Method and apparatus for determining a pilot signal pattern in a mobile network

Abstract

Described is a computing entity (203, 301, 501, 601, 701) for use in a mobile network (200, 300, 500, 600, 700, 800), the mobile network comprising a radio access network entity (201, 302, 502, 602, 702, 801) in communication with a user equipment device (202, 303, 503, 603, 703, 802), the computing entity being configured to: obtain (901) context information for the user equipment device; in dependence on the context information, consult (902) an address at a memory accessible to the computing entity containing one or more stored measurements; and in dependence on the stored measurements, determine (903) a pilot pattern in the frequency domain for uplink communications between the user equipment device and the radio access network entity. This may allow for the generation of a pilot pattern for uplink pilots in a frequency resources grid in a way that reduces the utilization of uplink signalling resources.

Description

[0001] METHOD AND APPARATUS FOR DETERMINING A PILOT SIGNAL PATTERN IN A MOBILE NETWORK

[0002] TECHNICAL FIELD

[0003] Embodiments of the present disclosure generally relate to the field of mobile networks, in particular to the sensing of pilot signals, such as sounding reference signals, between a user equipment device and a base station.

[0004] BACKGROUND

[0005] Sounding reference signals (SRS) are pilot signals sent in the uplink (UL) signals by the user equipment (UE). They allow a radio access network (RAN) entity, such as a base station (BS), to estimate the channel to perform downlink (DL) precoding and scheduling.

[0006] Figure 1 shows a typical scheme for the transmission of SRS pilots in the uplink in time division duplex (TDD) systems. In this example, a mobile network 100 comprises aBS 101 and aUE 102. The UE 102 can send UL reference signals L UL, which may be SRS signals, to the BS 101, as shown at 103. The BS 102 estimates the UL channel based on the UL reference signal. The BS 102 obtains DL channel state information (CSI) / / DI. based on channel reciprocity. The BS 102 completes DL scheduling and transmits data 104 based on the DL CSI.

[0007] One issue with SRS is that a UE has power limitations to transmit SRS across the whole frequency band, in particular for upcoming wideband solutions. The density of frequency sampling can be referred to as the comb. In the sense of SRS, a comb is referring to equally spaced signals in the frequency domain (thus resembling a brushing comb with equally spaced spikes), to allow for uniform sampling of the frequency domain. Higher comb width decreases the density of frequency sampling while lowering it increases the frequency information collected by the SRS signal. Thus, adjusting the frequency sampling affects the accuracy of the SRS measured. It is also important to note that adjusting the comb width can have an impact in the form of the number of available positions for other UEs to transmit their SRS without interfering.

[0008] A radiomap, is a representation of the radio frequency (RF) environment within a specific area. Radiomaps are instrumental in planning and optimizing wireless networks, ensuring reliable communication, and enhancing the efficiency of various RF- dependent applications. Traditionally, radiomaps have been used to keep signal strength, interference, localization and frequency usage information for the locations serviced by that BS. The radiomap can be stored at the BS to be useful in the event that it needs to perform various downlink (DL) operations when it has limited information provided from the UE.

[0009] Modem and advanced solutions can enable the storage of CSI with the possibility of using various compression methods. Such CSI information is useful in addressing the absence of SRS. However, radiomaps themselves can suffer from many issues when aiming to substitute SRS and are for this reason not generally useful in current implementations for DL precoding and scheduling.

[0010] It is desirable to develop an approach that may overcome at least some of the above issues.

[0011] SUMMARY

[0012] According to a first aspect, there is provided a computing entity for use in a mobile network, the mobile network comprising a radio access network entity in communication with a user equipment device, the computing entity being configured to: obtain context information for the user equipment device; in dependence on the context information, consult an address at a memory accessible to the computing entity containing one or more stored measurements; and in dependence on the stored measurements, determine a pilot pattern in the frequency domain for uplink communications between the user equipment device and the radio access network entity.

[0013] This may allow for the generation of a pilot pattern for uplink pilots in a frequency resources grid in a way that reduces the utilization of uplink signalling resources.

[0014] The computing entity may comprise the memory or may be communicatively connectable with it. This may allow channel state information to be stored at a convenient location at which it can be accessed and used by the computing entity.

[0015] The computing entity may be configured to determine a channel estimation based on the stored measurements and one or more pilot measurements received by the radio access network entity from the user equipment device according to the pilot pattern. Using the pilot measurements in combination with the prior information stored in the memory, the computing entity can optionally generate a final channel estimation, which the base station(s) can use to perform advanced communications processing.

[0016] The one or more stored measurements may comprise channel state information measurements for the link between the radio access network entity and the user equipment device. The address in memory may correspond to a location in a map of channel state information measurements for the link between the radio access network entity and the user equipment device. The map may be stored in the memory accessible to the computing entity. This may allow the computing entity to consult the memory to consult the appropriate part of the map to retrieve channel state information measurements that can be used to determine the pilot pattern.

[0017] The context information may comprise the address in the memory accessible to the computing entity. Where the context information includes a UE location, this may indicate a corresponding address in memory. For example, the location may indicate a corresponding location in a map of channel state information measurements. This may allow the computing entity to consult an address in memory containing appropriate information for the user equipment.

[0018] The context information for the user equipment device may be processed by the computing entity in one or more computations to extract one or more of the following to determine the pilot pattern: i) a maximum number of pilots that satisfy a specific signal to noise ratio of the link between the radio access network entity and the user equipment device; and (ii) a minimum number of pilots that provide a specific sampling granularity of the frequency domain. This may provide sufficient information for addressing the two main requirements of pilot accuracy or pilot resolution.

[0019] The computing entity may be configured to implement a weighting factor between the maximum number of pilots that satisfy a specific signal to noise ratio of the link between the radio access network entity and the user equipment device and the minimum number of pilots that provide a specific sampling granularity of the frequency domain to calculate a final number of pilots for the pilot pattern. This may provide flexibility in optimizing the number of pilots for a particular type of channel estimator that may prefer better frequency resolution or better signal-to-noise ratio.

