Neural network processing management
By introducing a measurement and management mechanism of neural network processing capabilities in wireless communication systems, the problems of increasing power demand and resource management of UE equipment are solved, signal transmission accuracy and battery life are improved, and task scheduling and resource utilization are optimized.
Patent Information
- Application Number
- CN202280095366.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2022-04-28
- Publication Date
- 2024-12-06
AI Technical Summary
In wireless communication systems, as the equipment functions increase, the power demand of user equipment (UE) devices increases, resulting in a shortening of battery life, and the prior art is difficult to effectively manage neural network processing resources, resulting in the accuracy and efficiency of signal transmission due to limit.
By introducing measurement and management mechanisms for neural network processing capabilities in wireless devices, UE devices are allowed to determine their neural network processing capabilities and report this information to the cellular base station. The base station schedules tasks based on this information to avoid tasks that exceed processing capabilities, and provides Priority framework to optimize task scheduling and resource occupancy.
It improves the signal transmission accuracy and efficiency of wireless communication systems, extends the battery life of UE equipment, and ensures the reasonable allocation of tasks and efficient utilization of resources.
Smart Images

Figure CN119096647A_ABST
Abstract
Description
[0001] The present application relates to wireless communications, and more particularly to systems, apparatus, and methods for managing neural network processing resources in wireless communication systems. Background Art
[0002] The use of wireless communication systems is growing rapidly. In recent years, wireless devices such as smart phones and tablet computers have become increasingly sophisticated. In addition to supporting phone calls, many mobile devices (i.e., user equipment devices or UEs) now also provide access to the Internet, email, text messaging, and navigation using a global positioning system (GPS), and are capable of operating complex and sophisticated applications that utilize these functions. In addition, there are many different wireless communication technologies and wireless communication standards. Some examples of wireless communication standards include GSM, UMTS (e.g., associated with WCDMA or TD-SCDMA air interfaces), LTE, Advanced LTE (LTE-A), NR, HSPA, 3GPP2 CDMA2000 (e.g., 1xRTT, 1xEV-DO, HRPD, eHRPD), IEEE 802.11 (WLAN or Wi-Fi), Bluetooth, etc. TM wait.
[0003] The introduction of an ever-increasing number of features and functions in wireless communication devices has also created a continuing need for improved wireless communications and improved wireless communication devices. In particular, it is important to ensure the accuracy of signals transmitted and received by user equipment (UE) devices (e.g., by wireless devices such as cellular phones, base stations, and relay stations used in wireless cellular communications). In addition, increasing the functionality of UE devices may place a great strain on the battery life of UE devices. Therefore, it is also very important to reduce the power requirements in UE device designs while allowing UE devices to maintain good transmission and reception capabilities to improve communications. Therefore, improvements are desired in this area. Summary of the invention
[0004] Embodiments of apparatus, systems, and methods for managing neural network processing resources in a wireless communication system are presented herein.
[0005] According to the techniques described herein, a wireless device may have the capability to perform neural network processing for any of a variety of tasks that may be related to wireless communication (such as for beam prediction, channel state information measurement, positioning estimation, etc.) and / or any of a variety of tasks that may be unrelated to wireless communication.
[0006] The wireless device may be able to determine its neural network processing capability, for example including a measure of a maximum number of units of neural network processing capability available to the wireless device (which may be referred to as neural network processing units), and / or any of a variety of other possible indicators of its neural network processing capability. The wireless device may provide information indicative of the neural network processing capability of the wireless device to a serving cellular base station in a cellular communication system.
[0007] The cellular base station (and / or other elements of the cellular network associated with the cellular base station) may use the neural network processing capability information for the wireless device to determine whether and when to schedule tasks or activities that may involve neural network processing based on the neural network processing capability of the wireless device. For example, the cellular base station may generally avoid scheduling neural network processing tasks that exceed the neural network processing capability of the wireless device at any given time.
[0008] Alternatively or in addition, it is possible to provide a priority framework for neural network processing tasks. In such a scenario, if the scheduled task has a higher priority than one or more currently scheduled neural network processing tasks for the wireless device, the cellular base station may schedule a neural network processing task for the wireless device that would exceed the indicated neural network processing capabilities of the wireless device.
[0009] Also described herein are techniques for determining occupancy times for various possible neural network processing tasks, which may allow the duration of neural network processing tasks to be consistently determined by both the cellular network and the wireless device, which may improve the ability of the cellular network to accurately determine when the neural network processing resources of the wireless device are available and when they are not available.
[0010] It should be noted that the techniques described herein may be implemented in and / or used with a number of different types of devices, including but not limited to base stations, access points, mobile phones, portable media players, tablet computers, wearable devices, unmanned aerial vehicles, unmanned flight controllers, automobiles and / or motor vehicles, and various other computing devices.
[0011] This disclosure is intended to provide a brief overview of some of the topics described in this document. Therefore, it should be understood that the above features are only examples and should not be construed as narrowing the scope or essence of the topics described herein in any way. Other features, aspects, and advantages of the topics described herein will become apparent through the following detailed description, drawings, and claims. BRIEF DESCRIPTION OF THE DRAWINGS
[0012] A better understanding of the present subject matter may be obtained when the following detailed description of various embodiments is considered in conjunction with the following drawings, in which:
[0013] Figure 1 illustrates an exemplary (and simplified) wireless communication system according to some embodiments;
[0014] Figure 2 illustrates an exemplary base station in communication with an exemplary wireless user equipment (UE) device according to some embodiments;
[0015] Figure 3 illustrates an exemplary block diagram of a UE according to some embodiments;
[0016] Figure 4 illustrates an exemplary block diagram of a base station according to some embodiments;
[0017] Figure 5 is a flow chart illustrating aspects of an exemplary possible method for managing neural network processing resources in a wireless communication system according to some embodiments;
[0018] Figure 6 illustrates possible example timing aspects of a wireless communication system that adheres to Z and Z' offset requirements when scheduling CSI reporting in accordance with some embodiments;
[0019] Figures 7 to 10 illustrates possible example timing aspects of a wireless communication system following various possible neural network processing occupancy rules according to some embodiments;
[0020] Fig.11 illustrates an example MAC CE format including a field for indicating a priority for a neural network processing task according to some embodiments;
[0021] Fig.12 illustrates example aspects of an NPU-DRX configuration according to some embodiments; and
[0022] Fig.13 Aspects of example scenarios in which dynamic NPU usage may be configurable according to some embodiments are illustrated.
[0023] Although the features described herein are susceptible to various modifications and alternative forms, specific embodiments thereof are shown by way of example in the drawings and described in detail herein. However, it should be understood that the drawings and detailed description thereof are not intended to be limited to the specific forms disclosed, but on the contrary, are intended to cover all modifications, equivalents, and alternatives falling within the spirit and scope of the subject matter as defined by the appended claims. DETAILED DESCRIPTION
[0024] Acronyms
[0025] Various acronyms are used throughout this disclosure. Definitions of the most commonly used acronyms that may appear throughout this disclosure are provided below:
[0026] UE: User Equipment
[0027] RF: Radio Frequency
[0028] BS: Base Station
[0029] GSM: Global System for Mobile Communications
[0030] UMTS: Universal Mobile Telecommunications System
[0031] LTE: Long Term Evolution
[0032] NR: New Radio
[0033] TX: Send / Send
[0034] RX: Receive / Receive
[0035] RAT: Radio Access Technology
[0036] ●TRP: Transmit Receive Point
[0037] ●DCI: Downlink Control Information
[0038] CORESET: Control resource set
[0039] QCL: quasi-co-located or quasi-co-located
[0040] ●CSI: Channel State Information
[0041] ●CSI-RS: Channel State Information Reference Signal
[0042] ●CSI-IM: Channel State Information Interference Management
[0043] ●CMR: Channel Measurement Resource
[0044] IMR: Interference Measurement Resource
[0045] ZP: Zero Power
[0046] NZP: Non-Zero Power
[0047] ●CQI: Channel Quality Indicator
[0048] ●PMI: Precoding matrix indicator
[0049] RI: Rank Indicator
[0050] the term
[0051] The following is a glossary of terms that will appear in this disclosure:
[0052] Memory medium—any of various types of non-transitory memory devices or storage devices. The term "memory medium" is intended to include installation media, such as CD-ROM, floppy disk, or tape devices; computer system memory or random access memory, such as DRAM, DDR RAM, SRAM, EDO RAM, Rambus RAM, etc.; non-volatile memory, such as flash memory, magnetic media, such as hard disk drives or optical storage devices; registers or other similar types of memory elements, etc. The memory medium may also include other types of non-transitory memory or combinations thereof. In addition, the memory medium may be located in a first computer system executing a program, or may be located in a different second computer system connected to the first computer system via a network such as the Internet. In the latter example, the second computer system may provide program instructions to the first computer system for execution. The term "memory medium" may include two or more memory media, which may reside in different locations in different computer systems connected, for example, via a network. The memory medium may store program instructions (e.g., embodied as a computer program) that may be executed by one or more processors.
