Methods, communications devices, and infrastructure equipment for beam quality prediction using ai / ML models

Abstract

A method of operating a communications device configured to transmit signals to and / or to receive signals from an infrastructure equipment via a channel between the communications device and the infrastructure equipment, the infrastructure equipment forming part of a wireless communications network is provided. The method comprises receiving, from the infrastructure equipment, an indication of a configuration in accordance with which the communications device is to perform one or more measurements, and performing the measurements in accordance with the indicated configuration. Here, the performed measurements are to be used as inputs into an artificial intelligence, AI, model as part of a training procedure, an inference procedure, or a monitoring procedure, the AI model being for use by the communications device in performing a prediction task.

Description

[0001] METHODS, COMMUNICATIONS DEVICES, AND INFRASTRUCTURE EQUIPMENT

[0002] BACKGROUND Field of Disclosure

[0003] The present disclosure relates to wireless communications, and particularly to communications devices, infrastructure equipment and methods for the transmission and / or reception of data by a communications device in a wireless communications network.

[0004] The present application claims the Paris Convention priority from European patent application number EP24204787.6, filed on 4 October 2024, the contents of which are hereby incorporated by reference.

[0005] Description of Related Art

[0006] The “background” description provided herein is for the purpose of generally presenting the context of the disclosure. Work of the presently named inventors, to the extent it is described in this background section, as well as aspects of the description which may not otherwise qualify as prior art at the time of filing, are neither expressly or impliedly admitted as prior art against the present invention.

[0007] Previous generation mobile telecommunication systems, such as those based on the 3GPP defined UMTS and Long Term Evolution (LTE) architecture, are able to support a wider range of services than simple voice and messaging services offered by previous generations of mobile telecommunication systems. For example, with the improved radio interface and enhanced data rates provided by LTE systems, a user is able to enjoy high data rate applications such as mobile video streaming and mobile video conferencing that would previously only have been available via a fixed line data connection. The demand to deploy such networks is therefore strong and the coverage area of these networks, i.e. geographic locations where access to the networks is possible, is expected to continue to increase rapidly.

[0008] Current and future wireless communications networks are expected to routinely and efficiently support communications with an ever-increasing range of devices associated with a wider range of data traffic profiles and types than existing systems are optimised to support. For example, it is expected future wireless communications networks will be expected to efficiently support communications with devices including reduced complexity devices, machine type communication (MTC) devices, high resolution video displays, virtual reality headsets, Extended Reality (XR) and so on. Some of these different types of devices may be deployed in very large numbers, for example low complexity devices for supporting the “The Internet of Things”, and may typically be associated with the transmissions of relatively small amounts of data with relatively high latency tolerance. Other types of device, for example supporting high-definition video streaming, may be associated with transmissions of relatively large amounts of data with relatively low latency tolerance. Other types of device, for example used for autonomous vehicle communications and for other critical applications, may be characterised by data that should be transmitted through the network with low latency and high reliability. A single device type might also be associated with different traffic profiles / characteristics depending on the application(s) it is running. For example, different considerations may apply for efficiently supporting data exchange with a smartphone when it is running a video streaming application (high downlink data) as compared to when it is running an Internet browsing application (sporadic uplink and downlink data) or being used for voice communications by an emergency responder in an emergency scenario (data subject to stringent reliability and latency requirements).

[0009] In view of this there is expected to be a desire for current wireless communications networks, for example those which may be referred to as 5G or new radio (NR) systems / new radio access technology (RAT) systems, or indeed future 6G wireless communications, as well as future iterations / releases of existing systems, to efficiently support connectivity for a wide range of devices associated with different applications and different characteristic data traffic profiles and requirements.

[0010] One example of a new service is referred to as Ultra Reliable Low Latency Communications (URLLC) services which, as its name suggests, requires that a data unit or packet be communicated with a high reliability and with a low communications delay. Another example of a new service is extended Reality (XR), which may be provided by various user equipment such as wearable devices. XR combines real- world and virtual environments, incorporating aspects such as augmented reality (AR), mixed reality (MR), and virtual reality (VR), and thus requires high quality and minimised interaction delay. Services such as URLLC and XR therefore represent a challenging example for both LTE type communications systems and 5G / NR communications systems, as well as future generation communications systems.

[0011] 5G NR has continuously evolved and the current work plan includes 5G-NR-advanced in which some further enhancements are expected, especially to support new use-cases / scenarios with higher requirements. One of these aspects includes the use of artificial intelligence (Al) or machine learning (ML) to model parameters which can be used to perform certain tasks. The desire to support these new use-cases and scenarios gives rise to new challenges for efficiently handling communications in wireless communications systems that need to be addressed.

[0012] SUMMARY OF THE DISCLOSURE

[0013] The present disclosure can help address or mitigate at least some of the issues discussed above.

[0014] Embodiments of the present technique can provide a method of operating a communications device configured to transmit signals to and / or to receive signals from an infrastructure equipment via a channel between the communications device and the infrastructure equipment, the infrastructure equipment forming part of a wireless communications network. The method comprises receiving, from the infrastructure equipment, an indication of a configuration in accordance with which the communications device is to perform one or more measurements, and performing the measurements in accordance with the indicated configuration. Here, the performed measurements are to be used as inputs into an artificial intelligence, Al, model as part of a training procedure, an inference procedure, or a monitoring procedure, the Al model being for use by the communications device in performing a prediction task.

[0015] Such embodiments of the present technique, which, in addition to methods of operating communications devices, relate to methods of operating infrastructure equipment of wireless communications networks, to such communications devices and infrastructure equipment, to circuitry for such communications devices and infrastructure equipment, to wireless communications systems, to computer programs, and to computer-readable storage mediums, can allow for the more efficient and effective performance of training and / or maintenance procedures of AI / ML models deployed at communications devices.

[0016] Respective aspects and features of the present disclosure are defined in the appended claims.

[0017] It is to be understood that both the foregoing general description and the following detailed description are exemplary, but are not restrictive, of the present technology. The described embodiments, together with further advantages, will be best understood by reference to the following detailed description taken in conjunction with the accompanying drawings.

[0018] BRIEF DESCRIPTION OF THE DRAWINGS

[0019] A more complete appreciation of the disclosure and many of the attendant advantages thereof will be readily obtained as the same becomes better understood by reference to the following detailed description when considered in connection with the accompanying drawings wherein like reference numerals designate identical or corresponding parts throughout the several views, and wherein:

[0020] Figure 1 schematically represents some aspects of an LTE-type wireless telecommunication system which may be configured to operate in accordance with certain embodiments of the present disclosure;

[0021] Figure 2 schematically represents some aspects of a new radio access technology (RAT) wireless telecommunications system which may be configured to operate in accordance with certain embodiments of the present disclosure;

[0022] Figure 3 is a schematic block diagram of an example infrastructure equipment and communications device which may be configured to operate in accordance with certain embodiments of the present disclosure;

[0023] Figure 4 illustrates an example of resource allocation process for Physical Downlink Shared Channels (PDSCH);

[0024] Figure 5 illustrates an example of resource allocation process for Physical Uplink Shared Channels (PUSCH);

[0025] Figure 6 schematically illustrates a life cycle management (LCM) architecture for an artificial intelligence (Al) model;

[0026] Figure 7 schematically illustrates an example of a training process for a reinforcement learning based AI / ML model;

[0027] Figure 8 illustrates an example of spatial beam prediction at a user equipment (UE);

[0028] Figure 9 illustrates an example of the time occurrence of synchronisation signal blocks (SSBs) pertaining to downlink beams;

[0029] Figure 10 shows a part schematic, part message flow diagram representation of an example process of communications in a communications system in accordance with embodiments of the present technique; and

[0030] Figure 11 shows a flow diagram illustrating an example process of communications in a communications system in accordance with embodiments of the present technique.

[0031] DETAILED DESCRIPTION OF THE EMBODIMENTS

[0032] Long Term Evolution Advanced Radio Access Technology (4G)

[0033] Figure 1 provides a schematic diagram illustrating some basic functionality of a mobile telecommunications network / system 6 operating generally in accordance with LTE principles, but which may also support other radio access technologies, and which may be adapted to implement embodiments of the disclosure as described herein. Various elements of Figure 1 and certain aspects of their respective modes of operation are well-known and defined in the relevant standards administered by the 3GPP (RTM) body, and also described in many books on the subject, for example, Holma H. and Toskala A [1], It will be appreciated that operational aspects of the telecommunications networks discussed herein which are not specifically described (for example in relation to specific communication protocols and physical channels for communicating between different elements) may be implemented in accordance with any known techniques, for example according to the relevant standards and known proposed modifications and additions to the relevant standards.

[0034] The network 6 includes a plurality of base stations 1 connected to a core network 2. Each base station provides a coverage area 3 (i.e. a cell) within which data can be communicated to and from communications devices 4. Although each base station 1 is shown in Figure 1 as a single entity, the skilled person will appreciate that some of the functions of the base station may be carried out by disparate, inter-connected elements, such as antennas (or antennae), remote radio heads, amplifiers, etc. Collectively, one or more base stations may form a radio access network. Data is transmitted from base stations 1 to communications devices 4 within their respective coverage areas 3 via a radio downlink (DL). Data is transmitted from communications devices 4 to the base stations 1 via a radio uplink (UL). The core network 2 routes data to and from the communications devices 4 via the respective base stations 1 and provides functions such as authentication, mobility management, charging and so on. Communications devices may also be referred to as mobile stations, user equipment (UEs), user terminals, mobile radios, mobile terminals, terminal devices, wireless transmit and receive units (WTRUs), and so forth. Services provided by the core network 2 may include connectivity to the internet or to external telephony services. The core network 2 may further track the location of the communications devices 4 so that it can efficiently contact (i.e. page) the communications devices 4 for transmitting downlink data towards the communications devices 4.

[0035] Base stations, which are an example of network infrastructure equipment, may also be referred to as transceiver stations, nodeBs, e-nodeBs, eNB, g-nodeBs, gNB and so forth. In this regard different terminology is often associated with different generations of wireless telecommunications systems for elements providing broadly comparable functionality. However, certain embodiments of the disclosure may be equally implemented in different generations of wireless telecommunications systems, and for simplicity certain terminology may be used regardless of the underlying network architecture. That is to say, the use of a specific term in relation to certain example implementations is not intended to indicate these implementations are limited to a certain generation of network that may be most associated with that particular terminology.

[0036] New Radio Access Technology (5G)

[0037] Systems incorporating NR technology are expected to support different services (or types of services), which may be characterised by different requirements for latency, data rate and / or reliability. For example, Enhanced Mobile Broadband (eMBB) services are characterised by high capacity with a requirement to support up to 20 Gb / s. The requirements for Ultra Reliable and Low Latency Communications (URLLC) services are for one transmission of a 32 byte packet to be transmitted from the radio protocol layer 2 / 3 SDU ingress point to the radio protocol layer 2 / 3 SDU egress point of the radio interface within 1 ms with a reliability of 1 - 10'5(99.999 %) or higher (99.9999%) [2],

[0038] Massive Machine Type Communications (mMTC) is another example of a service which may be supported by NR-based communications networks. In addition, systems may be expected to support further enhancements related to Industrial Internet of Things (IIoT) in order to support services with new requirements of high availability, high reliability, low latency, and in some cases, high-accuracy positioning.

[0039] An example configuration of a wireless communications network which uses some of the terminology proposed for and used in NR and 5G is shown in Figure 2. In Figure 2 a plurality of transmission and reception points (TRPs) 10 are connected to distributed control units (DUs) 41, 42 by a connection interface represented as a line 16. Each of the TRPs 10 is arranged to transmit and receive signals via a wireless access interface within a radio frequency bandwidth available to the wireless communications network. Thus, within a range for performing radio communications via the wireless access interface, each of the TRPs 10, forms a cell of the wireless communications network as represented by a circle 12. As such, wireless communications devices 14 which are within a radio communications range provided by the cells 12 can transmit and receive signals to and from the TRPs 10 via the wireless access interface. Each of the distributed units 41, 42 are connected to a central unit (CU) 40 (which may be referred to as a controlling node) via an interface 46. The central unit 40 is then connected to the core network 20 which may contain all other functions required to transmit data for communicating to and from the wireless communications devices and the core network 20 may be connected to other networks 25. The elements of the wireless access network shown in Figure 2 may operate in a similar way to corresponding elements of an LTE network as described with regard to the example of Figure 1. It will be appreciated that operational aspects of the telecommunications network represented in Figure 2, and of other networks discussed herein in accordance with embodiments of the disclosure, which are not specifically described (for example in relation to specific communication protocols and physical channels for communicating between different elements) may be implemented in accordance with any known techniques, for example according to currently used approaches for implementing such operational aspects of wireless telecommunications systems, e.g. in accordance with the relevant standards.

[0040] The TRPs 10 of Figure 2 may in part have a corresponding functionality to a base station or eNodeB of an LTE network or gNodeB of an NR network. Similarly, the communications devices 14 may have a functionality corresponding to the UE devices 4 known for operation with an LTE network. It will be appreciated therefore that operational aspects of a new RAT network (for example in relation to specific communication protocols and physical channels for communicating between different elements) may be different to those known from LTE or other known mobile telecommunications standards. However, it will also be appreciated that each of the core network component, base stations and communications devices of a new RAT network will be functionally similar to, respectively, the core network component, base stations and communications devices of an LTE wireless communications network.

[0041] In terms of broad top-level functionality, the core network 20 connected to the new RAT telecommunications system represented in Figure 2 may be broadly considered to correspond with the core network 2 represented in Figure 1, and the respective central units 40 and their associated distributed units / TRPs 10 may be broadly considered to provide functionality corresponding to the base stations 1 of Figure 1. The term network infrastructure equipment / access node may be used to encompass these elements and more conventional base station type elements of wireless telecommunications systems. Depending on the application at hand the responsibility for scheduling transmissions which are scheduled on the radio interface between the respective distributed units and the communications devices may lie with the controlling node / central unit and / or the distributed units / TRPs. A communications device 14 is represented in Figure 2 within the coverage area of the first communication cell 12. This communications device 14 may thus exchange signalling with the first central unit 40 in the first communication cell 12 via one of the distributed units / TRPs 10 associated with the first communication cell 12.

[0042] It will further be appreciated that Figure 2 represents merely one example of a proposed architecture for a new RAT based telecommunications system in which approaches in accordance with the principles described herein may be adopted, and the functionality disclosed herein may also be applied in respect of wireless telecommunications systems having different architectures.

[0043] Thus, certain embodiments of the disclosure as discussed herein may be implemented in wireless telecommunication systems / networks according to various different architectures, such as the example architectures shown in Figures 1 and 2. It will thus be appreciated the specific wireless telecommunications architecture in any given implementation is not of primary significance to the principles described herein. In this regard, certain embodiments of the disclosure may be described generally in the context of communications between network infrastructure equipment / access nodes and a communications device, wherein the specific nature of the network infrastructure equipment / access node and the communications device will depend on the network infrastructure for the implementation at hand. For example, in some scenarios the network infrastructure equipment / access node may comprise a base station, such as an LTE-type base station 1 as shown in Figure 1 which is adapted to provide functionality in accordance with the principles described herein, and in other examples the network infrastructure equipment may comprise a control unit / controlling node 40 and / or a TRP10 of the kind shown in Figure 2 which is adapted to provide functionality in accordance with the principles described herein.

[0044] A more detailed diagram of some of the components of the network shown in Figure 2 is provided by Figure 3. In Figure 3, a TRP10 as shown in Figure 2 comprises, as a simplified representation, a wireless transmitter 30, a wireless receiver 32 and a controller or controlling processor 34 which may operate to control the transmitter 30 and the wireless receiver 32 to transmit and receive radio signals to one or more UEs 14 within a cell 12 formed by the TRP10. As shown in Figure 3, an example UE 14 is shown to include a corresponding transmitter 49, a receiver 48 and a controller 44 which is configured to control the transmitter 49 and the receiver 48 to transmit signals representing uplink data to the wireless communications network via the wireless access interface formed by the TRP10 and to receive downlink data as signals transmitted by the transmitter 30 and received by the receiver 48 in accordance with the conventional operation.