[0020] The computing entity may be configured to request the context information for the user equipment device from the radio access network entity, another network entity of the mobile network or the user equipment device. This may allow the user equipment content to be retrieved by the computing entity in a manner suitable for a particular mobile network infrastructure.

[0021] The computing entity may be configured to send a request for and receive a response indicating one or more of the following for the user equipment device: (i) an antenna arrangement; (ii) a general user equipment orientation; and (iii) a geographical location of the user equipment device. This may allow information contained in the user equipment context to be used to determine the pilot pattern.

[0022] The computing entity may be an entity separate to the radio access network entity and is configured to directly communicate with multiple radio access network entities in the mobile network that have overlapping coverage. This may be a convenient implementation in some network configurations.

[0023] The computing entity may be embedded within a radio access network entity of the mobile network and is configured to communicate one or more other radio access network entities in the mobile network via one or more interfaces. This may allow the computing entity to be incorporated into existing network infrastructure.

[0024] The computing entity may be configured to communicate with a resource allocation entity of the radio access network entity to perform one or more of the following: (i) send the pilot pattern for the user equipment device to the resource allocation entity; (ii) receive one or more pilot measurements according to the pilot pattern; and (iii) send a determined channel estimation to the resource allocation entity. This may enable support of the computing entity from the radio access network entity.

[0025] The radio access network entity may be a base station. This may allow the approach to be used in common mobile communication networks.

[0026] According to a second aspect, there is provided a radio access network entity in a mobile network, the radio access network entity being configured for communication with a user equipment device, the radio access network entity being configured to send a downlink message to the user equipment device indicating a comb-like pilot pattern in the frequency domain with a comb width X, where X is between 2 and 512, the pilot pattern being determined in dependence on user equipment context information for the user equipment device. This may enable full resolution of frequency sampling, which is extremely beneficial in wide spectrum implementations. This may allow for allocating the optimal pilot signal structure, which may reduce interference.

[0027] The radio access network entity may be further configured to receive uplink messages from the user equipment device according to the pilot pattern. Compared to current mechanisms, this may allow for full liberty in equally spaced pilots in frequency domain, allowing for much more granular sampling of the frequency domain. Such sampling provides a method to gently modify the UL signalling to the UE power capabilities, and the channel sampling requirements.

[0028] The radio access network entity may be configured to communicate with the user equipment device and a computing entity in the mobile network to perform the following: receive the pilot pattern in the frequency domain from the computing entity for uplink communications between the user equipment device and the radio access network entity; transmit one or more pilot measurements received by the radio access network entity from the user equipment device according to the pilot pattern to the computing entity; and receive a determined channel estimation from the computing entity. This may allow the accuracy of channel estimations to be improved from baseline comb implementations.

[0029] According to another aspect, there may be provided a mobile network comprising the computing entity having any of the features described herein and the radio access network entity having any of the features described herein.

[0030] According to another aspect, there is provided a method of determining a pilot pattern in a frequency domain for uplink communications between a user equipment device and a radio access network entity in a mobile network, the method comprising : obtaining context information for the user equipment device; in dependence on the context information, consulting an address at a memory accessible to the computing entity containing one or more stored measurements; and in dependence on the stored measurements, determining the pilot pattern in the frequency domain for uplink communications between the user equipment device and the radio access network entity. This method may allow for the generation of a pilot pattern for uplink pilots in a frequency resources grid in a way that reduces the utilization of uplink signalling resources.

[0031] According to another aspect, there is provided a method for implementation at a radio access network entity in a mobile network, the radio access network entity being configured for communication with a user equipment device, the method comprising sending a downlink message to the user equipment device indicating a comb-like pilot pattern in the frequency domain with a comb width X, where X is between 2 and 512, the pilot pattern being determined in dependence on user equipment context information for the user equipment device. This method may enable full resolution of frequency sampling, which is extremely beneficial in wide spectrum implementations. This may allow for allocating the optimal pilot signal structure, which may reduce interference.

[0032] According to a further aspect, there is provided one or more computer programs for instructing a computer comprising one or more processors to implement the methods above.

[0033] According to a further aspect there is provided a data carrier storing in non-transitory form the one or more computer programs above.

[0034] BRIEF DESCRIPTION OF THE FIGURES

[0035] Figure 1 schematically illustrates a general implementation for the transmission of SRS pilots in the uplink in time division duplex systems.

[0036] Figure 2 schematically illustrates a pilot prediction entity in a mobile network comprising one base station and one user equipment device illustrating exemplary interactions between the entities.

[0037] Figure 3 schematically illustrates a further example of a mobile network comprising a pilot prediction entity.

[0038] Figure 4 schematically illustrates using stored channel state information measurements to assist with the prediction of pilot patterns.

[0039] Figure 5 schematically illustrates a closed loop implementation of a mobile network comprising a pilot prediction entity.

[0040] Figure 6 schematically illustrates an open loop implementation of a mobile network comprising a pilot prediction entity.

[0041] Figure 7 schematically illustrates a further example of a mobile network comprising a pilot prediction entity.

[0042] Figure 8 schematically illustrates an example where the pilot prediction entity is implemented at a gNB.

[0043] Figure 9 schematically illustrates the steps of an exemplary method of determining a pilot pattern in a frequency domain for uplink communications between a user equipment device and a radio access network entity in a mobile network.

[0044] Figure 10 schematically illustrates the steps of an exemplary method for implementation at a radio access network entity in a mobile network. DETAILED DESCRIPTION

[0045] The present disclosure relates to mobile networks. The approaches described herein may be implemented in such mobile networks as 3GPP 5G networks and other mobile communication networks that are currently available or developed in the future. The network may comprise a plurality of network entities (NEs). The NEs may be network function (NF s), which may be software-based. The NEs may alternatively be network apparatus (hardware-based).

