TW202239254A - Spatial inter-cell interference aware downlink coordination - Google Patents

Abstract

A method of wireless communication by a first network device includes predicting spatial inter-cell downlink interference experienced by a UE. The method also includes communicating with a second network device to reduce the spatial inter-cell downlink interference in a direction of the UE by protecting resources across selected resource sets.

Description

Interference-aware downlink coordination between spatial cells

This patent application claims priority to U.S. Patent Application No. 17/675,980, filed February 18, 2022, and entitled "SPATIAL INTER-CELL INTERFERENCE AWARE DOWNLINK COORDINATION," which claims The benefit of U.S. Provisional Patent Application No. 63/155,635, filed March 2 and entitled "SPATIAL INTER-CELL INTERFERENCE AWARE DOWNLINK COORDINATION," the disclosure of which is expressly incorporated by reference in its entirety middle.

Aspects of the subject matter of this case relate generally to wireless communications and, more specifically, to enhanced techniques and devices for downlink coordination aware of spatial intercellular interference.

Wireless communication systems are widely deployed to provide various telecommunication services such as telephony, video, data, messaging, and broadcasting. A typical wireless communication system may adopt a multiple access technology capable of supporting communication with multiple users by sharing available system resources (eg, bandwidth, transmit power, etc.). Examples of such multiple access techniques include Code Division Multiple Access (CDMA) systems, Time Division Multiple Access (TDMA) systems, Frequency Division Multiple Access (FDMA) systems, Orthogonal Frequency Division Multiple Access Access (OFDMA) system, single carrier frequency division multiple access (SC-FDMA) system, time division synchronous code division multiple access (TD-SCDMA) system and long-term evolution (LTE). LTE/LTE-Advanced is a set of enhancements to the Universal Mobile Telecommunications System (UMTS) mobile service standard promulgated by the Third Generation Partnership Project (3GPP).

A wireless communication network may include a plurality of base stations (BSs) capable of supporting communication for a plurality of user equipments (UEs). A user equipment (UE) can communicate with a base station (BS) via downlink and uplink. The downlink (or forward link) represents the communication link from the BS to the UE, and the uplink (or reverse link) represents the communication link from the UE to the BS. As will be described in more detail, a BS may be called a Node B, a gNB, an Access Point (AP), a radio head, a Transmission and Reception Point (TRP), a New Radio (NR) BS, a 5G Node B, and the like.

The above multiple access techniques have been adopted in various telecommunication standards to provide a common protocol enabling different user equipments to communicate at city, country, regional, and even global levels. New Radio (NR), which may also be referred to as 5G, is a set of enhancements to the LTE mobile service standard promulgated by the 3rd Generation Partnership Project (3GPP). NR is designed to improve spectral efficiency, reduce cost, improve service, utilize new spectrum, and use Orthogonal Frequency Division Multiplexing (OFDM) (CP-OFDM) with cyclic prefix (CP) on the downlink (DL). ), use CP-OFDM and/or SC-FDM (e.g. also known as discrete Fourier transform spread OFDM (DFT-s-OFDM)) on the uplink (UL) and support beamforming, multiple-input multiple Output (MIMO) antenna technology and carrier aggregation to better integrate with other open standards to better support mobile broadband Internet access.

An artificial neural network may include groups of interconnected artificial neurons (eg, neuron models). An artificial neural network may be a computing device or be represented as a method to be performed by a computing device. A convolutional neural network (such as a deep convolutional neural network) is a type of feed-forward artificial neural network. A convolutional neural network may include layers of neurons that may be arranged in tiled receptive fields. It would be desirable to apply neural network processing to wireless communications to achieve greater efficiency.

A method of wireless communication by a first network device includes predicting spatial intercellular downlink interference experienced by a UE. The method also includes communicating with a second network device to reduce spatial inter-cell downlink interference in the direction of the UE by protecting resources across the selected set of resources.