[0045] The transmitters 30, 49 and the receivers 32, 48 (as well as other transmitters, receivers and transceivers described in relation to examples and embodiments of the present disclosure) may include radio frequency filters and amplifiers as well as signal processing components and devices in order to transmit and receive radio signals in accordance for example with the 5G / NR standard. The controllers 34, 44 (as well as other controllers described in relation to examples and embodiments of the present disclosure) may be, for example, a microprocessor, a CPU, or a dedicated chipset, etc., configured to carry out instructions which are stored on a computer readable medium, such as a non-volatile memory. The processing steps described herein may be carried out by, for example, a microprocessor in conjunction with a random access memory, operating according to instructions stored on a computer readable medium. The transmitters, the receivers and the controllers are schematically shown in Figure 3 as separate elements for ease of representation. However, it will be appreciated that the functionality of these elements can be provided in various different ways, for example using one or more suitably programmed programmable computer(s), or one or more suitably configured application-specific integrated circuit(s) / circuitry / chip(s) / chipset(s). As will be appreciated the infrastructure equipment / TRP / base station as well as the UE / communications device will in general comprise various other elements associated with its operating functionality.

[0046] As shown in Figure 3, the TRP 10 also includes a network interface 50 which connects to the DU 42 via a physical interface 16. The network interface 50 therefore provides a communication link for data and signalling traffic from the TRP 10 via the DU 42 and the CU 40 to the core network 20.

[0047] The interface 46 between the DU 42 and the CU 40 is known as the F 1 interface which can be a physical or a logical interface. The Fl interface 46 between CU and DU may operate in accordance with specifications 3GPP TS 38.470 and 3GPP TS 38.473, and may be formed from a fibre optic or other wired or wireless high bandwidth connection. In one example the connection 16 from the TRP 10 to the DU 42 is via fibre optic. The connection between a TRP 10 and the core network 20 can be generally referred to as a backhaul, which comprises the interface 16 from the network interface 50 of the TRP 10 to the DU 42 and the Fl interface 46 from the DU 42 to the CU 40. The core network 20 is connected to the CU 40 via the N2 (also called NG-C) interface for carrying control data and via the N3 (also called NG- U) interface for carrying user data.

[0048] Although reference is made to 5G networks, the discussions in this specification apply equally to 6G networks (and beyond) where there is expected to be significantly higher throughput, lower latency and higher reliability utilising sub-THz frequencies. In the case of 6G, base stations, which are an example of network infrastructure equipment, may also referred to as 6G NB (6G Node B), 6G RAN node, and so forth. In 6G, the core network 20 may be one or more network functions.

[0049] Future 6G Wireless Communications

[0050] As described above, several generations of mobile communications have been standardised globally up to now, where each generation took approximately a decade from introduction before the development and introduction of another new generation. For example, generations of mobile communications have moved from the Global System for Mobile Communications (GSM) (2G) to Wideband Code Division Multiple Access (WCDMA) (3G), from WCDMA (3G) to LTE (4G), and most recently from LTE (4G) to NR (5G).

[0051] The latest generation of mobile communications is 5G, as discussed above with reference to the example configurations of Figures 2 and 3, where a significant number of additional features have been incorporated in different releases to provide new services and capabilities. Such services include eMBB, IIoT and URLLC as discussed above, but also include such services as 2-step Random Access (RACH), Unlicensed NR (NR-U), Cross-link Interference (CLI) handling for Time Division Duplexing (TDD), Positioning, Small Data Transmissions (SDT), Multicast and Broadcast Services (MBS), Reduced Capability UEs, Vehicular Communications (V2X), Integrated Access and Backhaul (IAB), UE power saving, Non Terrestrial Networks (NTN), NR operation up to 71GHz, loT over NTN, Non-public networks (NPN), and Radio Access Network (RAN) slicing.

[0052] Nevertheless, as in every decade, a new generation (e.g. 6G) is expected to be developed and deployed in the near future (around the year 2030), and will be expected to provide new services and capabilities that the current 5G cannot provide. There are discussions on technologies beyond 5G, i.e. 6G, that are expected to have significantly higher throughput, lower latency and higher reliability than 5G services, which are also expected to utilise sub-THz frequencies. One of the functionalities being considered for 6G is the utilisation of artificial intelligence / machine learning (AI / ML) models to assist the network and users with various tasks.

[0053] Transmission Scheduling

[0054] The lower layers (for example, the datalink and physical layers) of a mobile communication system are designed to schedule data transmissions between a transmitter and receiver at times, frequency bands and in beam directions at which propagation is not unduly negatively affected by the prevailing radio propagation conditions between the communicating gNB and the UE. The prevailing radio conditions determine the amount of data that can be transmitted (for example, transport block size); the modulation and coding rate that are to be used for the data; and the time, frequency and spatial resources that will be used for the transmission. Some of the parameters that determine these are included in the modulation and coding (MCS) configuration used for the transmission in the process of link adaptation via adaptive modulation and coding.

[0055] In LTE and NR, downlink radio conditions are measured on reference signals such as the Cell Reference Signal (CRS) and Channel State Information Reference Signal (CSI-RS), respectively. Thus, for the NR downlink, the gNB transmits CSI-RS of known configurations (timing, frequency, beam direction, etc.) to allow the UE to measure the channel conditions. The UE feeds back such measurements in the form of Channel State Information (CSI) reports to the gNB. The CSI report includes metrics such as:

[0056] • CQI (Channel Quality Information);

[0057] • PMI (Precoding Matrix Indicator);

[0058] • CRI (CSI-RS Resource Indicator); • SSBRI (SS / PBCH Resource Block Indicator);

[0059] • LI (Layer Indicator);

[0060] • RI (Rank Indicator) and / or Ll-RSRP (Reference Signal Received Power); and

[0061] • Capability Index or time-domain channel properties (TDCP).

[0062] The gNB uses these metrics to, amongst other uses, perform beam management and to determine scheduling information. Scheduling information includes resource allocation information, MCS, precoding matrix, etc. for the forthcoming downlink transmission (such as a PDSCH). The scheduling information is then sent to the UE in a scheduling Downlink Control Information (DCI) via a PDCCH. The resource allocation information includes the designation of resources both in time (time domain resource allocation or TDRA) and frequency (frequency domain resource allocation or FDRA) for a PDSCH transmission to the UE. The UE is then expected to receive a PDSCH on the downlink resources described in the DCI when they arrive. Figure 3 shows an example of this resource allocation process for Physical Downlink Shared Channels (PDSCH).

[0063] For UL channel measurements, the UE transmits Sounding Reference Signals (SRS) to the gNB, which the gNB uses for channel measurements in the slot in which it receives the SRS. When the UE wants to transmit data on the uplink, it sends a scheduling request to the gNB. The gNB then uses the channel measurement results done on the SRS to determine the MCS and resource allocation for the forthcoming uplink transmission (PUSCH). This resource allocation information is then sent to the UE in an uplink scheduling DCI via a PDCCH. The UE is then expected to transmit the PUSCH on the resources allocated in the DCI. Figure 4 illustrates an example of this resource allocation process for Physical Uplink Shared Channels (PUSCH).

[0064] In both cases, the channel measurements from which the CSI report is generated are carried out in one slot and the uplink or downlink data transmission happens in a later slot. Scheduling and resource allocation are done this way under the assumption that the UE to gNB channel propagation characteristics between the time the channel measurements are taken and the time the data transmission takes place are the same or quite similar. In this assumption, it is considered that the instant of measurement and the instant of transmission are both within the coherence time of the channel. The degree to which the channel measurements taken from a time before the actual transmission reflect the prevailing channel conditions during the time the physical channel transmission occurs depends on how far apart the two times are (TU-rin Figures 3 and 4). Other factors include the relative mobility of the UE and also the channel spatial consistency within the coverage area.

[0065] In NR, the time TM-T incorporates the processing time for the channel measurements, CSI report transmission time, PDCCH transmission and processing times, as well as kO (for downlink resource allocation) and k2 (for uplink resource allocation). Whilst most of the other components of this time are not under the control of the gNB, both kO and k2 are under the control of the gNB and thus need to be signalled in the scheduling DCI to the UE. Spatial consistency is the degree to which channel characteristics measured at one spot within the coverage footprint of a network are similar to those measured at another spot. More spatial consistency is expected for locations within the coverage footprint of a particular cell, whereas less spatial consistency is expected for locations that fall in the coverage footprint of different cells.

[0066] In a fast-changing channel for example, when there is high speed mobility between the UE and the gNB, the described method of channel measurement and resultant resource allocation is very likely to be sub- optimal. This happens because the channel characteristics at the time of measurement are more likely to have changed significantly by the time of data transmission due to the relative mobility between the UE and gNB. In this case, TM-T would be much greater than the coherence time of the channel. This may also happen if the channel measurements were carried out in a frequency band A within the operating component carrier and / or current bandwidth part (BWP), whilst the data transmission resources are allocated in a different frequency band B of the operating component carrier and / or current BWP, and where the channel characteristics are sufficiently different between bands A and B. In a frequency selective channel, channel characteristics are likely to be different between bands A and B if the frequency separation between bands A and B exceeds the coherence bandwidth of the channel.

[0067] In general, it is therefore desirable to schedule the transmissions in the same band of frequencies as the channel measurements were done. As for time, channel measurements for the transmission time window after TM-T can be predicted from present and past measurements.

[0068] As prediction of channel measurements is a multivariate problem, it is hoped that artificial intelligence / machine learning (AI / ML) can help in respect of accuracy and potentially, complexity.

[0069] Beam Management

[0070] When a wireless network such as 5G operates with multi-antennas both at the gNB and the UE, there is scope for beam forming by both the gNB via its TRPs and the UE and hence, the need for beam management. In beam forming, the gNB’s DL transmissions (and similarly, reception on the UL) can be performed via multiple time-swept beams radiating in different directions. There is an issue for a given UE in the service of such a gNB in respect of what DL beam should it receive from and what UL beam should it transmit in.

[0071] In 5G legacy, beam management comprises three processes, known as Pl, P2 and P3.

[0072] • Pl: Initial beam selection. The UE sweeps its receive beam while measuring the quality of transmit beams from the gNB which are also swept by the gNB / TRP. The aim of this is to select the best quality combination of transmit and receive beam directions between the gNB and the UE. This is often carried out with wide beams;

[0073] • P2: Transmit beam refinement. The UE points its reception beam direction towards the gNB and the gNB sweeps its DL beams as the UE measures the quality of all or a subset (e.g. the total set of beams is split into two subsets A and B) of the gNB’s DL beams. In this, the UE receive beam is often wide whilst the gNB DL beams are narrow. The intention is therefore to find the best quality DL narrow beam; and

[0074] • P3 : Receive beam refinement. The UE sweeps its receive beam while measuring a specific gNB DL narrow beam. The intention is to find the best quality UE receive beam for a particular gNB narrow transmit beam.

[0075] The resources that can be configured by the gNB for beam quality measurements include Synchronisation Signal Blocks (SSBs) and CSI-RS. For wide beam measurements, Reference Signal Received Power (RSRP) and Reference Signal Received Quality (RSRQ) measured on SSBs (SSB-RSRP and SSB- RSRQ) can be used. For narrow beam measurements, RSRP and RSRQ measured on CSI-RS (CSI- RSRP and CSI-RSRQ) can be used. These measurements (whether from SSBs or CSI-RS) are collectively referred to as LI -RSRP and LI -RSRQ, respectively.

[0076] The processes Pl, P2 and P3 as described above relate to the selection or refinement of beams in the spatial dimension. There are two issues that arise from this: • Complexity - the number of beams to measure and process in each case can be quite high. This increases computational complexity in respect of both the processing and the measurements. This complexity can be reduced if the number of beams to be measured and processed is reduced; and

[0077] • Radio propagation conditions change with time especially if there is relative movement between the UE and the gNB. This means that a beam with the best quality during the measurement window may not still be the best quality beam at a near future instant when the beams actually have to be used for transmission. One remedy for this is to use the measurements taken during the measurement window to predict the best quality beam during a future transmission window.

[0078] As these issues are multivariate in nature, it is hoped that AI / ML can help to improve on these issues with regards to beam selection / refinement.

[0079] Artificial Intelligence (Al)

[0080] The use of artificial intelligence / machine learning (AI / ML) models for applications such as beam management and CSI processing as described above, as well as positioning, was firstly studied in Release- 18 of the 3GPP standards and further developed in later release(s).

[0081] AI / ML-based tasks (such as positioning, beam management, and CSI processing, as noted above) tend to outperform legacy techniques because AI / ML models are configured to collect and process large amounts of data (e.g. positioning measurements, Ll-RSRP measurements, or CSI measurements). By performing model training (based on input data), AI / ML models learn and establish an understanding of the environment associated with these measurements. AI / ML models may then be deployed, for example at a UE or base-station (such as a gNB), in order to generate an output. This process / operation is also known as AI / ML model inference.

[0082] Lor positioning AI / ML models, the input to the AI / ML model may comprise positioning measurements (for example, measurements of one or more PRSs, or one or more SRSs, such as time of arrival and angle of arrival) and the output of the AI / ML model may comprise a position estimate of the communications device or improved positioning measurements.

[0083] Lor beam management and beam prediction AI / ML models, input data may comprise one or more of an Ll-RSRP and its associated beam / resource ID. The output data may be one or more of a predicted beam, narrow beam, improved beam prediction or a refined RSRP.

[0084] Lor CSI AI / ML models (e.g. CSI measurement and reporting enhancement, CSI compression, CSI processing or CSI prediction), input data may comprise one or more of CSI-RS quality measurements (csi-RSRP, csi-RSRQ), a signal to interference and noise ratio (SINR) estimation, a Modulation and Coding (MCS) index, Channel Quality Indicator (CQI) index, or a Precoding matrix index (PMI). The output data may be one or more of CSI-RS quality, a refined or predicted MCS, CQI index, PMI index, RI index, or a new type of output. In some examples of CSI compression AI / ML models, input data may be a full CSI compression report and output data may be a reduced size CSI compression report.

[0085] As examples, AI / ML models may utilise one or more of: supervised learning, generative Al, autoencoding, and reinforcement learning. These are explained in detail in the forthcoming paragraphs.

[0086] Supervised Learning

[0087] AI / ML models may implement a supervised machine learning model. The supervised learning model is trained using labelled training data to learn a function that maps inputs (typically provided as feature vectors) to outputs (i.e. labels). The labelled training data comprises pairs of inputs and corresponding output labels. The output labels are typically provided by an operator to indicate the desired output for each input. The supervised learning model processes the training data to produce an inferred function that can be used to map new (i.e. unseen) inputs to a label.

[0088] The input data (during training and / or inference) may comprise various types of data, such as numerical values, images, video, text, or audio. Raw input data may be pre-processed to obtain an appropriate feature vector used as input to the model - for example, features of an image or text input may be extracted to obtain a corresponding feature vector. It will be appreciated that the type of input data and techniques for pre-processing of the data (if required) may be selected based on the specific task the supervised learning model is used for.

[0089] Once prepared, the labelled training data set is used to train the supervised learning model. During training the model adjusts its internal parameters (e.g. weights) so as to optimize (e.g. minimize) an error or cost function, aiming to minimize the discrepancy between the model’s predicted outputs and the labels provided as part of the training data. In some cases, the error function may include a regularization penalty to reduce overfitting of the model to the training data set.

[0090] The supervised learning model may use one or more machine learning algorithms in order to learn a mapping between its inputs and outputs. Example suitable learning algorithms include linear regression, logistic regression, artificial neural networks, decision trees, support vector machines (SVM), random forests, and the K-nearest neighbour algorithm.

[0091] Once trained, the supervised learning model may be used for inference - i.e. for predicting outputs for previously unseen input data. The supervised learning model may perform classification and / or regression tasks. In a classification task, the supervised learning model predicts discrete class labels for input data, and / or assigns the input data into predetermined categories. In a regression task, the supervised learning model predicts labels that are continuous values.

[0092] In some cases, limited amounts of labelled data may be available for training of the model (e.g. because labelling of the data is expensive or impractical). In such cases, the supervised learning model may be extended to further use unlabelled data and / or to generate labelled data.

[0093] Considering using unlabelled data, the training data may comprise both labelled and unlabelled training data, and semi-supervised learning may be used to learn a mapping between the model’s inputs and outputs. For example, a graph-based method such as Laplacian regularization may be used to extend a SVM algorithm to Laplacian SVM in order to perform semi-supervised learning on the partially labelled training data.

[0094] Considering generating labelled data, an active learning model may be used in which the model actively queries an information source (such as a user, or operator) to label data points with the desired outputs. Labels are typically requested for only a subset of the training data set thus reducing the amount of labelling required as compared to fully supervised learning. The model may choose the examples for which labels are requested - for example, the model may request labels for data points that would most change the current model, or that would most reduce the model's generalization error. Semi-supervised learning algorithms may then be used to train the model based on the partially labelled data set. Generative Al

[0095] AI / ML models may implement generative artificial intelligence (Al).

[0096] A generative Al system learns patterns and structures in its input training data, in order to then generate new output data which exhibits similar characteristics to the training data.

