User equipment side machine learning model update triggering machine learning based positioning

CN118922735APending Publication Date: 2024-11-08NOKIA TECHNOLOGIES OY
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202180104064.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2021-09-14
Publication Date
2024-11-08

Smart Images

  • Figure CN118922735A_ABST
    Figure CN118922735A_ABST
Patent Text Reader

Abstract

Techniques for updating the ML model include performing such updates when the UE satisfies certain criteria. In some implementations, a UE uses an ML model to determine a location within a network. In some implementations, the criteria include a version number of an ML model used by the UE. In some implementations, the criteria include providing the time elapsed since the latest ML model update to the user equipment.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This specification relates to communications. Background Art

[0002] A communication system may be a facility that enables communication between two or more nodes or devices, such as fixed or mobile communication devices. Signals may be carried by wired or wireless carriers.

[0003] An example of a cellular communication system is an architecture standardized by the Third Generation Partnership Project (3GPP). The latest development in this field is generally referred to as the Long Term Evolution (LTE) of the Universal Mobile Telecommunications System (UMTS) radio access technology. E-UTRA (Evolved UMTS Terrestrial Radio Access) is the air interface of the LTE upgrade path of 3GPP for mobile networks. In LTE, a base station or access point (AP) referred to as an enhanced node AP (eNB) provides wireless access within a coverage area or cell. In LTE, a mobile device or mobile station is referred to as a user equipment (UE). LTE has included multiple improvements or developments.

[0004] For example, the global bandwidth shortage faced by wireless carriers has stimulated the consideration of underutilized millimeter wave (mmWave) spectrum for future broadband cellular communication networks. For example, mmWave (or extremely high frequency) can include a frequency range between 30 and 300 gigahertz (GHz). For example, the radio waves in this band can have a wavelength from 10 to 1 mm, so it is called millimeter band or millimeter wave. In the next few years, the amount of wireless data is likely to increase significantly. Various technologies have been used to try to solve this challenge, including obtaining more spectrum, having smaller cell sizes, and using improved technologies that can achieve higher bits / second / hertz. One element that can be used to obtain more spectrum is to move to higher frequencies, such as above 6GHz. For the fifth generation wireless system (5G), an access architecture for deploying cellular radio equipment using mmWave radio spectrum has been proposed. Other example spectrums, such as cmWave radio spectrum (e.g., 3-30GHz) can also be used. Summary of the invention

[0005] According to an example implementation, a method includes: receiving first inference input data from a network or direct measurement, the first inference input data representing an inference input to an inference operation of a specified radio measurement within the network, the inference operation including a machine learning model, the machine learning model being configured to predict a first value of a device parameter based on the specified radio measurement. The method also includes: sending indication data to a server connected to the network, the indication data representing an indication of the accuracy of the inference output based on the machine learning model when determining the value of the device parameter. The method also includes: receiving or not receiving an update to the machine learning model based on the indication data from the server, the update enabling the machine learning model to predict a second value of the device parameter based on the specified radio measurement, the second value being more accurate than the first value.

[0006] According to an example implementation, an apparatus includes: at least one processor and at least one memory including computer program code, the at least one memory and the computer program code being configured to, together with the at least one processor, cause the apparatus to at least: receive first inference input data from a network or direct measurement, the first inference input data representing an inference input to an inference operation of a specified radio measurement within the network, the inference operation including a machine learning model, the machine learning model being configured to predict a first value of a device parameter based on the specified radio measurement. The at least one memory and the computer program code are also configured to, together with the at least one processor, cause the apparatus to at least: send indication data to a server connected to the network, the indication data representing an indication of the accuracy of the inference output based on the machine learning model when determining the first value of the device parameter. The at least one memory and the computer program code are also configured to, together with the at least one processor, cause the apparatus to at least: receive or not receive an update to the machine learning model based on the indication data from the server, the update enabling the machine learning model to predict a second value of the device parameter based on the specified radio measurement, the second value being more accurate than the first value.

[0007] According to an example implementation, an apparatus includes: a component for receiving first inference input data from a network or direct measurement, the first inference input data representing an inference input to an inference operation of a specified radio measurement within the network, the inference operation including a machine learning model, the machine learning model being configured to predict a first value of a device parameter based on the specified radio measurement. The apparatus also includes: a component for sending indication data to a server connected to the network, the indication data representing an indication of the accuracy of the inference output based on the machine learning model when determining the value of the device parameter. The apparatus also includes: a component for receiving or not receiving an update to the machine learning model based on the indication data from the server, the update enabling the machine learning model to predict a second value of the device parameter based on the specified radio measurement, the second value being more accurate than the first value.