Means for wireless communication by a first network device are described. The apparatus includes means for predicting spatial intercellular downlink interference experienced by a UE. The apparatus also includes means for communicating with a second network device to reduce spatial inter-cell downlink interference in the direction of the UE by protecting resources across the selected set of resources.

The first network device includes a processor and a memory coupled with the processor. The first network device also includes instructions stored in memory. The instructions, when executed by the processor, the first network device is operable to predict spatial inter-cell downlink interference experienced by a UE. The first network device is also operable to communicate with the second network device to reduce spatial inter-cell downlink interference in the direction of the UE by protecting resources across the selected set of resources.

A non-transitory computer readable medium having program code recorded thereon is executed by a processor of the first network device. The non-transitory computer readable medium includes program code for predicting spatial intercellular downlink interference experienced by a UE. The non-transitory computer readable medium also includes code for communicating with the second network device to reduce spatial inter-cell downlink interference in the direction of the UE by protecting resources across the selected set of resources.

Aspects generally include methods, apparatus, systems, computer program products, non-transitory computer readable media, user equipment, base stations, A wireless communication device, and a processing system.

The foregoing has outlined rather broadly the features and technical advantages of examples according to the present disclosure so that the detailed description that follows may be better understood. Additional features and advantages will be described. The conception and specific example disclosed may be readily utilized as a basis for modifying or designing other structures for carrying out the same purposes of the teachings herein. Such equivalent constructions do not depart from the scope of the appended claims. The nature of the disclosed concepts, both their organization and method of operation, and associated advantages will be better understood from the following description when considered in conjunction with the accompanying drawings. Each of the drawings is provided for purposes of illustration and description, and not as a definition of the limitations of the claims.

Various aspects of the content of this case are more fully described below with reference to the accompanying drawings. This disclosure may, however, be embodied in many different forms and should not be construed as limited to any specific structure or function presented throughout this disclosure. Rather, these aspects are provided so that this disclosure will be thorough and complete, and will fully convey the scope of this disclosure to those having ordinary knowledge in the art to which this invention pertains. Based on the teachings, those of ordinary skill in the art to which the present invention pertains should understand that the scope of the subject matter is intended to cover any aspect of the subject matter, regardless of whether the aspect is implemented independently of any other aspect of the subject matter or in conjunction with any other aspect of the subject matter. It is realized by combining other aspects. For example, an apparatus may be implemented or a method may be practiced using any number of aspects set forth. Furthermore, the scope of the present disclosure is intended to cover such devices or devices practiced with other structures, functions, or both, in addition to or other than the various aspects of the stated disclosure. method. It should be understood that any aspect of the disclosed disclosure may be embodied by one or more elements of a claim.

Several aspects of telecommunications systems will now be provided with reference to various devices and technologies. These devices and techniques will be described in the following detailed description and illustrated in the accompanying drawings via various blocks, modules, components, circuits, steps, procedures, algorithms, etc. (collectively referred to as "elements"). These elements may be implemented using hardware, software or a combination thereof. Whether such elements are implemented as hardware or software depends upon the particular application and design constraints imposed on the overall system.

It should be noted that although aspects may be described using terms commonly associated with 5G and beyond wireless technologies, aspects of this disclosure may apply to communication systems based on other generations, such as and including 3G and / or 4G technology.

Cell-to-cell interference can degrade the user's signal (eg, signal-to-interference-plus-noise ratio (SINR)). The user's signal degradation may be especially noticeable when the user is at the edge of the cell. Additionally, due to the introduction of massive multiple-input multiple-output (MIMO) antennas in next-generation Node Bs (gNBs), this inter-cell interference may cause significant SINR degradation for users. For example, a narrow downlink transmit beam may be highly directional and highly variable over time, causing SINR degradation. In particular, highly directional interference at the cell edge can reduce data rates and negatively impact user experience. In addition, high variability in interference makes it more difficult to perform link self-tuning, such as predicting supported modulation and coding schemes (MCS). This is challenging for latency-sensitive applications with limited latency budgets.

