3 perform admission control.

[0112] At 1506-1, target gNB 170-2 sends a handover request acknowledgment to source gNB 170, and at 1506-2, one or more other potential (multiple) target gNBs 170-3 send a handover request acknowledgment to source gNB 170. At 1507, source gNB 170 sends an RRC reconfiguration message to UE 110, and at 1508, UE 110 sends an RRC reconfiguration complete message to source gNB 170.

[0113] At 1508-a, source gNB 170 sends an Early State Transfer message to other potential (multiple) target gNB 170-3. At 1566, UE 110 evaluates CHO conditions. At 1568, UE 110 detaches from the old cell and synchronizes to the new cell. At 1570, user data is sent from UPF 190-2 to source gNB 170, and then from source gNB 170 to other potential (multiple) target gNB 170-3. At 1509, UE 110, source gNB 170, and target gNB 170-2 perform CHO Handover Complete.

[0114] At 1509-a, the target gNB 170-2 sends a Handover Success message to the source gNB 170, and at 1509-b, the source gNB 170 sends an SN State Transfer message to the target gNB 170-2. At 1572, user data is sent from the one or more UPFs 190-2 to the source gNB 170, and then the user data is sent from the source gNB 170 to the target gNB 170-2. At 1509-3, the source gNB 170 sends a Handover Cancel message to the target gNB 170-2 and the other potential (multiple) target gNBs 170-3. At 1574, the UE 110, the source gNB 170, the target gNB 170-2, the one or more other potential (multiple) target gNBs 170-3, the AMF 190-1, and the one or more UPFs 190-2 perform 3GPP TS 38.300. Fig. 9 .2.3.2.1-1 steps 9-12.

[0115] The examples described herein have several technical effects and advantages. Using the solutions described herein, trajectory prediction can better capture UE mobility behavior, where there may typically be several optional trajectories that the UE can take, but predicting which one may be impossible. A vehicle UE at a fork is such an example. Trajectory prediction that predicts only one trajectory may give wrong predictions, which may have a significant impact on the performance of functions that utilize the prediction. On the other hand, it can be predicted that the UE takes one of the possible trajectories, which is very important information. Therefore, a method for making such a forked trajectory prediction and a method for presenting a forked trajectory prediction and a configuration of information elements are described herein.

[0116] Another advantage of the examples described herein is that in a use case it is possible to exploit multiple possible and probable future trajectories, for example to reserve resources, which are committed or released depending on whether the UE takes that path. Such examples are conditional handover and beam selection optimization.

[0117] The examples described herein affect the X2 / Xn interfaces, including the interfaces described within the TR37.817, TS36.423, and TS38.423 specifications through the indicated standards-affected and discussed IEs.

[0118] Fig.16 An example apparatus 1600 may be implemented in hardware configured to implement the examples described herein. The apparatus 1600 includes at least one processor 1602 (e.g., an FPGA and / or a CPU), at least one memory 1604 including computer program code 1605, wherein the at least one memory 1604 and the computer program code 1605 are configured to, together with the at least one processor 1602, cause the apparatus 1600 to implement circuits, processes, components, modules, or functions (collectively referred to as controls 1606) to implement the examples described herein, including confidence-based advanced trajectory prediction. The memory 1604 may be a non-transitory memory, a transient memory, a volatile memory (e.g., a RAM), or a non-volatile memory (e.g., a ROM).

[0119] The device 1600 optionally includes a display and / or I / O interface 1608, which can be used to display various aspects or states of the methods described herein (e.g., when one of the methods is being performed or at a subsequent time), or to receive input from a user, such as using a keypad, camera, touch screen, touch area, microphone, biometrics, etc. The (multiple) communication I / F 1610 can be wired and / or wireless and communicate over the Internet / other (multiple) networks via any communication technology. The (multiple) communication I / F 1610 can include one or more transmitters and one or more receivers. The (multiple) communication I / F 1610 may include standard well-known components such as amplifiers, filters, frequency converters, (de)modulators, and encoder / decoder circuit systems and one or more antennas.

[0120] The apparatus 1600 for implementing the functionality of the control 1606 may be a UE 110, a RAN node 170 (e.g., a gNB), or a network element 190. Thus, the processor 1602 may correspond to the processor 120, the processor 152, and / or the processor 175, the memory 1604 may correspond to the memory 125, the memory 155, and / or the memory 171, the computer program code 1605 may correspond to the computer program code 123, the module 140-1, the module 140-2, and / or the computer program code 153, the module 150-1, the module 150-2, and / or the computer program code 173, and the (multiple) communication I / Fs may correspond to the transceiver 130, the (multiple) antenna 128, the transceiver 160, the (multiple) antenna 158, the (multiple) N / WI / F 161, and / or the (multiple) N / WI / F 180. Alternatively, apparatus 1600 may not correspond to any of UE 110, RAN node 170, or network element(s) 190, as apparatus 1600 may be part of a Self-Organizing / Optimizing Network (SON) node, such as in the cloud.