[0046] A mobile network generally comprises a Radio Access Network (RAN) and a Core Network (CN). The RAN handles the wireless aspects, while the CN handles the management and control aspects. Both the RAN and CN have a User Plane (UP) to transmit traffic. A Control Plane (CP) can carry signalling traffic.

[0047] A BS in a mobile network provides communication services to one or more wireless UEs on downlink (BS to UE information flow), uplink (UE to BS information flow) and sidelink (UE to UE information flow) communications. Examples of BSs include, but are not limited to, a 3GPP 5G next-generation evolved node B (gNB) and an IEEE 802.11 access point (AP).

[0048] In 3GPP 5G networks, a gNB is a RAN node providing new radio (NR) user plane and control plane protocol terminations towards the UE. A gNB is connected via the NG interface to the 5GC. The gNB may in some implementations operate as defined in 3GPP TS 38.300. Other implementations are possible, for example according to future specifications.

[0049] Pilot signals are known signals both to transmitter and receiver used in modem communication systems to perform procedures such as channel estimation, Multiple-Input Multiple-Output (MIMO) precoding, Adaptive Modulation and Coding (AMC), scheduling, beam-management, and other procedures related to adapting the transmission to the current channel conditions. These signals may be scrambled with data signals in time and frequency domains so that the channel conditions experienced by pilots and data are as identical as possible.

[0050] SRS are pilot signals used in TDD and sent by the UE to the BS. SRS are typically allocated in frequency regions not currently used for UL data transmission between a particular UE and BS. This information, together with the assumption of channel reciprocity in TDD systems, allows the BS to know the channel conditions in frequency bands outside of the current UL channel conditions for one UE. With this information, the BS can perform DL precoding, scheduling, etc., without the need of explicit CSI feedback from the UE.

[0051] A radiomap is a memory component that contains CSI data for a specified and limited geographical region covered by one or multiple BSs. It, is a representation of the radio frequency (RF) environment within the region. Each point in the memory of CSI data is associated with a geographical location. Radiomaps can contain CSI information related to the frequency characteristics of the channel throughout its bandwidth, and considering multiple antenna elements of the host network entity, the spatial characteristics of the channel. The temporal effects can be also included. With this information, BSs can reduce or remove the need for SRS in applications such as DL precoding and scheduling.

[0052] UE context information may comprise information indicating an address in memory where historic CSI data is stored for the link between the UE and a BS. For example, UE context information may be location information that describes the position of the UE in a radiomap for direct referencing (such as location, orientation, and antenna arrangement) or indirect referencing (i.e. a piece of CSI information to be checked for similarity in the radiomap).

[0053] Embodiments of the present disclosure can address a technical problem of pilot signal design in the presence of an entity having access to a historic CSI information, for example in the form of a radiomap, for a BS or multiple BSs. One challenge of improving the CSI information for the active UE-BS communication link, while reducing the utilization of resources for each pilot signal in the UL, comes with exemplary challenges of low transmission power of a UE, radiomap estimation inaccuracies based on location only, and changes in the channel information depending on frequency within a bandwidth of a carrier.

[0054] In embodiments of the present disclosure, a computing entity referred to herein as a pilot prediction entity (PPE) is associated with a BS or multiple BSs and has access to a memory storing CSI information for communication between UEs and BSs and communicates with a BS or group of BSs, and / or other network entities, to inquire UE context information. For example, the PPE may store a radiomap in memory. The memory may be part of the PPE or communicatively connected with it. The PPE may be a separate entity to the BS or may be integrated with it. The functions of the PPE may be split over two of more physical or logical entities, for example, some operations may be performed by the BS and some operations by a separate entity to the BS.

[0055] Using the information stored in memory and UE context information, the PPE generates a pattern for uplink pilots in the frequency resources grid in a way that reduces the utilization on uplink signalling resources while also improving channel estimation. The PPE can also further receive (from the BS or BSs) the measured channel properties through the UE uplink pilots. Using this information in combination with the prior information stored in the memory, the PPE can optionally generate a final channel estimation, which the BS(s) can use to perform advanced communications processing.

[0056] The PPE can calculate the UE’s optimal pilot pattern based on UE context information. It can also perform hybrid channel estimation based on UE context information and the received pilot signals.

[0057] The present disclosure considers an area under the management of a PPE entity that is associated with a wireless communication system (for example, at one or more BSs), or co-located with an entity that oversees the operation of one or more BSs, used to transmit and receive information to one or more UEs.

[0058] The PPE may coordinate multiple BSs directly to avoid incompatibility with pilot signals configured by each BS that overlaps with coverage but does not coordinate with the PPE. The PPE may exist as an individual entity separate to the BS or other RAN entity that directly communicates with multiple BSs that have overlapping coverage, or the PPE can be embedded within a BS, or other RAN entity, and communicate with other BSs in the mobile network through provided interfaces.

[0059] A summary of the present system is schematically illustrated in Figure 2. A mobile network 200 comprises a BS 201 and a UE 202. The PPE is shown at 203.

[0060] The PPE 203 obtains context information 204 for the UE 202. The PPE 203 may be configured to request the context information for the UE from the BS 201, another network entity of the mobile network or the UE directly.

[0061] In dependence on the context information, the PPE 203 consults an address in a memory 205 containing one or more stored measurements. In the example shown in Figure 2, the memory 205 stores a radiomap. In dependence on the stored measurements at the memory address, the PPE 203 determines a pilot pattern in the frequency domain for uplink communications between the UE 202 and the BS 201. The PPE 203 allocates the pilot pattern to the UE 202.

[0062] At the UE 202, a pilot of the pilot pattern is selected at 208. For each selected pilot, the UE 202 can perform resource mapping 209, modulation 210, precoding 211 and add the pilots P 212. The pilots are then passed through the radiofrequency components (RF) 213 at the UE transmitter and passed through the wireless channel 214. The pilots are received at RF 215 at the BS receiver. If multiple precoded streams are used, the streams are combined at 216. The pilots are then extracted at 217. The extracted pilots at 217 can then be passed to the PPE 203.