[0053] Carrier Media—Memory media as described above, as well as physical transmission media such as a bus, network, and / or other physical transmission media that convey signals such as electrical, electromagnetic, or digital signals.
[0054] Computer system (or computer)—any of various types of computing or processing systems, including a personal computer system (PC), a mainframe computer system, a workstation, a network appliance, an Internet appliance, a personal digital assistant (PDA), a television system, a grid computing system, or other devices or combinations of devices. In general, the term "computer system" may be broadly defined as any device (or combination of devices) that includes at least one processor that executes instructions from a memory medium.
[0055] User Equipment (UE) (or "UE device") - any of various types of computer systems or devices that are mobile or portable and that perform wireless communications. Examples of UE devices include mobile phones or smart phones (e.g., iPhone TM , based on Android TM phones), tablet computers (e.g., iPad TM 、Samsung Galaxy TM ), portable gaming devices (e.g., Nintendo DS TM , PlayStation Portable TM、Gameboy Advance TM , iPhone TM ), wearable devices (e.g., smart watches, smart glasses), laptops, PDAs, portable Internet devices, music players, data storage devices, other handheld devices, automobiles and / or motor vehicles, unmanned aerial vehicles (UAVs) (e.g., drones), UAV controllers (UACs), etc. In general, the term "UE" or "UE device" can be broadly defined to cover any electronic device, computing device, and / or telecommunication device (or a combination of these devices) that is easily transportable by a user and capable of wireless communication.
[0056] Wireless Device—Any of various types of computer systems or devices that perform wireless communications. A wireless device may be portable (or mobile), or may be stationary or fixed at a location. A UE is an example of a wireless device.
[0057] Communication device - any of various types of computer systems or devices that perform communication, where the communication may be wired or wireless. A communication device may be portable (or mobile), or may be stationary or fixed at a location. A wireless device is an example of a communication device. A UE is another example of a communication device.
[0058] Base Station (BS)—The term "base station" has the full breadth of its ordinary meaning and includes at least a wireless communication station installed at a fixed location and used to communicate as part of a wireless telephone system or radio system.
[0059] Processing element (or processor) - refers to various elements or combinations of elements that are capable of performing functions in a device (e.g., a user equipment device or a cellular network device). Processing elements may include, for example, processors and associated memory, portions or circuits of individual processor cores, entire processor cores, processor arrays, circuits such as ASICs (application specific integrated circuits), programmable hardware elements such as field programmable gate arrays (FPGAs), and any of the above combinations.
[0060] Wi-Fi—The term "Wi-Fi" has the full scope of its usual meaning and includes at least wireless communication networks or RATs that are served by wireless LAN (WLAN) access points and provide connectivity to the Internet through these access points. Most modern Wi-Fi networks (or WLAN networks) are based on the IEEE 802.11 standard and are marketed under the name "Wi-Fi". Wi-Fi (WLAN) networks are distinct from cellular networks.
[0061] Automatically—refers to an action or operation being performed by a computer system (e.g., software executed by a computer system) or a device (e.g., a circuit, a programmable hardware element, an ASIC, etc.) without the action or operation being directly specified or performed by a user input. Thus, the term "automatically" is in contrast to an operation being manually performed or specified by a user, where the user provides input to directly perform the operation. An automatic process may be initiated by input provided by a user, but subsequent actions performed "automatically" are not specified by the user, i.e., are not performed "manually," where the user specifies each action to be performed. For example, a user filling out an electronic form by selecting each field and providing input specifying information (e.g., by typing in information, selecting checkboxes, radio selections, etc.) is manually filling out the form, even though the computer system must update the form in response to the user action. The form may be automatically filled out by a computer system, where the computer system (e.g., software executed on the computer system) analyzes the fields of the form and fills out the form without any user input specifying the answers to the fields. As indicated above, a user may invoke automatic filling out of a form, but not participate in the actual filling out of the form (e.g., the user does not manually specify the answers to the fields but rather they are automatically completed). This specification provides various examples of operations that are automatically performed in response to actions that a user has taken.
[0062] Configured to—Various components may be described as being “configured to” perform one or more tasks. In such contexts, “configured to” is a broad statement that generally means “having a structure” that performs one or more tasks during operation. Thus, a component may be configured to perform a task even when the component is not currently performing the task (e.g., a set of electrical conductors may be configured to electrically connect a module to another module even when the two modules are not connected). In some contexts, “configured to” may be a broad statement that generally means “having a structure” that performs one or more tasks during operation. Thus, a component may be configured to perform a task even when the component is not currently turned on. Typically, circuitry that forms a structure corresponding to “configured to” may include hardware circuitry.
[0063] For ease of description, various components may be described as performing one or more tasks. Such descriptions should be interpreted as including the phrase "configured to". The statement that a component is configured to perform one or more tasks is expressly intended not to invoke the interpretation of the sixth paragraph of Section 112 of Title 35 of the United States Code on that component.
[0064] Figure 1 and Figure 2 -Exemplary Communication System
[0065] Figure 1 An exemplary (and simplified) wireless communication system that can implement various aspects of the present disclosure according to some embodiments is illustrated. Note that Figure 1 The system is only one example of a possible system, and the embodiments may be implemented in any of a variety of systems as desired.
[0066] As shown, the exemplary wireless communication system includes a base station 102, which communicates with one or more (e.g., any number) user equipment 106A, 106B, etc. to 106N through a transmission medium. Each user equipment may be referred to as a "user equipment" (UE) or a UE device in this document. Therefore, the user equipment 106 is referred to as a UE or a UE device.
[0067] The base station 102 may be a base transceiver station (BTS) or a cell site, and may include hardware and / or software that implements wireless communication with UEs 106A to 106N. If the base station 102 is implemented in the context of LTE, it may be referred to as an "eNodeB" or "eNB". If the base station 102 is implemented in the context of 5G NR, it may alternatively be referred to as a "gNodeB" or "gNB". The base station 102 may also be equipped to communicate with a network 100 (e.g., a core network of a cellular service provider, a telecommunications network such as a public switched telephone network (PSTN), and / or the Internet, and various possible networks). Therefore, the base station 102 may facilitate communication between user devices and / or between user devices and the network 100. The communication area (or coverage area) of a base station may be referred to as a "cell". Also as used herein, with respect to a UE, sometimes a base station may be considered to represent a network in consideration of the uplink and downlink communications of the UE. Therefore, a UE communicating with one or more base stations in a network may also be understood as a UE communicating with a network.
[0068] The base station 102 and the user equipment may be configured to communicate over a transmission medium using any of a variety of radio access technologies (RATs), also referred to as wireless communication technologies or telecommunication standards, such as GSM, UMTS (WCDMA), LTE, Advanced LTE (LTE-A), LAA / LTE-U, 5G NR, 3GPP2 CDMA2000 (e.g., 1xRTT, 1xEV-DO, HRPD, eHRPD), Wi-Fi, etc.
[0069] Base station 102 and other similar base stations operating according to the same or different cellular communication standards may thus be provided as one or more cell networks that may provide continuous or nearly continuous overlapping service to UE 106 and similar devices over a geographic area via one or more cellular communication standards.
[0070] Note that UE 106 is capable of communicating using multiple wireless communication standards. For example, UE 106 may be configured to communicate using either or both of the 3GPP cellular communication standard or the 3GPP2 cellular communication standard. In some embodiments, UE 106 may be configured to perform techniques for managing neural network processing resources in a wireless communication system, such as according to the various methods described herein. UE 106 may also or alternatively be configured to communicate using WLAN, Bluetooth, or the like. TM , one or more global navigation satellite systems (GNSS, such as GPS or GLONASS), one and / or more mobile television broadcast standards (e.g., ATSC-M / H), etc. Other combinations of wireless communication standards (including more than two wireless communication standards) are also possible.
[0071] Figure 2 An exemplary user equipment 106 (e.g., one of devices 106A to 106N) communicating with a base station 102 according to some embodiments is illustrated. UE 106 may be a device with wireless network connectivity, such as a mobile phone, a handheld device, a wearable device, a computer or tablet computer, an unmanned aerial vehicle (UAV), an unmanned flight controller (UAC), a car, or almost any type of wireless device. UE 106 may include a processor (processing element) configured to execute program instructions stored in a memory. UE 106 may perform any method implementation in the method implementation described herein by executing such stored instructions. Alternatively or in addition, UE 106 may include a programmable hardware element, such as an FPGA (field programmable gate array), an integrated circuit, and / or any of various other possible hardware components configured to perform (e.g., individually or in combination) any method implementation in the method implementation described herein or any part of any method implementation in the method implementation described herein. UE 106 may be configured to communicate using any of a plurality of wireless communication protocols. For example, UE 106 may be configured to communicate using two or more of CDMA2000, LTE, LTE-A, 5G NR, WLAN, or GNSS. Other combinations of wireless communication standards are also possible.