In some aspects of this case, a neural network is trained to infer the impact of any potential transmit beam from a neighboring base station (e.g., gNB) on any potential victim User Equipment (UE). In these aspects of the present case, the neural network is trained over time using UE channel state information (CSI) reference signal (CSI-RS) measurement reports, beam information, and UE location. In other aspects of the subject matter, the database stores information about the impact of any potential transmit beam of a neighboring base station (eg, gNB) on any potential victim UE. In these aspects of the content of the case, the database stores UE channel state information (CSI) reference signal (CSI-RS) measurement report, beam information and UE location to enable database inspection to determine the location of the sub-adjacent base station Potential impact on any potential victim UE.

Once the neural network is trained to infer the impact of the neighbor base station's transmit beam on the victim UE, the network device (e.g., gNB) coordinates with the neighbor base station. In these aspects of the subject matter, network device coordination may prohibit downlink transmit beam communication by neighbor base stations during time/frequency resources used to serve victim UEs. For example, the first user equipment (UE1) may experience significant interference from the downlink transmit beam k of the neighbor base station (gNB2). In this example, spatial inter-cell interference is mitigated when the neighbor base station gNB2 avoids sending energy in the direction of the downlink transmit beam k on resources serving the first UE.

According to aspects of this disclosure, the serving cell predicts, for the served UE, experienced downlink interference from different potential downlink transmit beams of neighboring cells. For example, a serving cell may identify a subset of served UEs where the potential negative impact caused by inter-cell downlink interference exceeds a predetermined UE interference threshold. This identification of the subset of potential victims of the UE may also include additional criteria, such as the type of transmission the UE is receiving. For example, UEs receiving delay sensitive traffic and/or high reliability traffic may be selected as part of the subset of potential victims of UEs.

In these aspects of the present disclosure, the serving cell sends request messages to potentially interfering neighbor cells for each of a subset of potential victims of UEs. The request message may include a forbidden list of beam indices from which potentially interfering neighbor cells are requested to avoid. Furthermore, the request message may indicate the requested time/frequency resource (eg, time slot/resource block (RB)) for which protection is requested. In some implementations, a predefined set of time/frequency resources is configured per cell such that signaling can simply represent an index of the proposed set of resources. Alternatively, time/frequency resources are left to neighbor cells to decide.

In some aspects of the subject matter, learning spatial inter-cell downlink interference and predicting a victim subset of UEs is performed at interfering cells (which may be referred to as aggressor neighbor cells). In these aspects of the subject matter, the serving cell identifies potentially vulnerable UEs based on their location, traffic type, or other selection criteria. Once identified, the serving cell sends a request message to the aggressor neighbor cell. The request message may indicate a subset of: (1) location of vulnerable UEs; (2) UE interference tolerance threshold; and/or (3) protected resource requirements. In other aspects of this disclosure, an indication of the traffic load of the victim UE is provided to represent the desired amount of protected resources.

Request messages sent by service cells can omit resource requirements, which are optional. When the request message omits the resource requirement, the protected resource is determined by the neighbor cells. Furthermore, there may be a predefined set of time/frequency resources configured for each cell. In this configuration, the signaling of the request message may simply represent the index of the proposed set of resources that satisfy the resource requirements of the UE.

In other implementations, learning and prediction are performed at a centralized coordinating node, such as a network controller. For example, the serving cell identifies potentially vulnerable UEs based on their location, traffic type, or other selection criteria. Once identified, the serving cell sends information about the identified vulnerable UEs to the centralized coordinating node. This information may indicate all or a subset of: (1) the location of the vulnerable UE; (2) the interference threshold for the vulnerable UE; and (3) the traffic load (e.g., the transmission of updates volume demand). Some implementations including central node coordination may be beneficial when multiple aggressor neighbor cells are potentially causing interference to the victim UE.