[0008] According to an example implementation, a computer program product includes a computer-readable storage medium and stores executable code, which, when executed by at least one data processing device, is configured to cause the at least one data processing device to: receive first inference input data from a network or direct measurement, the first inference input data representing an inference input to an inference operation of a specified radio measurement within the network, the inference operation including a machine learning model, the machine learning model being configured to predict a first value of a device parameter based on the specified radio measurement. The executable code, when executed by the at least one data processing device, is also configured to cause the at least one data processing device to: send indication data to a server connected to the network, the indication data representing an indication of the accuracy of the inference output based on the machine learning model when determining the first value of the device parameter. The executable code, when executed by the at least one data processing device, is also configured to cause the at least one data processing device to: receive or not receive an update to the machine learning model based on the indication data from the server, the update enabling the machine learning model to predict a second value of the device parameter based on the specified radio measurement, the second value being more accurate than the first value.

[0009] According to an example implementation, a method includes: receiving indication data from a user device in a network, the indication data representing an indication of the accuracy of an inferred output based on a machine learning model in determining a value of a device parameter. The method also includes: determining whether to send an update to the machine learning model to the user device based on the indication data.

[0010] According to an example implementation, an apparatus includes: at least one processor and at least one memory including computer program code, the at least one memory and the computer program code being configured to, together with the at least one processor, cause the apparatus to at least: receive indication data from a user device in a network, the indication data representing an indication of the accuracy of an inference output based on a machine learning model when determining a value of a device parameter. The at least one memory and the computer program code are further configured to: determine, based on the indication data, whether to send an update to the machine learning model to the user device.

[0011] According to an example implementation, an apparatus includes: a component for receiving indication data from a user device in a network, the indication data representing an indication of the accuracy of an inferred output based on a machine learning model when determining a value of a device parameter. The apparatus also includes: a component for determining whether to send an update to the machine learning model to the user device based on the indication data.

[0012] According to an example implementation, a computer program product includes a computer-readable storage medium and stores executable code, which, when executed by at least one data processing device, is configured to cause the at least one data processing device to: receive indication data from a user device in a network, the indication data representing an indication of the accuracy of an inference output based on a machine learning model when determining a value of a device parameter. The executable code, when executed by the at least one data processing device, is also configured to cause the at least one data processing device to: determine whether to send an update to the machine learning model to the user device based on the indication data.

[0013] The details of one or more examples of implementations are set forth in the accompanying drawings and the description below. Other features will be apparent from the description and drawings, and from the claims. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 is a block diagram of a digital communication network according to an example implementation.

[0015] Figure 2A is a diagram illustrating a scenario in which a server updates all UEs served by it regardless of ML model usage according to an example implementation.

[0016] Figure 2B is a diagram illustrating a scenario in which a server updates its served UEs depending on ML model usage according to an example implementation.

[0017] Figure 3 is a sequence diagram illustrating explicit versioning according to an example implementation.

[0018] Figure 4 is a flow diagram illustrating a process for updating an ML model without version control according to an example implementation.

[0019] Figure 5 is a sequence diagram illustrating a process of updating an ML model without version control according to an example implementation.

[0020] Figure 6 is a flow chart illustrating a process of updating an ML model according to an example implementation.

[0021] Figure 7 is a flow chart illustrating a process of updating an ML model according to an example implementation.

[0022] Figure 8 is a block diagram of a node or wireless station (eg, a base station / access point, a relay node, or a mobile station / user equipment) according to an example implementation. DETAILED DESCRIPTION

[0023] The principles of the present disclosure will now be described with reference to some example embodiments. It should be understood that these embodiments are described only for illustrative purposes and to help those skilled in the art understand and implement the present disclosure without indicating any limitation to the scope of the present disclosure. The disclosure described herein can be implemented in various ways except for the ways described below.

[0024] The terms used herein are for the purpose of describing specific embodiments only and are not intended to limit the exemplary embodiments. As used herein, the singular forms "a", "an", and "the" are also intended to include the plural forms, unless the context clearly indicates otherwise. It should also be understood that the terms "comprises", "comprising", "has", "having", "includes", and / or "including" when used herein specify the presence of the features, elements, and / or components, etc., but do not exclude the presence or addition of one or more other features, elements, components, and / or combinations thereof.

[0025] Figure 1 is a block diagram of a digital communication system such as wireless network 130 according to an example implementation. Figure 1 In a wireless network 130, user equipment 131, 132, and 133, which may also be referred to as mobile stations (MS) or user equipment (UE), may be connected (and communicate) with a base station (BS) 134, which may also be referred to as an access point (AP), an enhanced node B (eNB), a gNB (which may be a 5G base station), or a network node. At least some of the functionality of an access point (AP), a base station (BS), or an (e) node B (eNB) may also be implemented by any node, server, or host operably coupled to a transceiver (such as a remote radio head). BS (or AP) 134 provides wireless coverage within a cell 136, which includes user equipment 131, 132, and 133. Although only three user equipment are shown connected or attached to BS 134, any number of user equipment may be provided. BS 134 is also connected to a core network 150 via an interface 151. This is just one simple example of a wireless network, and other examples may also be used.