[0072] UE 106 may include one or more antennas for communicating using one or more wireless communication protocols according to one or more RAT standards. In some embodiments, UE 106 may share one or more parts of a receive chain and / or a transmit chain between multiple wireless communication standards. The shared radio component may include a single antenna, or may include multiple antennas (e.g., for a multiple-input, multiple-output or "MIMO" antenna system) for performing wireless communications. Typically, the radio component may include any combination of a baseband processor, an analog RF signal processing circuit (e.g., including filters, mixers, oscillators, amplifiers, etc.), or a digital processing circuit (e.g., for digital modulation and other digital processing). Similarly, the radio component may use the aforementioned hardware to implement one or more receive chains and transmit chains. For example, UE 106 may share one or more parts of a receive chain and / or a transmit chain between multiple wireless communication technologies such as those discussed above.
[0073] In some embodiments, UE 106 may include any number of antennas and may be configured to use the antennas to send and / or receive directional wireless signals (e.g., beams). Similarly, BS 102 may also include any number of antennas and may be configured to use the antennas to send and / or receive directional wireless signals (e.g., beams). In order to receive and / or send such directional signals, the antennas of UE 106 and / or BS 102 may be configured to apply different "weights" to different antennas. The process of applying these different weights may be referred to as "pre-coding".
[0074] In some embodiments, the UE 106 may include a separate transmit chain and / or receive chain (e.g., including separate antennas and other radio components) for each wireless communication protocol with which it is configured to communicate. As another possibility, the UE 106 may include one or more radio components shared between multiple wireless communication protocols, and one or more radio components used uniquely by a single wireless communication protocol. For example, the UE 106 may include a shared radio component for communicating using any of LTE or CDMA2000 1xRTT (or LTE or NR, or LTE or GSM), and a shared radio component for communicating using Wi-Fi and Bluetooth. TM Each of the radio components communicates with each other. Other configurations are also possible.
[0075] Figure 3 - Block diagram of an exemplary UE device
[0076] Figure 3A block diagram of an exemplary UE 106 according to some embodiments is illustrated. As shown, the UE 106 may include a system on chip (SOC) 300, which may include parts for various purposes. For example, as shown, the SOC 300 may include a processor 302 that can execute program instructions for the UE 106, and a display circuit 304 that can perform graphics processing and provide display signals to a display 360. The SOC 300 may also include a sensor circuit 370, which may include a component for sensing or measuring any of various possible characteristics or parameters of the UE 106. For example, the sensor circuit 370 may include a motion sensing circuit that is configured to detect the motion of the UE 106, for example, using a gyroscope, an accelerometer, and / or any of various other motion sensing components. As another possibility, the sensor circuit 370 may include one or more temperature sensing components, which are used, for example, to measure the temperature of each of one or more antenna panels and / or other components of the UE 106. Any of a variety of other possible types of sensor circuits may also or alternatively be included in the UE 106, as desired. The processor 302 may also be coupled to a memory management unit (MMU) 340, which may be configured to receive addresses from the processor 302 and translate those addresses into locations in memory (e.g., memory 306, read-only memory (ROM) 350, NAND flash memory 310) and / or other circuits or devices, such as display circuit 304, radio 330, connector I / F 320, and / or display 360. The MMU 340 may be configured to perform memory protection and page table translation or setup. In some embodiments, the MMU 340 may be included as part of the processor 302.
[0077] As shown, SOC 300 can be coupled to various other circuits of UE 106. For example, UE 106 may include various types of memory (e.g., including NAND flash memory 310), connector interface 320 (e.g., for coupling to a computer system, a docking station, a charging station, etc.), display 360, and wireless communication circuitry 330 (e.g., for LTE, LTE-A, NR, CDMA2000, Bluetooth TM, Wi-Fi, GPS, etc.). The UE device 106 may include or be coupled to at least one antenna (e.g., 335a), and may include multiple antennas (e.g., as shown by antennas 335a and 335b) for performing wireless communications with base stations and / or other devices. Antennas 335a and 335b are shown by way of example, and the UE device 106 may include fewer or more antennas. In general, one or more antennas are collectively referred to as antennas 335. For example, the UE device 106 may use antenna 335 with the aid of radio circuitry 330 to perform wireless communications. The wireless communication circuit may include multiple receive chains and / or multiple transmit chains for receiving and / or transmitting multiple spatial streams, such as in a multiple-input multiple-output (MIMO) configuration. As mentioned above, in some embodiments, the UE may be configured to perform wireless communications using multiple wireless communication standards.
[0078] UE 106 may include hardware and software components for implementing methods for UE 106 to perform techniques for managing neural network processing resources in a wireless communication system, such as those described further herein. Processor 302 of UE device 106 may be configured to implement part or all of the methods described herein, for example, by executing program instructions stored on a memory medium (e.g., a non-transitory computer-readable memory medium). In other embodiments, processor 302 may be configured as a programmable hardware element, such as an FPGA (field programmable gate array) or as an ASIC (application-specific integrated circuit). In addition, as Figure 3 As shown, processor 302 may be coupled to other components and / or may interoperate with other components to perform techniques for managing neural network processing resources in a wireless communication system in accordance with various embodiments disclosed herein. Processor 302 may also implement various other applications and / or end-user applications running on UE 106.
[0079] In some embodiments, radio 330 may include a separate controller dedicated to controlling communications for each corresponding RAT standard. Figure 3 As shown, the radio component 330 may include a Wi-Fi controller 352, a cellular controller (e.g., an LTE and / or LTE-A controller) 354, and a Bluetooth TM Controller 356, and in at least some embodiments, one or more or all of these controllers may be implemented as respective integrated circuits (referred to as ICs or chips) that communicate with each other and with SOC 300 (more specifically, with processor 302). For example, Wi-Fi controller 352 may communicate with cellular controller 354 via a cell-ISM link or WCI interface, and / or Bluetooth controller 354 may communicate with the cellular controller 354 via a cell-ISM link or WCI interface. TMThe controller 356 may communicate with the cellular controller 354 via a cell-ISM link, etc. Although three separate controllers are shown within the radio 330, other embodiments may be implemented in the UE device 106 having fewer or more similar controllers for various different RATs.
[0080] Additionally, embodiments are contemplated in which the controller may implement functionality associated with multiple radio access technologies. For example, according to some embodiments, in addition to hardware and / or software components for performing cellular communications, the cellular controller 354 may also include hardware and / or software components for performing one or more activities associated with Wi-Fi, such as Wi-Fi preamble detection, and / or generation and transmission of Wi-Fi physical layer preamble signals.
[0081] Figure 4 - Block diagram of an exemplary base station
[0082] Figure 4 1 illustrates a block diagram of an exemplary base station 102 according to some embodiments. Note that Figure 4 The base station of is only one example of a possible base station. As shown, the base station 102 may include a processor 404 that may execute program instructions for the base station 102. The processor 404 may also be coupled to a memory management unit (MMU) 440 or other circuit or device, which may be configured to receive addresses from the processor 404 and convert these addresses to locations in memory (e.g., memory 460 and read-only memory (ROM) 450).
[0083] Base station 102 may include at least one network port 470. Network port 470 may be configured to couple to a telephone network and provide access to the telephone network described above. Figure 1 and Figure 2 106. The network port 470 (or an additional network port) may also or alternatively be configured to couple to a cellular network, such as a core network of a cellular service provider. The core network may provide mobility-related services and / or other services to multiple devices, such as the UE device 106. In some cases, the network port 470 may be coupled to the telephone network via the core network, and / or the core network may provide the telephone network (e.g., in other UE devices served by the cellular service provider).
[0084] In some embodiments, base station 102 may be a next generation base station, e.g., a 5G New Radio (5G NR) base station, or "gNB". In such embodiments, base station 102 may be connected to a legacy evolved packet core (EPC) network and / or to an NR core (NRC) network. Furthermore, base station 102 may be considered a 5G NR cell and may include one or more transmit and receive points (TRPs). Furthermore, a UE capable of operating in accordance with 5G NR may be connected to one or more TRPs within one or more gNBs.
[0085] The base station 102 may include at least one antenna 434 and possibly multiple antennas. One or more antennas 434 may be configured to operate as a wireless transceiver and may be further configured to communicate with the UE device 106 via the radio component 430. The antenna 434 communicates with the radio component 430 via a communication chain 432. The communication chain 432 may be a receive chain, a transmit chain, or both. The radio component 430 may be designed to communicate via various wireless telecommunication standards, including but not limited to 5G NR, 5G NR SAT, LTE, LTE-A, GSM, UMTS, CDMA2000, Wi-Fi, etc.
[0086] The base station 102 may be configured to perform wireless communications using multiple wireless communication standards. In some cases, the base station 102 may include multiple radio components that enable the base station 102 to communicate according to multiple wireless communication technologies. For example, as one possibility, the base station 102 may include an LTE radio component for performing communications according to LTE and a 5G NR radio component for performing communications according to 5G NR. In this case, the base station 102 may be able to operate as both an LTE base station and a 5G NR base station. As another possibility, the base station 102 may include a multimode radio component that can perform communications according to any one of a plurality of wireless communication technologies (e.g., 5G NR and Wi-Fi, 5G NR SAT and Wi-Fi, LTE and Wi-Fi, LTE and UMTS, LTE and CDMA2000, UMTS and GSM, etc.).