[0063] The PPE 203 can optionally determine a channel estimation 218 based on the stored measurements in memory and one or more pilot measurements 217 received by the BS 201 from the UE 202 according to the pilot pattern.

[0064] In the examples described herein, the pilots may be SRSs.

[0065] Figure 3 schematically illustrates a mobile network 300 comprising a PPE 301 with one BS 302 and one UE 303 to further illustrate the exemplary interactions. The BS 302 can also communicate with an access and mobility management function (AMF) 304 of the CN of the mobile network.

[0066] In a typical scenario, the UE 303 registers itself for the wireless communication services along with a report for its capability to support the comb-X type configurations.

[0067] Given this reported capability of the UE 303, the PPE 301 can decide whether or not to involve the PPE capabilities. The PPE can then send the UE association decision with the BS 302 and informs the BS whether to forward measurements and whether to expect to receive the channel estimation from the PPE 301.

[0068] A messaging mechanism between a PPE and a network entity (UE, another BS, or another network entity,) can be used for exchanging contextual information, relevant to the PPE process. This allows the PPE to receive relevant information for the initialization step. The PPE 301 can (i) generate a request message 311 and (ii) receive a response 312 containing information regarding one or multiple pieces of the following information: antenna arrangement (i.e. distance and direction from primary), general UE orientation (i.e. pitch and yaw from SSB) and geo-location (i.e. x,y,z coordinates).

[0069] The PPE 301 can then request context information from the BS 302 or other entity, or from the UE 303 itself (for example, through the control channels of the BS or BSs to which the UE is configured to communicate with).

[0070] In one particular implementation, the UE context information comprises the orientation of the UE (for example, the pitch, yaw and roll) relative to primary polarization orientations (that can be derived from synchronization signal blocks (SSB)), and antenna arrangement (for example, (H, V, V, H, (+0.5, +0.5), (0,0), (-1,-1), (-1.5, 0.5), where H is horizontal and V is vertical) relative to primary antenna in 2, and location (for example, longitude, latitude, elevation). This may be received as, for example:

[0071] ULOrientedAntenna-r20-IEs ::= SEQUENCE { orientation sequence {INTEGER (0..255 ), INTEGER (0..255), INTEGER (0..255),} arrangeAntennasPolarisation sequence {binary, binary, binary... N antennas) arrangeAntennasDistance sequence {INTEGER, INTEGER, INTEGER ... N antennas) }

[0072] The UE context information 310 received by the PPE 301 is used to generate a pilot signal in a manner that improves the channel estimation. This can be performed at a pilot prediction entity 311 of the PPE 301. In a first exemplary implementation, the pilot pattern can be chosen in an open-loop way that benefits channel estimation that is performed by the BS 302 itself, for which the BS 302 can inform the PPE 301 of the specific algorithm it utilizes (for example, least squares with linear interpolation). In a second exemplary implementation, the pilot pattern can be chosen in a closed-loop manner that it benefits the PPE implemented channel estimation. The closed-loop implementation may provide better results, while the open-loop implementation may provide better compatibility with existing infrastructure. These implementations will be described in further detail below.

[0073] As described above, the PPE 301 calculates the UE’s optimal pilot pattern based on UE context information and uses stored CSI measurements in memory 312, which may be in the form of a radiomap or other data structure, to extract the optimal pattern of pilots for the UE.

[0074] The UE context information may indicate an address in memory (which may, for example, correspond to a location in a radiomap of channel state information measurements for the link between the BS and the UE that is stored in memory accessible by the PPE. The memory may be a memory of the PPE. The context information may comprise or indicate the address in the memory. For example, the context information may comprise a location for the UE (for example, the geolocation). The location may correspond to an address in the memory. The location may in some cases directly correspond to an address in memory storing CSI measurements. In other implementations, the directly corresponding address in memory may be empty and in such cases, the CSI measurements may be interpolated from nearby points.

[0075] Assuming UE association with the PPE 301, the context information of the UE 303 is given to the PPE 301 which can use it in a series of computations to extract the expected signal to noise of the BS-UE link, and its expected frequency characteristics.

[0076] For example, the context information for the UE can be processed by the PPE in one or more computations to extract one or more of the following to determine the pilot pattern:

[0077] (i) a maximum number of pilots that satisfy a specific signal to noise ratio of the link between the BS and the UE; and

[0078] (ii) a minimum number of pilots that provide a specific sampling granularity of the frequency domain.

[0079] The PPE may implement a weighting factor between the maximum number of pilots that satisfy a specific signal to noise ratio of the link between the BS and the UE and the minimum number of pilots that provide a specific sampling granularity of the frequency domain to calculate a final number of pilots for the pilot pattern.

[0080] A messaging mechanism between a BS and a UE can be implemented to configure and communicate a comb-like pilot pattern with flexible, custom width (i.e. comb-X, with X as a whole number between 2 and 512 inclusive). The comb width may also be referred to as the comb period in the frequency domain. Herein, a comb-like pilot pattern comprises equally spaced signals in the frequency domain. This can allow for uniform sampling of the frequency domain. Increasing the comb width X decreases the density of frequency sampling while lowering it increases the frequency information collected by the signal.

[0081] The messaging scheme may comprise two parts. Firstly, one or more downlink messages 307 are sent from the PPE 301 to the UE 303 (for example, to a custom pilot resource allocator entity 350) to inform the UE 303 of a custom pilot pattern. Secondly, one or more uplink messages 308 are sent between the UE 303 (for example, the custom pilot resource allocator entity 350) and the BS 302 with the custom pilot pattern. In one example, the messaging mechanism contains two elements: a downlink message transmitted over control channel to inform the UE of the new X as the width (in frequency domain) of the comb and an uplink message of the UE that contains SRS pilots in a pattern that is equally spaced with an exact width of X.