[0097] The generative Al system may generate output data based on an input prompt. The prompt may comprise various types of data, such as images, video, text, or audio. The prompt may be of the same or different datatype to the model’s training and / or output data.

[0098] The generative AI / ML model may comprise a generative model trained to learn a probability distribution of the input training data, and generate new output data based on this learned distribution. For example, for a set of data instances / observable variables (X) and a set of labels / target variables (Y) in the training data set, the generative model may learn a joint probability distribution of data instances and labels p(X,Y), and / or a probability distribution of the data instances p(X) (for example where no labels are available).

[0099] Example suitable generative models for learning a probability distribution of the input training data include Variational Autoencoder (VAEs), transformer-based models, diffusion models (e.g. de-noising diffusion probabilistic models (DDPMs)), Reinforcement Learning (RL), and Generative Adversarial Networks (GANs). The choice of generative model may depend on the specific task performed by the generative AI / ML.

[0100] The generative model may comprise one or more artificial neural networks. For example, a Variational Autoencoder (VAE) may comprise a pair of neural networks acting as an encoder and a decoder to and from a reduced (i.e. latent space) representation of the training data respectively, and a Generative Adversarial Network (GAN) may comprise a first ‘generator’ neural network that generates new data and a second ‘discriminator’ neural network that learns to discriminate between generated data and real data. The one or more constituent neural networks of the generative model may be trained together or separately.

[0101] During training the generative model may adjust its internal parameters (e.g. neural network weights) so as to optimize (e.g. minimize) a loss / error function, aiming to minimize discrepancy between the generated output data and desired output data. It will be appreciated that the specific loss function, and algorithm used to optimize the function may vary depending on the nature of the generative model, and its intended application. For example, a mean squared error loss function may be used for an image generation task, and a cross-entropy loss function may be used for a text generation task. These loss functions may be optimized using various existing optimization algorithms, such as gradient descent.

[0102] Once trained, the generative model may be used to generate new output data based on an input prompt. The input prompt may be provided by a user, or by an appropriate device (e.g. using an application programming interface (API)). Thus, the generative Al / ML model allows generating new based on only a prompt and without requiring detailed instructions for doing so.

[0103] Autoencoders

[0104] AI / ML models may implement autoencoding.

[0105] An autoencoder is a type of an unsupervised machine learning model that uses one or more artificial neural networks to learn an efficient representation of unlabelled input data. The autoencoder may be used to encode various types of data, such as images, video, text, audio, or positioning / CSI measurements.

[0106] The autoencoder may comprise an encoder neural network that encodes input data into a reduced representation (also called a “latent space”), and a decoder neural network that aims to recreate the input data from the encoded reduced representation. The latent space is typically of a lower-dimension than the input data - thus, the latent space generated by the encoder typically provides a more efficient, compressed representation of the input data that requires less memory storage than the original input data.

[0107] The encoder neural network may comprise one or more layers that transform input data into a reduced representation. The encoder neural network receives input data, and the final layer of the encoder neural network outputs a reduced representation of the input data, i.e. a latent space (also termed a “bottleneck layer”).

[0108] The decoder neural network comprises one or more layers that transform data from the latent space into output data of the same dimensionality as the data input to the encoder. The decoder aims to reconstruct the data originally input to the encoder neural network from the latent space representation of the data.

[0109] The encoder and / or decoder neural networks typically comprise a plurality of hidden layers. For example, an encoder may comprise a plurality of hidden layers that progressively extract further reduced representations of the input data. Using deeper neural networks (i.e. with a higher number of hidden layers) for the encoder and / or the decoder may improve performance of the autoencoder, and in some cases may reduce the amount of training data that is required.

[0110] The encoder and decoder neural networks are typically trained together. During training the autoencoder may adjust its internal parameters (e.g. weights and biases of the encoder and decoder neural networks) so as to optimize (e.g. minimize) a loss / error function, aiming to minimize discrepancy between the data input to the encoder and the output reconstructed data generated by the decoder. It will be appreciated that the specific loss function, and algorithm used to optimize the function may vary depending on the nature of the autoencoder model, and its intended application. In an example, a mean squared error loss function optimized using gradient descent may be used. In some cases, a sparse autoencoder may be used in order to promote sparsity of the latent representation (as compared to the input) and to prevent the autoencoder from learning the identity function - for example, a sparse autoencoder may be implemented by modifying the loss function to include a sparsity regularization penalty.

[0111] In some cases, the autoencoder may be a Variational Autoencoder (VAE). The VAE is a specific type of auto-encoder in which a probability model is imposed on the encoded representation by the training process (in that deviations from the probability model are penalised by the training process). The VAE may be used for generative artificial intelligence applications to generate new output data which exhibits similar characteristics to the input encoded data by sampling from the learned latent space.

[0112] Reinforcement Learning

[0113] AI / ML models may implement reinforcement learning (RL).

[0114] Reinforcement learning is a type of machine learning directed to training an artificial intelligence agent to take actions in an environment that maximize the notion of a cumulative reward. During reinforcement learning, the agent interacts with the environment, and learns from the results of its actions, thus allowing the agent to progressively improve its decision-making. An RL model typically comprises an action-reward feedback loop. The feedback loop comprises: an environment, state, agent, policy, action, and reward. The environment is the system with which the agent interacts and in which the agent operates - for example, the environment may be a virtual environment of a video game. The state represents the current conditions in the environment. The agent receives the state as an input and takes an action which may affect the environment and change the state of the environment. The agent takes the action based on its policy which is a mapping from states of the environment to actions of the agent. The policy may be deterministic or stochastic. The reward represents feedback from the environment to the action taken by the agent. The reward provides an indication (typically in the form of a numerical value) of the desirability of the result of the agent’s action. The reward may comprise positive signals to reward desirable behaviour of the agent and / or negative signals to penalize undesirable behaviour of the agent.

[0115] Through multiple iterations of action-reward feedback loop, the agent aims to maximise the total cumulative reward it receives, thus learning how to take optimal actions in the environment. The reinforcement learning process thus allows the agent to learn an optimal policy that maximizes the cumulative reward. The cumulative award may be estimated using a value function which estimates the expected return starting from a given state or from a given state and action. Using the cumulative reward in the reinforcement learning process allows the agent to consider long-term effects of its policy.

[0116] A reinforcement learning algorithm may be used to refine the agent’s policy and the value function over iterations of the action-reward feedback loop. The learning algorithm may rely on a model of the environment (e.g. based on Markov Decision Processes (MDPs)) or be model-free. Example suitable model-free reinforcement learning algorithms include Q-leaming, State-Action-Reward-State-Action (SARSA), Deep Q-Networks (DQNs), or Deep Deterministic Policy Gradient (DDPG).

[0117] It will be appreciated that the agent will typically engage in both exploration and exploitation of the environment in which it operates. In exploration, the agent takes typically random actions to gather information about the environment and identify potentially desirable actions (i.e. actions that maximise cumulative reward). In exploitation, the agent takes actions that are expected to maximise reward (e.g. by selecting the action based on the agent’s latest policy). Various techniques may be used to control the proportion of explorative and exploitative actions taken by the agent - for example, a predetermined probability of taking an explorative action in a given iteration of the feedback loop may be set (and optionally reduced over time to allow the agent to shifts more towards exploitation over time to maximise cumulative reward in view of diminishing returns for further exploration).

[0118] In some cases, the RL model may be configured to learn from feedback provided by a user. Utilising user feedback in this way may allow the agent to improve its choice of actions and better align with user preferences. For example, reinforcement learning from human feedback (RLHF) techniques may be used. RLHF includes training a reward model based on user feedback and using this model for determining the reward in the reinforcement learning process described above. The user feedback may be received in various forms depending on the specific reinforcement learning problem being solved - for example, the feedback may be received in the form of a user ranking of instances of the agent’s actions. RLHF thus allows incorporating user feedback into the reinforcement learning process. RLHF approaches may be advantageous where it is easier for a user than for an algorithm to assess the quality of the machine learning model’s output (e.g. for generative artificial intelligence RL models).

[0119] Life Cycle Management (LCM) Architecture for an AI / ML Model

[0120] 3GPP has identified a general AI / ML framework for the NR air interface to facilitate different radio frequency applications, including positioning applications, beam management, and CSI feedback enhancements applications. The aim of the framework is to cover a general architecture addressing the whole Al model life cycle, Life Cycle Management (LCM), including such steps as data collection, model training, etc. Here, LCM for AI / ML for the NR air interface is described in the below paragraphs. As those skilled in the art would appreciate, the following description is also applicable to 6G and other wireless communications systems.

[0121] An example of a life cycle management (LCM) architecture for an artificial intelligence model is schematically illustrated in Figure 6 which has been reproduced from [3], the contents of which are hereby incorporated by reference in their entirety. As shown in Figure 6, LCM architecture comprises a data collection function 60, an inference function 62, a management function 64, a model training function 66 and a model storage function 68.

[0122] The data collection function 60 is configured to provide training data to the model training function 66, to provide monitoring data to the management function 64 and to provide inference data to the inference function 62. The data collection function 60 is a process / function of collecting data by the network nodes, management entity, or UE for the purpose of AI / ML model training, data analytics and inference.

[0123] The inference function 62 is configured to provide an inference output to the management function 64, to receive a management instruction from the management function 64 and to receive a model from the model storage function 68. The inference function 62 is a process / function of using a trained AI / ML model to produce a set of outputs based on a set of inputs.

[0124] The management function 64 is configured to provide performance feedback and / a retraining request to the model training function 66, and to provide a model delivery request to the model storage unit 68.

[0125] The model training function 66 is configured to provide an updated model to the model storage unit 68. The model training function 66 is a process / function to train an AI / ML Model (e.g., by learning the input / output relationship) in a data driven manner and obtain the trained AI / ML Model for inference.

[0126] As shown in Figure 6, the management function 64 represents the core of the LCM architecture. The function of the management function 64 is to monitor model performance at different entities (such as UEs or gNBs) and request delivery of updated models. Model management comprises two procedures - model switching and model updating.

[0127] Technical Problem

[0128] 3GPP wishes to standardize the use of AI / ML in beam management and CSI processing, as well as specifying aspects of the training of an AI / ML model. In this approach the taking of channel measurements and their use by the various AI / ML models are to be standardized. Channel measurements will be used for model training, model inference and model monitoring.

[0129] Reinforcement learning-based model training requires that measurements are provided as inputs into the model under training. An example of such model training is shown in Figure 7. The training starts with a rudimentary model 74 which is improved in each iteration of the training. In each iteration, the model under training 74 executes on current and previous inputs 70 and produces an output which is then compared with an expected value (referred to as the ground truth) 72. This comparison uses a cost function 76 which measures the distance / difference between the expected value 72 and the output of the model 74. During each iteration of the training, adaptation 78 of the model 76 is performed based on the evaluated cost function 76, with the aim of minimizing the cost function 76, before the adapted model 74 again executes with respect to inputs 70 during the next iteration. When the cost function 76 is minimized, the model 74 can be considered fully trained and is then deployed.

[0130] Using a trained model to achieve what it was trained for is known as inference. During inference, inputs of the model are derived from measurements of the same quantities that were measured and used for its training. At inference, a model is presented some inputs and it produces an output. As described above with respect to Figure 6, model monitoring can be performed in parallel with inference in order to ensure continued good performance of the model. In model monitoring, there are also expected outputs (ground truths) of the model. These can be taken from measurements of the same quantities that were measured and used for model training at the time or spatial direction that the model output is supposed to relate to.

[0131] For example, in spatial domain beam prediction, during inference, the Ll-RSRP of a set of beams chosen from selected directions are measured and then input into a fully trained spatial beam prediction AI / ML model to find the best beam amongst another (different) set of beam directions.

[0132] An example of spatial domain beam prediction is illustrated in Figure 8. In the example of Figure 8, eight beams (numbered 1 to 8) are radiated from a gNB. Ll-RSRP measurements from the even numbered beams, for example, are used to predict the best quality DL beam (from the point of view of the UE 81) from amongst the odd numbered beams. Using only measurements from some - but not all - of the beams reduces the computational complexity of both the training and the inference when contrasted with using measurements from all the beams. The ground truth to be used for either (or both) reinforcementbased training or performance monitoring of the spatial beam prediction model is the actual measured Ll- RSRP measured in the direction of the predicted beam - in the example of Figure 8, this is the Ll-RSRP measured from each of the odd numbered beams (i.e. the set of beams whose quality is being predicted).

[0133] Time-based beam prediction can also be performed, using SSB-RSRP as the beam quality measurement, and an example of this is described with respect to Figure 9. Beam quality measurements are taken during a time interval TM (for example during the first SSB burst shown in the example of Figure 8) but only on a set B of beams - e.g. all the even numbered beams / SSBs: {2, 4, 6, 8}. These measurements are then used to predict the beam qualities of another set A of beams- e.g. the odd numbered beams / SSBs: {1,3, 5, 7} - at a future time interval TT (for example during the third SSB burst shown in the example of Figure 8).

[0134] In this example, measuring the quality of only a fraction (i.e. of set B) of the beams reduces the number of measurements to be taken and processed, whilst predicting the quality of only a fraction of the beams also reduces complexity and so saves power. Based on the predicted beam qualities for the future time interval TT, one or more future best beams (for example - from amongst the odd numbered beams) can be selected for use in DL transmissions to the UE after time interval TT. This is possible because some of the even numbered beams are close spatially to some odd numbered beams whose measurements are used in the prediction. For example, with reference to Figure 7, beam 2 is straddled spatially by beams 1 and 3. The ground truths to be used for either reinforcement-based learning or performance monitoring of the beam prediction model are the actual beam qualities (in this example, the SSB-RSRP of the even numbered beams) measured during time interval TT.

[0135] In another example of time-based CSI prediction, channel measurements are taken on CSI-RS during a time interval TM and then input into a model which uses these measurements to predict the CSI at a future time interval TT. The model output is the predicted CSI at the designated instant within the time window TT. The ground truth to be used for either reinforcement-based training or performance monitoring of the model is the actual CSI measured at the designated instant within the time window TT. In all of the above examples, the gNB has to configure the measurement resources to be measured for the model input as well as resources to be measured for the ground truth in each case as follows:

[0136] • In the spatial beam prediction example, the gNB has to configure beam quality measurement resources in all the beams within the measurement window. In the example of Figure 8, the measurements taken on the even numbered beams are used for inference input into the model whilst the measurements taken on each of the odd numbered beams are used as ground truths in training. Only the measurements taken from the predicted best quality beams (for the Figure 8 example: from amongst the odd numbered beams) are used as ground truths in monitoring. However, given that the gNB has no pre-knowledge of the best beams prior to the execution of the UE spatial beam prediction AI / ML model, monitoring measurement resources have to be preconfigured to all potential best beams which in practice in the case illustrated in Figure 8 means all the odd numbered beams or all the set A beams;

[0137] • In the time-based beam prediction example, the gNB has to configure beam quality measurement resources in all the set B beams (which, in the example of Figure 9, are the even numbered SSBs) during time interval TM and configure monitoring resources in all the set A beams (which, in the example of Figure 9, are the odd numbered SSBs) during time interval TT. In this example, the measurements taken on the set B beams during time interval TM are used for inference input into the time-based beam prediction AI / ML model whilst the measurements taken on each of the set A beams during time interval TT are used as ground truths in training. For monitoring, the measurements taken from the predicted best quality beams amongst the set A beams are used as the ground truths; and

[0138] • In the time-based CSI prediction example, the gNB has to configure CSI measurement resources at time TM and also CSI measurement resources at time TT. In this example, the CSI measurements taken at time TMare used for inference input into the time-based CSI prediction model whilst the measurements taken at time TT are used as ground truths for training and monitoring.

[0139] Various arrangements relating to the use of AI / ML models during inference by a UE to predict channel properties for certain resources are described in co-pending European patent application number EP24175304.5 [4], the contents of which are hereby incorporated by reference. However, in view of what is described above, in order for such AI / ML models to be effective during inference, they must be properly and effectively trained and maintained.

[0140] Thus, technical problems to be solved as identified by the present disclosure relate to how measurement resources (e.g. such as CSI-RS and SSBs) can be configured for measurement for both the model input during training and the ground truth for monitoring (during inference), and how and when these measurements are reported by the UE to the network. Embodiments of the present technique therefore define further arrangements to those described in [4] in view of these technical problems, so as to enable the efficient and effective training and maintenance of AI / ML models deployed at UEs and used for tasks such as beam management and CSI processing.