[0087] As further described later herein, BS 102 may include hardware and software components for implementing or supporting a specific implementation of the features described herein. The processor 404 of the base station 102 may be configured to implement and / or support a specific implementation of part or all of the methods described herein, for example, by executing program instructions stored on a memory medium (e.g., a non-transitory computer-readable memory medium). Alternatively, the processor 404 may be configured as a programmable hardware element such as an FPGA (field programmable gate array), or as an ASIC (application-specific integrated circuit) or a combination thereof. In the case of certain RATs (e.g., Wi-Fi), the base station 102 may be designed as an access point (AP), in which case the network port 470 may be implemented to provide access to a wide area network and / or one or more local area networks, for example, it may include at least one Ethernet port, and the radio component 430 may be designed to communicate according to the Wi-Fi standard.
[0088] In addition, as described herein, the one or more processors 404 may include one or more processing elements. Thus, the processor 404 may include one or more integrated circuits (ICs) configured to perform the functions of the processor 404. In addition, each integrated circuit may include circuits (e.g., first circuits, second circuits, etc.) configured to perform the functions of the one or more processors 404.
[0089] In addition, as described herein, radio 430 may include one or more processing elements. Thus, radio 430 may include one or more integrated circuits (ICs) configured to perform the functions of radio 430. In addition, each integrated circuit may include a circuit (e.g., a first circuit, a second circuit, etc.) configured to perform the functions of radio 430.
[0090] Reference signal
[0091] A wireless device (such as a user equipment) may be configured to perform various tasks including using reference signals (RS) provided by one or more cellular base stations. For example, initial access and beam measurements of the wireless device may be performed based at least in part on synchronization signal blocks (SSBs) provided by one or more cells provided by one or more cellular base stations within the communication range of the wireless device. Another type of reference signal commonly provided in a cellular communication system may include a channel state information (CSI) RS. Various types of CSI-RS may be provided for tracking (e.g., for time and frequency offset tracking), beam management (e.g., CSI-RS configured with repetitions to help determine one or more beams for uplink and / or downlink communications), and / or channel measurement (e.g., CSI-RS configured in a resource set for measuring the quality of a downlink channel and reporting information related to the quality measurement to a base station), as well as various possibilities. For example, in the case where the CSI-RS is used for CSI acquisition, the UE may periodically perform channel measurements and transmit channel state information (CSI) to the BS. The base station may then receive and use the channel state information during communication with the wireless device to determine adjustments to various parameters. In particular, the BS may use the received channel state information to adjust the coding of its downlink transmissions to improve downlink channel quality.
[0092] In many cellular communication systems, a base station may periodically transmit some or all such reference signals (or pilot signals), such as SSB and / or CSI-RS. In some cases, aperiodic reference signals (e.g., aperiodic reference signals for aperiodic CSI reporting) may also or alternatively be provided.
[0093] As a detailed example, at least according to some embodiments, in the 3GPP NR cellular communication standard, the channel state information fed back from the UE based on the CSI-RS for CSI acquisition may include one or more of a channel quality indicator (CQI), a precoding matrix indicator (PMI), a rank indicator (RI), a CSI-RS resource indicator (CRI), a SSBRI (SS / PBCH resource block indicator, and a layer indicator (LI).
[0094] Channel quality information may be provided to the base station for link adaptation, for example, to provide guidance on which modulation and coding scheme (MCS) the base station should use when sending data. For example, when the downlink channel communication quality between the base station and the UE is determined to be high, the UE may feedback a high CQI value, which may cause the base station to send data using a relatively high modulation order and / or a low channel coding rate. For another example, when the downlink channel communication quality between the base station and the UE is determined to be low, the UE may feedback a low CQI value, which may cause the base station to send data using a relatively low modulation order and / or a high channel coding rate.
[0095] The PMI feedback may include preferred precoding matrix information and may be provided to the base station to indicate which MIMO precoding scheme the base station should use. In other words, the UE may measure the quality of the downlink MIMO channel between the base station and the UE based on the pilot signal received on the channel, and may recommend which MIMO precoding the desired base station applies through PMI feedback. In some cellular systems, the PMI configuration is represented in matrix form, which provides linear MIMO precoding. The base station and the UE may share a codebook consisting of multiple precoding matrices, wherein each MIMO precoding matrix in the codebook may have a unique index. Therefore, as part of the channel state information fed back by the UE, the PMI may include an index (or possibly multiple indexes) corresponding to the most preferred MIMO precoding matrix (or multiple matrices) in the codebook. This may enable the UE to minimize the amount of feedback information. Therefore, at least according to some embodiments, the PMI may indicate which precoding matrix from the codebook should be used for transmission to the UE.
[0096] For example, when the base station and UE have multiple antennas, the rank indicator information (RI feedback) may indicate the number of transmission layers that the UE determines can be supported by the channel, which can achieve multi-layer transmission through spatial multiplexing. RI and PMI together may allow the base station to know which precoding needs to be applied to which layer, for example, depending on the number of transmission layers.
[0097] In some cellular systems, the PMI codebook is defined based on the number of transmission layers. In other words, for R layer transmission, N N codebooks can be defined. t ×R matrix (for example, where R represents the number of layers, N t denotes the number of transmitter antenna ports, and N denotes the size of the codebook). In such a scenario, the number of transmission layers (R) may conform to the rank value of the precoding matrix (N t ×R matrix), and thus R in this context may be referred to as a “rank indicator (RI)”.
[0098] Thus, the channel state information may include an assigned rank (e.g., a rank indicator or RI). For example, a MIMO-enabled UE communicating with a BS may include four receiver chains, e.g., may include four antennas. The BS may also include four or more antennas to implement MIMO communication (e.g., 4×4 MIMO). Thus, the UE is able to simultaneously receive up to four (or more) signals (e.g., layers) from the BS. Layer-to-antenna mapping may be applied, e.g., each layer may be mapped to any number of antenna ports (e.g., antennas). Each antenna port may transmit and / or receive information associated with one or more layers. The rank may include multiple bits and may indicate the number of signals that the BS may transmit to the UE in an upcoming time period (e.g., during an upcoming transmit time interval or TTI). For example, an indication of rank 4 may indicate that the BS will transmit 4 signals to the UE. As one possibility, the length of the RI may be two bits (e.g., because two bits are sufficient to distinguish 4 different rank values). Note that other numbers and / or configurations of antennas (e.g., at either or both of the UE or BS) and / or other numbers of data layers are also possible according to various embodiments.
[0099] Figure 5 -Neural network processing management
[0100] Introducing machine learning techniques (e.g., including the use of neural networks) into wireless communication systems may have significant benefits. For example, at least in some embodiments, such techniques may provide the potential for greater accuracy, greater speed, and / or reduced power costs when used to perform certain wireless communication tasks. Some areas where machine learning may be introduced to facilitate wireless communications, at least according to some embodiments, may include beam prediction in the time and / or spatial domains, channel state information measurement and reporting, and / or positioning estimation.
[0101] As machine learning techniques become more prevalent, there is also the possibility that such machine learning tools may be used by wireless devices for any of a variety of possible non-wireless communication purposes, such as by other applications active on the wireless device.
[0102] Therefore, it may be useful to provide techniques for managing machine learning processing resources of a wireless device, for example to help prevent or reduce the likelihood that more demands may be placed on the machine learning processing resources of a wireless device than the wireless device is able to handle. To illustrate one set of such possible techniques, Figure 5 is a flow chart illustrating a method for managing neural network processing resources in a wireless communication system according to at least some embodiments.
[0103] Figure 5Aspects of the methods of can be implemented by a wireless device, for example, in conjunction with one or more cellular base stations (such as UE 106 and BS 102 shown and described with respect to various figures herein), or more generally, as desired, in conjunction with any of the computer circuits, systems, devices, elements or components, etc. shown in the above figures. For example, the processor (and / or other hardware) of such a device can be configured to cause the device to perform any combination of the method elements shown and / or other method elements.
[0104] It is noted that although the present invention is described in a manner involving the use of communication techniques and / or features associated with 3GPP and / or NR specification documents, Figure 5 However, this description is not intended to limit the present disclosure and may be used in any suitable wireless communication system as needed. Figure 5 In various embodiments, some of the elements of the method shown may be performed simultaneously in an order different from that shown, may be replaced by other method elements, or may be omitted. Additional method elements may also be performed as needed. As shown in the figure, Figure 5 The method can be operated as follows.