FIG. 1 is a schematic diagram illustrating a network 100 in which aspects of the present disclosure may be practiced. Network 100 may be a 5G or NR network or some other wireless network such as an LTE network. Wireless network 100 may include multiple BSs 110 (shown as BS 110a, BS 110b, BS 110c, and BS 110d) and other network entities. A BS is an entity that communicates with a User Equipment (UE) and may also be referred to as a Base Station, NR BS, Node B, gNB, 5G Node B (NB), Access Point, Transmitting and Receiving Point (TRP), etc. Each BS can provide communication coverage for a specific geographic area. In 3GPP, the term "cell" can refer to a coverage area of a BS and/or a BS subsystem serving that coverage area, depending on the context in which the term is used.

A BS may provide communication coverage for macrocells, picocells, femtocells, and/or another type of cell. A macrocell may cover a relatively large geographic area (eg, several kilometers in radius) and may allow unrestricted access by UEs with service subscription. A picocell may cover a relatively small geographic area and may allow unrestricted access by UEs with a service subscription. A femtocell may cover a relatively small geographic area (e.g., a residence) and may allow restricted access by UEs that have an association with the femtocell (e.g., UEs in a Closed Subscriber Group (CSG)) . The BS for macrocells may be referred to as macro BS. A BS for a pico cell may be referred to as a pico BS. A BS for a femto cell may be called a femto BS or a home BS. In the example shown in FIG. 1, BS 110a may be a macro BS for macro cell 102a, BS 110b may be a pico BS for pico cell 102b, and BS 110c may be a pico BS for femto cell 102c. Femto BS. A BS can support one or more (eg, three) cells. The terms "eNB", "base station", "NR BS", "gNB", "TRP", "AP", "Node B", "5G NB" and "cell" are used interchangeably.

In some aspects, the cell may not necessarily be stationary, and the geographic area of the cell may move depending on the location of the mobile BS. In some aspects, BSs may interconnect with each other and/or with one of wireless networks 100 via various types of backhaul interfaces (eg, direct physical connections, virtual networks, etc.) using any suitable transport network. or multiple other BSs or network nodes (not shown) are interconnected.

The wireless network 100 may also include relay stations. A relay station is an entity that may receive data transmissions from upstream stations (eg, BS or UE) and send data transmissions to downstream stations (eg, UE or BS). A relay station may also be a UE capable of relaying transmissions to other UEs. In the example shown in FIG. 1, a relay station 11Od may communicate with a macro BS 110a and a UE 12Od in order to facilitate communication between the BS 110a and the UE 12Od. A relay station may also be called a relay BS, a relay base station, a repeater, etc.

The wireless network 100 may be a heterogeneous network including different types of BSs (eg, macro BSs, pico BSs, femto BSs, relay BSs, etc.). These different types of BSs may have different transmit power levels, different coverage areas, and different effects on interference in the wireless network 100 . For example, macro BSs may have high transmit power levels (eg, 5 to 40 watts), while pico, femto, and relay BSs may have lower transmit power levels (eg, 0.1 to 2 watts).

A network controller 130 can couple to a group of BSs and can provide coordination and control for these BSs. The network controller 130 can communicate with the BS via backhaul. BSs can also communicate with each other via wireless or wired backhaul (eg, directly or indirectly).

UEs 120 (eg, 120a, 120b, 120c) may be dispersed throughout wireless network 100, and each UE may be stationary or mobile. A UE may also be called an access terminal, terminal, mobile station, subscriber unit, station, and the like. A UE may be a cellular phone (e.g., a smartphone), a personal digital assistant (PDA), a wireless modem, a wireless communication device, a handheld device, a laptop computer, a wireless phone, a wireless local loop (WLL) station, a tablet device , cameras, game devices, small notebooks, smart computers, ultrabooks, medical equipment or devices, biometric sensors/devices, wearable devices (smart watches, smart clothing, smart glasses, smart wristbands, smart jewelry ( For example, smart rings, smart bracelets, etc.)), entertainment equipment (such as music or video equipment, or satellite radio units, etc.), vehicle components or sensors, smart meters/sensors, industrial manufacturing equipment, global positioning system device or any other suitable device configured to communicate via wireless or wired media.