[0026] User equipment (user terminal, user equipment (UE)) may refer to a portable computing device, including wireless mobile communication devices operating with or without a subscriber identity module (SIM), including but not limited to the following types of devices: as examples, mobile stations (MS), mobile phones, cellular phones, smart phones, personal digital assistants (PDAs), handhelds, devices using wireless modems (alarm or measurement equipment, etc.), laptop and / or touch screen computers, tablet computers, phablets, game consoles, notebooks, vehicles, and multimedia devices. It should be understood that user equipment may also be almost exclusively uplink-only devices, an example of which is a camera or video camera that loads images or video clips to a network.

[0027] In LTE (as an example), the core network 150 may be referred to as an evolved packet core (EPC), which may include: a mobility management entity (MME), which may handle or assist with mobility / serving cell changes of user equipment between BSs; one or more gateways, which may forward data and control signals between the BSs and a packet data network or the Internet; and other control functions or boxes.

[0028] Various example implementations may be applied to various wireless technologies, wireless networks, such as LTE, LTE-A, 5G (new radio or NR), cmWave and / or mmWave band networks, or any other wireless network or use case. LTE, 5G, cmWave and mmWave band networks are provided as illustrative examples only, and various example implementations may be applied to any wireless technology / wireless network. Various example implementations may also be applied to a variety of different applications, services, or use cases, such as ultra-reliable low latency communications (URLLC), Internet of Things (IoT), time-sensitive communications (TSC), enhanced mobile broadband (eMBB), massive machine type communications (MMTC), vehicle-to-vehicle (V2V), vehicle-to-device, etc. Each of these use cases or UE types may have its own set of requirements.

[0029] Machine learning (ML) will be widely used in 5G, 5G Advanced, and Beyond 5G networks to optimize various network functions, including user equipment (UE) positioning, active handover control, uplink power control, and load balancing, among others. Although many trained ML models can use more computing resources hosted on the network side, ML models can also be hosted on the UE side to reduce latency. For example, while UE positioning functions can be performed on the network side (e.g., using a location server), if the UE's precise location will be used for applications with low latency, the UE may prefer positioning inference on the UE side. Industrial robotics is an example of such a use case, where latency requirements are very strict and UE positioning requirements can reach 1cm accuracy. UE manufacturers are constantly seeking to improve UE capabilities for hosting ML-trained models in the network, and it is expected that UE capability enhancements will leverage the use of artificial intelligence.

[0030] In one example, the network trains an ML model based on radio measurements, beam RSRP being one such example (angle of arrival (AoA) could be another example). The trained model is then transmitted via the network radio link to the UE, which performs real-time inference of the UE's positioning.

[0031] When the ML model is hosted at the UE for inference, continuous model evaluation and retraining of the model is required if the performance of the model degrades significantly due to changing radio conditions. To trigger the retraining of the ML model, the inference statistics of all UEs are taken into account. Due to the availability of more computing resources and input data at the network, the retraining of the ML model is performed again at the network side.

[0032] After a new ML model is calculated (periodically or based on a trigger), it needs to be transmitted to all UEs again using the wireless communication link. Therefore, model training and subsequent transmission to UEs can be divided into a two-step process:

[0033] 1. Retrain the ML model

[0034] Receive positioning inference results from multiple UEs.

[0035] Evaluate ML model inference accuracy based on various statistics of UE inference results.

[0036] Deciding to retrain the ML model in the network.

[0037] 2. Transmit the retrained ML model to the UE

[0038] After the trained model is available at the network location server, it needs to be transmitted back to the UE for local inference.

[0039] In conventional ML model updates, it is assumed that both steps are performed sequentially, i.e., whenever an updated ML model is available on the network, it is immediately transmitted to all UEs configured to use the model.

[0040] Conventional ML model updates can burden already congested radio network links (downlink and uplink, U-plane and C-plane) if the frequency of model retraining / updating is too high and the number of UEs to be updated is considerable.

[0041] Compared to the conventional method of updating the ML model described above, an improved technique for updating the ML model includes performing such an update when the UE meets certain criteria. In some implementations, the UE uses the ML model to determine the location within the network. In some implementations, the criteria include a version number of the ML model used by the UE. In some implementations, the criteria include the time that has passed since the ML model update was last provided to the user equipment.

[0042] Advantageously, the above-described improved technique for traditional ML model updating reduces the burden on the network by only updating asynchronously and in response to meeting specified criteria.

[0043] The improved technology includes a method for triggering a local model update at the UE for improved UE location accuracy without unnecessary model updates while achieving signaling overhead reduction and traffic reduction (U-plane and C-plane). Figure 2B As shown, UE selection for ML model update can be performed asynchronously, and only UEs performing active inference can be selected for ML model update.