[0105] In 502, the wireless device may establish a wireless link with a cellular base station. According to some embodiments, the wireless link may include a cellular link according to 5G NR. For example, the wireless device may establish a session with an AMF entity of a cellular network through one or more gNBs that provide radio access to the cellular network. As another possibility, the wireless link may include a cellular link according to LTE. For example, the wireless device may establish a session with a mobility management entity of a cellular network through an eNB that provides radio access to the cellular network. According to various embodiments, other types of cellular links are also possible, and the cellular network may also or alternatively operate according to another cellular communication technology (e.g., UMTS, CDMA2000, GSM, etc.).
[0106] Establishing a wireless link may include establishing an RRC connection with a serving cellular base station according to at least some embodiments. Establishing a first RRC connection may include configuring various parameters for communication between the wireless device and the cellular base station, establishing environmental information of the wireless device, and / or any of a variety of other possible features, for example, involving establishing an air interface of the wireless device for cellular communication with a cellular network associated with the cellular base station. After establishing the RRC connection, the wireless device may operate in an RRC connected state. In some instances, the RRC connection may also be released (e.g., after a certain period of inactivity relative to data communication), in which case the wireless device may operate in an RRC idle state or an RRC inactive state. In some instances, the wireless device may perform a handover (e.g., when in an RRC connected mode) or a cell reselection (e.g., when in an RRC idle mode or an RRC inactive mode) to a new serving cell, for example, due to wireless device mobility, changes in wireless medium conditions, and / or any of a variety of other possible reasons.
[0107] According to at least some embodiments, a wireless device may establish multiple wireless links according to a multi-TRP configuration, for example, with multiple TRPs of a cellular network. In this case, the wireless device may be configured (e.g., via RRC signaling) with one or more transmit control indicators (TCIs), for example, which may correspond to various beams that may be used to communicate with the TRPs. In addition, there may be situations where one or more configured TCI states may be activated by a medium access control (MAC) control element (CE) of the wireless device at a particular time.
[0108] In at least some instances, establishing the wireless link may include the wireless device providing capability information of the wireless device. Such capability information may include information related to any of a plurality of types of wireless device capabilities. As further described later herein, one such type of capability information may relate to a neural network processing capability of the wireless device.
[0109] At 504, the wireless device may determine neural network processing capability information for the wireless device. The neural network processing capability information may include a number of supported neural network processing units (NPUs) for the wireless device. The number of supported NPUs for the wireless device may represent a maximum total neural network processing capability for the wireless device, or may represent an amount of neural network processing capability available for wireless communications with a cellular base station.
[0110] In some embodiments, it is also possible or alternatively possible to determine the neural network processing capability information at a finer granularity. For example, neural network processing capability information, potentially including the number of supported NPUs, may be determined for each of one or more specific types of neural network processing tasks that may be configured in a wireless communication system, such as beam prediction in the spatial domain, beam prediction in the time domain, channel state information measurement and reporting, and / or positioning estimation.
[0111] In some cases, the neural network processing capability information may include availability information of neural network processing resources for the wireless device. For example, the neural network processing capability information may indicate that an indicated number of supported NPUs for the wireless device are always available for neural network processing tasks associated with the wireless link, for example, where the neural network processing resources of the wireless device are dedicated to such purpose and limiting power consumption is not a high priority consideration. As another example, the neural network processing capability information may indicate one or more of a preferred on duration, a preferred off duration, a preferred periodicity, and / or other information that facilitates shared use (e.g., shared use with other wireless or non-wireless applications) and / or power conservation according to a periodic discontinuous availability pattern of the indicated number of supported NPUs for the wireless device. As a further possibility, a hybrid approach may be used, for example where a particular portion of the NPUs of the wireless device are always available for neural network processing tasks associated with the wireless link, while another portion of the NPUs of the wireless device are available for neural network processing tasks associated with the wireless link, in which case the neural network processing capability information may indicate the hybrid availability.
[0112] As another possibility, it may be the case that the neural network processing capability information indicates additional neural network processing capability information for one or more NPUs of the wireless device, such as the (e.g., maximum) number of layers and the (e.g., maximum) number of nodes per layer with which the wireless device can perform neural network processing for the multiple NPUs. According to various embodiments, such capability information may be indicated to the NPUs of the wireless device collectively, or individually, or in groups (e.g., where each group of NPUs is associated with a type or use of the group of NPUs, such as NPUs for beam prediction, NPUs for CSI measurement, NPUs for positioning, etc.).
[0113] The wireless device may provide the neural network processing capability information to the cellular base station at 506. According to various embodiments, the neural network processing capability information may be provided when the RRC connection is established, or when the RRC connection has been established.
[0114] The cellular base station may determine whether and when to schedule tasks that may involve neural network processing for the wireless device, and may proceed to configure one or more neural network processing activities for the wireless device based at least in part on the neural network processing capability information for the wireless device.
[0115] In some cases, such as if the neural network processing capability information for the wireless device includes neural network processing availability information indicating limited availability of some or all NPUs of the wireless device for tasks associated with the wireless link, the cellular base station may determine an availability mode for the NPUs of the wireless device, and may provide an indication of the determined availability mode for the NPUs of the wireless device to the wireless device. In some cases, the availability mode may include a discontinuous availability mode, e.g., in which discontinuously available NPUs of the wireless device may be available for neural network processing tasks associated with the wireless link (or perhaps more generally for wireless communication tasks) according to some or all of an on-duration, an off-duration, and a periodicity. As another possibility, the availability mode may include a dynamic availability mode, e.g., in which various time windows with various numbers of NPUs of the wireless device may be available for neural network processing tasks associated with the wireless link (or perhaps more generally for wireless communication tasks). In some cases, such a dynamic availability mode may be repeated according to a configured periodicity, last for a certain duration, and / or enabled at a configured offset after being signaled.
[0116] In some cases, the cellular base station may limit the number and type of neural network processing activities configured for the wireless device at any given time to within the neural network processing capabilities for the wireless device. For example, in accordance with some embodiments, the cellular base station may allocate neural network processing tasks that adhere to the maximum number of layers and the maximum number of nodes per layer supported by the wireless device and indicated in the neural network processing capabilities information. Similarly, in some embodiments, the cellular base station may allocate neural network processing tasks that require no more than the maximum number of NPUs supported by the wireless device and indicated in the neural network processing capabilities information. Alternatively, in some cases, the cellular base station may exceed the neural network processing capabilities for the wireless device in terms of the number and / or type of neural network processing activities configured for the wireless device, for example, if there are priority indications and / or rules that the wireless device can use to determine which of the configured tasks to execute and which tasks to discard.
[0117] According to various possibilities, such priority indication may be provided using RRC signaling (e.g., configured for each NPU-related task), configured by MAC CE (e.g., using a triggered MAC CE for semi-persistent reporting), or provided by a triggered DCI. As another possibility, static or semi-static priority rules may be configured or predefined so that reports triggered later have higher priorities (e.g., as one possibility, non-periodic reports may have a higher priority than semi-persistent reports, which may have a higher priority than periodic reports). Other priority sorting rules (e.g., for determining the priority order of tasks with the same priority indicator (e.g., tasks associated with the primary cell have a higher priority than tasks associated with the secondary cell; tasks associated with beam measurement have a higher priority than tasks associated with channel state information measurement, which in turn have a higher priority than tasks associated with positioning estimation; tasks started earlier have a higher priority than tasks started later; and / or any of various other possibilities)) are also possible.
[0118] If the neural network processing activity configured by the cellular base station is greater than the neural network processing capabilities for the wireless device, the wireless device may recognize such a situation and determine one or more neural network processing activities to discard, for example based at least in part on the neural network processing activity configured by the cellular base station being greater than the neural network processing capabilities for the wireless device and possibly based at least in part on a priority indication provided for the neural network processing activities. According to at least some embodiments, the one or more neural network processing activities discarded by the wireless device may include those neural network processing activities of the lowest priority, and it may be the case that only enough neural network processing activities are discarded so that the remaining neural network processing activities are within the capabilities of the wireless device to perform.
[0119] Another aspect of determining whether and when to schedule a neural network processing task may include determining the duration that a neural network processing unit is occupied by a given neural network processing task, which may also be referred to as the NPU occupancy of the neural network processing task. Various rules may be configured (e.g., via broadcast system information or dedicated signaling) and / or specified (e.g., in a 3GPP technical specification) to facilitate consistent determination of the NPU occupancy for various possible neural network processing tasks.
[0120] As one possible NPU occupation rule, for artificial intelligence (e.g., neural network) based beam prediction, for periodic reporting or semi-persistent reporting, the NPU may be considered occupied starting from the first symbol of the first downlink reference signal used for measurement until the last symbol of the uplink channel having reporting information about the predicted beam. Alternatively, in such a scenario, the NPU may be considered occupied starting from a certain number of symbols after the last symbol of the last downlink reference signal used for measurement (e.g., a "first" offset) until a certain number of symbols before the first symbol of the uplink channel having reporting information about the predicted beam (e.g., a "second" offset). For artificial intelligence based beam prediction, for non-periodic reporting, the NPU may be considered occupied starting from the first symbol of the control channel transmission (e.g., a physical downlink control channel (PDCCH) transmission) that triggers the report until the last symbol of the uplink channel having reporting information about the predicted beam. Alternatively, in such a scenario, the NPU may be considered to be occupied starting from a certain number of symbols (e.g., a first offset or another offset) after the last symbol of the control channel transmission that triggered the report until a certain number of symbols (e.g., a second offset or another offset) before the first symbol of the uplink channel having report information about the predicted beam. Note that according to various embodiments, the offset value may be determined based on wireless device capability information or may be predefined (e.g., in 3GPP technical specifications).