[0082] Generally, the BS 302 can receive the pilot pattern in the frequency domain from the PPE 301 for uplink communications between the UE 303 and the BS 302. The BS 302 can transmit one or more pilot measurements received by the radio access network entity from the user equipment device according to the pilot pattern to the computing entity. The BS 302 can then receive a determined channel estimation from the PPE 301. For example, the PPE may be configured to communicate with a resource allocation entity of the BS (or other RAN entity) to perform one or more of the following: (i) send the pilot pattern for the user equipment device to the resource allocation entity; (ii) receive one or more pilot measurements according to the pilot pattern; and (iii) send a determined channel estimation to the resource allocation entity.

[0083] Compared to current mechanisms, this mechanism can allow for full liberty in equally spaced pilots in frequency domain, allowing for much more granular sampling of the frequency domain. Such sampling provides a method to gently modify the UL signalling to the UE power capabilities, and the channel sampling requirements.

[0084] A function at the BS 302 can allow message exchanges with the PPE 301 and can facilitate message exchanges between the PPE 301 and the UE 303. The BS can receive the PPE-driven pilot configuration, as shown at 321. This can be performed similarly to the information regarding the X in the comb-X configuration that the UE receives for SRS configuration. Pilot measured CSI samples are extracted at 332 and transmitted from the BS 302 to the PPE 301, as shown at 322. In this step, the SRS pilots that measure the channel properties as per the configured comb-X or other configuration are transmitted to the PPE for further estimation. The estimated channel properties determined by channel estimation entity 313 can be transmitted by the PPE 301 to the BS 302, as shown at 323. This can serve as the final step of the interaction process to provide a trustworthy estimate of the channel properties to the BS 302 using the stored data (for example, the radiomap).

[0085] The BS 302 can store and forward CSI information, shown at 331. With this information, the BS can perform DL precoding at 333 without the need of explicit CSI feedback from the UE 303. Pre-coded data can then be sent from the BS 302 to the UE 303, as shown at 334. The CSI can also be sent to the AMF 304 via upper layers 335 of the BS 302.

[0086] Communication messages between the BS and UE may be performed as shown in Figure 1. That is, the UE 303 can send UL reference signals HVL, which may be SRS signals, to the BS 302. The BS 302 estimates the UL channel based on the UL reference signal. The BS 302 obtains DL channel state information (CSI) / / DI. based on channel reciprocity. The BS 302 completes DL scheduling and transmits data based on the DL CSI.

[0087] An exemplary approach will now be described mathematically.

[0088] A first step is to extract the location of the UE LUEfrom the context information and calibrate it per its orientation and antenna characteristics. The PPE can then extract an estimated channel to arrive to a predicted channel properties hpredusing the first function f (LUE) of the process.

[0089] As schematically illustrated in Figure 4, the first function establishes a radius of raaround the location LUE(that can be parametrized to be relevant to each PPE entity). Collected from memory are Narelevant stored channel properties himem) tfor each chosen element i of the Na. The predicted channel can be calculated using the following Cubic-Distance weighted Radius Nearest Neighbour (3DwRNN) function:

[0090] Using hpred, the PPE proceeds with two more functions f2 and f3 to calculate two separate parameters characterizing the UE-

[0091] BS link. hpredinto its sinusoid components and the penod T of the strongest one is chosen. used to calculate the SNR estimation error budget, giving the optimal number of pilots supported

[0092] K for a targeted SNRtarg.

[0093] Finally, using the aforementioned two parameters r and K, and a further function f4, the exact parameter to be configured to the UE can be extracted with:

[0094] Comb where, fftsize is the number of subcarriers directly dependent on the size of the bandwidth, M is the weighted preference over full selectivity sampling or SNR priority as a real number between 0 and 1.

[0095] In open-loop implementations more significance is allocated to r, i.e. closer to 1, while in closed-loop implementations more significance is allocated to the SNR requirements, i.e. M closer to 0. The resulting Comb_X parameter is then requested from the SRS. In an open-loop implementation the processing would stop here.

[0096] In a closed-loop implementation the resulting channel properties information PPUEfrom the configured UE is collected by the BS and forwarded to the PPE. Using a set of Nbpieces of CSI from the memory that are closer than the predefined radius di b(LUE, ^i) <rb the channel charted distance is calculated d.cc i(PPUE,himem) i) as the similarity of the channel measured by the UE and each entry in the memory of Nbentries. Dedic-Distance weighted Radius Nearest Neighbor (lODwRNN) can be used to arrive at the final channel estimation:

[0097] In one particular implementation, hybrid channel estimation is performed with prediction distances of dcc l. This embodiment of the disclosure covers an implementation with modified calculation for distance. In particular where the context information is extracted through indirect referencing. In this case it is the measurement of a pilot signal across the initial bandwidth PPuE ntt - In this manner, di a(LUE, Lt= dcc i(PPUE init, himem) i) . as a replacement for only the pilot prediction. Moreover, to maintain a dimensional consistency the Cubic-Distance weighted Radius Nearest Neighbor (3DwRNN) may be substituted with Dedic-Distance weighted Radius Nearest Neighbor (lODwRNN).

[0098] In another implementation, hybrid channel estimation is performed with estimation distances of dcc l. This embodiment of the disclosure covers an implementation with modified calculation for distance. In particular with indirect referencing where the pilot pattern measurement PPUEof the UE is used to calculate the distance in terms of channel charting distance. In this manner, di,b LuE’ Li — dCCi PPUE,h(mem i), as a replacement for only the pilot prediction.

[0099] Hybrid channel estimation in a closed-loop implementation can result in a great improvement of channel estimation performance for DL precoding, with an additional benefit of significant reduction in UL signalling overhead.