[0044] In some implementations, model updates for UEs are provided based on the time history of recent use of the ML model for UE positioning and the time elapsed since the previous update of its ML model. That is, the server selects UEs for model updates that satisfy the following two conditions:

[0045] In the current network, the latest model version is M, and the UE uses the old model version number M', for example, based on the model version number comparison M' <M。

[0046] the number of inferences (i.e., deductions from measurements) is greater than a threshold number of inferences in the time window,

[0047] For example a counter can be used to extrapolate the number of times within the timer T.

[0048] The improved technique helps reduce congestion on the radio link without compromising the accuracy of the positioning inference of the UE. For example, it is possible that the UE is idle for a long time, uses an outdated ML model for positioning inference, and sleeps again for a long time. In traditional ML model updates, the ML model of the UE may need to update the ML model immediately based on (possibly) bad inference results, but updating the ML model for this particular UE is not efficient for radio link usage. Therefore, this particular UE can tolerate performance degradation due to bad inference to improve overall network efficiency. In fact, if the UE does not use the ML model for a long time, the network may have updated several versions of the ML model, and the UE can skip the update of all versions without any performance loss.

[0049] Figure 3 is a sequence diagram illustrating explicit versioning. With explicit versioning, each ML model version has an ID and an associated average positional accuracy average (positional accuracy).

[0050] Whenever an ML model is updated (retrained) at the network location server, its model ID is incremented. The associated average positioning accuracy may change or remain the same as before the model update.

[0051] Input data for ML model retraining is provided from UE reports.

[0052] At 301 , the location server updates the ML model based on the joint inference results from the UEs, and distributes a model version to each updated ML model.

[0053] At 302, the location server configures the NG-RAN to enable an inference process.

[0054] At 303, the NG-RAN configures a threshold number of inferences based on the estimated number of UE position reports within a certain time window. A higher threshold results in fewer ML model updates and a greater likelihood of incorrect positioning estimates; a lower threshold results in more ML model updates and a lower likelihood of incorrect positioning estimates.

[0055] At 304, the NG-RAN provides inputs such as reference signal received power (RSRP), angle of arrival (AOA), etc. to the UE for measurement and inference.

[0056] At 305, the UE performs inference on the measurements of the provided input.

[0057] At 306, in response to the number of inferences made by the UE being greater than a threshold configured by the NG-RAN, the UE sends an ML model identifier (eg, version number) to the location server.

[0058] Note that, in some implementations, the UE sends the ML model identifier periodically rather than in response to an event. In some implementations, the transmission period is configured by the NG-RAN.

[0059] At 307, the location server compares each model ID (M’) received from the UE with the latest (current) model ID (M), i.e., checks if M’ < M. In some implementations, the location server checks (additionally or alternatively) the time of the most recent update against a time threshold elapsed.

[0060] At 308, the location server determines that an ML model update is needed and, if so, reconfigures the NG-RAN.

[0061] At 309, the location server determines to provide an ML model update to the UE based on the data received from the UE.

[0062] Figure 4 It is a flowchart of process 400 for updating an ML model by showing an explicit accuracy comparison between the UE local model and the network's most recently updated model without explicit version control.

[0063] At 401, the UE performs an inference based on the measurements provided to it by the NG-RAN. The inference is performed using a version of the ML model that determines the values of device parameters (e.g., location, uplink power control, load balancing, etc.). It is assumed herein that the device parameter is the UE location.

[0064] At 402, the UE evaluates soft conditions such as its recent inference history, mobility state changes, etc. Fast mobility and rapidly changing locations in the recent history may trigger the need to check the ML model.

[0065] At 403, the UE determines whether the soft conditions indicate a need for an ML model update.

[0066] At 404, the UE determines that the soft conditions indicate a need for an ML model update and, in response, the UE sends a model update request to the network location server by signaling a model ID M’ = 0. This request is a soft signal / trigger that does not necessarily initiate a model update because fast mobility or poor (past) inference decisions are not always due to an outdated ML model at the UE.

[0067] At 405, the location server receives the model ID M’ = 0 and checks the condition that the number of inferences performed by the UE within the time window T is greater than a specified inference threshold. The NG-RAN tracks the UE location information and can deduce the number of inferences. Note that the network location server configures (on the network side) the inference threshold based on the estimated number of UE location reports for a particular UE within a certain time window.

[0068] In some implementations, the location server checks (additionally or alternatively) the time of the last update against an elapsed time threshold.

[0069] At 406, in response to the condition being met, the network location server requests from the UE report a report containing the most recently estimated inference accuracy metric and the corresponding input data used.

[0070] At 407 , the location server performs inference using the updated ML model, and determines whether to send the updated ML model to the UE based on comparison with the inference result from the UE.