[0121] As another possible NPU occupancy rule, for artificial intelligence-based CSI measurement and reporting, for periodic reporting or semi-persistent reporting, the NPU may be considered occupied starting from the first symbol of the first channel measurement resource (CMR) or interference measurement resource (IMR) until the last symbol of the uplink channel with CSI report. Alternatively, in such a scenario, the NPU may be considered occupied starting from a certain number of symbols (e.g., a first offset or another offset) after the last symbol of the last CMR or IMR until a certain number of symbols (e.g., a second offset or another offset) before the first symbol of the uplink channel with CSI report. For artificial intelligence-based CSI measurement and reporting, for non-periodic reporting, the NPU may be considered occupied starting from the first symbol of the control channel transmission that triggered the report until the last symbol of the uplink channel with CSI report. Alternatively, in such a scenario, the NPU may be considered to be occupied starting from a certain number of symbols (e.g., a first offset or another offset) after the last symbol of the control channel transmission that triggered the report until a certain number of symbols (e.g., a second offset or another offset) before the first symbol of the uplink channel with the CSI report. As in the aforementioned example, according to various embodiments, the offset value may be determined based on the wireless device capability information, or may be predefined.
[0122] As yet another possible NPU occupancy rule, for artificial intelligence-based positioning estimates, the NPU may be considered to be occupied starting from the first symbol of a positioning reference signal (PRS) until a certain number (e.g., a "third" offset) of symbols or slots have passed. Again, according to various embodiments, the offset value may be determined based on wireless device capability information, or may be predefined.
[0123] Note that many other NPU occupancy rules are possible. NPU occupancy rules may be used to determine how long NPUs assigned to scheduled tasks are occupied, and therefore determine when those NPUs are available again. Additionally or alternatively, in scenarios where some or all of the NPUs of a wireless device have discontinuous availability for tasks associated with a wireless link, NPU occupancy rules may be used to determine whether the wireless device has a sufficient duration and number of NPU availability to complete neural network processing tasks that the cellular base station may schedule before entering an off-duration period. Various other uses of NPU occupancy rules are also possible.
[0124] Thus, at least according to some embodiments, at least in some cases, Figure 5 The method can be used to provide a framework according to which a wireless device and a cellular network can manage the machine learning processing resources of the wireless device for wireless communications with the cellular network and for other purposes.
[0125] Figures 6 to 13 and additional information
[0126] Figures 6 to 13 Shown can be combined if necessary Figure 5 However, it should be noted that Figures 6 to 13 Shown in and about Figures 6 to 13 The exemplary details described are not intended to limit the disclosure as a whole: many variations and alternatives to the details provided below are possible and should be considered within the scope of the disclosure.
[0127] In 3GPP Release 15, for Channel State Information (CSI) processing, there are minimum processing delays and CSI processing unit occupancy rules defined with respect to different kinds of UE capabilities for CSI measurement and reporting. For example, it may be the case that for CSI reporting, a gNB scheduling the report may follow certain minimum Z and Z' requirements. Figure 6 Possible example timing aspects of a wireless communication system that follows such Z and Z' requirements when scheduling CSI reports in accordance with some embodiments are illustrated. As one possibility, Z may be defined as the offset (including timing advance) between the last symbol of the PDCCH 602 that schedules the CSI report and the first symbol of the CSI report 608. As one possibility, Z' may be defined as the offset (including timing advance) between the last symbol of the channel measurement resources (CMR) 604 and the interference measurement resources (IMR) 606 and the first symbol of the CSI report 608. According to at least some embodiments, if the gNB schedules a CSI report with a smaller offset than the defined minimum value of Z or Z', the UE may report the quality of an outdated (e.g., previous) CSI measurement.
[0128] A CSI processing unit (CPU) may be used as a measurement unit for the CSI processing capabilities of a UE in a cellular communication system (such as a system operating according to a particular 3GPP technical specification). In some cases, the UE may report its CPU capabilities (e.g., in terms of the number of CPUs), and the UE may be able to process one or more CSI measurements and reports, wherein the number and / or type of measurements and reports that the UE is able to make may be based at least in part on the CPU capabilities of the UE. The CPU occupancy rule may specify a period during which the CPU of the UE is considered to be occupied when CSI measurements and reports are configured. For example, it may be the case that a periodic or semi-persistent CSI report (excluding an initial semi-persistent CSI report on the PUSCH after a PDCCH triggered report) occupies the CPU starting from the first symbol of the earliest one of each CSI-RS / CSI-IM / SSB resource used for channel or interference measurement, with the corresponding latest CSI-RS / CSI-IM / SSB timing no later than the corresponding CSI reference resource until the last symbol of the configured PUSCH / PUCCH carries the report. As an example, it may be the case that an aperiodic CSI report occupies the CPU starting from the first symbol after the PDCCH triggers the CSI report until the last symbol of the scheduled PUSCH carries the report. As a further example, it may be the case that an initial semi-persistent CSI report on the PUSCH after a PDCCH trigger occupies the CPU starting from the first symbol after the PDCCH until the last symbol of the scheduled PUSCH carries the report. Further possible CPU definition information may be provided in 3GPP TS 38.214 v.17.0.0 section 5.2.1.6, at least according to some embodiments.
[0129] There is increasing interest in utilizing machine learning techniques in 3GPP based communication systems (e.g., as one possibility, potentially in 3GPP Release 18). For example, machine learning may be introduced to facilitate wireless communications in beam prediction in the time and / or spatial domains, in CSI measurements and reporting, in positioning estimation, and / or in various other possible aspects of wireless communications. Thus, it is possible that a UE may need or be expected to perform parallel machine learning processing for such areas and / or other potential areas where machine learning techniques may be applicable.
[0130] Neural networks are one possible type of machine learning tool that can be used in this context. There may be the possibility that neural network processing capabilities (e.g., a neural network processing unit or NPU) may be provided in a UE device and may be used in conjunction with such machine learning techniques. According to various embodiments, such NPU capabilities in the UE may be dedicated to wireless communications or may be shared with other applications or UE resources.
[0131] Given such potential uses for machine learning techniques (such as use of neural networks), it may be useful to provide techniques for managing UE machine learning resources and / or capabilities. For example, a framework for UE NPU capabilities and NPU occupancy rules may be provided, and / or for discontinuous use schemes for NPUs (e.g., similar to discontinuous reception schemes), which may be shared among UEs for non-wireless communication purposes, for example. In some cases, such a framework may include defining the NPU as a specific or alternative type of CPU (e.g., as one possibility, a Type 2 CSI processing unit or a Type 2 CPU).
[0132] As described above, one aspect of the machine learning resource management framework may include UE capability reporting. In one example, this may include the UE reporting its capability regarding the maximum number of supported NPUs. According to various embodiments, the capability may be reported per UE or per band combination or per band. With respect to different processing complexities for different areas, the UE may report the maximum number of supported NPUs for different types of areas. For example, the UE may report the maximum number of supported NPUs for beam prediction in the spatial domain, the maximum number of supported NPUs for beam prediction in the time domain, the maximum number of supported NPUs for CSI reporting, the maximum number of supported NPUs for positioning, and the like. As another possibility, the number of occupied NPUs for different use cases may be defined in the 3GPP technical specifications, and a single NPU processing capability may be reported by a given UE. It may be the case that, for all cases, it is not expected that the UE is configured with a number of occupied NPUs that exceeds the value reported by the UE capability at the same time.
[0133] Additionally, it is possible that the UE may report a preferred on / off duration and / or periodicity for the NPU. For example, if the NPU is dedicated to wireless communication, it may be the case that it may be considered to be always on. As another example, if the NPU may be shared with other applications, it may be the case that it may be considered to be on for a certain period of time, or may be interrupted by another application, and not always on. For each NPU or for each type of NPU for a use case, the UE may report maximum capability information for neural network processing, such as the maximum number of layers and the maximum number of nodes per layer.
[0134] The NPU occupancy rule may be defined differently for different possible areas for neural network processing. As one possibility of the NPU occupancy rule for machine learning-based beam prediction, for periodic reporting or semi-persistent reporting, the NPU may be occupied starting from the first symbol of the first downlink (DL) reference signal (RS) used for measurement until the last symbol of the uplink channel having report information about the predicted beam. As another possibility of the NPU occupancy rule for machine learning-based beam prediction, for periodic reporting or semi-persistent reporting, the NPU may be occupied starting from "Y1" symbols after the last symbol of the last DL RS until "Y2" symbols before the first symbol of the uplink channel having report information about the predicted beam. For non-periodic reporting, as one possibility, the NPU may be occupied starting from the first symbol of the PDCCH that triggers the report until the last symbol of the uplink channel having report information about the predicted beam, or as another possibility, may be occupied starting from Y1 symbols after the last symbol of the PDCCH that triggers the report until Y2 symbols before the first symbol of the uplink channel having report information about the predicted beam. Among various possibilities, the parameters Y1 and Y2 may be predefined or reported as part of the UE capability information. Note that variations or alternatives to these examples of NPU occupancy rules for machine learning based beam prediction are also possible.