[0121] The apparatus 1600 may also be distributed throughout the network (eg 100 ), including within and between the apparatus 1600 and any network elements (such as a network control element (NCE) 190 and / or a RAN node 170 and / or a UE 110 ).

[0122] like Fig.16 As shown, interface 1612 enables data communication between various items of device 1600. For example, interface 1612 can be one or more buses, such as an address bus, a data bus, or a control bus, and can include any interconnection mechanism, such as a series of lines on a motherboard or integrated circuit, an optical fiber or other optical communication device, etc. Computer program code 1605 including control 1606 can include object-oriented software configured to pass data / messages between objects within computer program code 1605. Device 1600 need not include each of the features mentioned, or may also include other features.

[0123] Fig.17 Schematic representations of non-volatile memory media 1700a (e.g., a computer disk (CD) or digital versatile disk (DVD)) and 1700b (e.g., a universal serial bus (USB) memory stick) storing instructions and / or parameters 1702 that, when executed by a processor, enable the processor to perform one or more steps of the previously described method.

[0124] It should be noted that the example embodiments may be implemented as circuit systems, software, hardware, application logic, or a combination of software, hardware, and application logic. In the example embodiments, the application logic, software, or instruction set is maintained on any computer-readable medium. In the context of this document, a "computer-readable medium" may be any medium or component that can contain, store, communicate, propagate, or transmit instructions for use by or in conjunction with an instruction execution system, apparatus, or device (such as a base station or user equipment of the above example embodiments).

[0125] Fig.18 is an example method 1800 for implementing the example embodiments described herein. At 1810, the method includes receiving a sequence of past service locations of a user device as an input to a machine learning model. At 1820, the method includes determining a confidence vector for a predicted next location for the user device using the machine learning model. At 1830, the method includes creating at least one fork for at least one of the predicted next locations having a confidence exceeding a fork threshold. At 1840, the method includes iteratively determining a plurality of next locations for the user device using at least one predicted next location having a confidence exceeding the fork threshold as input. The method 1800 may be performed using a network node (e.g., a RAN node 170, a source gNB 170).

[0126] Fig.19 An example method 1900 for implementing an example embodiment described herein is provided. At 1910, the method includes receiving a handover request from a network node. At 1920, the method includes receiving a predicted confidence that a user equipment follows a trajectory via the handover request from the network node. At 1930, the method includes determining whether to perform at least one of admission control, contention-free random access, or resource allocation based on the received prediction confidence and one or more predicted trajectory forks. The method 1900 may be performed using a network node (e.g., a RAN node 170, a target node 170-2, or other potential (multiple) target gNBs 170-3).

[0127] Fig. 20An example method 2000 for implementing an example embodiment described herein is provided. At 2010, the method includes forming a trajectory prediction information element configured to store information for predicting a trajectory of a user device in a set of at least two candidate trajectories. At 2020, the method includes using at least one predicted next position with a confidence exceeding a fork threshold as an input to a machine learning model, and iteratively determining at least two candidate trajectories by creating at least one fork for at least one predicted next position with a confidence exceeding the fork threshold. At 2030, the method includes forming a trajectory information element configured to store information related to a trajectory of the user device. At 2040, the method includes wherein the trajectory prediction information element is composed of at least one trajectory information element in the trajectory information element. At 2050, the method includes forming a trajectory node information element, the trajectory node information element representing at least one location in the trajectory of the user device. The method 2000 may be performed using a network node (e.g., RAN node 170, source gNB 170). The following examples (1-30) are provided and described herein.

[0128] Example 1. A device comprising: at least one processor; and at least one memory comprising computer program code; wherein the at least one memory and the computer program code are configured to, together with the at least one processor, cause the device to at least: receive a sequence of past service locations of a user device as input to a machine learning model; determine a confidence vector for a predicted next location for the user device using the machine learning model; create at least one fork for at least one predicted next location among the predicted next locations having a confidence level exceeding a fork threshold; and iteratively determine multiple next locations of the user device using at least one predicted next location having a confidence level exceeding the fork threshold as input.

[0129] Example 2. An apparatus according to Example 1, wherein at least one memory and computer program code are configured to, together with at least one processor, cause the apparatus to at least: determine, using a machine learning model, at least one predicted next location having a confidence level exceeding a bifurcation threshold as input, at least one second confidence vector for a candidate next location for a user device; wherein multiple next locations are determined using at least one second confidence vector for a candidate next location for the user device as input.

[0130] Example 3. An apparatus according to Example 2, wherein the sequence of past positions, predicted next positions, and candidate next positions include a serving beam, a serving cell, a location pixel, or a location quantum.

[0131] Example 4. An apparatus according to any one of Examples 2 to 3, wherein at least one memory and computer program code are configured to, together with at least one processor, cause the apparatus to at least: predict at least one next location for a user device among a plurality of next locations based in part on at least one second confidence vector.