[0071] Figure 5 is a sequence diagram illustrating the process of updating an ML model without version control.

[0072] At 501 , the location server configures the NG-RAN to enable the inference process and configures the inference threshold and error tolerance.

[0073] At 502, the NG-RAN provides inputs such as reference signal received power (RSRP), angle of arrival (AOA), etc. to the UE for measurement and inference.

[0074] At 503, the UE performs inference on the measurements of the provided inputs and evaluates soft conditions such as mobility state, location change, and the like.

[0075] At 504 and 505, in response to the soft conditions being met, the UE sends a model update request to the network location server by signaling model ID M' = 0. This request is a soft signal / trigger that does not necessarily initiate a model update, since fast mobility or bad (past) inference decisions are not always due to an outdated ML model at the UE.

[0076] At 506, the location server receives the model ID M'=0 and checks the condition that the number of inferences performed by the UE within the time window T is greater than a specified inference threshold.

[0077] At 507, in response to the condition being met, the location server requests from the UE report a report containing the most recently estimated inference accuracy metric and the corresponding input data used.

[0078] At 508, the UE sends its inference results and input data to the location server.

[0079] At 509, the location server performs inferences for UE positioning using its most recently trained model version M, and compares the positioning accuracy with the accuracy associated with version M, e.g., checking that positioning accuracy (model M) - positioning accuracy (M') > tolerance. In some implementations, the accuracy is calculated by assuming that more non-ML based, more accurate positioning methods are available. Otherwise, the ML model version M is assumed to be more accurate and error-free (because it is the most recently updated model), and simple differences in positioning predictions are determined by evaluating positioning inference (model M) - positioning inference (M') > tolerance.

[0080] At 510 and 511 , in response to the condition at 510 being met, the location server begins updating the ML model and, if necessary, reconfiguring the NG-RAN.

[0081] At 512, the location server sends the updated ML model to the UE.

[0082] Example 1-1: Figure 6 6 is a flow diagram illustrating a process 600 for updating an ML model. Operation 610 includes: receiving first inference input data from a network or direct measurement, the first inference input data representing an inference input to an inference operation of a specified radio measurement within the network, the inference operation including a machine learning model, the machine learning model configured to predict a first value of a device parameter based on the specified radio measurement. Operation 620 includes: sending indication data to a server connected to the network, the indication data representing an indication of the accuracy of the inference output based on the machine learning model when determining the value of the device parameter. Operation 630 includes: receiving or not receiving an update to the machine learning model based on the indication data from the server, the update enabling the machine learning model to predict a second value of the device parameter based on the specified radio measurement, the second value being more accurate than the first value.

[0083] Example 1-2: An example implementation according to Example 1-1, wherein the device parameter is a user device location within a network, and the server is a location server.

[0084] Example 1-3: An example implementation according to Example 1-2, wherein the specified radio measurement includes a reference signal received power.

[0085] Example 1-4: An example implementation according to Examples 1-2 and 1-3, wherein the indication data includes a version number of the machine learning model.

[0086] Example 1-5: An example implementation according to Example 1-4, wherein the indication data is sent periodically according to a specified period.

[0087] Example 1-6: An example implementation according to Example 1-5, wherein the specified period is specified by the network.

[0088] Example 1-7: An example implementation according to Examples 1-4 to 1-6, wherein the indication data is sent in response to a condition being met.

[0089] Example 1-8: According to the example implementation of Example 1-7, the condition satisfied includes: the number of inference operations performed by the device within a specified time window is greater than a specified inference threshold.

[0090] Example 1-9: The example implementation of Example 1-8 further includes receiving threshold data representing the inferred threshold from the network.

[0091] Example 1-10: An example implementation according to Examples 1-7 to 1-9, wherein the conditions met include that the time when the machine learning model was last updated is greater than a threshold time.

[0092] Example 1-11: The example implementation according to Examples 1-1 to 1-2, wherein the condition satisfied includes any one of: the inferred frequency is greater than a threshold, and the frequency of the estimated position change is greater than a threshold.

[0093] Example 1-12: The example implementation according to Example 1-11 further includes sending first inference input data used in the inference operation to a server.

[0094] Example 1-13: An example implementation according to Example 1-11 or 1-12, wherein the indication data includes a trigger for the server to determine whether to send an update to the machine learning model to the device.

[0095] Example 1-14: An apparatus comprising components for performing the method of any one of Examples 1-1 to 1-13.

[0096] Example 1-15: A computer program product, comprising a non-transitory computer-readable storage medium and storing executable code, which, when executed by at least one data processing device, is configured to cause the at least one data processing device to perform any of the methods of Examples 1-1 to 1-13.