[0135] As one possibility for NPU occupancy rule for machine learning based CSI measurement and reporting, for periodic reporting or semi-persistent reporting, the NPU may be occupied starting from the first symbol of the CMR and / or IMR until the last symbol of the uplink channel with CSI reporting. Figure 7 Possible example timing aspects of a wireless communication system following such rules according to some embodiments are illustrated. As another possibility for NPU occupancy rules for machine learning based CSI measurement and reporting, for periodic reporting or semi-persistent reporting, the NPU may be occupied starting from "Y1" symbols after the last symbol of the last CMR / IMR until "Y2" symbols before the first symbol of the uplink channel with CSI report. Figure 8 Possible example timing aspects of a wireless communication system following such rules according to some embodiments are illustrated.For aperiodic reporting, as one possibility, the NPU may be occupied starting from the first symbol of the PDCCH triggering the report until the last symbol of the uplink channel with the CSI report. Fig. 9Possible example timing aspects of a wireless communication system following such rules according to some embodiments are illustrated. Alternatively, for aperiodic reporting, as another possibility, the NPU may be occupied starting from Y1 symbols after the last symbol of the PDCCH triggering the report until Y2 symbols before the first symbol of the uplink channel with the CSI report. Fig.10 Possible example timing aspects of a wireless communication system that follows such rules according to some embodiments are illustrated. Among various possibilities, parameters Y1 and Y2 may be predefined or reported as part of the UE capability information. Note that, according to various embodiments, the same (e.g., consistent Y1 and / or Y2) or different (e.g., Y1 to Y1' and / or Y2 to Y2') values of Y1 and Y2 may be used for NPU occupancy rules for machine learning-based beam prediction and NPU occupancy rules for machine learning-based CSI measurement and reporting. Note also that variations or alternatives to these examples of NPU occupancy rules for machine learning-based CSI reporting are possible.
[0136] As one possibility for an NPU occupancy rule for machine learning based positioning estimation, the NPU may be occupied starting from the first symbol of a positioning reference signal (PRS) until "X" symbols or slots have passed. Similar to the Y1 and Y2 parameters previously described herein, in various possibilities, X may be predefined or reported as part of the UE capability information.
[0137] In the event that all NPUs are occupied and the UE may not be able to process all assigned tasks, and / or for other possible purposes, it may be useful to introduce a priority indicator for jobs associated with NPUs. For example, with such a framework in place, if all NPUs are occupied for a UE and a job with a higher priority is triggered, the UE may be able to stop processing with the lowest priority, for example, to be able to process the job with the higher priority. For jobs configured with the same priority indicator, the UE may select the job to be discarded (or interrupted) based on the start time, job type, component carrier type and index, and / or based on any of a variety of other possible considerations. As an example, the UE may discard jobs based on a specified or configured priority order (such as primary cell>secondary cell, beam management>channel state information>positioning, and earlier start>later start).
[0138] As one possibility, the priority indicator may be configured by RRC signaling. For example, a priority indicator may be configured for each NPU-related job (e.g., in a CSI-reportConfig information element). As another possibility, the priority indicator may be configured by a medium access control (MAC) control element (CE). For example, for a MAC CE for semi-persistent reporting, the priority may be indicated by a triggering MAC CE. A 2-bit priority field may be used to indicate the priority for the report. In some embodiments, the remainder of the fields of the MAC CE may be as indicated in 3GPP 38.321 v.16.7.0 section 6.1.3.16. Fig.11 An example MAC CE format including such a priority field according to some embodiments is illustrated. As yet another possibility, the priority indicator may be indicated by a triggering DCI. As a further possibility, a rule for determining priority may be configured, such as a rule according to which a report triggered later always has a higher priority (e.g., aperiodic report > semi-persistent report > periodic report).
[0139] In the case of scenarios where the UE's NPU may be shared with other applications, the gNB may be able to configure "NPU-DRX" through higher layer signaling. When NPU-DRX is in the OFF duration, it may be the case that the UE may use the NPU for other purposes. When NPU-DRX is in the ON duration, it may be the case that the UE may not use the NPU for other purposes. NPU-DRX may be defined / configured for each NPU of a UE or for the UE (e.g. as a whole). Fig.12 Example aspects of such NPU-DRX configurations according to some embodiments are illustrated.
[0140] As another possibility, the gNB may be able to configure the number N of NPUs reserved for wireless communications to the UE during a time period. For example, for a UE with a maximum number of NPUs “N max "UE, the UE may be able to transfer the remaining N max - N NPUs are used for other purposes. The offset, periodicity and duration for this time period may be configured through higher layer signaling. Fig.13 Aspects of example scenarios in which such dynamic NPU usage may be configured according to some embodiments are illustrated. In the illustrated example scenario, for a UE, the gNB may reserve 5 NPUs in a first time window, 3 NPUs in a second time window, and 1 NPU in a third time window, where the time windows repeat according to a configured periodic pattern.
[0141] In the following, further exemplary embodiments are provided.
[0142] A set of embodiments may include a method comprising: by a wireless device: establishing a wireless link with a cellular base station; determining neural network processing capability information for the wireless device; and providing the neural network processing capability information for the wireless device to the cellular base station.
[0143] In accordance with some embodiments, the neural network processing capability information includes an indication of the number of supported neural network processing units (NPUs) for the wireless device.
[0144] In accordance with some embodiments, the neural network processing capability information includes an indication of the number of supported neural network processing units (NPUs) for the wireless device for each of one or more of: beam prediction in the spatial domain; beam prediction in the time domain; channel state information measurement and reporting; or positioning estimation.
[0145] According to some embodiments, the neural network processing capability information includes availability information of a neural network processing unit (NPU) for the wireless device.
[0146] According to some embodiments, the availability information for the NPU of the wireless device includes one or more of the following: an indication that one or more NPUs for the wireless device are always available for neural network processing tasks associated with the wireless link; or an indication of one or more of a preferred on duration, a preferred off duration, or a preferred periodicity for the availability of one or more NPUs for the wireless device.
[0147] According to some embodiments, the neural network processing capability information includes an indication of the number of layers and the number of nodes in each layer that the wireless device can utilize to perform neural network processing for one or more neural network processing units (NPUs) of the wireless device.
[0148] According to some embodiments, the method also includes: receiving information from the cellular base station to configure the wireless device to perform a neural network processing activity; determining a neural network processing unit occupancy duration for the neural network processing activity; and executing the configured neural network processing activity.
[0149] According to some embodiments, the neural network processing activity includes one or more of: beam prediction and reporting; channel state information measurement and reporting; or wireless device location estimation.
[0150] According to some embodiments, the method also includes: determining that a neural network processing activity is greater than a neural network processing capability for the wireless device configured by the cellular base station; and determining one or more neural network processing activities to be discarded based at least in part on the neural network processing activity configured by the cellular base station being greater than the neural network processing capability for the wireless device.
[0151] According to some embodiments, the one or more neural network processing activities to be discarded are determined at least in part based on priority indications for the one or more neural network processing activities, wherein the neural network processing activities with the lowest priority indications are discarded until the remaining number of neural network processing activities configured by the cellular base station is within the neural network processing capacity for the wireless device.
[0152] In accordance with some embodiments, the priority indication for the one or more neural network processing activities is received by the wireless device via one or more of: radio resource control (RRC) signaling; medium access control (MAC) control element (CE) signaling; or downlink control information (DCI) signaling.
[0153] Another set of embodiments may include a wireless device comprising: one or more processors; and a memory having instructions stored thereon that, when executed by the one or more processors, perform the steps of the method described in any of the foregoing examples.
[0154] Yet another set of embodiments may include a method comprising: by a cellular base station: establishing a wireless link with a wireless device; receiving neural network processing capability information from the wireless device; and configuring one or more neural network processing activities for the wireless device, wherein the one or more neural network processing activities are configured at least in part based on the neural network processing capability information.
[0155] According to some embodiments, the neural network processing capability information includes an indication of a number of supported neural network processing units (NPUs) for the wireless device, wherein configuring one or more neural network processing activities for the wireless device based at least in part on the neural network processing capability information includes limiting the number and type of neural network processing activities configured for the wireless device to within the neural network processing capability for the wireless device.
[0156] According to some embodiments, the method also includes: determining a neural network processing unit occupancy duration of the one or more neural network processing activities configured for the wireless device, wherein limiting the number and type of neural network processing activities configured for the wireless device to within the neural network processing capabilities of the wireless device is further based at least in part on the determined neural network processing unit occupancy duration of the one or more neural network processing activities configured for the wireless device.