[0100] In a further implementation, machine learning for open-loop hybrid channel estimation may be used. This embodiment covers an optimization technique for the generation of Comb_ using machine learning technique to enable the PPE entity to generate a for applications that require age of information minimization, as in the case of multi-static sensing. In order to start using the machine learning, an artificial neural network is trained. To train the artificial neural network, the output is compared with the true optimal pilot pattern channel estimation with a mean square error loss function: where Y,- is the true maximal measured channel information, and Ytis the measured channel information through using the artificial neural network. Using gradient descent, the loss L is propagated through the artificial neural network using the backwards propagation algorithm to correct for the errors in each layer of the artificial neural network. The process is repeated to improve the performance of the machine learning algorithm and can be stopped whenever the performance of the estimator is deemed satisfactory.

[0101] Hybrid channel estimation in an open-loop implementation can result in a significant improvement of channel estimation performance for DL precoding and the like. This can result in a reduction in UL signalling overhead, and the open-loop approach may generally have lower complexity than the closed-loop approach.

[0102] An example of a closed-loop implementation will now be described with reference to Figure 5. A mobile network 500 comprises a PPE 501, BS 502 and UE 503.

[0103] The PPE 501 specifies the UE’s context information that it requires (for example, spatial / spectral), as shown at 510. In this example, the context information is the UE’s location coordinates [x,y,z], which may be determined using positioning or UE- location techniques.

[0104] The sourcing of the UE context information is not shown in Figure 5. The context information can be obtained directly from the UE, from the BS or from another entity such as an AMF. The context information can be preliminary CSI [Rx_antennnas x Tx_antennas], for example using SRS or physical random access channel (PRACH) preambles.

[0105] In dependence on the UE context information (in this example, the location), the PPE 501 consults an address in memory storing CSI information, shown as a CSI database 511 in Figure 5. The address corresponds to the location of the UE.

[0106] Based on the CSI information stored in the CSI database at the memory address, the PPE 501 calculates the SNR normalized estimation error budget and creates a pilot pattern with a custom comb width that samples the spectrum. The resulting width can be a custom number between 512 and 2 (i.e. comb-512. . . comb-x . . . comb-2). The pilot prediction is shown at 512 in Figure 5.

[0107] The PPE 501 creates a repeatable pattern that satisfies the sampling period (Comb-X) and configures the UE with the SRS pilot pattern, as shown at 513. This downlink signal may be sent to the UE 503 via a physical downlink control channel (PDCCH).

[0108] The UE custom pilot resource allocator entity 514 maps the pilots and transmits them in uplink from the UE 503 to the BS 502, as shown at 515. This uplink signal may be sent via a physical uplink shared channel (PUSCH).

[0109] The BS 502 accepts this pattern and extracts the pilots at 516. The BS 502 proceeds to use the extracted pilot to calculate the subcarrier-distance in featurespace (CSI). The pilots are sent to the PPE at 517. In this example, using Dedic-Distance weighted Radius Nearest Neighbour (lODwRNN), the BS uses the memory of stored channel responses to perform a prediction of the estimated channel H, as shown at 518. The PPE then forwards the estimated channel state information (CSI) 519 to the BS module for managing and forwarding CSI information 520 to other entities within and outside the BS 502. Dedic weighting is a 10-degree polynomial to measure the featurespace distance between two points in said featurespace.

[0110] In this example, the featurespace is cross-correlation, not location. Given that in this example at least 12 pilots are submitted, this allows for more than 12*256 features to be sampled for comparison. Therefore, location information is not needed.

[0111] An example of a closed-loop implementation will now be described with reference to Figure 6. A mobile network 600 comprises a PPE 601, BS 602, UE 603 and AMF 604.

[0112] The PPE 601 specifies the UE’s context information 610 that it requires (for example, spatial / spectral). In this example, the context information is the UE’s location coordinates [x,y,z], which may be determined using positioning or UE-location techniques. The context information can be obtained directly from the UE, from the BS or from another entity such as an AMF. The context information can be preliminary CSI [Rx_antennnas x Tx_antennas], for example using SRS or PRACH preambles.

[0113] In dependence on the UE context information (in this example, the location), the PPE 601 consults an address in memory storing CSI information, shown as a CSI database 611 in Figure 6. The address corresponds to the location of the UE.

[0114] At 612, pilot prediction is performed. In this example, Cubic-Distance weighted Radius Nearest Neighbour (3DwRNN) uses the memory of stored channel responses at 611 to perform a prediction. In this example, a third degree polynomial dA3 is used for calculating the weight between neighbours. For example, the radius can be set to a maximum of 3 meters. The dominant component of 3DwRNN output is sampled to create a custom width between 512 and 2 (i.e. comb-512... comb-x ... comb-2).

[0115] The PPE 601 creates a repeatable pattern that satisfies the sampling period (Comb-X) and configures the UE with the SRS pilot pattern, as shown at 613. This downlink signal may be sent to the UE 603 via a PDCCH.

[0116] The UE custom pilot resource allocator entity 614 maps the pilots and transmits them in uplink from the UE 603 to the BS 602, as shown at 615. This uplink signal may be sent via a PUSCH.

[0117] The BS 602 accepts this pattern and extracts the pilots, as shown at 616.

[0118] The BS proceeds to build a channel estimate H , as shown at 617, as per its active implementation, in example least squares with linear interpolation - LS(lin)

[0119] The BS 602 can store and forward CSI information, shown at 618. With this information, the BS can perform DL precoding at 619 without the need of explicit CSI feedback from the UE 603. Pre-coded data can then be sent between the BS 602 and UE 603, as shown at 620. The CSI can also be sent to the AMF 604 via upper layers 621 of the BS 602.

[0120] A further example is shown in Figure 7. A mobile network 700 comprises a PPE 701, a RAN entity in the form of a first gNB, gNB 1, 702 and a UE 703.

[0121] At 704, the PPE 702 sends a UE context request to the UE 703. The UE 703 provides a response 705 to the PPE 701. These messages are sent as part of the UE-PPE link contextualization. The UE’s location is retrieved from the gNB 702 at 706. This is transferred to the PPE 701 via an xn interface.