[0141] CSI and Beam Prediction AI / ML Model Training and Monitoring

[0142] Figure 10 shows a part schematic, part message flow diagram representation of a wireless communications system comprising a communications device 101 (e.g. a UE 14) and an infrastructure equipment 102 (e.g. a gNB / TRP 10) forming part of a wireless communications network in accordance with at least some embodiments of the present technique. The communications device 101 may be configured to transmit signals to and / or receive signals from the wireless communications network, for example, to and from the infrastructure equipment 102 via a channel between the communications device 101 and the infrastructure equipment 102. Specifically, the communications device 101 may be configured to transmit data to and / or receive data from the wireless communications network (e.g. to / from the infrastructure equipment 102) via a wireless radio interface provided by the wireless communications network (e.g. a Uu interface between the communications device 101 and the Radio Access Network (RAN), which includes the infrastructure equipment 102). The communications device 101 and the infrastructure equipment 102 each comprise a transceiver (or transceiver circuitry) 101.1, 102.1, and a controller (or controller circuitry) 101.2, 102.2. Each of the controllers 101.2, 102.2 may be, for example, a microprocessor, a CPU, or a dedicated chipset, etc. The controllers 101.2, 102.2 may also each be equipped with a memory unit (which is not shown in Figure 10).

[0143] As shown in the example of Figure 10, the controller 101.2 of the communications device 101 is configured to control the transceiver 101.1 of the communications device 101 to receive 104, from the infrastructure equipment 102, an indication of a configuration in accordance with which the communications device 101 is to perform one or more measurements, and to perform 106 the measurements in accordance with the indicated configuration. Here, the performed measurements 106 are to be used as inputs into an artificial intelligence, Al, model as part of a training procedure, an inference procedure, or a monitoring procedure, the Al model being for use by the communications device 101 in performing a prediction task.

[0144] Here, where reference is made to the prediction task performed by the communications device 101, such a prediction task is generally described in accordance with the below arrangements of embodiments of the present technique as either a beam management / prediction task, or a CSI processing / prediction task. However, embodiments of the present technique may be applicable for any appropriate task performed by the communications device 101 (or network) for which utilization of the AI / ML model aids such a task, for example, a positioning task.

[0145] As the Al model is deployed at the communications device 101, the performed measurements 106 may be input into the Al model (for training, inference, or monitoring purposes) by the communications device 101 itself after performing 106 those measurements. Alternatively, if the Al model is being trained on the network side prior to deployment at the communications device 101, or the monitoring of the Al model deployed at the communications device 101 is being carried out on the network side, the communications device 101 may (alternatively or additionally) report those performed measurements 106 to the infrastructure equipment 102, for inputting to the Al model on the network side as part of the training procedure. Some arrangements of embodiments of the present technique relating to the reporting of measurements or the reporting of outputs (i.e. predictions) from the Al model are described in more detail in later paragraphs. In other words then, the communications device 101 may be configured to input the performed measurements into the Al model as part of the training procedure, the inference procedure, or the monitoring procedure. Alternatively (or in addition), the indicated configuration may indicate that the communications device 101 is to transmit an indication of the performed measurements to the infrastructure equipment 102, and so the communications device 101 may be configured to transmit, to the infrastructure equipment 102, the indication of the performed measurements. Here, the infrastructure equipment 102 may then be configured to input the indicated performed measurements into the Al model as part of the training procedure or use the indicated performed measurements in the monitoring procedure. Here, the indicated configuration 104 received by the communications device 101 may in accordance with arrangements of embodiments of the present technique configure one or more (or indeed all) of:

[0146] • the measurements to be performed by the communications device 101 for input (by the communications device 101 or infrastructure equipment 102) into the Al model;

[0147] • the reporting of such measurements performed by the communications device 101 that are used as input to the Al model;

[0148] • the measurements to be performed by the communications device 101 to be used as the ground truth for the Al model; and

[0149] • the reporting of such measurements performed by the communications device 101 that are to be used as the ground truth for the Al model.

[0150] That is, in accordance with any of the below-described arrangements of embodiments of the present technique, the same measurement object may be configured for both model input measurements and ground truth measurements (and their reporting). Alternatively, in accordance with any of such below- described arrangements, separate measurement objects may be configured - for example, one each for model input measurements (for example, this may be the indicated configuration 104 shown in Figure 10) and ground truth measurements - where the reporting of such input / ground truth measurements may be configured by the same separate measurement objects as those configuring the measurements, or may be configured by further separate configurations.

[0151] Furthermore, the configuration of further reporting by the communications device 101 - such as the model outputs or processed measurements - may be configured by the same single indicated configuration 104 or by yet further separate configurations.

[0152] Essentially then, embodiments of the present disclosure, as exemplified by the example wireless communications system of Figure 10 for example, propose techniques for the appropriate configuration of resources and signals (e.g. CSI-RSs or SSBs) for measurement by UEs for the purposes of training or monitoring an Al model used for a prediction task at that UE, as well as for the appropriate configuration of how such measurements and outputs of the Al model are reported. Such embodiments of the present disclosure thus allow for the efficient and effective performance of training and / or maintenance procedures of AI / ML models deployed at a UE. As those skilled in the art would understand, CSI and beam quality measurements are derived from channel quality measurements of both CSI-RS and SSB resources. CSI-RS are configured and transmitted on particular DL beams for this purpose, whilst SSBs are always-on signals beam-swept over DL beams for the purposes of DL synchronization and measurement. CSI-RS transmission resources can either be periodic, semi-persistent or aperiodic whilst SSBs are always periodic.

[0153] In some arrangements of embodiments of the present technique, where the prediction task relates to beam management, SSBs are configured to be measured and these measurements input to the beam prediction model for training, inference, and monitoring. In other words, the indicated configuration indicates that the communications device is to perform the one or more measurements on a configured set of synchronisation signal blocks, SSBs, the set of SSBs being received by the communications device as a broadcast from the infrastructure equipment using at least some of a plurality of beams radiated by the infrastructure equipment in different spatial directions, wherein each of the plurality of beams forms part of either a first subset or a second subset. SSBs are broadcast with a configured periodicity chosen from the set {5, 10, 20, 40, 80, 160} ms. Each SSB lasts for 4 OFDM symbols. Each SSB burst contains 4, 8 or 64 SSBs with each SSB in a burst transmitted on a separate beam. In 5G NR, the number of beams or SSBs in a burst dependents on the component carrier frequency band: 4 SSBs per burst for component carrier frequencies less than 3GHz, 8 SSBs per burst for frequencies between 3GHz and 6GHz and, 64 SSBs per burst for frequencies above 6GHz. Measurements of the SSB used for beam prediction can be SSB-RSRP or SSB-RSRQ.

[0154] In such arrangements, the measurement period for input into model training, inference and / or monitoring can be configured based on the SSB burst period. The measurement window duration is configured as a multiple of the SSB burst period. In other words, the indicated configuration may indicate a measurement time window during which measurements are to be performed on SSBs broadcast using beams of the first subset and on SSBs that are broadcast using beams of the second subset, the measurement time window being configured based on a period over which the set of SSBs are received.

[0155] In accordance with various arrangements of embodiments of the present technique for beam prediction, the network configures the UE with the beam indices of the set A and set B beams. Here, measurements performed on set B beams (also referred to herein as beams of the first subset) are input into the Al model as part of the training or inference procedure, while measurements performed on set A beams (also referred to herein as beams of the second subset) are used during the training or monitoring procedure as the ground truth, in order to improve performance of the Al model through comparison with the predictions output by the Al model. In other words, the communications device or infrastructure equipment may be configured (e.g. by the indicated configuration) to input measurements performed on SSBs broadcast using beams of the first subset into the Al model in order to predict a quality of the beams of the second subset as the prediction task, and the communications device / infrastructure equipment may be configured (e.g. by the indicated configuration) to use the measurements performed on SSBs broadcast using beams of the second subset as a ground truth for the Al model.

[0156] For time based-beam prediction using SSBs in such arrangements, the network will - in accordance with some arrangements of embodiments of the present technique - configure the future time to which the beams should be predicted. This time is an offset Tp between the time the set B measurements are taken and the time to which the prediction is done. In other words, the indicated configuration may indicate a time offset between the performance of the measurements and the inputting of the measurements performed on SSBs broadcast using beams of the first subset into the Al model in order to predict a quality of the beams of the second subset at a future time (e.g. during a future transmission window) as the prediction task. The duration is configured as NT SSB burst periods where NT > 2 and the duration of NT SSB burst periods is at least as long as Tp. This ensures that there will be at least two SSB bursts within the measurement window. The SSBs of the set B beams within the first or any early SSB burst within the window are measured by the UE and used for either training or inference. The SSBs of the set A beams within the second or any later burst Tp seconds later from the set B measurements are measured by the UE as the ground truth.

[0157] Above arrangements relate to where the prediction task is a time-based beam prediction task. As noted above however, arrangements of embodiments of the present technique are also directed to spatial domain-based beam prediction. In other words, the predication task may comprise predicting a quality of beams of the second subset using measurements performed on beams of the first subset.

[0158] For spatial domain-based beam prediction using SSBs in accordance with some arrangements of embodiments of the present technique, the measurement window duration is configured as one SSB burst period. All measurements for both set A and set B beams are taken during the same burst. In other words, the indicated configuration may indicate that the communications device is to perform measurements on SSBs broadcast using the beams of the first subset and the beams of the second subset during a single burst period of the received set of SSBs. This can however entail carrying out up to 64 (the maximum number of SSB beams per component carrier) measurements within a burst duration of 5ms in frequency range 2 (FR2, which includes bands from 24.25 GHz to 71 GHz) for example. Therefore, in some other arrangements of embodiments of the present technique, the time offset Tp between the time set B measurements are taken and the time set A measurements are taken is configured. In other words, the indicated configuration may indicate a time offset between a first time at which the communications device is to perform measurements on SSBs broadcast using the beams of the first subset and a second time at which the communications device is to perform measurements on SSBs broadcast using the beams of the second subset.

[0159] The duration of the measurement window is configured as NT SSB burst periods where the duration of NT SSB bursts is substantially equal to TP. The measurements of set B are therefore taken in one SSB burst and those of set A are taken in a later SSB burst. In such arrangements, the window duration and Tp should be as short as possible (NT is small for example taken from {2,3,4}) so as to ensure that the two sets of beam measurements are within the coherence time of the channel.

[0160] SSBs are broadcast in wide beams. The spatial granularity of SSB beams is thus much lower than that achievable with CSI-RS for spatial beam prediction since each CSI-RS is transmitted in a specific narrow beamwidth direction. Thus, whilst SSBs can be used for Pl (initial beam selection) procedure, CSI-RS is better for use in P2 and P3 (beam refinement) procedures. Thus, in accordance with at least some arrangements of embodiments of the present technique, the indicated configuration may indicate that the communications device is to perform the one or more measurements on channel state information reference signals, CSI-RS, the CSI-RS being received by the communications device from the infrastructure equipment using at least some of a plurality of beams radiated by the infrastructure equipment in different spatial directions, wherein each of the plurality of beams forms part of either a first subset or a second subset. CSI-RS may be periodic, aperiodic, or semi-persistent. Periodic CSI-RS, once configured occur, at regular intervals; aperiodic CSI-RS have to be scheduled by DCI; whilst semi- persistent CSI-RS are configured to occur periodically but need to be activated by DCI before transmission.

[0161] In some arrangements of embodiments of the present technique, measurement resources for model training, inference and monitoring for beam prediction are configured as periodic CSI-RS. In other words, the CSI-RS may be received by the communications device from the infrastructure equipment in accordance with a configured periodicity.

[0162] In time-based beam prediction, CSI-RS on set B beams are measured at a time TM and are used as model inputs for training and / or inference. That is, the indicated configuration may indicate that measurements performed on a first set of the CSI-RS that are received at a first time using beams of the first subset are to be input into the Al model in order to predict a quality of the beams of the second subset at a future time as the prediction task. At a later time TT, CSI-RS on another set A of beams are measured as ground truths. That is, the indicated configuration may indicate that measurements performed on a second set of the CSI-RS that are received at a second time using beams of the second subset are to be used as ground truths for the Al model. The following arrangements of embodiments of the present technique cover the configuration of the measurement window (Tp - TM), which are applicable to periodic (and semi- persistently activated periodic) CSI-RS.

[0163] In some arrangements of embodiments of the present technique, the measurement period is configured to match N CSI-RS periods for periodic CSI-RS where N is a small integer. In other words, the indicated configuration may indicate a measurement time window during which measurements are to be performed on CSI-RS that are received using beams of the first subset and on CSI-RS that are transmitted using beams of the second subset, the measurement time window being configured to match a specified number of periods of the CSI-RS. An Al model can be trained such that the measurement periods for training and inference can be different.

[0164] In some arrangements of embodiments of the present technique, the measurement period is fixed or already chosen - for example in a case where the beam prediction model is trained with a fixed measurement period. In other words, a measurement time window during which measurements are to be performed both on CSI-RS that are received using beams of the first subset and on CSI-RS that are received using beams of the second subset is preconfigured and known to the communications device. In such arrangements, the measurement period for inference is the same as the measurement period for training. In such arrangements, the CSI-RS period is configured to match or be a factor of the training measurement period. In Rell8, only a limited choice of CSI-RS periods {4, 5, 8, 10, 16, 20, 32, 40, 64, 80, 120, 320, 640} in slots is available. The beam prediction model can be trained with any of these periods. In some implementations of such arrangements, new more granular periods can also be configured. This would require changes in RRC signalling (CSI-ResourceConfig IE) to support new periods such as {2, 6, 12, 14, 18 etc} slots.

[0165] Semi-persistent CSI-RS are first configured via RRC for transmission by the gNB periodically, but the gNB has to send activation signalling to the UE via DCI before they are transmitted. In other words, the CSI-RS may be received by the communications device from the infrastructure equipment in accordance with the configured periodicity upon the communications device receiving an activation signal from the infrastructure equipment. Thus, in arrangements of embodiments of the present technique in which beam prediction is done using semi-persistent CSI-RS after activation, during configuration, the semi-persistent CSI-RS period and hence the measurement period will be selected in accordance with the arrangements described in the previous paragraphs (i.e. it may be configured to match N CSI-RS periods or may be fixed or otherwise already chosen). Thus, just before set B measurements, the gNB would activate the transmission of the semi-persistent CSI-RS by sending the activation signalling to the UE which will carry out the measurements on the CSI-RS that follow. Activation would last until at least after the measurement period is exceeded and so include the time for CSI measurements of set A beams.

[0166] The transmission of aperiodic CSI-RS is signalled to UEs via DCI and then the CSI-RS are inserted on the DL by the gNB at a time that is indicated in the DCI. In other words, the CSI-RS may be dynamically scheduled by the infrastructure equipment and are received by the communications device from the infrastructure equipment in accordance with the dynamic scheduling. In some arrangements of embodiments of the present technique which uses aperiodic CSI-RS for AI / ML beam prediction, the gNB will signal the transmission of CSI-RS for set B measurements. For the scheduling and transmission of CSI-RS for set A measurements, any of the following arrangements may be used:

[0167] • Set A measurement CSI-RS are scheduled by a separate DCI that comes after the CSI-RS for set B measurements are transmitted to the UE. While this improves flexibility for set A measurement timing, the two DCIs use up more signalling resources;

[0168] • Set A measurement CSI-RS are scheduled and transmitted at a time determined by specification following the transmission of CSI-RS for set B measurements; or

[0169] • Set A measurement CSI-RS are scheduled by the same DCI that schedules the CSI-RS for set B measurements. An extra scheduling field is included in the DCI for this additional CSI-RS scheduling. The extra scheduling field can be an offset from the first scheduling field.

[0170] As particular CSI-RS are transmitted in particular beam directions, CSI-RS for different beams can occur together even in the same OFDM symbol. In some arrangements of the present technique therefore, relating to spatial beam prediction with CSI-RS, both set B and set A beam measurements for spatial beam prediction can be taken within a short measurement window, such as one CSI-RS period. Thus, here, the prediction task may comprise predicting a quality of beams of the second subset using measurements performed on beams of the first subset. Such arrangements may be applicable for all forms of CSI-RS: periodic, semi-persistent, and aperiodic.