[0157] According to some embodiments, the neural network processing capability information includes availability information of a neural network processing unit (NPU) for the wireless device, wherein the method further includes: determining an availability mode of the NPU for the wireless device based at least in part on the availability information of the NPU for the wireless device; and providing an indication of the availability mode of the NPU for the wireless device to the wireless device.
[0158] According to some embodiments, the neural network processing capability information includes an indication of the number of layers and the number of nodes in each layer that the wireless device can utilize to perform neural network processing for one or more neural network processing units (NPUs) of the wireless device.
[0159] According to some embodiments, the method also includes: providing priority information for the one or more neural network processing activities to the wireless device, wherein the priority information for the one or more neural network processing activities is provided via one or more of the following: radio resource control (RRC) signaling; medium access control (MAC) control element (CE) signaling; or downlink control information (DCI) signaling.
[0160] Still another set of embodiments may include a cellular base station comprising: one or more processors; and a memory having instructions stored thereon that, when executed by the one or more processors, perform the steps of the method described in any of the foregoing examples.
[0161] Another set of embodiments may include a computer program product comprising computer instructions that, when executed by one or more processors, perform the steps of the method according to any of the preceding examples.
[0162] Yet another example embodiment may include a method comprising: performing, by a wireless device, any or all of the foregoing examples.
[0163] Another exemplary embodiment may include a device comprising: an antenna; a radio component coupled to the antenna; and a processing element operably coupled to the radio component, wherein the device is configured to implement any or all portions of the foregoing examples.
[0164] Another exemplary set of embodiments may include a non-transitory computer-accessible storage medium including program instructions that, when executed at a device, cause the device to implement any or all portions of any of the foregoing examples.
[0165] Another exemplary set of embodiments may include a computer program comprising instructions for performing any or all portions of any of the foregoing examples.
[0166] Another exemplary set of embodiments may include an apparatus including means for performing any or all of the elements of any of the preceding examples.
[0167] Another exemplary set of embodiments may include an apparatus comprising a processing element configured to cause a wireless device to perform any or all elements of any of the foregoing examples.
[0168] It is understood that the use of personally identifiable information should be subject to privacy policies and practices that are generally recognized to meet or exceed industry or government requirements for maintaining user privacy. Specifically, personally identifiable information data should be managed and processed to minimize the risk of unintentional or unauthorized access or use, and the nature of the authorized use should be clearly stated to users.
[0169] By interpreting each message / signal X received by a user equipment (UE) in the downlink as a message / signal X sent by the base station, and interpreting each message / signal Y sent by the UE in the uplink as a message / signal Y received by the base station, any of the methods for operating a UE described herein may become the basis for the corresponding method for operating a base station.
[0170] The embodiments of the present disclosure may be implemented in any of a variety of forms. For example, in some embodiments, the subject matter may be implemented as a computer-implemented method, a computer-readable storage medium, or a computer system. In other embodiments, one or more custom-designed hardware devices such as ASICs may be used to implement the subject matter. In other embodiments, one or more programmable hardware elements such as FPGAs may be used to implement the subject matter.
[0171] In some embodiments, a non-transitory computer-readable storage medium (e.g., a non-transitory storage element) may be configured to store program instructions and / or data, wherein if the program instructions are executed by a computer system, the computer system is caused to perform a method, such as any method implementation of the method embodiments described herein, or any combination of the method implementations described herein, or any subset of any method implementation of the method embodiments described herein, or any combination of such subsets.
[0172] In some embodiments, a device (e.g., a UE) may be configured to include a processor (or a group of processors) and a memory medium (or a memory element), wherein the memory medium stores program instructions, wherein the processor is configured to read and execute these program instructions from the memory medium, wherein these program instructions are executable to implement any method implementation in the various method implementations described herein (or any combination of the method implementations described herein, or any subset of any method implementation in the method implementations described herein, or any combination of such subsets). The device may be implemented in any of various forms.
[0173] Although the above embodiments have been described in considerable detail, numerous variations and modifications will become apparent to those skilled in the art once the above disclosure is fully appreciated. It is intended that the following claims be interpreted to include all such variations and modifications.
Claims
1. A method, comprising: By wireless device: Establishing a wireless link with a cellular base station; determining neural network processing capability information for the wireless device; as well as The neural network processing capability information for the wireless device is provided to the cellular base station.
2. The method according to claim 1, The neural network processing capability information includes an indication of the number of supported neural network processing units (NPUs) for the wireless device.
3. The method according to claim 1, wherein the neural network processing capability information comprises an indication of a number of supported neural network processing units (NPUs) for the wireless device for each of one or more of: Beam prediction in the spatial domain; Beam prediction in the time domain; Channel state information measurement and reporting; or Position estimation.
4. The method according to claim 1, The neural network processing capability information includes availability information of a neural network processing unit (NPU) for the wireless device.
5. The method of claim 4, wherein the availability information for the NPU of the wireless device comprises one or more of: an indication that one or more NPUs of the wireless device are always available for neural network processing tasks associated with the wireless link; or An indication of one or more of a preferred on-duration, a preferred off-duration, or a preferred periodicity of availability of one or more NPUs for the wireless device.
6. The method according to claim 1, The neural network processing capability information includes an indication of the number of layers and the number of nodes in each layer that the wireless device can utilize to perform neural network processing for one or more neural network processing units (NPUs) of the wireless device.
7. The method according to claim 1, wherein the method further comprises: receiving information from the cellular base station configuring the wireless device to perform a neural network processing activity; determining a neural network processing unit occupation duration for the neural network processing activity; as well as Executes the configured neural network processing activity.
8. The method of claim 7, wherein the neural network processing activity comprises one or more of: beam prediction and reporting; Channel state information measurement and reporting; or Wireless device location estimation.
9. The method according to claim 7, wherein the method further comprises: determining that neural network processing activity greater than a neural network processing capability for the wireless device is configured by the cellular base station; as well as One or more neural network processing activities are determined to be discarded based at least in part on the neural network processing activity configured by the cellular base station being greater than the neural network processing capability for the wireless device.
10. The method according to claim 9, The one or more neural network processing activities to be discarded are determined at least in part based on priority indications for the one or more neural network processing activities, wherein the neural network processing activities with the lowest priority indications are discarded until a remaining number of neural network processing activities configured by the cellular base station is within the neural network processing capability for the wireless device.
11. The method according to claim 10, wherein the priority indication for the one or more neural network processing activities is received by the wireless device via one or more of: Radio Resource Control (RRC) signaling; Medium Access Control (MAC) Control Element (CE) signaling; or Downlink Control Information (DCI) signaling.
12. A wireless device, comprising: one or more processors; and A memory having instructions stored thereon which, when executed by the one or more processors, perform the steps of the method according to any one of claims 1 to 11.
13. A method comprising: By cellular base station: Establishing a wireless link with a wireless device; receiving neural network processing capability information from the wireless device; as well as One or more neural network processing activities are configured for the wireless device, wherein the one or more neural network processing activities are configured based at least in part on the neural network processing capability information.
14. The method according to claim 13, wherein the neural network processing capability information comprises an indication of a number of supported neural network processing units (NPUs) for the wireless device, Wherein configuring one or more neural network processing activities for the wireless device based at least in part on the neural network processing capability information includes limiting the number and type of neural network processing activities configured for the wireless device to within the neural network processing capability for the wireless device.
15. The method according to claim 14, wherein the method further comprises: determining a neural network processing unit occupancy duration for the one or more neural network processing activities configured for the wireless device, Wherein limiting the number and type of neural network processing activities configured for the wireless device to within the neural network processing capabilities of the wireless device is further based at least in part on a determined neural network processing unit occupancy duration of the one or more neural network processing activities configured for the wireless device.
16. The method according to claim 13, The neural network processing capability information includes availability information of a neural network processing unit (NPU) for the wireless device, wherein the method further comprises: determining an availability mode of an NPU for the wireless device based at least in part on the availability information of the NPU for the wireless device; as well as An indication of the availability mode of the NPU for the wireless device is provided to the wireless device.
17. The method according to claim 13, The neural network processing capability information includes an indication of the number of layers and the number of nodes in each layer that the wireless device can utilize to perform neural network processing for one or more neural network processing units (NPUs) of the wireless device.
18. The method according to claim 13, wherein the method further comprises: providing priority information for the one or more neural network processing activities to the wireless device, wherein the priority information for the one or more neural network processing activities is provided via one or more of: Radio Resource Control (RRC) signaling; Medium Access Control (MAC) Control Element (CE) signaling; or Downlink Control Information (DCI) signaling.
19. A cellular base station, comprising: one or more processors; and A memory having stored thereon instructions which, when executed by the one or more processors, perform the steps of the method according to any one of claims 13 to 18.
20. A computer program product comprising computer instructions which, when executed by one or more processors, perform the steps of the method according to any one of claims 1 to 11 or 13 to 18.