[0132] Example 5. An apparatus according to Example 1, wherein a first one of the predicted next positions is determined to include a first confidence, and wherein a second one of the predicted next positions is determined to include a second confidence; wherein using the first one of the predicted next positions as input, a first one of the candidate next positions is determined to include a third confidence; wherein using the second one of the predicted next positions as input, a second one of the candidate next positions is determined to include a fourth confidence; wherein a first trajectory confidence is determined by multiplying the first confidence and the third confidence, and a second trajectory confidence is determined by multiplying the second confidence and the fourth confidence.

[0133] Example 6. The apparatus of Example 5, wherein the at least one memory and the computer program code are configured to, together with the at least one processor, cause the apparatus to at least: determine a bifurcation confidence that the user device follows the first trajectory or the second trajectory based on the first trajectory confidence and the second trajectory confidence.

[0134] Example 7. An apparatus according to any one of Examples 1 to 6, wherein at least one memory and computer program code are configured to, together with at least one processor, cause the apparatus to at least: determine at least one confidence corresponding to at least one candidate next position in at least one second confidence vector for the candidate next position; compare at least one confidence corresponding to at least one candidate next position with a bifurcation threshold; and in response to at least one confidence corresponding to at least one candidate next position exceeding the bifurcation threshold, create at least one bifurcation for at least one candidate next position.

[0135] Example 8. An apparatus according to Example 7, wherein at least one memory and computer program code are configured to, together with at least one processor, cause the apparatus to at least: determine, using a machine learning model, a first trajectory confidence that a user device follows a first trajectory based on one of at least one fork for at least one candidate next location; and determine, using a machine learning model, a second trajectory confidence that the user device follows a second trajectory based on another of at least one fork for at least one candidate next location.

[0136] Example 9. An apparatus according to Example 8, wherein at least one memory and computer program code are configured to, together with at least one processor, cause the apparatus to: determine a first bifurcation confidence that the user device follows the first trajectory using at least the first trajectory confidence; and determine a second bifurcation confidence that the user device follows the second trajectory using at least the second trajectory confidence.

[0137] Example 10. The apparatus of Example 9, wherein the at least one memory and the computer program code are configured to, with the at least one processor, cause the apparatus to: determine a cumulative bifurcation confidence that the user device follows the first trajectory or the second trajectory using the first bifurcation confidence and the second bifurcation confidence.

[0138] Example 11. An apparatus according to any one of Examples 2 to 10, wherein at least one memory and computer program code are configured to cause the apparatus, together with at least one processor, to: use a machine learning model to determine a trajectory confidence that a user device follows a trajectory using one of at least one second confidence vector; compare the trajectory confidence with a second bifurcation threshold; in response to the trajectory confidence being higher than the second bifurcation threshold, determine to create an additional bifurcation for one of the candidate positions; and in response to the trajectory confidence being lower than the second bifurcation threshold, determine not to create an additional bifurcation for one of the candidate positions.

[0139] Example 12. An apparatus according to any one of Examples 1 to 11, wherein at least one memory and computer program code are configured to, together with at least one processor, cause the apparatus to: determine to create at least one fork for one of the candidate next positions in response to the number of forks created being less than a prediction window threshold; and determine not to create at least one fork for one of the candidate next positions in response to the number of forks created being greater than or equal to the prediction window threshold.

[0140] Example 13. An apparatus according to any one of Examples 1 to 12, wherein the machine learning model includes a long short-term memory recurrent neural network.

[0141] Example 14. An apparatus according to any one of Examples 2 to 13, wherein at least one memory and computer program code are configured to, together with at least one processor, cause the apparatus to at least: predict a sequence of one or more possible next service beams for a user device based at least in part on a confidence vector or a second confidence vector; and send a request to prepare for a conditional switch to a network node hosting the predicted potential next service beam in the sequence of the one or more next service beams.

[0142] Example 15. An apparatus according to any one of Examples 5 to 14, wherein at least one memory and computer program code are configured to, together with at least one processor, cause the apparatus to at least: send a first handover request to a first network node associated with a first candidate target cell in response to a first trajectory confidence exceeding a first threshold; and send a second handover request to a second network node associated with a second candidate target cell in response to a second trajectory confidence exceeding a second threshold; wherein the user equipment selects one of the first candidate target cell or the second candidate target cell based on at least one measurement and a configured conditional handover condition.

[0143] Example 16. A device comprising: at least one processor; and at least one memory, including computer program code; wherein the at least one memory and the computer program code are configured to, together with the at least one processor, enable the device to at least: receive a switching request from a network node; receive, through the switching request from the network node, a predicted confidence that a user device follows a trajectory; and determine whether to perform at least one of admission control, contention-free random access, or resource allocation based on the received prediction confidence and one or more predicted trajectory forks.