[0171] This can however entail the treatment of a large number of measurements over the two sets B and A of beams within a short time. Thus, in some arrangements of embodiments of the present technique, the measurement period may be configured to match N CSI-RS periods for periodic CSI-RS, where N is a very small integer. That is, the indicated configuration may indicate that a measurement time window during which the communications device can perform measurements on CSI-RS that are received using the beams of the first subset and the beams of the second subset, where here the measurement time window may be configured to match a specified (and small) number of periods of the CSI-RS. Here, N has to be small enough so that the time between set B and set A beam measurements does not exceed the coherence time of the channel. For semi-persistent CSI-RS, the measurement period may similarly be configured to match N periods of semi-persistent CSI-RS. Activation is done by DCI just before the onset of set B measurements. The network ensures that activation remains for the duration of the measurement window (N periods of semi-persistent CSI-RS) during which set A measurements are completed, and then the semi-persistent CSI-RS may be deactivated through transmission by the network of a deactivation signal. The measurement window is therefore defined by the time between activation and deactivation. For aperiodic CSI-RS, CSI-RS are scheduled for set B and set A measurements. The scheduling can either be explicit in the specification, combined in one DCI, or use two separate DCIs.

[0172] As those skilled in the art would appreciate, in respect of all of the above-described arrangements of embodiments of the present technique, the interval between set A and set B measurements needs to not exceed the coherence time of the channel.

[0173] Arrangements of embodiments of the present technique described in the above paragraphs thus relate to beam prediction (both time-based and spatial domain-based) using AI / ML models, and relate to the training, inference, or monitoring procedures in respect of such models. However, as previously noted, the other issue to consider is AI / ML CSI prediction models for use in other processes such as resource allocation. Arrangements of embodiments of the present technique thus furthermore relate to the prediction of CSI in respect of predicting measured values of CSI-RS to be received at a future point in time based on measurements performed on CSI-RS currently being received. In other words, the indicated configuration may indicate that the communications device is to perform the one or more measurements on channel state information reference signals, CSI-RS, the CSI-RS being received by the communications device from the infrastructure equipment, and wherein the prediction task comprises predicting a quality of CSI-RS that are to be received from the infrastructure equipment at a future time using the measurements performed on the received CSI-RS.

[0174] In some such arrangements of embodiments of the present technique, the measurement window is configured by the network with its length configured as a number of periods of the CSI-RS for periodic and semi-persistent CSI-RS. With aperiodic CSI-RS, the CSI-RS are scheduled explicitly by DCI. In the measurement window, CSI-RS are measured at the beginning of the window and used as input into the prediction model to predict the CSI-RS at a configured time TT in the future with TT being a small multiple of the CSI-RS period within the measurement window. For periodic and semi-persistent CSI- RS, CSI-RS occurring at time TTare measured and used as ground truths for the model predicted CSI-RS. With aperiodic CSI-RS, CSI-RS are also scheduled at time TT for measurement and use as ground truths. The predicted CSI-RS can then be further processed to generate CSI information for scheduling. In other words, in respect of such arrangements, the indicated configuration may indicate a measurement time window during which both the measurements are performed on the received CSI-RS and measurements are to be performed on the CSI-RS that are to be received at the future time, the future time being within the measurement time window, where here, the measurements performed on the received CSI-RS are to be input into the Al model as raw CSI-RS measurements (and / or used as the ground truth).

[0175] In other such arrangements of embodiments of the present technique, the measurement window is configured by the network in the same way as in the previously described arrangements. However, when the CSI-RS measurements are taken, these are processed for example to extract representations of the channel such as a channel matrix, eigenvalues of the channel, or a channel transfer function etc. This is most efficient especially in the case where CSI-RS from one period to the other do not occur at the same frequencies. The extracted representation of the channel is input into the AI / ML CSI prediction model. The model output would then be a predicted channel representation at time TT in the future. Similarly to the previously described arrangements, CSI-RS occurring or scheduled at time TT will also be measured and processed to extract the equivalent channel representation at time TT to provide the ground truth. In other words, in respect of such arrangements, the indicated configuration may indicate a measurement time window during which the measurements are to be performed on the received CSI-RS that are to be, processed to obtain a representation of the channel which is to be input into the Al model in order to predict CSI-RS that are to be received from the infrastructure equipment at a future time (and / or used as the ground truth) where the future time is within the measurement time window, and where measurements are to be performed on CSI-RS that are subsequently received at the future time within the measurement time window.

[0176] In both of the above described arrangements of embodiments of the present technique, i.e. where raw CSI-RS values are input into the Al model and where processed CSI-RS values (i.e. the extracted channel representation) are input into the model, the indicated configuration may indicate that the communications device is to perform measurements on the CSI-RS that are to be received at the future time, wherein the measurements to be performed on the CSI-RS that are to be received at the future time are to be used as a ground truth for the Al model.

[0177] During inference, monitoring can be carried out by converting the ground truth CSI-RS measurements to a channel representation of the same kind as the predicted channel representation. The quality metric between the predicted channel representation and the ground truth one is calculated to determine the efficacy of the prediction model. In other words, the communications device may be configured to perform measurements on CSI-RS received at the future time, to process the measurements performed on the CSI-RS received at the future time to obtain an actual representation of the channel at the future time, and to determine a quality metric of the Al model based on a comparison of the predicted representation of the channel at the future time and the actual representation of the channel at the future time. In some implementations, the quality metric may be the normalized mean square error (NMSE) between the two. In some other implementations, the metric may be the Squared Generalized Cosine Similarity (SGCS). Alternatively, the predicted channel representation may be deconvolved into CSI-RS measurements and these are then directly compared with the ground truth measurements using NMSE or the like to determine the efficacy of the model. In other words, the communications device may be configured to perform measurements on CSI-RS received at the future time, to deconvolve the predicted representation of the channel at the future time to obtain predicted measurements on CSI-RS to be received at the future time, and to determine a quality metric of the Al model based on a comparison of the predicted measurements on the CSI-RS to be received at the future time and the performed measurements on the CSI-RS received at the future time. In the arrangements described in the paragraph above, the communications device (i.e. UE) determines the efficacy of the Al model. However, in other arrangements - which correspond to those described in the paragraph above - the network (i.e. the infrastructure equipment / gNB) may determine the model efficacy based on the measurements received from the UE. That is, in other words, the infrastructure equipment may be configured to receive, from the communications device, an indication of measurements performed on the CSI-RS transmitted at the future time, to process the indicated measurements performed on the CSI-RS transmitted at the future time to obtain an actual representation of the channel at the future time, and to determine a quality metric of the Al model based on a comparison of the predicted representation of the channel at the future time and the actual representation of the channel at the future time. Alternatively, the infrastructure equipment may be configured to receive, from the communications device, an indication of measurements performed on the CSI-RS transmitted at the future time, to deconvolve the predicted representation of the channel at the future time to obtain predicted measurements on CSI-RS to be received at the future time, and to determine a quality metric of the Al model based on a comparison of the predicted measurements on the CSI-RS to be transmitted at the future time and the performed measurements on the CSI-RS transmitted at the future time.

[0178] As mentioned above, further arrangements of embodiments of the present technique relate to how and what the UE reports to the network after performing measurements. In other words, in accordance with such arrangements, the indicated configuration may indicate that the communications device is to transmit an indication of the performed measurements to the infrastructure equipment, and the communications device is thus configured to transmit, to the infrastructure equipment, the indication of the performed measurements. The following paragraphs describe such arrangements in more detail.

[0179] In some arrangements of embodiments of the present technique relating to beam prediction, during AI / ML model training in which the training happens at the network side, measurement reporting is configured to report set B beam measurements (for model training at the network side) and set A beam measurements (for use as ground truth information in model training and monitoring). In other words, the indication of the performed measurements may comprise measurements performed on both of a first subset and a second subset of a plurality of beams radiated by the infrastructure equipment in different spatial directions. This is essentially a data collection fortraining exercise.

[0180] In some other arrangements of embodiments of the present technique relating to beam prediction however, during AI / ML model training in which the training happens at the UE side, measurement reporting is not configured as set B and set A beam measurements are used for training at the UE only.

[0181] In some arrangements of embodiments of the present technique relating to CSI prediction, during AI / ML model training in which the training happens at the network side, measurement reporting is configured to report raw CSI-RS measurements both for the training model input and the ground truth, where here the Al model can determine the channel transfer function. In other words, the indication of the performed measurements may comprise one or more CSI-RS measurements performed on one or more reference signals received by the communications device from the infrastructure equipment.

[0182] In some arrangements of embodiments of the present technique relating to CSI prediction and in which the training input is a channel representation, the UE converts both the input and ground truth CSI measurements into the appropriate channel representation before reporting to the network. In other words, the indication of the performed measurements may comprise a representation of the channel, the representation of the channel being obtained by processing one or more CSI-RS measurements performed on one or more reference signals received by the communications device from the infrastructure equipment. In some other arrangements of embodiments of the present technique relating to CSI prediction however, during AI / ML model training in which the training happens at the UE side, measurement reporting is not configured as both the input CSI-RS measurements and the ground truths are used only at the UE.

[0183] In some arrangements of embodiments of the present technique relating to beam prediction, during AI / ML model inference, measurement reporting is configured to report the indices of the best or top K predicted beams where the value of K may be configured by the network to the UE with a configurable range from 1 to the total number of beams in set A. In other words, the indication of the performed measurements may comprise indices of a best one or more of a plurality of beams radiated by the infrastructure equipment in different spatial directions, wherein the one or more best (i.e. highest predicted quality) beams are determined by the communications device through use of the Al model. In some such arrangements, both the top K indices and their predicted Ll-RSRP are included in the report. In other words, the indication of the performed measurements may comprise an indication of a predicted quality of the one or more best beams - where this predicted quality may be Ll-RSRP or Ll-RSRQ as predicted by the Al model. Here, as noted above, K may be configured by the network. That is, the communications device may be configured to receive, from the infrastructure equipment, an indication of the number of the one or more best beams to be determined by the communications device through use of the Al model. The configuration of K may be comprised in the measurement reporting configuration, or may be configured separately.

[0184] In some arrangements of embodiments of the present technique relating to CSI prediction, during AI / ML model inference, measurement reporting is configured to report the predicted CSI. Prediction may be to various time instances. In this case, the report will include the one or more CSI predictions (and possibly the time to which each is predicted). In other words, the indication of the performed measurements may comprise an indication of one or more predicted CSI-RS values, wherein the one or more predicted CSI- RS values are determined by the communications device through use of the Al model, and - optionally - the one or more predicted CSI-RS values may comprise an indication of a time at which each of the predicted CSI-RS values are predicted.

[0185] In arrangements of embodiments of the present technique relating to beam prediction, for model monitoring during inference, measurement reporting depends on how monitoring is organized.

[0186] That is, in some such arrangements, the measured Ll-RSRPs of the set A beams of the same indices as the top K beams are added to the report of the top K indices and their predicted Ll-RSRP. In other words, the one or more best beams form part of a first subset of the plurality of beams, and wherein the indication of the performed measurements may comprise an indication of a measured quality of one or more beams of a second subset of the plurality of beams, the one or more beams of the second subset having indices corresponding to the indices of the one or more best predicted beams. Lor each of the top K beams, the network may compute a model quality metric for example, the difference between the predicted Ll-RSRP and its ground truth (the equivalent set A measured Ll-RSRP). The network may then determine model efficacy by comparing the magnitude of the calculated differences with a threshold - for example, if the magnitude of the differences is below the threshold, then the network may determine that the model is working effectively. In other words, the infrastructure equipment may be configured to determine a quality metric of the Al model based on a comparison of a predicted quality of the one or more best beams and the indicated measured quality of the one or more beams of the second subset, and to compare the determined quality metric of the Al model with a preconfigured threshold. In other such arrangements, model quality metrics are calculated at the UE and added to the report of the top K indices. In other words, the one or more best beams form part of a first subset of the plurality of beams, and wherein the indication of the performed measurements may comprise an indication of a quality metric of the Al model, the quality metric of the Al model being determined by the communications device based on a comparison of a predicted quality of the one or more best beams and a measured quality of one or more beams of a second subset of the plurality of beams, the one or more beams of the second subset having indices corresponding to the indices of the one or more best beams. In such arrangements, the network may then compare the reported metrics against model performance thresholds and then determine the efficacy of the current model. In other words, the infrastructure equipment may be configured to compare the indicated quality metric of the Al model with a preconfigured threshold.

[0187] In other such arrangements, model quality metrics are calculated at the UE and compared to a network configured threshold at the UE and the results of the comparison (i.e. a measure of the model’s efficacy) are added to the report of the top K indices. In other words, the one or more best beams form part of a first subset of the plurality of beams, and wherein the indication of the performed measurements may comprise an indication of a comparison of a quality metric of the Al model with a preconfigured threshold (i.e. an indication of the model’s efficacy), the quality metric of the Al model being determined by the communications device based on a comparisons of a predicted quality of the one or more best beams and a measured quality of one or more beams of a second subset of the plurality of beams, the one or more beams of the second subset having indices corresponding to the indices of the one or more best beams.

[0188] In arrangements of embodiments of the present technique relating to CSI prediction, for model monitoring during inference, measurement reporting depends on how monitoring is organized.

[0189] That is, in some such arrangements, the UE computes the quality metric between the predicted and measured CSI and includes this difference in the measurement reporting. In other words, the indication of the performed measurements may comprise an indication of a quality metric of the Al model, the quality metric of the Al model being determined by the communications device based on a comparison of one or more measured CSI-RS values and the one or more predicted CSI-RS values (e.g. where these are for CSI-RS received at the same time; one being measured by the UE, and the other being predicted by the Al model). The network may then compare the difference against a threshold and decide on the efficacy of the CSI prediction model. In other words, the infrastructure equipment may be configured to compare the indicated quality metric of the Al model with a preconfigured threshold.

[0190] In other such arrangements, after calculating the quality metric between the predicted and measured CSI the UE compares the difference against a threshold configured previously by the network and sends the result of this comparison (below, equal, above threshold) to the network to decide on the efficacy of the CSI prediction model. In other words, the indication of the performed measurements may comprise an indication of a comparison of a quality metric of the Al model with a preconfigured threshold, the quality metric of the Al model being determined by the communications device based on a comparison of one or more measured CSI-RS values and the one or more predicted CSI-RS values.

[0191] In accordance with any of the above-described arrangements, if the model efficacy is determined on the basis of the quality metric being above the threshold, the network may consider that the model is not working. The model may therefore be disconnected and legacy calculations (rather than Al model-based predictions) may be used, or the network may determine that the Al model should be updated, trained further, or swapped with a different model in a model pool which may comprise models designed by different model designers.

[0192] Alternatively, in accordance with any of the above-described arrangements, the calculation of the quality metric between the predicted and measured CSI / L1-RSRP or the like may be a measure of similarity rather than difference. Thus, in some implementations of such arrangements, the model may be considered to be not working if the model quality metric is below (rather than above) the threshold. Thus, in accordance with any of the arrangements described herein, the measure of the model efficacy may be determined based on the model quality metric being above, below, or the same as the threshold, with any of such relations to the threshold being considered to be good or bad depending on the measure of the quality metric being used.

[0193] In some arrangements of embodiments of the present technique relating to beam prediction, the top K beams information (indices and / or Ll-RSRPs), their ground truths or model quality metrics or model quality efficacy are sent to the network in the same measurement report. In other arrangements especially when the time between set B and set A measurements is long, the top K beam information is sent in an earlier report and their ground truths or model quality metrics or model quality efficacy are sent in a separate report.

[0194] In some arrangements of embodiments of the present technique relating to CSI prediction, the predicted CSI and the result of monitoring is delivered in a single CSI report from UE to network. In other arrangements, the predicted CSI and its monitoring result are delivered in different measurement objects at different times.

[0195] Figure 11 shows a flow diagram illustrating an example process of communications in a communications system in accordance with embodiments of the present technique. The process shown by Figure 11 is specifically a method of operating a communications device (e.g. a UE) configured to transmit signals to and / or to receive signals from an infrastructure equipment (e.g. a gNB) via a channel between the communications device and the infrastructure equipment, the infrastructure equipment forming part of a wireless communications network.

[0196] The method begins in step SI. The method comprises, in step S2, receiving, from the infrastructure equipment, an indication of a configuration in accordance with which the communications device is to perform one or more measurements. In step S3, the process comprises performing the measurements in accordance with the indicated configuration. Here, the performed measurements are to be used as inputs into an artificial intelligence, Al, model (and are input by either the communications device or the infrastructure equipment in the case that the communications device reports the performed measurements to the infrastructure equipment) as part of a training procedure, an inference procedure, or a monitoring procedure, the Al model being for use by the communications device in performing a prediction task. The process ends in step S4.

[0197] Those skilled in the art would appreciate that the method shown by Figure 11 may be adapted in accordance with embodiments of the present technique. For example, other intermediate steps may be included in such a method, or the steps may be performed in any logical order. Though embodiments of the present technique have been described largely by way of the example communications system shown in Figure 10, it would be clear to those skilled in the art that they could be equally applied to other systems to those described herein, provided that these are within the scope of the claims. Those skilled in the art would further appreciate that such infrastructure equipment and / or communications devices as herein defined may be further defined in accordance with the various arrangements and embodiments discussed in the preceding paragraphs. It would be further appreciated by those skilled in the art that such infrastructure equipment and communications devices as herein defined and described may form part of communications systems other than those defined by the present disclosure, provided that these are within the scope of the claims.