The base station 110 may include a neural processing engine 150 . For brevity, only the base station 110 a is shown as including the neural processing engine 150 , but neighboring base stations may also include the neural processing engine 150 . The neural processing engine 150 may predict the spatial inter-cell downlink interference experienced by the UE. The neural processing engine 150 may also communicate with the second network device to reduce spatial inter-cell downlink interference in the direction of the UE by protecting resources across the selected set of resources.

Network controller 130 may include neural processing engine 160 . The neural processing engine 160 may predict the spatial intercellular downlink interference experienced by the UE. The neural processing engine 160 may also communicate with the second network device to reduce spatial inter-cell downlink interference in the direction of the UE by protecting resources across the selected set of resources.

Some UEs may be considered as machine type communication (MTC) or evolved or enhanced machine type communication (eMTC) UEs. MTC and eMTC UEs include, for example, robots, drones, remote devices, sensors, meters, monitors, location tags, etc., which may interact with a base station, another device (e.g., a remote device), or some other entity communication. A wireless node may provide connectivity to or to a network (eg, a wide area network such as the Internet or a cellular network), eg, via a wired or wireless communication link. Some UEs may be considered Internet of Things (IoT) devices, and/or may be implemented as NB-IoT (Narrow Band Internet of Things) devices. Some UEs may be considered as Customer Premises Equipment (CPE). The UE 120 may be included inside a housing housing components of the UE 120 such as processor components, memory components, and the like.

In general, any number of wireless networks can be deployed in a given geographic area. Each wireless network may support a specific RAT and may operate on one or more frequencies. A RAT may also be referred to as a radio technology, an air interface, and the like. Frequency may also be called carrier, channel, etc. Each frequency can support a single RAT in a given geographic area in order to avoid interference between wireless networks of different RATs. In some cases, NR or 5G RAT networks can be deployed.

In some aspects, two or more UEs 120 (eg, shown as UE 120a and UE 120e ) may communicate directly using one or more sidelink channels (eg, without using base station 110 as a intermediary for communicating with each other). For example, UE 120 may use peer-to-peer (P2P) communication, device-to-device (D2D) communication, vehicle-to-everything (V2X) protocol (which may include, for example, vehicle-to-vehicle (V2V) protocol, vehicle-to-infrastructure (V2I) protocol etc.), mesh network, etc. for communication. In this case, UE 120 may perform scheduling operations, resource selection operations, and/or other operations described elsewhere herein as being performed by base station 110 . For example, the base station 110 may pass downlink control information (DCI), radio resource control (RRC) signaling, medium access control-control element (MAC-CE) or system information (such as system information block (SIB) )), to configure UE 120.

As noted above, Figure 1 is provided as an example only. Other examples may differ from those described with respect to FIG. 1 .

2 shows a block diagram of a design 200 of base station 110 and UE 120 (which may be one of the base stations and one of the UEs in FIG. 1 ). Base station 110 may be equipped with T antennas 234a through 234t, and UE 120 may be equipped with R antennas 252a through 252r, where generally, T≧1 and R≧1.

At base station 110, transmit processor 220 may receive data for one or more UEs from data source 212, select a channel quality indicator (CQI) for each UE based at least in part on a channel quality indicator (CQI) received from the UE for that UE or multiple modulation and coding schemes (MCS), process (eg, encode and modulate) data for each UE based at least in part on the MCS selected for that UE, and provide data symbols for all UEs. Lowering the MCS reduces the throughput, but increases the reliability of the transmission. The transmit processor 220 can also process system information (eg, for semi-static resource partitioning information (SRPI), etc.) and control information (eg, CQI requests, grants, upper layer signaling, etc.), and provide management burden symbols and control symbols. The transmit processor 220 may also generate reference symbols for reference signals (eg, cell-specific reference signal (CRS)) and synchronization signals (eg, primary synchronization signal (PSS) and secondary synchronization signal (SSS)). Transmit (TX) multiple-input multiple-output (MIMO) processor 230 may perform spatial processing (e.g., precoding) on data symbols, control symbols, administrative burden symbols, and/or reference symbols, as applicable, and may send data to T Modulators (MOD) 232a through 232t provide T output symbol streams. Each modulator 232 may process a corresponding output symbol stream (eg, for Orthogonal Frequency Division Multiplexing (OFDM), etc.) to obtain an output sample stream. Each modulator 232 may further process (eg, convert to analog, amplify, filter, and upconvert) the output sample stream to obtain a downlink signal. The T downlink signals from the modulators 232a to 232t may be transmitted via the T antennas 234a to 234t, respectively. According to various aspects described in more detail below, positional encoding may be used to generate synchronization signals to convey additional information.