[0144] Example 17. An apparatus according to Example 16, wherein at least one memory and computer program code are configured to, together with at least one processor, cause the apparatus to at least: receive, via a switching request from a network node, a second predicted confidence that the user equipment follows the second trajectory; compare the predicted confidence that the user equipment follows the trajectory with the second predicted confidence that the user equipment follows the second trajectory; and based on the comparison, determine whether to perform at least one of admission control, contention-free random access, or resource allocation.

[0145] Example 18. A device comprising: at least one processor; and at least one memory comprising computer program code; wherein the at least one memory and the computer program code are configured to, together with the at least one processor, cause the device to at least: form a trajectory prediction information element, the trajectory prediction information element being configured to store information for predicting a trajectory of a user device in a set of at least two candidate trajectories; wherein at least one predicted next position having a confidence level exceeding a bifurcation threshold is used as an input to a machine learning model, and at least two candidate trajectories are iteratively determined by creating at least one bifurcation for at least one predicted next position having a confidence level exceeding the bifurcation threshold; form a trajectory information element, the trajectory information element being configured to store information related to the trajectory of the user device; wherein the trajectory prediction information element is composed of at least one of the trajectory information elements; and form a trajectory node information element, the trajectory node information element representing at least one position in the trajectory of the user device.

[0146] Example 19. The apparatus according to Example 18, wherein the trajectory prediction information element comprises: a prediction confidence of a predicted trajectory of the user device; a list of predicted trajectories; and a trajectory tree storing the prediction branches as a tree structure, the tree structure being used to predict the trajectory of the user device.

[0147] Example 20. An apparatus according to any one of Examples 18 to 19, wherein the trajectory information element comprises: a list of location identifiers in the order of the trajectory of the user equipment, wherein the location identifiers comprise beam identifiers, cell identifiers, or user equipment coordinates; and a trajectory confidence of the predicted trajectory of the user equipment.

[0148] Example 21. An apparatus according to any one of Examples 18 to 20, wherein the trajectory node information element comprises: a pointer to a previous trajectory node information element; a list of pointers to multiple trajectory nodes in a prediction of a trajectory of a user device; and a list of confidences for the multiple trajectory nodes, stored in an order corresponding to the order of the pointer list.

[0149] Example 22. A method comprising: receiving a sequence of past service locations of a user device as input to a machine learning model; determining a confidence vector for a predicted next location for the user device using the machine learning model; creating at least one fork for at least one predicted next location in the predicted next locations having a confidence exceeding a fork threshold; and iteratively determining multiple next locations for the user device using at least one predicted next location having a confidence exceeding the fork threshold as input.

[0150] Example 23. A method comprising: receiving a handover request from a network node; receiving, via the handover request from the network node, a predicted confidence that a user device follows a trajectory; and determining whether to perform at least one of admission control, contention-free random access, or resource allocation based on the received prediction confidence and one or more predicted trajectory forks.

[0151] Example 24. A method, comprising: forming a trajectory prediction information element, the trajectory prediction information element being configured to store information for predicting a trajectory of a user device in a set of at least two candidate trajectories; wherein at least one predicted next position having a confidence exceeding a bifurcation threshold is used as an input to a machine learning model, and at least two candidate trajectories are iteratively determined by creating at least one bifurcation for at least one predicted next position having a confidence exceeding the bifurcation threshold; forming a trajectory information element, the trajectory information element being configured to store information related to the trajectory of the user device; wherein the trajectory prediction information element is composed of at least one of the trajectory information elements; and forming a trajectory node information element, the trajectory node information element representing at least one position in the trajectory of the user device.

[0152] Example 25. An apparatus comprising: a component for receiving a sequence of past service locations of a user device as input to a machine learning model; a component for determining a confidence vector for a predicted next location for the user device using the machine learning model; a component for creating at least one fork for at least one of the predicted next locations having a confidence exceeding a fork threshold; and a component for iteratively determining multiple next locations of the user device using at least one predicted next location having a confidence exceeding the fork threshold as input.

[0153] Example 26. An apparatus comprising: a component for receiving a handover request from a network node; a component for receiving a predicted confidence that a user device follows a trajectory through the handover request from the network node; and a component for determining whether to perform at least one of admission control, contention-free random access, or resource allocation based on the received prediction confidence and one or more predicted trajectory forks.

[0154] Example 27. An apparatus comprising: a component for forming a trajectory prediction information element, the trajectory prediction information element being configured to store information for predicting a trajectory of a user device in a set of at least two candidate trajectories; wherein at least one predicted next position having a confidence exceeding a bifurcation threshold is used as an input to a machine learning model, and at least two candidate trajectories are iteratively determined by creating at least one bifurcation for at least one predicted next position having a confidence exceeding the bifurcation threshold; a component for forming a trajectory information element, the trajectory information element being configured to store information related to the trajectory of the user device; wherein the trajectory prediction information element is composed of at least one of the trajectory information elements; and a component for forming a trajectory node information element, the trajectory node information element representing at least one position in the trajectory of the user device.