[0198] The following numbered paragraphs provide further example aspects and features of the present technique:

[0199] Paragraph 1. A method of operating a communications device configured to transmit signals to and / or to receive signals from an infrastructure equipment via a channel between the communications device and the infrastructure equipment, the infrastructure equipment forming part of a wireless communications network, the method comprising receiving, from the infrastructure equipment, an indication of a configuration in accordance with which the communications device is to perform one or more measurements, and performing the measurements in accordance with the indicated configuration, wherein the performed measurements are to be used as inputs into an artificial intelligence, Al, model as part of a training procedure, an inference procedure, or a monitoring procedure, the Al model being for use by the communications device in performing a prediction task.

[0200] Paragraph 2. A method according to Paragraph 1, comprising inputting the performed measurements into the Al model as part of the training procedure, the inference procedure, or the monitoring procedure.

[0201] Paragraph 3. A method according to Paragraph 1 or Paragraph 2, wherein the indicated configuration indicates that the communications device is to perform the one or more measurements on a configured set of synchronisation signal blocks, SSBs, the set of SSBs being received by the communications device as a broadcast from the infrastructure equipment using at least some of a plurality of beams radiated by the infrastructure equipment in different spatial directions, wherein each of the plurality of beams forms part of either a first subset or a second subset.

[0202] Paragraph 4. A method according to Paragraph 3, wherein the indicated configuration indicates a measurement time window during which measurements are to be performed both on SSBs broadcast using beams of the first subset and on SSBs broadcast using beams of the second subset, the measurement time window being configured based on a period over which the set of SSBs are received.

[0203] Paragraph 5. A method according to Paragraph 3 or Paragraph 4, wherein the indicated configuration indicates a time offset between the performance of the measurements and the inputting of the measurements performed on SSBs broadcast using beams of the first subset into the Al model in order to predict a quality of the beams of the second subset at a future time as the prediction task.

[0204] Paragraph 6. A method according to Paragraph 5, wherein the indicated configuration indicates that the measurements performed on SSBs broadcast using beams of the second subset are to be used as a ground truth for the Al model.

[0205] Paragraph 7. A method according to any of Paragraphs 3 to 6, wherein the prediction task comprises predicting a quality of beams of the second subset using measurements performed on beams of the first subset.

[0206] Paragraph 8. A method according to Paragraph 7, wherein the indicated configuration indicates that the communications device is to perform measurements on SSBs broadcast using the beams of the first subset and the beams of the second subset during a single burst period of the received set of SSBs.

[0207] Paragraph 9. A method according to Paragraph 7 or Paragraph 8, wherein the indicated configuration indicates a time offset between a first time at which the communications device is to perform measurements on SSBs broadcast using the beams of the first subset and a second time at which the communications device is to perform measurements on SSBs broadcast using the beams of the second subset.

[0208] Paragraph 10. A method according to any of Paragraphs 1 to 9, wherein the indicated configuration indicates that the communications device is to perform the one or more measurements on channel state information reference signals, CSI-RS, the CSI-RS being received by the communications device from the infrastructure equipment using at least some of a plurality of beams radiated by the infrastructure equipment in different spatial directions, wherein each of the plurality of beams forms part of either a first subset or a second subset.

[0209] Paragraph 11. A method according to Paragraph 10, wherein the prediction task comprises predicting a quality of the beams of the second subset at a future time using measurements performed on the beams of the first subset, and wherein the CSI-RS are received by the communications device from the infrastructure equipment in accordance with a configured periodicity.

[0210] Paragraph 12. A method according to Paragraph 11, wherein the CSI-RS are received by the communications device from the infrastructure equipment in accordance with the configured periodicity upon the communications device receiving an activation signal from the infrastructure equipment. Paragraph 13. A method according to Paragraph 11 or Paragraph 12, wherein the indicated configuration indicates a measurement time window during which measurements are to be performed on CSI-RS that are received using beams of the first subset and on CSI-RS that are received using beams of the second subset, the measurement time window being configured to match a specified number of periods of the CSI-RS.

[0211] Paragraph 14. A method according to any of Paragraphs 11 to 13, wherein a measurement time window during which measurements are to be performed both on CSI-RS that are received using beams of the first subset and on CSI-RS that are received using beams of the second subset is preconfigured and known to the communications device.

[0212] Paragraph 15. A method according to any of Paragraphs 10 to 14, wherein the prediction task comprises predicting a quality of the beams of the second subset at a future time using measurements performed on the beams of the first subset, and wherein the CSI-RS are dynamically scheduled by the infrastructure equipment and are received by the communications device from the infrastructure equipment in accordance with the dynamic scheduling.

[0213] Paragraph 16. A method according to any of Paragraphs 10 to 15, wherein the indicated configuration indicates that measurements performed on a first set of the CSI-RS that are received at a first time using beams of the first subset are to be input into the Al model in order to predict a quality of the beams of the second subset at a future time as the prediction task, and that measurements performed on a second set of the CSI-RS that are received at a second time using beams of the second subset are to be used as a ground truth for the Al model.

[0214] Paragraph 17. A method according to any of Paragraphs 10 to 16, wherein the prediction task comprises predicting a quality of beams of the second subset using measurements performed on beams of the first subset, and wherein the indicated configuration indicates a measurement time window during which the communications device can perform measurements on CSI-RS that are received using the beams of the first subset and the beams of the second subset.

[0215] Paragraph 18. A method according to any of Paragraphs 1 to 17, wherein the indicated configuration indicates that the communications device is to perform the one or more measurements on channel state information reference signals, CSI-RS, the CSI-RS being received by the communications device from the infrastructure equipment, and wherein the prediction task comprises predicting a quality of CSI-RS that are to be received from the infrastructure equipment at a future time using the measurements performed on the received CSI-RS.

[0216] Paragraph 19. A method according to Paragraph 18, wherein the indicated configuration indicates a measurement time window during which both the measurements are performed on the received CSI-RS and measurements are to be performed on the CSI-RS that are to be received at the future time, the future time being within the measurement time window.

[0217] Paragraph 20. A method according to Paragraph 19, wherein the measurements performed on the received CSI-RS are to be input into the Al model as raw CSI-RS measurements.

[0218] Paragraph 21. A method according to Paragraph 19 or Paragraph 20, wherein the measurements performed on the received CSI-RS are to be processed to obtain a representation of the channel which is to be input into the Al model to predict a representation of the channel at a future time.

[0219] Paragraph 22. A method according to Paragraph 21, comprising performing measurements on CSI-RS received at the future time, processing the measurements performed on the CSI-RS received at the future time to obtain an actual representation of the channel at the future time, and determining a quality metric of the Al model based on a comparison of the predicted representation of the channel at the future time and the actual representation of the channel at the future time.

[0220] Paragraph 23. A method according to Paragraph 21 or Paragraph 22, comprising performing measurements on CSI-RS received at the future time, deconvolving the predicted representation of the channel at the future time to obtain predicted measurements on CSI-RS to be received at the future time, and determining a quality metric of the Al model based on a comparison of the predicted measurements on the CSI-RS to be received at the future time and the performed measurements on the CSI-RS received at the future time.

[0221] Paragraph 24. A method according to any of Paragraphs 19 to 23, wherein the indicated configuration indicates that the communications device is to perform measurements on the CSI-RS that are to be received at the future time, wherein the measurements to be performed on the CSI-RS that are to be received at the future time are to be used as a ground truth for the Al model.

[0222] Paragraph 25. A method according to any of Paragraphs 1 to 24, wherein the indicated configuration indicates that the communications device is to transmit an indication of the performed measurements to the infrastructure equipment, and wherein the method comprises transmitting, to the infrastructure equipment, the indication of the performed measurements. Paragraph 26. A method according to Paragraph 25, wherein the indication of the performed measurements comprises measurements performed on both of a first subset and a second subset of a plurality of beams radiated by the infrastructure equipment in different spatial directions. Paragraph 27. A method according to Paragraph 25 or Paragraph 26, wherein the indication of the performed measurements comprises one or more CSI-RS measurements performed on one or more reference signals received by the communications device from the infrastructure equipment. Paragraph 28. A method according to any of Paragraphs 25 to 27, wherein the indication of the performed measurements comprises a representation of the channel, the representation of the channel being obtained by processing one or more CSI-RS measurements performed on one or more reference signals received by the communications device from the infrastructure equipment. Paragraph 29. A method according to any of Paragraphs 25 to 28, wherein the indication of the performed measurements comprises indices of a best one or more of a plurality of beams radiated by the infrastructure equipment in different spatial directions, wherein the one or more best beams are determined by the communications device through use of the Al model. Paragraph 30. A method according to Paragraph 29, comprising receiving, from the infrastructure equipment, an indication of the number of the one or more best beams to be determined by the communications device through use of the Al model. Paragraph 31. A method according to Paragraph 29 or Paragraph 30, wherein the indication of the performed measurements comprises an indication of a predicted quality of the one or more best beams. Paragraph 32. A method according to any of Paragraphs 29 to 31, wherein the one or more best beams form part of a first subset of the plurality of beams, and wherein the indication of the performed measurements comprises an indication of a measured quality of one or more beams of a second subset of the plurality of beams, the one or more beams of the second subset having indices corresponding to the indices of the one or more best beams.

[0223] Paragraph 33. A method according to any of Paragraphs 29 to 32, wherein the one or more best beams form part of a first subset of the plurality of beams, and wherein the indication of the performed measurements comprises an indication of a quality metric of the Al model, the quality metric of the Al model being determined by the communications device based on a comparison of a predicted quality of the one or more best beams and a measured quality of one or more beams of a second subset of the plurality of beams, the one or more beams of the second subset having indices corresponding to the indices of the one or more best beams.

[0224] Paragraph 34. A method according to any of Paragraphs 29 to 33, wherein the one or more best beams form part of a first subset of the plurality of beams, and wherein the indication of the performed measurements comprises an indication of a comparison of a quality metric of the Al model with a preconfigured threshold, the quality metric of the Al model being determined by the communications device based on a comparison of a predicted quality of the one or more best beams and a measured quality of one or more beams of a second subset of the plurality of beams, the one or more beams of the second subset having indices corresponding to the indices of the one or more best beams.

[0225] Paragraph 35. A method according to any of Paragraphs 25 to 34, wherein the indication of the performed measurements comprises an indication of one or more predicted CSI-RS values, wherein the one or more predicted CSI-RS values are determined by the communications device through use of the Al model.

[0226] Paragraph 36. A method according to Paragraph 35, wherein the one or more predicted CSI-RS values comprises an indication of a time at which each of the predicted CSI-RS values are predicted.

[0227] Paragraph 37. A method according to Paragraph 35 or Paragraph 36, wherein the indication of the performed measurements comprises an indication of a quality metric of the Al model, the quality metric of the Al model being determined by the communications device based on a comparison of one or more measured CSI-RS values and the one or more predicted CSI-RS values.

[0228] Paragraph 38. A method according to any of Paragraphs 35 to 37, wherein the indication of the performed measurements comprises an indication of a comparison of a quality metric of the Al model with a preconfigured threshold, the quality metric of the Al model being determined by the communications device based on a comparison of one or more measured CSI-RS values and the one or more predicted CSI-RS values.

[0229] Paragraph 39. A method according to any of Paragraphs 1 to 38, wherein the indicated configuration further indicates that the communications device is to perform one or more measurements that are to be used as a ground truth for the Al model.

[0230] Paragraph 40. A method according to any of Paragraphs 1 to 39, comprising receiving, by the communications device, an indication of a second configuration in accordance with which the communications device is to perform one or more measurements that are to be used as a ground truth for the Al model.

[0231] Paragraph 41. A communications device comprising a transceiver configured to transmit signals to and / or to receive signals from an infrastructure equipment via a channel between the communications device and the infrastructure equipment, the infrastructure equipment forming part of a wireless communications network, a controller configured in combination with the transceiver to receive, from the infrastructure equipment, an indication of a configuration in accordance with which the communications device is to perform one or more measurements, and to perform the measurements in accordance with the indicated configuration, wherein the performed measurements are to be used as inputs into an artificial intelligence, Al, model as part of a training procedure, an inference procedure, or a monitoring procedure, the Al model being for use by the communications device in performing a prediction task.

[0232] Paragraph 42. Circuitry for a communications device, the circuitry comprising transceiver circuitry configured to transmit signals to and / or to receive signals from an infrastructure equipment via a channel between the communications device and the infrastructure equipment, the infrastructure equipment forming part of a wireless communications network, controller circuitry configured in combination with the transceiver circuitry to receive, from the infrastructure equipment, an indication of a configuration in accordance with which the communications device is to perform one or more measurements, and to perform the measurements in accordance with the indicated configuration, wherein the performed measurements are to be used as inputs into an artificial intelligence, Al, model as part of a training procedure, an inference procedure, or a monitoring procedure, the Al model being for use by the communications device in performing a prediction task.

[0233] Paragraph 43. A method of operating an infrastructure equipment forming part of a wireless communications network configured to transmit signals to and / or to receive signals from a communications device via a channel between the communications device and the infrastructure equipment, the method comprising transmitting, to the communications device, an indication of a configuration in accordance with which the communications device is to perform one or more measurements, and wherein the measurements are to be used as inputs into an artificial intelligence, Al, model as part of a training procedure, an inference procedure, or a monitoring procedure, the Al model being for use by the communications device in performing a prediction task.

[0234] Paragraph 44. A method according to Paragraph 43, wherein the indicated configuration indicates that the communications device is to perform the one or more measurements on a configured set of synchronisation signal blocks, SSBs, the set of SSBs being broadcast by the infrastructure equipment using at least some of a plurality of beams radiated by the infrastructure equipment in different spatial directions, wherein each of the plurality of beams forms part of either a first subset or a second subset. Paragraph 45. A method according to Paragraph 44, wherein the indicated configuration indicates a measurement time window during which measurements are to be performed both on SSBs broadcast using beams of the first subset and on SSBs broadcast using beams of the second subset, the measurement time window being configured based on a period over which the set of SSBs are received.

[0235] Paragraph 46. A method according to Paragraph 44 or Paragraph 45, wherein the indicated configuration indicates a time offset between the performance of the measurements and the inputting of the measurements performed on SSBs broadcast using beams of the first subset into the Al model in order to predict a quality of the beams of the second subset at a future time as the prediction task.

[0236] Paragraph 47. A method according to Paragraph 46, wherein the indicated configuration indicates that the measurements performed on SSBs broadcast using beams of the second subset are to be used as a ground truth for the Al model.

[0237] Paragraph 48. A method according to any of Paragraphs 44 to 47, wherein the prediction task comprises predicting a quality of beams of the second subset using measurements performed on beams of the first subset.

[0238] Paragraph 49. A method according to Paragraph 48, wherein the indicated configuration indicates that the communications device is to perform measurements on SSBs broadcast using the beams of the first subset and the beams of the second subset during a single burst period of the received set of SSBs. Paragraph 50. A method according to Paragraph 48 or Paragraph 49, wherein the indicated configuration indicates a time offset between a first time at which the communications device is to perform measurements on SSBs broadcast using the beams of the first subset and a second time at which the communications device is to perform measurements on SSBs broadcast using the beams of the second subset.

[0239] Paragraph 51. A method according to any of Paragraphs 43 to 50, wherein the indicated configuration indicates that the communications device is to perform the one or more measurements on channel state information reference signals, CSI-RS, the CSI-RS being transmitted by the infrastructure equipment to the communications device using at least some of a plurality of beams radiated by the infrastructure equipment in different spatial directions, wherein each of the plurality of beams forms part of either a first subset or a second subset.

[0240] Paragraph 52. A method according to Paragraph 51, wherein the prediction task comprises predicting a quality of the beams of the second subset at a future time using measurements performed on the beams of the first subset, and wherein the CSI-RS are transmitted by the infrastructure equipment in accordance with a configured periodicity.

[0241] Paragraph 53. A method according to Paragraph 52, wherein the CSI-RS are transmitted by the infrastructure equipment in accordance with the configured periodicity upon the infrastructure equipment transmitting an activation signal to the communications device.

[0242] Paragraph 54. A method according to Paragraph 52 or Paragraph 53, wherein the indicated configuration indicates a measurement time window during which measurements are to be performed both on CSI-RS that are transmitted using beams of the first subset and on CSI-RS that are transmitted using beams of the second subset, the measurement time window being configured to match a specified number of periods of the CSI-RS.