At UE 120, antennas 252a through 252r may receive downlink signals from base station 110 and/or other base stations and may provide received signals to demodulators (DEMOD) 254a through 254r, respectively. Each demodulator 254 may condition (eg, filter, amplify, downconvert, and digitize) the received signal to obtain input samples. Each demodulator 254 may further process the input samples (eg, for OFDM, etc.) to obtain received symbols. A MIMO detector 256 may obtain received symbols from all R demodulators 254a through 254r, perform MIMO detection on the received symbols (if applicable), and provide detected symbols. Receive processor 258 may process (e.g., demodulate and decode) detected symbols, provide decoded data for UE 120 to data slot 260, and provide decoded control information and system information to controller/processor 280 . The channel processor can determine Reference Signal Received Power (RSRP), Received Signal Strength Indicator (RSSI), Reference Signal Received Quality (RSRQ), Channel Quality Indicator (CQI), etc. In some aspects, one or more components of UE 120 may be included in a housing.

On the uplink, at UE 120, transmit processor 264 may receive and process data from data source 262 and control information from controller/processor 280 (e.g., for information including RSRP, RSSI, RSRQ, CQI, etc. Report). The transmit processor 264 may also generate reference symbols for one or more reference signals. Symbols from transmit processor 264 may be precoded by TX MIMO processor 266 (if applicable), further processed by modulators 254a through 254r (e.g., for DFT-s-OFDM, CP-OFDM, etc.), and transmitted Give 110 to the base station. At base station 110, uplink signals from UE 120 and other UEs may be received by antenna 234, processed by demodulator 254, detected by MIMO detector 236 (if applicable), and processed by receive processor 238. Further processing to obtain decoded data and control information sent by UE 120 . Receive processor 238 may provide decoded data to data slot 239 and decoded control information to controller/processor 240 . The base station 110 may include a communication unit 244 and communicate with the network controller 130 via the communication unit 244 . The network controller 130 may include a communication unit 294 , a controller/processor 290 and a memory 292 .

Controller/processor 240 of base station 110 and/or controller/processor 280 of UE 120 of FIG. One or more techniques, as described in more detail elsewhere. For example, controller/processor 280 of UE 120 of FIG. 2 may perform or direct the operation of, for example, the procedure of FIG. 8 and/or other procedures described. Additionally, controller/processor 240 of base station 110 of FIG. 2 may execute or direct the operation of, for example, the procedures of FIGS. 9-12 and/or other procedures described. Memories 242 and 282 may store data and program codes for base station 110 and UE 120, respectively. The scheduler 246 can schedule the UE to perform data transmission on the downlink and/or uplink.

In some aspects, the base station 110 and the network controller 130 may include a unit for predicting, a unit for selecting, a unit for transmitting, a unit for receiving, a unit for updating and/or a unit for unit for communication. Such units may include one or more components of the network controller 130 or base station 110 described in connection with FIG. 2 .

As noted above, Figure 2 is provided as an example only. Other examples may differ from those described with respect to FIG. 2 .

In some cases, different types of devices supporting different types of applications and/or services can coexist in a cell. Examples of different types of devices include UE handsets, customer premises equipment (CPE), vehicles, Internet of Things (IoT) devices, etc. Examples of different types of applications include ultra-reliable low-latency communication (URLLC) applications, massive machine-type communication (mMTC) applications, enhanced mobile broadband (eMBB) applications, vehicle-to-everything (V2X) applications, etc. Furthermore, in some cases, a single device may support different applications or services at the same time.