[0155] Example 28. A machine-readable, non-transitory program storage device tangibly embodying an instruction program executable with a machine to perform operations, the operations comprising: receiving a sequence of past service locations of a user device as input to a machine learning model; determining a confidence vector for a predicted next location of the user device using the machine learning model; creating at least one fork for at least one of the predicted next locations having a confidence exceeding a fork threshold; and iteratively determining multiple next locations of the user device using at least one predicted next location having a confidence exceeding the fork threshold as input.

[0156] Example 29. A machine-readable, non-transitory program storage device tangibly embodying an instruction program executable with the machine to perform operations, the operations comprising: receiving a switching request from a network node; receiving, via the switching request from the network node, a predicted confidence that a user device follows a trajectory; and determining whether to perform at least one of admission control, contention-free random access, or resource allocation based on the received prediction confidence and one or more predicted trajectory forks.

[0157] Example 30. A machine-readable, non-transitory program storage device, tangibly embodying an instruction program executable with a machine to perform operations, the operations comprising: forming a trajectory prediction information element, the trajectory prediction information element being configured to store information for predicting a trajectory of a user device in a set of at least two candidate trajectories; wherein at least one predicted next position having a confidence exceeding a bifurcation threshold is used as an input to a machine learning model, and at least two candidate trajectories are iteratively determined by creating at least one bifurcation for at least one predicted next position having a confidence exceeding the bifurcation threshold; forming a trajectory information element, the trajectory information element being configured to store information related to the trajectory of the user device; wherein the trajectory prediction information element is composed of at least one of the trajectory information elements; and forming a trajectory node information element, the trajectory node information element representing at least one position in the trajectory of the user device.

[0158] References to "computers," "processors," and the like should be understood to include not only computers having different architectures, such as single / multi-processor architectures and sequential or parallel architectures, but also special purpose circuits such as field programmable gate arrays (FPGAs), application specific circuits (ASICs), signal processing devices, and other processing circuit systems. References to computer programs, instructions, code, and the like should be understood to include software or firmware for programmable processors, such as, for example, the programmable content of a hardware device, whether instructions for a processor, or configuration settings for a fixed function device, gate array, or programmable logic device, and the like.

[0159] The memory(s) as described herein may be implemented using any suitable data storage technology, such as semiconductor-based memory devices, flash memory, magnetic memory devices and systems, optical memory devices and systems, non-volatile memory, transient memory, fixed memory, and removable memory. The memory(s) may include a database for storing data.

[0160] As used herein, the term "circuitry" may refer to the following: (a) hardware circuit implementations, such as those found in analog and / or digital circuits, and (b) combinations of circuits and software (and / or firmware), such as, as applicable: (i) a combination of processors or (ii) portions of processors / software, including digital signal processors, software, and memory, which work together to enable the device to perform various functions; and (c) circuits, such as microprocessors or portions of microprocessors, that require software or firmware for operation, even if the software or firmware is not physically present. As another example, the term "circuitry" as used herein would also encompass implementations of only a processor (or processors) or a portion of a processor and its (or their) accompanying software and / or firmware. The term "circuitry" would also encompass, for example and if applicable to the particular element, a baseband integrated circuit or an application processor integrated circuit for a mobile phone or a similar integrated circuit in a server, cellular network device, or another network device.

[0161] In the figures, arrows between blocks indicate operational couplings therebetween and the direction of data flow over those couplings.

[0162] It should be understood that the above description is illustrative only. Various alternatives and modifications may be devised by those skilled in the art. For example, the features described in the various dependent claims may be combined with each other in any suitable (multiple) combination. In addition, features from the above-described different example embodiments may be selectively combined into new example embodiments. Therefore, this specification is intended to cover all such alternatives, modifications and variations that fall within the scope of the appended claims.