[0097] Example 2-1: Figure 7 is a flow chart illustrating a process 700 for updating an ML model. Operation 710 includes receiving indication data from a user device in a network, the indication data representing an indication of the accuracy of an inferred output based on a machine learning model in determining a value of a device parameter. Operation 720 includes determining, based on the indication data, whether to send an update to the machine learning model to the user device.

[0098] Example 2-2: The example implementation of Example 2-1, wherein the device parameter is a user device located within the network, and the apparatus includes a location server.

[0099] Example 2-3: An example implementation according to Example 2-2, wherein the specified radio measurement includes a reference signal received power.

[0100] Example 2-4: An example implementation according to Examples 2-2 to 2-3, wherein the indication data is sent periodically according to a period.

[0101] Example 2-5: The example implementation of Example 2-4 further includes specifying the period based on an estimated number of location reports sent by the user equipment within the specified time window.

[0102] Example 2-6: An example implementation according to Example 2-5, wherein the indication data is sent in response to a condition being met.

[0103] Example 2-7: An example implementation according to Examples 2-4 to 2-6, wherein the conditions met include the time when the machine learning model was last updated being greater than a threshold time.

[0104] Example 2-8: An example implementation according to Examples 2-4 to 2-7, wherein the condition satisfied includes any one of: the inferred frequency is greater than a threshold, and the frequency of position changes is greater than a threshold.

[0105] Example 2-9: An example implementation according to Examples 2-1 to 2-8, wherein the indication data includes a version number of the machine learning model.

[0106] Example 2-10: An apparatus comprising components for performing the method described in any one of Examples 2-1 to 2-9.

[0107] Example 2-11: A computer program product comprising a non-transitory computer-readable storage medium and storing executable code, which, when executed by at least one data processing device, is configured to cause the at least one data processing device to perform any of the methods described in Examples 2-1 to 2-9.

[0108] Example abbreviation list:

[0109] CN(5G) Core Network

[0110] LMF Location Mobility Function

[0111] FL Federated Learning

[0112] ML Machine Learning

[0113] NG-RAN Next Generation Radio Access Network

[0114] UE User Equipment

[0115] Figure 8800 is a block diagram of a wireless station (e.g., AP, BS, e / gNB, NB-IoT UE, UE, or user equipment) according to an example implementation. The wireless station 800 may include, for example, one or more RF (radio frequency) or wireless transceivers 802A, 802B, each of which includes a transmitter for transmitting signals (or data) and a receiver for receiving signals (or data). The wireless station also includes: a processor or control unit / entity (controller) 804, which is used to execute instructions or software and control the transmission and reception of signals; and a memory 806, which is used to store data and / or instructions.

[0116] The processor 804 may also make decisions or determinations, generate time slots, subframes, packets or messages for transmission, decode received time slots, subframes, packets or messages for further processing, and other tasks or functions described herein. For example, the processor 804 may be a baseband processor that may generate messages, packets, frames or other signals for transmission via the wireless transceiver 802 (802A or 802B). The processor 804 may control the transmission of signals or messages via a wireless network, and may control the reception of signals or messages, etc. via a wireless network (e.g., after being down-converted by the wireless transceiver 802, for example). The processor 804 may be programmable and capable of executing software or other instructions stored on a memory or other computer medium to perform the various tasks and functions described above, such as one or more of the tasks or methods described above. The processor 804 may be (or may include) a programmable processor such as hardware, programmable logic, execution software or firmware, and / or any combination of these. For example, using other terms, the processor 804 and the transceiver 802 may be considered together as a wireless transmitter / receiver system.

[0117] In addition, reference Figure 8 , the controller (or processor) 808 can execute software and instructions and can provide overall control for the station 800 and can Figure 8 Other systems not shown provide controls, such as controlling input / output devices (e.g., display, keyboard), and / or may execute software for one or more applications that may be provided on wireless station 800, such as, for example, an email program, audio / video applications, a word processor, an IP voice application, or other applications or software.

[0118] Additionally, a storage medium may be provided that includes stored instructions that, when executed by a controller or processor, may cause the processor 804 or other controller or processor to perform one or more of the functions or tasks described above.

[0119] According to another example implementation, the RF or (multiple) wireless transceiver 802A / 802B can receive signals or data and / or send or transmit signals or data. The processor 804 (and possibly the transceiver 802A / 802B) can control the RF or wireless transceiver 802A or 802B to receive, issue, broadcast or transmit signals or data.

[0120] However, the embodiments are not limited to the systems given as examples, but those skilled in the art may apply the solution to other communication systems. Another example of a suitable communication system is the 5G concept. It is assumed that the network architecture in 5G will be very similar to the network architecture of LTE Advanced. 5G uses multiple-input-multiple-output (MIMO) antennas, many more base stations or nodes than LTE (the so-called small cell concept), including macro stations that operate in cooperation with smaller stations, and may also adopt various radio technologies to obtain better coverage and enhanced data rates.