[0243] Paragraph 55. A method according to any of Paragraphs 52 to 54, wherein a measurement time window during which measurements are to be performed both on CSI-RS that are transmitted using beams of the first subset and on CSI-RS that are transmitted using beams of the second subset is preconfigured and known to the communications device.

[0244] Paragraph 56. A method according to any of Paragraphs 51 to 55, wherein the prediction task comprises predicting a quality of the beams of the second subset at a future time using measurements performed on the beams of the first subset, and wherein the CSI-RS are dynamically scheduled by the infrastructure equipment and are transmitted by the infrastructure equipment to the communications device in accordance with the dynamic scheduling.

[0245] Paragraph 57. A method according to any of Paragraphs 51 to 56, wherein the indicated configuration indicates that measurements performed on a first set of the CSI-RS that are received by the communications device at a first time using beams of the first subset are to be input into the Al model in order to predict a quality of the beams of the second subset at a future time as the prediction task, and that measurements performed on a second set of the CSI-RS that are received by the communications device at a second time using beams of the second subset are to be used as a ground truth for the Al model.

[0246] Paragraph 58. A method according to any of Paragraphs 51 to 57, wherein the prediction task comprises predicting a quality of beams of the second subset using measurements performed on beams of the first subset, and wherein the indicated configuration indicates a measurement time window during which the communications device can perform measurements on CSI-RS that are received using the beams of the first subset and the beams of the second subset.

[0247] Paragraph 59. A method according to any of Paragraphs 43 to 58, wherein the indicated configured indicates that the communications device is to perform the one or more measurements on channel state information reference signals, CSI-RS, the CSI-RS being transmitted by the infrastructure equipment to the communications device, and wherein the prediction task comprises predicting a quality of CSI-RS that are to be received at the communications device from the infrastructure equipment at a future time using the measurements performed on the received CSI-RS.

[0248] Paragraph 60. A method according to Paragraph 59, wherein the indicated configuration indicates a measurement time window during which both the measurements are performed on the transmitted CSI- RS and measurements are to be performed on the CSI-RS that are to be transmitted at the future time, the future time being within the measurement time window.

[0249] Paragraph 61. A method according to Paragraph 60, wherein the measurements performed on the transmitted CSI-RS are to be input into the Al model as raw CSI-RS measurements.

[0250] Paragraph 62. A method according to Paragraph 60 or Paragraph 61, wherein the measurements performed on the transmitted CSI-RS are to be processed to obtain a representation of the channel which is to be input into the Al model to predict a representation of the channel at a future time.

[0251] Paragraph 63. A method according to any of Paragraphs 60 to 62, wherein the indicated configuration indicates that the communications device is to perform measurements on the CSI-RS that are to be received at the future time, wherein the measurements to be performed on the CSI-RS that are to be received at the future time are to be used as a ground truth for the Al model.

[0252] Paragraph 64. A method according to any of Paragraphs 43 to 63, wherein the indicated configuration indicates that the communications device is to transmit an indication of the performed measurements to the infrastructure equipment, and wherein the method comprises receiving, from the communications device, the indication of the performed measurements.

[0253] Paragraph 65. A method according to Paragraph 64, comprising inputting the indicated performed measurements into the Al model as part of the training procedure, or using the indicated performed measurements as part of the monitoring procedure.

[0254] Paragraph 66. A method according to Paragraph 64 or Paragraph 65, wherein the indication of the performed measurements comprises measurements performed on both of a first subset and a second subset of a plurality of beams radiated by the infrastructure equipment in different spatial directions. Paragraph 67. A method according to any of Paragraphs 64 to 66, wherein the indication of the performed measurements comprises one or more CSI-RS measurements performed on one or more reference signals received by the communications device from the infrastructure equipment.

[0255] Paragraph 68. A method according to any of Paragraphs 64 to 67, wherein the indication of the performed measurements comprises a representation of the channel, the representation of the channel being obtained by processing one or more CSI-RS measurements performed on one or more reference signals received by the communications device from the infrastructure equipment.

[0256] Paragraph 69. A method according to any of Paragraphs 64 to 68, wherein the indication of the performed measurements comprises indices of a best one or more of a plurality of beams radiated by the infrastructure equipment in different spatial directions, wherein the one or more best beams are determined by the communications device through use of the Al model.

[0257] Paragraph 70. A method according to Paragraph 69, comprising transmitting, to the communications device, an indication of the number of the one or more best beams to be determined by the communications device through use of the Al model.

[0258] Paragraph 71. A method according to Paragraph 69 or Paragraph 70, wherein the indication of the performed measurements comprises an indication of a predicted quality of the one or more best beams. Paragraph 72. A method according to any of Paragraphs 69 to 71, wherein the one or more best beams form part of a first subset of the plurality of beams, and wherein the indication of the performed measurements comprises an indication of a measured quality of one or more beams of a second subset of the plurality of beams, the one or more beams of the second subset having indices corresponding to the indices of the one or more best beams.

[0259] Paragraph 73. A method according to Paragraph 72, comprising determining a quality metric of the Al model based on a comparison of a predicted quality of the one or more best beams and the indicated measured quality of the one or more beams of the second subset, and comparing the determined quality metric of the Al model with a preconfigured threshold. Paragraph 74. A method according to any of Paragraphs 69 to 73, wherein the one or more best beams form part of a first subset of the plurality of beams, and wherein the indication of the performed measurements comprises an indication of a quality metric of the Al model, the quality metric of the Al model being determined by the communications device based on a comparison of a predicted quality of the one or more best beams and a measured quality of one or more beams of a second subset of the plurality of beams, the one or more beams of the second subset having indices corresponding to the indices of the one or more best beams.

[0260] Paragraph 75. A method according to Paragraph 74, comprising comparing the indicated quality metric of the Al model with a preconfigured threshold.

[0261] Paragraph 76. A method according to any of Paragraphs 69 to 75, wherein the one or more best beams form part of a first subset of the plurality of beams, and wherein the indication of the performed measurements comprises an indication of a comparison of a quality metric of the Al model with a preconfigured threshold, the quality metric of the Al model being determined by the communications device based on a comparison of a predicted quality of the one or more best beams and a measured quality of one or more beams of a second subset of the plurality of beams, the one or more beams of the second subset having indices corresponding to the indices of the one or more best beams.

[0262] Paragraph 77. A method according to any of Paragraphs 64 to 76, wherein the indication of the performed measurements comprises an indication of one or more predicted CSI-RS values, wherein the one or more predicted CSI-RS values are determined by the communications device through use of the Al model.

[0263] Paragraph 78. A method according to Paragraph 77, wherein the one or more predicted CSI-RS values comprises an indication of a time at which each of the predicted CSI-RS values are predicted.

[0264] Paragraph 79. A method according to Paragraph 77 or Paragraph 78, wherein the indication of the performed measurements comprises an indication of a quality metric of the Al model, the quality metric of the Al model being determined by the communications device based on a comparison of one or more measured CSI-RS values and the one or more predicted CSI-RS values.

[0265] Paragraph 80. A method according to Paragraph 79, comprising comparing the indicated quality metric of the Al model with a preconfigured threshold.

[0266] Paragraph 81. A method according to any of Paragraphs 77 to 80, wherein the indication of the performed measurements comprises an indication of a comparison of a quality metric of the Al model with a preconfigured threshold, the quality metric of the Al model being determined by the communications device based on a comparison of one or more measured CSI-RS values and the one or more predicted CSI-RS values.

[0267] Paragraph 82. A method according to any of Paragraphs 43 to 81, wherein the indicated configuration further indicates that the communications device is to perform one or more measurements that are to be used as a ground truth for the Al model.

[0268] Paragraph 83. A method according to any of Paragraphs 43 to 82, comprising transmitting, to the communications device, an indication of a second configuration in accordance with which the communications device is to perform one or more measurements that are to be used as a ground truth for the Al model.

[0269] Paragraph 84. An infrastructure equipment forming part of a wireless communications network, the infrastructure equipment comprising a transceiver configured to transmit signals to and / or to receive signals from a communications device via a channel between the communications device and the infrastructure equipment, and a controller configured in combination with the transceiver to transmit, to the communications device, an indication of a configuration in accordance with which the communications device is to perform one or more measurements, and wherein the measurements are to be used as inputs into an artificial intelligence, Al, model as part of a training procedure, an inference procedure, or a monitoring procedure, the Al model being for use by the communications device in performing a prediction task.

[0270] Paragraph 85. Circuitry for an infrastructure equipment forming part of a wireless communications network, the circuitry comprising transceiver circuitry configured to transmit signals to and / or to receive signals from a communications device via a channel between the communications device and the infrastructure equipment, and controller circuitry configured in combination with the transceiver circuitry to transmit, to the communications device, an indication of a configuration in accordance with which the communications device is to perform one or more measurements, and wherein the measurements are to be used as inputs into an artificial intelligence, Al, model as part of a training procedure, an inference procedure, or a monitoring procedure, the Al model being for use by the communications device in performing a prediction task.

[0271] Paragraph 86. A wireless communications system comprising a communications device according to Paragraph 41 and an infrastructure equipment according to Paragraph 84.

[0272] Paragraph 87. A computer program comprising instructions which, when loaded onto a computer, cause the computer to perform a method according to any of Paragraphs 1 to 40 or Paragraphs 43 to 83.

[0273] Paragraph 88. A non-transitory computer-readable storage medium storing a computer program according to Paragraph 87.

[0274] It will be appreciated that the above description for clarity has described embodiments with reference to different functional units, circuitry and / or processors. However, it will be apparent that any suitable distribution of functionality between different functional units, circuitry and / or processors may be used without detracting from the embodiments.

[0275] Described embodiments may be implemented in any suitable form including hardware, software, firmware or any combination of these. Described embodiments may optionally be implemented at least partly as computer software running on one or more data processors and / or digital signal processors. The elements and components of any embodiment may be physically, functionally and logically implemented in any suitable way. Indeed, the functionality may be implemented in a single unit, in a plurality of units or as part of other functional units. As such, the disclosed embodiments may be implemented in a single unit or may be physically and functionally distributed between different units, circuitry and / or processors.

[0276] Although the present disclosure has been described in connection with some embodiments, it is not intended to be limited to the specific form set forth herein. Additionally, although a feature may appear to be described in connection with particular embodiments, one skilled in the art would recognise that various features of the described embodiments may be combined in any manner suitable to implement the technique.

[0277] References

[0278] [1] Holma H. and Toskala A, “LTE for UMTS OFDMA and SC-FDMA based radio access”, John Wiley and Sons, 2009.

[0279] [2] TR 38.913, “3rdGeneration Partnership Project; Technical Specification Group Radio Access Network; Study on Scenarios and Requirements for Next Generation Access Technologies (Release 14)”, 3GPP, vl4.3.0, August 2017.

[0280] [3] TR 38.843, “Study on Artificial Intelligence (AI)ZMachine Learning (ML) for NR air interface (Release 18)”, 3GPP, vl8.0.0, December 2023.

[0281] [4] European patent application number EP24175304.5.

Claims

CLAIMSWhat is claimed is:

1. A method of operating a communications device configured to transmit signals to and / or to receive signals from an infrastructure equipment via a channel between the communications device and the infrastructure equipment, the infrastructure equipment forming part of a wireless communications network, the method comprising receiving, from the infrastructure equipment, an indication of a configuration in accordance with which the communications device is to perform one or more measurements, and performing the measurements in accordance with the indicated configuration, wherein the performed measurements are to be used as inputs into an artificial intelligence, Al, model as part of a training procedure, an inference procedure, or a monitoring procedure, the Al model being for use by the communications device in performing a prediction task.

2. A method according to Claim 1, comprising inputting the performed measurements into the Al model as part of the training procedure, the inference procedure, or the monitoring procedure.

3. A method according to Claim 1, wherein the indicated configuration indicates that the communications device is to perform the one or more measurements on a configured set of synchronisation signal blocks, SSBs, the set of SSBs being received by the communications device as a broadcast from the infrastructure equipment using at least some of a plurality of beams radiated by the infrastructure equipment in different spatial directions, wherein each of the plurality of beams forms part of either a first subset or a second subset.

4. A method according to Claim 3, wherein the indicated configuration indicates a measurement time window during which measurements are to be performed both on SSBs broadcast using beams of the first subset and on SSBs broadcast using beams of the second subset, the measurement time window being configured based on a period over which the set of SSBs are received.

5. A method according to Claim 3, wherein the indicated configuration indicates a time offset between the performance of the measurements and the inputting of the measurements performed on SSBs broadcast using beams of the first subset into the Al model in order to predict a quality of the beams of the second subset at a future time as the prediction task.

6. A method according to Claim 5, wherein the indicated configuration indicates that the measurements performed on SSBs broadcast using beams of the second subset are to be used as a ground truth for the Al model.

7. A method according to Claim 3, wherein the prediction task comprises predicting a quality of beams of the second subset using measurements performed on beams of the first subset.

8. A method according to Claim 7, wherein the indicated configuration indicates that the communications device is to perform measurements on SSBs broadcast using the beams of the first subset and the beams of the second subset during a single burst period of the received set of SSBs.

9. A method according to Claim 7, wherein the indicated configuration indicates a time offset between a first time at which the communications device is to perform measurements on SSBs broadcastusing the beams of the first subset and a second time at which the communications device is to perform measurements on SSBs broadcast using the beams of the second subset.

10. A method according to Claim 1, wherein the indicated configuration indicates that the communications device is to perform the one or more measurements on channel state information reference signals, CSI-RS, the CSI-RS being received by the communications device from the infrastructure equipment using at least some of a plurality of beams radiated by the infrastructure equipment in different spatial directions, wherein each of the plurality of beams forms part of either a first subset or a second subset.

11. A method according to Claim 10, wherein the prediction task comprises predicting a quality of the beams of the second subset at a future time using measurements performed on the beams of the first subset, and wherein the CSI-RS are received by the communications device from the infrastructure equipment in accordance with a configured periodicity.

12. A method according to Claim 11, wherein the CSI-RS are received by the communications device from the infrastructure equipment in accordance with the configured periodicity upon the communications device receiving an activation signal from the infrastructure equipment.

13. A method according to Claim 11, wherein the indicated configuration indicates a measurement time window during which measurements are to be performed on CSI-RS that are received using beams of the first subset and on CSI-RS that are received using beams of the second subset, the measurement time window being configured to match a specified number of periods of the CSI-RS.

14. A method according to Claim 11, wherein a measurement time window during which measurements are to be performed both on CSI-RS that are received using beams of the first subset and on CSI-RS that are received using beams of the second subset is preconfigured and known to the communications device.

15. A method according to Claim 10, wherein the prediction task comprises predicting a quality of the beams of the second subset at a future time using measurements performed on the beams of the first subset, and wherein the CSI-RS are dynamically scheduled by the infrastructure equipment and are received by the communications device from the infrastructure equipment in accordance with the dynamic scheduling.

16. A method according to Claim 10, wherein the indicated configuration indicates that measurements performed on a first set of the CSI-RS that are received at a first time using beams of the first subset are to be input into the Al model in order to predict a quality of the beams of the second subset at a future time as the prediction task, and that measurements performed on a second set of the CSI-RS that are received at a second time using beams of the second subset are to be used as a ground truth for the Al model.

17. A method according to Claim 10, wherein the prediction task comprises predicting a quality of beams of the second subset using measurements performed on beams of the first subset, and wherein the indicated configuration indicates a measurement time window during which the communications device can perform measurements on CSI-RS that are received using the beams of the first subset and the beams of the second subset.

18. A method according to Claim 1, wherein the indicated configuration indicates that the communications device is to perform the one or more measurements on channel state information reference signals, CSI-RS, the CSI-RS being received by the communications device from the infrastructure equipment, and wherein the prediction task comprises predicting a quality of CSI-RS that are to be received from the infrastructure equipment at a future time using the measurements performed on the received CSI-RS.

19. A method according to Claim 18, wherein the indicated configuration indicates a measurement time window during which both the measurements are performed on the received CSI-RS and measurements are to be performed on the CSI-RS that are to be received at the future time, the future time being within the measurement time window.

20. A method according to Claim 19, wherein the measurements performed on the received CSI-RS are to be input into the Al model as raw CSI-RS measurements.

21. A method according to Claim 19, wherein the measurements performed on the received CSI-RS are to be processed to obtain a representation of the channel which is to be input into the Al model to predict a representation of the channel at a future time.

22. A method according to Claim 21, comprising performing measurements on CSI-RS received at the future time, processing the measurements performed on the CSI-RS received at the future time to obtain an actual representation of the channel at the future time, and determining a quality metric of the Al model based on a comparison of the predicted representation of the channel at the future time and the actual representation of the channel at the future time.