[0122] At 707, the pilot pattern, PP=L2PP is predicted by the PPE 701. The PPE configures the gNB 702 with the pilot pattern at 708. At 709, the gNB 702 sends a downlink message to the UE 703 to inform the UE 703 of the SRS pilot pattern. In this example, the downlink message is transmitted over the PDCCH control channel to inform the UE 703 of the new X as the width (in frequency domain) of the comb.

[0123] At 710, the UE 703 sends an uplink message to the gNB with the custom SRS pilot pattern. The uplink message contains the SRS pilots in a pattern that is equally spaced with an exact width of X. The SRS pilots are also sent on to the PPE 701, which can perform estimation of the channel H = PP2H(P)~ .

[0124] The estimated channel H can be used by the gNB 702 to perform procedures such as MIMO precoding, as shown at 712. Downlink data can then be sent from the gNB 702 to the UE 701 using the estimated channel H, as shown at 713.

[0125] The current 3GPP 5G standard implements limiting and dense SRS sampling, according to the following: nrSRSConfig - transmissionComb kTC = 2; % Comb number (2,4,8) kBarTC = 1; % Comb offset (0...kTC-l)

[0126] The radio resource control (RRC) downlink control channel information element (IE) can be modified to support the allocation of comb-X, allowing the use of custom equidistant width combs that enable larger sparsity in frequency.

[0127] An exemplary modification is shown below:

[0128] SRS-Resource ::= SEQUENCE { srs-Resourceld SRS-Resourceld, nrofSRS-Ports ENUMERATED {portl, ports2, ports4}, ptrs-Portlndex ENUMERATED {nO, nl } OPTIONAL, - NeedR transmissionComb CHOICE { n2 SEQUENCE } combOffset-n2 INTEGER (0..1), cyclicShift-n2 INTEGER (0..7) } , n4 SEQUENCE } combOffset-n4 INTEGER (0..3), cyclicShift-n4 INTEGER (0..11) nX SEQUENCE { customCombX INTEGER(0..512) combOffset-n4 INTEGER (0..3), cyclicShift-n4 INTEGER (0..11) resourceMapping SEQUENCE { startPosition INTEGER (0..5), nrofSymbols ENUMERATED {nl, n2, n4}, repetitionFactor ENUMERATED {nl, n2, n4}

[0129] }, freqDomainPosition INTEGER (0..67), freqDomainShift INTEGER (0..268), c-SRS INTEGER (0..63), b-SRS INTEGER (0..3), b-hop INTEGER (0..3)

[0130] }, groupOrSequenceHopping ENUMERATED { neither, groupHopping, sequenceHopping }, resourceType CHOICE { aperiodic SEQUENCE {

[0131] }, semi-persistent SEQUENCE { periodicityAndOffset-sp SRS-PeriodicityAndOffset,

[0132] }, periodic SEQUENCE { periodicityAndOffset-p SRS-PeriodicityAndOffset,

[0133] }

[0134] }, sequenceld INTEGER (0..1023), spatialRelationlnfo SRS-SpatialRelationlnfo OPTIONAL, — Need R

[0135] }

[0136] In a further implementation, as schematically illustrated in Figure 8, a mobile network 800 comprises a RAN entity in the form of a gNB 801 and a UE 802. In this example, the PPE is implemented at the gNB 801. The gNB 801 estimates the pilot pattern at 803, as described above. At 804, the custom pilot is configured and the custom pilots are later sent back to the gNB at 805.

[0137] The PPE may use the memory of previous pilots transmitted to construct a radiomap to discover the best locations on the resource grid to place pilots with full flexibility (irregular) or limited flexibility (regular).

[0138] The memory-reference (which may be in the form of a channel chart) can source from the UE’s geolocation, estimated reference location (positioning), or simply signalling on prior non-wideband communications.

[0139] Full flexibility implementations may allow for hyper-optimized pilot placement. Thus, in this case, the pilot placement optimization process may memorize the full channel properties.

[0140] Limited flexibility implementations may follow a more memory conservative approach and use two key parameters from the memory: the frequency selectivity and the SNR of the uplink pilots.

[0141] Based on the predicted uplink noise budget and the selectivity, the following implementations may be used:

[0142] High selectivity and high SNR budget - allow full sampling Low selectivity and high SNR budget - allow over-sampling

[0143] High selectivity and low SNR budget - limit sampling

[0144] Low selectivity and low snr budget - allow full sampling

[0145] Given the adequate implementation selectivity and SNR relationship can be interpreted, the BS can assign a pilot pattern for the next physical resource block (PRB).

[0146] Figure 9 shows exemplary steps of a method of determining a pilot pattern in a frequency domain for uplink communications between a user equipment device and a radio access network entity in a mobile network. At step 901, the method comprises obtaining context information for the user equipment device. At step 902, the method comprises, in dependence on the context information, consulting an address at a memory accessible to the computing entity containing one or more stored measurements. At step 903, the method comprises, in dependence on the stored measurements (at the address), determining the pilot pattern in the frequency domain for uplink communications between the user equipment device and the radio access network entity.

[0147] Figure 10 shows an exemplary method for implementation at a radio access network entity in a mobile network, the radio access network entity being configured for communication with a user equipment device. At step 1001, the method comprises sending a downlink message to the user equipment device indicating a comb-like pilot pattern in the frequency domain with a width X, where X is between 2 and 512, the pilot pattern being determined in dependence on user equipment context information for the user equipment device.

[0148] Using the PPE, the accuracy of channel estimations may be drastically improved from the baseline comb implementations.

[0149] The messaging scheme between the BS and the UE to configure and communicate a comb-like pilot pattern enables the operation of the PPE, while maintaining very low signalling overhead. This may further potentially reduce interference.

[0150] The described messaging scheme with request and response for UE-PPE contextualization can enable the calibration of the information between the radiomap and the pilot information.

[0151] The described messaging schemes between the PPE and (i) the UE and (ii) the BS can enable the support of the PPE from the UE and BS respectively.

[0152] The use of pilot patterns in the frequency domain with a comb width comb-X can enable full resolution of frequency sampling, which is extremely beneficial in wide spectrum implementations. This may allow for allocating the optimal pilot signal structure, which may reduce interference.