[0121] It should be understood that future networks will most likely utilize network function virtualization (NFV), which is a network architecture concept that proposes virtualizing network node functions into "building blocks" or entities that are operably connected or linked together to provide services. A virtualized network function (VNF) may include one or more virtual machines that use standard or general-purpose type servers rather than custom hardware to run computer program code. Cloud computing or data storage may also be utilized. In radio communications, this may mean that node operations may be implemented at least in part in a server, host, or node that is operably coupled to a remote radio head. Node operations may also be distributed across multiple servers, nodes, or hosts. It should also be understood that the distribution of labor between core network operations and base station operations may be different from LTE, or even non-existent.

[0122] The implementation of the various techniques described herein can be implemented in a digital electronic circuit system, or in computer hardware, firmware, software, or in a combination thereof. The implementation can be implemented as a computer program product, i.e., a computer program tangibly embodied in an information carrier (e.g., a machine-readable storage device or a propagation signal) for execution or control of its operation by a data processing device (e.g., a programmable processor, a computer, or multiple computers). The implementation can also be provided on a computer-readable medium or a computer-readable storage medium, which can be a non-transient medium. The implementation of various techniques can also include an implementation provided via a transient signal or medium, and / or a program and / or software implementation that can be downloaded via the Internet or (multiple) other networks (wired networks and / or wireless networks). In addition, the implementation can be provided via machine type communication (MTC), and can also be provided via the Internet of Things (IOT).

[0123] A computer program may be in source code form, object code form, or some intermediate form, and it may be stored in some carrier, distribution medium, or computer readable medium, which may be any entity or device capable of carrying the program. Such carriers include, for example, recording media, computer memory, read-only memory, optical and / or electrical carrier signals, telecommunication signals, and software distribution packets. Depending on the processing power required, a computer program may be executed in a single electronic digital computer, or may be distributed among several computers.

[0124] In addition, implementations of the various techniques described herein may use cyber-physical systems (CPS) (systems of collaborative computing elements that control physical entities). CPS can implement and utilize a large number of interconnected ICT devices (sensors, actuators, processor microcontrollers, etc.) embedded in physical objects in different locations. Mobile cyber-physical systems are a subclass of cyber-physical systems in which the physical systems in question have inherent mobility. Examples of mobile physical systems include mobile robots and electronic devices transported by humans or animals. The popularity of smartphones has increased interest in the field of mobile cyber-physical systems. Therefore, various implementations of the techniques described herein may be provided via one or more of these technologies.

[0125] Computer programs, such as the computer program(s) described above, can be written in any form of programming language, including compiled or interpreted languages, and can be deployed in any form, including as a stand-alone program or as a module, component, subroutine, or other unit or part thereof suitable for use in a computing environment. A computer program can be deployed to be executed on one computer or multiple computers at one station or distributed across multiple stations and interconnected by a communication network.

[0126] The method steps may be performed by one or more programmable processors executing a computer program or portion of a computer program to perform functions by operating on input data and generating output. The method steps may also be performed by special purpose logic circuitry, and the apparatus may be implemented as special purpose logic circuitry, such as an FPGA (field programmable gate array) or an ASIC (application specific integrated circuit).

[0127] Processors suitable for executing computer programs include, for example, general and special purpose microprocessors, and any one or more processors of any type of digital computer, chip or chipset. Typically, the processor will receive instructions and data from a read-only memory or a random access memory or both. The elements of a computer may include at least one processor for executing instructions and one or more memory devices for storing instructions and data. Typically, a computer may also include one or more mass storage devices for storing data, such as magnetic, magneto-optical or optical disks, or operatively coupled to the computer to receive data from or transfer data to or both. Information carriers suitable for implementing computer program instructions and data include all forms of non-volatile memory, including, for example, semiconductor memory devices, such as EPROM, EEPROM and flash memory devices; magnetic disks, such as internal hard disks or removable disks; magneto-optical disks; and CD-ROM and DVD-ROM disks. The processor and memory may be supplemented by or incorporated into a dedicated logic circuit system.

[0128] To provide interaction with a user, implementations may be implemented on a computer having a display device (e.g., a cathode ray tube (CRT) or liquid crystal display (LCD) monitor) for displaying information to the user and a user interface (e.g., a keyboard and a pointing device (e.g., a mouse or trackball)) through which the user can provide input to the computer. Other types of devices may also be used to provide interaction with the user; for example, feedback provided to the user may be any form of sensory feedback, such as visual feedback, auditory feedback, or tactile feedback; and input from the user may be received in any form, including sound, voice, or tactile input.

[0129] The implementation may be implemented in a computing system that includes a back-end component, such as a data server, or includes a middleware component, such as an application server, or that includes a front-end component, such as a client computer with a graphical user interface or a web browser, through which a user can interact with the implementation, or any combination of such back-end, middleware, or front-end components. The components may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include local area networks (LANs) and wide area networks (WANs), such as the Internet.