23. A method according to Claim 21 , comprising performing measurements on CSI-RS received at the future time, deconvolving the predicted representation of the channel at the future time to obtain predicted measurements on CSI-RS to be received at the future time, and determining a quality metric of the Al model based on a comparison of the predicted measurements on the CSI-RS to be received at the future time and the performed measurements on the CSI-RS received at the future time.

24. A method according to Claim 19, wherein the indicated configuration indicates that the communications device is to perform measurements on the CSI-RS that are to be received at the future time, wherein the measurements to be performed on the CSI-RS that are to be received at the future time are to be used as a ground truth for the Al model.

25. A method according to Claim 1, wherein the indicated configuration indicates that the communications device is to transmit an indication of the performed measurements to the infrastructure equipment, and wherein the method comprises transmitting, to the infrastructure equipment, the indication of the performed measurements.

26. A method according to Claim 25, wherein the indication of the performed measurements comprises measurements performed on both of a first subset and a second subset of a plurality of beams radiated by the infrastructure equipment in different spatial directions.

27. A method according to Claim 25, wherein the indication of the performed measurements comprises one or more CSI-RS measurements performed on one or more reference signals received by the communications device from the infrastructure equipment.

28. A method according to Claim 25, wherein the indication of the performed measurements comprises a representation of the channel, the representation of the channel being obtained by processing one or more CSI-RS measurements performed on one or more reference signals received by the communications device from the infrastructure equipment.

29. A method according to Claim 25, wherein the indication of the performed measurements comprises indices of a best one or more of a plurality of beams radiated by the infrastructure equipment in different spatial directions, wherein the one or more best beams are determined by the communications device through use of the Al model.

30. A method according to Claim 29, comprising receiving, from the infrastructure equipment, an indication of the number of the one or more best beams to be determined by the communications device through use of the Al model.

31. A method according to Claim 29, wherein the indication of the performed measurements comprises an indication of a predicted quality of the one or more best beams.

32. A method according to Claim 29, wherein the one or more best beams form part of a first subset of the plurality of beams, and wherein the indication of the performed measurements comprises an indication of a measured quality of one or more beams of a second subset of the plurality of beams, the one or more beams of the second subset having indices corresponding to the indices of the one or more best beams.

33. A method according to Claim 29, wherein the one or more best beams form part of a first subset of the plurality of beams, and wherein the indication of the performed measurements comprises an indication of a quality metric of the Al model, the quality metric of the Al model being determined by the communications device based on a comparison of a predicted quality of the one or more best beams and a measured quality of one or more beams of a second subset of the plurality of beams, the one or more beams of the second subset having indices corresponding to the indices of the one or more best beams.

34. A method according to Claim 29, wherein the one or more best beams form part of a first subset of the plurality of beams, and wherein the indication of the performed measurements comprises an indication of a comparison of a quality metric of the Al model with a preconfigured threshold, the quality metric of the Al model being determined by the communications device based on a comparison of a predicted quality of the one or more best beams and a measured quality of one or more beams of a second subset of the plurality of beams, the one or more beams of the second subset having indices corresponding to the indices of the one or more best beams.

35. A method according to Claim 25, wherein the indication of the performed measurements comprises an indication of one or more predicted CSI-RS values, wherein the one or more predicted CSI- RS values are determined by the communications device through use of the Al model.

36. A method according to Claim 35, wherein the one or more predicted CSI-RS values comprises an indication of a time at which each of the predicted CSI-RS values are predicted.

37. A method according to Claim 35, wherein the indication of the performed measurements comprises an indication of a quality metric of the Al model, the quality metric of the Al model being determined by the communications device based on a comparison of one or more measured CSI-RS values and the one or more predicted CSI-RS values.

38. A method according to Claim 35, wherein the indication of the performed measurements comprises an indication of a comparison of a quality metric of the Al model with a preconfigured threshold, the quality metric of the Al model being determined by the communications device based on a comparison of one or more measured CSI-RS values and the one or more predicted CSI-RS values.

39. A method according to Claim 1, wherein the indicated configuration further indicates that the communications device is to perform one or more measurements that are to be used as a ground truth for the Al model.

40. A method according to Claim 1, comprising receiving, by the communications device, an indication of a second configuration in accordance with which the communications device is to perform one or more measurements that are to be used as a ground truth for the Al model.

41. A communications device comprising a transceiver configured to transmit signals to and / or to receive signals from an infrastructure equipment via a channel between the communications device and the infrastructure equipment, the infrastructure equipment forming part of a wireless communications network, a controller configured in combination with the transceiver to receive, from the infrastructure equipment, an indication of a configuration in accordance with which the communications device is to perform one or more measurements, and to perform the measurements in accordance with the indicated configuration, wherein the performed measurements are to be used as inputs into an artificial intelligence, Al, model as part of a training procedure, an inference procedure, or a monitoring procedure, the Al model being for use by the communications device in performing a prediction task.

42. Circuitry for a communications device, the circuitry comprising transceiver circuitry configured to transmit signals to and / or to receive signals from an infrastructure equipment via a channel between the communications device and the infrastructure equipment, the infrastructure equipment forming part of a wireless communications network, controller circuitry configured in combination with the transceiver circuitry to receive, from the infrastructure equipment, an indication of a configuration in accordance with which the communications device is to perform one or more measurements, and to perform the measurements in accordance with the indicated configuration, wherein the performed measurements are to be used as inputs into an artificial intelligence, Al, model as part of a training procedure, an inference procedure, or a monitoring procedure, the Al model being for use by the communications device in performing a prediction task.

43. A method of operating an infrastructure equipment forming part of a wireless communications network configured to transmit signals to and / or to receive signals from a communications device via a channel between the communications device and the infrastructure equipment, the method comprising transmitting, to the communications device, an indication of a configuration in accordance with which the communications device is to perform one or more measurements, andwherein the measurements are to be used as inputs into an artificial intelligence, Al, model as part of a training procedure, an inference procedure, or a monitoring procedure, the Al model being for use by the communications device in performing a prediction task.

44. A method according to Claim 43, wherein the indicated configuration indicates that the communications device is to perform the one or more measurements on a configured set of synchronisation signal blocks, SSBs, the set of SSBs being broadcast by the infrastructure equipment using at least some of a plurality of beams radiated by the infrastructure equipment in different spatial directions, wherein each of the plurality of beams forms part of either a first subset or a second subset.

45. A method according to Claim 44, wherein the indicated configuration indicates a measurement time window during which measurements are to be performed both on SSBs broadcast using beams of the first subset and on SSBs broadcast using beams of the second subset, the measurement time window being configured based on a period over which the set of SSBs are received.

46. A method according to Claim 44, wherein the indicated configuration indicates a time offset between the performance of the measurements and the inputting of the measurements performed on SSBs broadcast using beams of the first subset into the Al model in order to predict a quality of the beams of the second subset at a future time as the prediction task.

47. A method according to Claim 46, wherein the indicated configuration indicates that the measurements performed on SSBs broadcast using beams of the second subset are to be used as a ground truth for the Al model.

48. A method according to Claim 44, wherein the prediction task comprises predicting a quality of beams of the second subset using measurements performed on beams of the first subset.

49. A method according to Claim 48, wherein the indicated configuration indicates that the communications device is to perform measurements on SSBs broadcast using the beams of the first subset and the beams of the second subset during a single burst period of the received set of SSBs.

50. A method according to Claim 48, wherein the indicated configuration indicates a time offset between a first time at which the communications device is to perform measurements on SSBs broadcast using the beams of the first subset and a second time at which the communications device is to perform measurements on SSBs broadcast using the beams of the second subset.

51. A method according to Claim 43, wherein the indicated configuration indicates that the communications device is to perform the one or more measurements on channel state information reference signals, CSI-RS, the CSI-RS being transmitted by the infrastructure equipment to the communications device using at least some of a plurality of beams radiated by the infrastructure equipment in different spatial directions, wherein each of the plurality of beams forms part of either a first subset or a second subset.

52. A method according to Claim 51, wherein the prediction task comprises predicting a quality of the beams of the second subset at a future time using measurements performed on the beams of the first subset, and wherein the CSI-RS are transmitted by the infrastructure equipment in accordance with a configured periodicity.

53. A method according to Claim 52, wherein the CSI-RS are transmitted by the infrastructure equipment in accordance with the configured periodicity upon the infrastructure equipment transmitting an activation signal to the communications device.

54. A method according to Claim 52, wherein the indicated configuration indicates a measurement time window during which measurements are to be performed both on CSI-RS that are transmitted using beams of the first subset and on CSI-RS that are transmitted using beams of the second subset, the measurement time window being configured to match a specified number of periods of the CSI-RS.

55. A method according to Claim 52, wherein a measurement time window during which measurements are to be performed both on CSI-RS that are transmitted using beams of the first subset and on CSI-RS that are transmitted using beams of the second subset is preconfigured and known to the communications device.

56. A method according to Claim 51, wherein the prediction task comprises predicting a quality of the beams of the second subset at a future time using measurements performed on the beams of the first subset, and wherein the CSI-RS are dynamically scheduled by the infrastructure equipment and are transmitted by the infrastructure equipment to the communications device in accordance with the dynamic scheduling.

57. A method according to Claim 51, wherein the indicated configuration indicates that measurements performed on a first set of the CSI-RS that are received by the communications device at a first time using beams of the first subset are to be input into the Al model in order to predict a quality of the beams of the second subset at a future time as the prediction task, and that measurements performed on a second set of the CSI-RS that are received by the communications device at a second time using beams of the second subset are to be used as a ground truth for the Al model.

58. A method according to Claim 51, wherein the prediction task comprises predicting a quality of beams of the second subset using measurements performed on beams of the first subset, and wherein the indicated configuration indicates a measurement time window during which the communications device can perform measurements on CSI-RS that are received using the beams of the first subset and the beams of the second subset.

59. A method according to Claim 43, wherein the indicated configured indicates that the communications device is to perform the one or more measurements on channel state information reference signals, CSI-RS, the CSI-RS being transmitted by the infrastructure equipment to the communications device, and wherein the prediction task comprises predicting a quality of CSI-RS that are to be received at the communications device from the infrastructure equipment at a future time using the measurements performed on the received CSI-RS.

60. A method according to Claim 59, wherein the indicated configuration indicates a measurement time window during which both the measurements are performed on the transmitted CSI-RS and measurements are to be performed on the CSI-RS that are to be transmitted at the future time, the future time being within the measurement time window.

61. A method according to Claim 60, wherein the measurements performed on the transmitted CSI- RS are to be input into the Al model as raw CSI-RS measurements.

62. A method according to Claim 60, wherein the measurements performed on the transmitted CSI- RS are to be processed to obtain a representation of the channel which is to be input into the Al model to predict a representation of the channel at a future time.

63. A method according to Claim 60, wherein the indicated configuration indicates that the communications device is to perform measurements on the CSI-RS that are to be received at the future time, wherein the measurements to be performed on the CSI-RS that are to be received at the future time are to be used as a ground truth for the Al model.

64. A method according to Claim 43, wherein the indicated configuration indicates that the communications device is to transmit an indication of the performed measurements to the infrastructure equipment, and wherein the method comprises receiving, from the communications device, the indication of the performed measurements.

65. A method according to Claim 64, comprising inputting the indicated performed measurements into the Al model as part of the training procedure, or using the indicated performed measurements as part of the monitoring procedure.

66. A method according to Claim 64, wherein the indication of the performed measurements comprises measurements performed on both of a first subset and a second subset of a plurality of beams radiated by the infrastructure equipment in different spatial directions.

67. A method according to Claim 64, wherein the indication of the performed measurements comprises one or more CSI-RS measurements performed on one or more reference signals received by the communications device from the infrastructure equipment.

68. A method according to Claim 64, wherein the indication of the performed measurements comprises a representation of the channel, the representation of the channel being obtained by processing one or more CSI-RS measurements performed on one or more reference signals received by the communications device from the infrastructure equipment.

69. A method according to Claim 64, wherein the indication of the performed measurements comprises indices of a best one or more of a plurality of beams radiated by the infrastructure equipment in different spatial directions, wherein the one or more best beams are determined by the communications device through use of the Al model.

70. A method according to Claim 69, comprising transmitting, to the communications device, an indication of the number of the one or more best beams to be determined by the communications device through use of the Al model.

71. A method according to Claim 69, wherein the indication of the performed measurements comprises an indication of a predicted quality of the one or more best beams.

72. A method according to Claim 69, wherein the one or more best beams form part of a first subset of the plurality of beams, and wherein the indication of the performed measurements comprises an indication of a measured quality of one or more beams of a second subset of the plurality of beams, the one or more beams of the second subset having indices corresponding to the indices of the one or more best beams.

73. A method according to Claim 72, comprising determining a quality metric of the Al model based on a comparison of a predicted quality of the one or more best beams and the indicated measured quality of the one or more beams of the second subset, and comparing the determined quality metric of the Al model with a preconfigured threshold.

74. A method according to Claim 69, wherein the one or more best beams form part of a first subset of the plurality of beams, and wherein the indication of the performed measurements comprises an indication of a quality metric of the Al model, the quality metric of the Al model being determined by the communications device based on a comparison of a predicted quality of the one or more best beams and a measured quality of one or more beams of a second subset of the plurality of beams, the one or more beams of the second subset having indices corresponding to the indices of the one or more best beams.

75. A method according to Claim 74, comprising comparing the indicated quality metric of the Al model with a preconfigured threshold.

76. A method according to Claim 69, wherein the one or more best beams form part of a first subset of the plurality of beams, and wherein the indication of the performed measurements comprises an indication of a comparison of a quality metric of the Al model with a preconfigured threshold, the quality metric of the Al model being determined by the communications device based on a comparison of a predicted quality of the one or more best beams and a measured quality of one or more beams of a second subset of the plurality of beams, the one or more beams of the second subset having indices corresponding to the indices of the one or more best beams.

77. A method according to Claim 64, wherein the indication of the performed measurements comprises an indication of one or more predicted CSI-RS values, wherein the one or more predicted CSI- RS values are determined by the communications device through use of the Al model.

78. A method according to Claim 77, wherein the one or more predicted CSI-RS values comprises an indication of a time at which each of the predicted CSI-RS values are predicted.

79. A method according to Claim 77, wherein the indication of the performed measurements comprises an indication of a quality metric of the Al model, the quality metric of the Al model being determined by the communications device based on a comparison of one or more measured CSI-RS values and the one or more predicted CSI-RS values.

80. A method according to Claim 79, comprising comparing the indicated quality metric of the Al model with a preconfigured threshold.

81. A method according to Claim 77, wherein the indication of the performed measurements comprises an indication of a comparison of a quality metric of the Al model with a preconfigured threshold, the quality metric of the Al model being determined by the communications device based on a comparison of one or more measured CSI-RS values and the one or more predicted CSI-RS values.

82. A method according to Claim 43, wherein the indicated configuration further indicates that the communications device is to perform one or more measurements that are to be used as a ground truth for the Al model.

83. A method according to Claim 43, comprising transmitting, to the communications device, an indication of a second configuration in accordance with which the communications device is to perform one or more measurements that are to be used as a ground truth for the Al model.

84. An infrastructure equipment forming part of a wireless communications network, the infrastructure equipment comprising a transceiver configured to transmit signals to and / or to receive signals from a communications device via a channel between the communications device and the infrastructure equipment, and a controller configured in combination with the transceiver to transmit, to the communications device, an indication of a configuration in accordance with which the communications device is to perform one or more measurements, and wherein the measurements are to be used as inputs into an artificial intelligence, Al, model as part of a training procedure, an inference procedure, or a monitoring procedure, the Al model being for use by the communications device in performing a prediction task.

85. Circuitry for an infrastructure equipment forming part of a wireless communications network, the circuitry comprising transceiver circuitry configured to transmit signals to and / or to receive signals from a communications device via a channel between the communications device and the infrastructure equipment, and controller circuitry configured in combination with the transceiver circuitry to transmit, to the communications device, an indication of a configuration in accordance with which the communications device is to perform one or more measurements, and wherein the measurements are to be used as inputs into an artificial intelligence, Al, model as part of a training procedure, an inference procedure, or a monitoring procedure, the Al model being for use by the communications device in performing a prediction task.

86. A wireless communications system comprising a communications device according to Claim 41 and an infrastructure equipment according to Claim 84.

87. A computer program comprising instructions which, when loaded onto a computer, cause the computer to perform a method according to Claim 1 or Claim 43.

88. A non-transitory computer-readable storage medium storing a computer program according to Claim 87.

Images