[0163] The following acronyms and abbreviations that may be found in the specification and / or drawings are defined as follows (abbreviations and acronyms may be appended to each other or appended with other characters using, for example, a dash or hyphen): 3GPP Third Generation Partnership Project 4G Fourth Generation 5G Fifth Generation 5GC 5G Core Network A3 is triggered when the neighboring cell is better offset than the serving cell. AMF Access and Mobility Management Function ASIC Application-Specific Integrated Circuit CBRA Contention-based random access C F Bifurcation Confidence CFRA Contention-free random access CHO Condition Switch Conf confidence CPU Central Processing Unit C T Trajectory confidence CU Central Unit or Centralized Unit DSP Digital Signal Processor DU Distributed Unit eNB Evolved Node B (e.g., LTE base station) EN-DC E-UTRAN New Radio - Dual Connectivity en-gNB provides NR user plane and control plane protocol termination to UE and acts as a secondary node in EN-DC E-UTRA Evolved Universal Terrestrial Radio Access, also known as LTE radio access technology E-UTRAN E-UTRA Network F1 Interface between CU and DU FPGA Field Programmable Gate Array gNB is a base station for 5G / NR, which provides NR user plane and control plane protocol terminals to UE and is connected to 5GC via NG interface. IAB Integrated Access and Backhaul ID Identifier IE Information Elements I / F Interface I / O Input / Output L1 Layer 1 L2 Layer 2 L3 Layer 3 LMF Location Management Function LSTM Long Short-Term Memory LTE Long Term Evolution (4G) M Mandatory MAC Media Access Control ML Machine Learning MME Mobility Management Entity MRM Multi-RAT Mobility MRO mobility robustness optimization NCE Network Control Element ng or NG new generation ng-eNB Next Generation eNB NG-RAN Next Generation Radio Access Network NLP Natural Language Processing NR New Radio (5G) N / W Network O Optional OAM Operations, Administration and Maintenance PDA Personal Digital Assistant PDCP Packet Data Convergence Protocol PHY Physical Layer RAM Random Access Memory RAN Radio Access Network RAN3 RAN Meeting Rel- Release- RIC RAN Intelligent Controller RLC Radio Link Control RNN Recurrent Neural Network ROM Read Only Memory RRC Radio Resource Control (protocol) RRH Remote Radio Head RSRP Reference Signal Received Power RT Real-time RU Radio Unit Rx Receiver or Receiver SDAP Service Data Adaptation Protocol SGW Service Gateway SI Research Projects SMF session management functions SN Secondary Node SON self-organizing / optimizing network T1 recorded traces T2P predicted trajectory TRP Send and / or Receive Point TS Technical Specifications TSG Technical Specification Group Tx Transmitter or Send UAV UE User Equipment (e.g., wireless, usually mobile) UML Unified Modeling Language UPF User Plane Function WG Working Group WI Work Item X2 Network interface between RAN nodes and between RAN and core network Xn Network interface between NG-RAN nodes

Claims

1. A device comprising: at least one processor; as well as at least one memory including computer program code; wherein the at least one memory and the computer program code are configured to, together with the at least one processor, cause the apparatus to at least: receiving a sequence of past service locations of a user device as input to a machine learning model; determining a confidence vector for a predicted next location of the user device using the machine learning model; creating at least one fork for at least one of the predicted next positions having a confidence level exceeding a fork threshold; as well as A plurality of next positions of the user equipment are iteratively determined using the at least one predicted next position having the confidence exceeding the bifurcation threshold as input.

2. The apparatus of claim 1 , wherein the at least one memory and the computer program code are configured to, together with the at least one processor, cause the apparatus to at least: determining, using the machine learning model, at least one second confidence vector for a candidate next location for the user device using as input the at least one predicted next location having the confidence exceeding the bifurcation threshold; The plurality of next positions are determined using the at least one second confidence vector for the candidate next position of the user equipment as input.

3. The apparatus of claim 2, wherein the sequence of past positions, the predicted next position, and the candidate next position comprise a serving beam, a serving cell, a location pixel, or a location quantum.

4. The apparatus according to any one of claims 2 to 3, wherein the at least one memory and the computer program code are configured to, together with the at least one processor, cause the apparatus to at least: At least one next location for the user equipment in the plurality of next locations is predicted based in part on the at least one second confidence vector.

5. The device according to claim 1, wherein a first one of the predicted next positions is determined to include a first confidence level, and wherein a second one of the predicted next positions is determined to include a second confidence level; wherein using the first of the predicted next positions as input, the first of the candidate next positions is determined to include a third confidence level; wherein using the second one of the predicted next positions as input, the second one of the candidate next positions is determined to include a fourth confidence level; A first trajectory confidence factor is determined by multiplying the first confidence factor and the third confidence factor, and a second trajectory confidence factor is determined by multiplying the second confidence factor and the fourth confidence factor.

6. The apparatus of claim 5, wherein the at least one memory and the computer program code are configured to, together with the at least one processor, cause the apparatus to at least: Based on the first trajectory confidence and the second trajectory confidence, a bifurcation confidence that the user equipment follows the first trajectory or the second trajectory is determined.

7. The apparatus according to any one of claims 1 to 6, wherein the at least one memory and the computer program code are configured to, together with the at least one processor, cause the apparatus to at least: determining at least one confidence corresponding to at least one candidate next position in the at least one second confidence vector for the candidate next position; comparing the at least one confidence level corresponding to the at least one candidate next position to the bifurcation threshold; as well as In response to the at least one confidence level corresponding to at least one candidate next position exceeding the bifurcation threshold, at least one bifurcation is created for the at least one candidate next position.

8. The apparatus of claim 7, wherein the at least one memory and the computer program code are configured to, with the at least one processor, cause the apparatus to at least: determining, using the machine learning model, a first trajectory confidence that the user device follows a first trajectory based on one of the at least one fork for the at least one candidate next location; and A second trajectory confidence that the user device follows a second trajectory is determined based on another one of the at least one fork for the at least one candidate next location using the machine learning model.