[0130] While certain features of the described implementations have been illustrated as described herein, many modifications, substitutions, changes and equivalents will now occur to those skilled in the art. It should therefore be understood that the appended claims are intended to cover all such modifications and changes in various embodiments.

Claims

1. A device comprising: at least one processor; as well as at least one memory including computer program code; The at least one memory and the computer program code are configured to cause the apparatus to at least: receiving first inference input data from a network or direct measurement, the first inference input data representing an inference input to an inference operation of a specified radio measurement within the network, the inference operation comprising a machine learning model configured to predict a first value of a device parameter based on the specified radio measurement; sending, to a server connected to the network, indication data representing an indication of the accuracy of an inferred output based on the machine learning model in determining the first value of the device parameter; as well as Receive or not receive from the server an update to the machine learning model based on the indication data, the update enabling the machine learning model to predict a second value of the device parameter based on the specified radio measurement, the second value being more accurate than the first value. 2 . The apparatus of claim 1 , wherein the device parameter is a user device location within the network and the server is a location server.

3. The apparatus of claim 2, wherein the specified radio measurement comprises a reference signal received power.

4. An apparatus according to claim 2, wherein the indicative data includes a version number of the machine learning model. The apparatus according to claim 4 , wherein the indication data is sent periodically according to a specified period. The apparatus according to claim 5 , wherein the designated period is designated by the network. The apparatus of claim 4 , wherein the indication data is sent in response to a condition being satisfied.

8. The apparatus according to claim 7, wherein the conditions satisfied include: The number of inference operations performed by the device within a specified time window is greater than a specified inference threshold.

9. The apparatus of claim 8, wherein the at least one memory and the computer program code are further configured to cause the apparatus to at least: Threshold data representing the inferred threshold is received from the network.

10. The apparatus according to claim 7, wherein the conditions satisfied include: The time when the machine learning model was last updated is greater than a threshold time.

11. The apparatus according to claim 7, wherein the condition satisfied comprises any one of: an inferred frequency being greater than a threshold, and an estimated frequency of position change being greater than a threshold.

12. The apparatus of claim 11, wherein the at least one memory and the computer program code are further configured to cause the apparatus to at least: The first inference input data used in the inference operation is sent to the server.

13. The apparatus according to claim 11, wherein the indication data comprises: A trigger for the server to determine whether to send an update to the machine learning model to the device.

14. A method comprising: receiving first inference input data from a network or direct measurement, the first inference input data representing an inference input to an inference operation of a specified radio measurement within the network, the inference operation comprising a machine learning model configured to predict a first value of a device parameter based on the specified radio measurement; sending, to a server connected to the network, indication data representing an indication of the accuracy of an inferred output based on the machine learning model in determining the value of the device parameter; as well as Receive or not receive from the server an update to the machine learning model based on the indication data, the update enabling the machine learning model to predict a second value of the device parameter based on the specified radio measurement, the second value being more accurate than the first value.

15. An apparatus comprising: at least one processor; as well as at least one memory including computer program code; The at least one memory and the computer program code are configured to cause the apparatus to at least: receiving, from a user device in the network, indication data representing an indication of the accuracy of an inferred output based on a machine learning model in determining a value of a device parameter; as well as Based on the indication data, determine whether to send an update to the machine learning model to the user device.

16. The apparatus of claim 15, wherein the device parameter is a user equipment located within the network, and the apparatus comprises a location server.

17. The apparatus of claim 16, wherein the specified radio measurement comprises a reference signal received power. The apparatus according to claim 16 , wherein the indication data is sent periodically according to a period.

19. The apparatus of claim 18, wherein the at least one memory and the computer program code are further configured to cause the apparatus to at least: The period is specified based on an estimated number of location reports sent by the user equipment within a specified time window.

20. The apparatus of claim 19, wherein the indication data is sent in response to a condition being satisfied.

21. An apparatus according to claim 20, wherein the condition satisfied includes that the time when the machine learning model was last updated is greater than a threshold time.

22. The apparatus according to claim 20, wherein the condition satisfied comprises any one of: a frequency of inference being greater than a threshold, and a frequency of position change being greater than a threshold.

23. An apparatus according to claim 20, wherein the indicative data includes a version number of the machine learning model.

24. A method comprising: receiving, from a user device in the network, indication data representing an indication of the accuracy of an inferred output based on a machine learning model in determining a value of a device parameter; as well as Based on the indication data, determine whether to send an update to the machine learning model to the user device.

25. A computer program product comprising a non-transitory computer-readable storage medium and storing executable code, which, when executed by at least one data processing device, is configured to cause the at least one data processing device to perform the method according to claim 14.

26. An apparatus comprising means for performing the method of claim 14.