[0153] The applicant hereby discloses in isolation each individual feature described herein and any combination of two or more such features, to the extent that such features or combinations are capable of being carried out based on the present specification as a whole in the light of the common general knowledge of a person skilled in the art, irrespective of whether such features or combinations of features solve any problems disclosed herein. The applicant indicates that aspects of the present disclosure may consist of any such individual feature or combination of features. In view of the foregoing description, it will be evident to a person skilled in the art that various modifications may be made within the scope of the disclosure.

Claims

CLAIMS1. A computing entity (203, 301, 501, 601, 701) for use in a mobile network (200, 300, 500, 600, 700, 800), the mobile network comprising a radio access network entity (201 , 302, 502, 602, 702, 801 ) in communication with a user equipment device (202, 303, 503, 603, 703, 802), the computing entity being configured to: obtain (901) context information for the user equipment device; in dependence on the context information, consult (902) an address at a memory accessible to the computing entity containing one or more stored measurements; and in dependence on the stored measurements, determine (903) a pilot pattern in the frequency domain for uplink communications between the user equipment device and the radio access network entity.

2. The computing entity as claimed in claim 1, wherein the computing entity is configured to determine a channel estimation (218, 313, 518) based on the stored measurements and one or more pilot measurements received by the radio access network entity from the user equipment device according to the pilot pattern.

3. The computing entity as claimed in claim 1 or claim 2, wherein the address in the memory corresponds to a location in a map of channel state information measurements for the link between the radio access network entity and the user equipment device, the map being stored in the memory accessible to the computing entity.

4. The computing entity as claimed in any preceding claim, wherein the context information comprises the address in the memory accessible to the computing entity.

5. The computing entity as claimed in any preceding claim, wherein the context information for the user equipment device is processed by the computing entity in one or more computations to extract one or more of the following to determine the pilot pattern: i) a maximum number of pilots that satisfy a specific signal to noise ratio of the link between the radio access network entity and the user equipment device; and (ii) a minimum number of pilots that provide a specific sampling granularity of the frequency domain.

6. The computing entity as claimed in claim 5, wherein the computing entity is configured to implement a weighting factor between the maximum number of pilots that satisfy a specific signal to noise ratio of the link between the radio access network entity and the user equipment device and the minimum number of pilots that provide a specific sampling granularity of the frequency domain to calculate a final number of pilots for the pilot pattern.

7. The computing entity as claimed in any preceding claim, wherein the computing entity is configured to request the context information for the user equipment device from the radio access network entity, another network entity of the mobile network or the user equipment device.

8. The computing entity as claimed in any preceding claim, wherein the computing entity is configured to send a request for and receive a response indicating one or more of the following for the user equipment device: (i) an antenna arrangement; (ii) a general user equipment orientation; and (iii) a geographical location of the user equipment device.

9. The computing entity as claimed in any preceding claim, wherein the computing entity is an entity separate to the radio access network entity and is configured to directly communicate with multiple radio access network entities in the mobile network that have overlapping coverage.

10. The computing entity as claimed in any of claims 1 to 8, wherein the computing entity is embedded within a radio access network entity of the mobile network and is configured to communicate one or more other radio access network entities in the mobile network via one or more interfaces.

11. The computing entity as claimed in any preceding claim, wherein the computing entity is configured to communicate with a resource allocation entity of the radio access network entity to perform one or more of the following: (i) send the pilot pattern for the user equipment device to the resource allocation entity; (ii) receive one or more pilot measurements according to the pilot pattern; and (iii) send a determined channel estimation to the resource allocation entity.

12. The computing entity as claimed in any preceding claim, wherein the radio access network entity is a base station.

13. A radio access network entity (201, 302, 502, 602, 702, 801) in a mobile network (200, 300, 500, 600, 700, 800), the radio access network entity being configured for communication with a user equipment device (202, 303, 503, 603, 703, 802), the radio access network entity being configured to send (1001) a downlink message to the user equipment device indicating a comb-like pilot pattern in the frequency domain with a width X, where X is between 2 and 512, the pilot pattern being determined in dependence on user equipment context information for the user equipment device.

14. The radio access network entity as claimed in claim 13, wherein the radio access network entity is further configured to receive uplink messages from the user equipment device according to the pilot pattern.

15. The radio access network entity as claimed in claim 13 or claim 14, wherein the radio access network entity is configured to communicate with the user equipment device and a computing entity in the mobile network to perform the following: receive the pilot pattern in the frequency domain from the computing entity for uplink communications between the user equipment device and the radio access network entity; transmit one or more pilot measurements received by the radio access network entity from the user equipment device according to the pilot pattern to the computing entity; and receive a determined channel estimation from the computing entity.

16. A mobile network (200, 300, 500, 600, 700, 800) comprising the computing entity of any of claims 1 to 12 and the radio access network entity of any of claims 13 to 15.

17. A method (900) of determining a pilot pattern in a frequency domain for uplink communications between a user equipment device (202, 303, 503, 603, 703, 802) and a radio access network entity (201, 302, 502, 602, 702, 801) in a mobile network (200, 300, 500, 600, 700, 800), the method comprising: obtaining (901) context information for the user equipment device; in dependence on the context information, consulting (902) an address at a memory accessible to the computing entity containing one or more stored measurements; and in dependence on the stored measurements, determining (903) the pilot pattern in the frequency domain for uplink communications between the user equipment device and the radio access network entity.

18. A method (1000) for implementation at a radio access network entity (201, 302, 502, 602, 702, 801) in a mobile network (200, 300, 500, 600, 700, 800), the radio access network entity being configured for communication with a user equipment device (202, 303, 503, 603, 703, 802), the method comprising sending (1001) a downlink message to the user equipment device indicating a comb-like pilot pattern in the frequency domain with a width X, where X is between 2 and 512, the pilot pattern being determined in dependence on user equipment context information for the user equipment device.

Images