9. The apparatus of claim 8, wherein the at least one memory and the computer program code are configured to, together with the at least one processor, cause the apparatus to: determining a first bifurcation confidence that the user equipment follows the first trajectory using at least the first trajectory confidence; and A second bifurcation confidence that the user equipment follows the second trajectory is determined using at least the second trajectory confidence.

10. The apparatus of claim 9, wherein the at least one memory and the computer program code are configured to, together with the at least one processor, cause the apparatus to: The first bifurcation confidence and the second bifurcation confidence are used to determine a cumulative bifurcation confidence that the user equipment follows the first trajectory or the second trajectory.

11. The apparatus according to any one of claims 2 to 10, wherein the at least one memory and the computer program code are configured to, together with the at least one processor, cause the apparatus to: using the machine learning model, determining a trajectory confidence that the user device is following the trajectory using one of the at least one second confidence vectors; comparing the trajectory confidence with a second bifurcation threshold; In response to the trajectory confidence being higher than the second bifurcation threshold, determining to create an additional bifurcation for one of the candidate positions; as well as In response to the trajectory confidence being lower than the second fork threshold, it is determined not to create the additional fork for the one of the candidate positions.

12. The apparatus according to any one of claims 1 to 11, wherein the at least one memory and the computer program code are configured to, together with the at least one processor, cause the apparatus to: In response to the number of created forks being less than a prediction window threshold, determining to create at least one fork for one of the candidate next positions; and In response to the number of created forks being greater than or equal to the prediction window threshold, it is determined not to create the at least one fork for the one of the candidate next positions.

13. An apparatus according to any one of claims 1 to 12, wherein the machine learning model comprises a long short-term memory recurrent neural network.

14. The apparatus according to any one of claims 2 to 13, wherein the at least one memory and the computer program code are configured to, together with the at least one processor, cause the apparatus to at least: predicting a sequence of one or more possible next serving beams for the user equipment based at least in part on the confidence vector or the second confidence vector; and A request to prepare for a conditional handover is sent to a network node hosting a predicted potential next serving beam in the sequence of the one or more next serving beams.

15. The apparatus according to any one of claims 5 to 14, wherein the at least one memory and the computer program code are configured to, together with the at least one processor, cause the apparatus to at least: In response to the first trajectory confidence exceeding a first threshold, sending a first handover request to a first network node associated with a first candidate target cell; and In response to the second trajectory confidence exceeding a second threshold, sending a second handover request to the second network node associated with a second candidate target cell; The user equipment selects one of the first candidate target cell or the second candidate target cell based on at least one measurement and a configured conditional handover condition.

16. An apparatus comprising: at least one processor; as well as at least one memory including computer program code; wherein the at least one memory and the computer program code are configured to, together with the at least one processor, cause the apparatus to at least: receiving a switching request from a network node; receiving, via the handover request from the network node, a predicted confidence that the user equipment follows the trajectory; as well as Based on the received prediction confidence and one or more predicted trajectory divergences, it is determined whether to perform at least one of admission control, contention-free random access, or resource allocation.

17. The apparatus of claim 16, wherein the at least one memory and the computer program code are configured to, with the at least one processor, cause the apparatus to at least: receiving, via the handover request from the network node, a second prediction confidence that the user equipment follows a second trajectory; comparing the prediction confidence that the user device follows the trajectory with the second prediction confidence that the user device follows the second trajectory; as well as Based on the comparison, it is determined whether to perform at least one of admission control, contention-free random access, or resource allocation.

18. An apparatus comprising: at least one processor; as well as at least one memory including computer program code; wherein the at least one memory and the computer program code are configured to, together with the at least one processor, cause the apparatus to at least: forming a trajectory prediction information element configured to store information for predicting a trajectory of a user device in a set of at least two candidate trajectories; wherein at least one predicted next position having a confidence exceeding a bifurcation threshold is used as an input to the machine learning model, and the at least two candidate trajectories are iteratively determined by creating at least one bifurcation for the at least one predicted next position having a confidence exceeding the bifurcation threshold; forming a trajectory information element, the trajectory information element being configured to store information related to the trajectory of the user device; wherein the trajectory prediction information element is composed of at least one of the trajectory information elements; and A trajectory node information element is formed, the trajectory node information element representing at least one location in the trajectory of the user equipment.

19. The apparatus according to claim 18, wherein the trajectory prediction information element comprises: A prediction confidence of the predicted trajectory of the user equipment; a list of predicted trajectories; as well as The predicted branches are stored as a trajectory tree in a tree structure, where the tree structure is used to predict the trajectory of the user equipment.

20. The apparatus according to any one of claims 18 to 19, wherein the trajectory information element comprises: a list of location identifiers in order of the trajectory of the user equipment, wherein the location identifiers comprise beam identifiers, cell identifiers or user equipment coordinates; as well as A trajectory confidence of the predicted trajectory of the user equipment.

21. The apparatus according to any one of claims 18 to 20, wherein the trajectory node information element comprises: Pointer to the previous trajectory node information element; a list of pointers to a plurality of trajectory nodes in the prediction of the trajectory of the user equipment; as well as The confidence lists for the plurality of trajectory nodes are stored in an order corresponding to the order of the pointer lists.

22. A method comprising: receiving a sequence of past service locations of a user device as input to a machine learning model; determining a confidence vector for a predicted next location of the user device using the machine learning model; creating at least one fork for at least one of the predicted next positions having a confidence level exceeding a fork threshold; as well as A plurality of next positions of the user equipment are iteratively determined using the at least one predicted next position having the confidence exceeding the bifurcation threshold as input.

23. A method comprising: receiving a switching request from a network node; receiving, via the handover request from the network node, a predicted confidence that the user equipment follows the trajectory; as well as Based on the received prediction confidence and one or more predicted trajectory divergences, it is determined whether to perform at least one of admission control, contention-free random access, or resource allocation.

24. A method comprising: forming a trajectory prediction information element configured to store information for predicting a trajectory of a user device in a set of at least two candidate trajectories; wherein at least one predicted next position having a confidence exceeding a bifurcation threshold is used as an input to the machine learning model, and the at least two candidate trajectories are iteratively determined by creating at least one bifurcation for the at least one predicted next position having a confidence exceeding the bifurcation threshold; forming a trajectory information element, the trajectory information element being configured to store information related to the trajectory of the user device; wherein the trajectory prediction information element is composed of at least one of the trajectory information elements; and A trajectory node information element is formed, the trajectory node information element representing at least one location in the trajectory of the user equipment.

25. An apparatus comprising: means for receiving a sequence of past service locations of a user device as input to a machine learning model; means for determining a confidence vector for a predicted next location of the user device using a machine learning model; means for creating at least one bifurcation for at least one of the predicted next positions having a confidence level exceeding a bifurcation threshold; as well as Means for iteratively determining a plurality of next locations of the user equipment using the at least one predicted next location having the confidence exceeding the bifurcation threshold as input.

26. An apparatus comprising: means for receiving a handover request from a network node; means for receiving, via said handover request from said network node, a predicted confidence that a user equipment is following a trajectory; as well as Means for determining whether to perform at least one of admission control, contention-free random access, or resource allocation based on the received prediction confidence and one or more predicted trajectory divergences.

27. An apparatus comprising: means for forming a trajectory prediction information element, the trajectory prediction information element being configured to store information for predicting a trajectory of a user device in a set of at least two candidate trajectories; wherein at least one predicted next position having a confidence exceeding a bifurcation threshold is used as an input to the machine learning model, and the at least two candidate trajectories are iteratively determined by creating at least one bifurcation for the at least one predicted next position having a confidence exceeding the bifurcation threshold; means for forming a trajectory information element, the trajectory information element being configured to store information related to the trajectory of the user equipment; wherein the trajectory prediction information element is composed of at least one of the trajectory information elements; and Means for forming a trajectory node information element representing at least one position in the trajectory of the user equipment.

28. A non-transitory program storage device readable by a machine, tangibly embodying a program of instructions executable with the machine to perform operations comprising: receiving a sequence of past service locations of a user device as input to a machine learning model; determining a confidence vector for a predicted next location of the user device using the machine learning model; creating at least one fork for at least one of the predicted next positions having a confidence level exceeding a fork threshold; as well as A plurality of next positions of the user equipment are iteratively determined using the at least one predicted next position having the confidence exceeding the bifurcation threshold as input.

29. A non-transitory program storage device readable by a machine, tangibly embodying a program of instructions executable with the machine to perform operations comprising: receiving a switching request from a network node; receiving, via the handover request from the network node, a predicted confidence that the user equipment follows the trajectory; as well as Based on the received prediction confidence and one or more predicted trajectory divergences, it is determined whether to perform at least one of admission control, contention-free random access, or resource allocation.

30. A non-transitory program storage device readable by a machine, tangibly embodying a program of instructions executable with the machine to perform operations comprising: forming a trajectory prediction information element configured to store information for predicting a trajectory of a user device in a set of at least two candidate trajectories; wherein at least one predicted next position having a confidence exceeding a bifurcation threshold is used as an input to the machine learning model, and the at least two candidate trajectories are iteratively determined by creating at least one bifurcation for the at least one predicted next position having a confidence exceeding the bifurcation threshold; forming a trajectory information element, the trajectory information element being configured to store information related to the trajectory of the user device; wherein the trajectory prediction information element is composed of at least one of the trajectory information elements; and A trajectory node information element is formed, the trajectory node information element representing at least one location in the trajectory of the user equipment.

Images