3 illustrates an example implementation of a system-on-chip (SOC) 300 that may include a central processing unit (CPU) 302 configured to generate gradients for neural network training or Multi-core CPUs. SOC 300 may be included in base station 110 or UE 120 . Variables (e.g., neural signals and synaptic weights), system parameters associated with computing devices (e.g., neural networks with weights), delays, frequency information, and task information may be stored in a neural processing unit (NPU) 308 associated memory block, memory block associated with CPU 302, memory block associated with graphics processing unit (GPU) 304, memory block associated with digital signal processor (DSP) 306 in memory block 318, or may be distributed across multiple blocks. Instructions executed at CPU 302 may be loaded from program memory associated with CPU 302 or may be loaded from memory block 318 .

SOC 300 may also include additional processing blocks customized for specific functions, such as GPU 304, DSP 306, connectivity block 310 (which may include fifth generation (5G) connectivity, fourth generation long term evolution (4G LTE) connectivity, Wi-Fi -Fi connection, USB connection, Bluetooth connection, etc.) and the multimedia processor 312 (which can detect and recognize gestures, for example). In one implementation, the NPU is implemented in a CPU, DSP and/or GPU. SOC 300 may also include sensor processor 314 , image signal processor (ISP) 316 and/or navigation module 320 (which may include a global positioning system).

SOC 300 may be based on the ARM instruction set. In an aspect of this disclosure, the instructions loaded into the general purpose processor 302 may include: code for predicting spatial inter-cell downlink interference experienced by the UE; and for communicating with the second network device Code is communicated to reduce spatial inter-cell downlink interference in the direction of a UE by protecting resources across a selected set of resources.

Deep learning architectures can perform object recognition tasks by learning to represent inputs at successively higher levels of abstraction in each layer, thereby building useful feature representations of the input data. In this way, deep learning solves a major bottleneck of traditional machine learning. Before the advent of deep learning, machine learning approaches to object recognition problems likely relied heavily on human-engineered features, perhaps combined with shallow classifiers. A shallow classifier can be a two-class linear classifier, for example, where a weighted sum of feature vector components can be compared to a threshold to predict which class the input belongs to. Human engineering features can be templates or kernels customized for a specific problem domain by engineers with domain expertise. In contrast, deep learning architectures can learn to represent features similar to those a human engineer might design, but learn through training. In addition, deep networks can learn to represent and recognize new types of features that humans may not have considered.

Deep learning architectures can learn hierarchies of features. For example, the first layer can learn to recognize relatively simple features in an input stream, such as edges, if given visual data. In another example, the first layer can learn to recognize spectral power in specific frequencies if given auditory data. A second layer, which takes as input the output of the first layer, can learn to recognize combinations of features, such as simple shapes for visual material or combinations of sounds for auditory material. For example, higher layers can learn to represent complex shapes in visual material, or words in auditory material. Higher layers can learn to recognize common visual objects or spoken phrases.

Deep learning architectures can perform particularly well when applied to problems with a natural hierarchy. For example, classifying motor vehicles could benefit from first learning to recognize wheels, windshields, and other features. These features can be combined at higher layers in different ways to recognize cars, trucks and planes.

Neural networks can be designed using a variety of connectivity patterns. In a feed-forward network, information is passed from lower layers to higher layers, where each neuron in a given layer communicates to a neuron in a higher layer. Hierarchical representations can be built in successive layers of feed-forward networks, as described above. Neural networks can also have recurrent or feedback (also known as top-down) connections. In recurrent connections, the output from a neuron in a given layer can be passed to another neuron in the same layer. The recurrent architecture can help identify patterns across more than one block of input data that is passed sequentially to the neural network. Connections from neurons in a given layer to neurons in lower layers are called feedback (or top-down) connections. Networks with many feedback connections may be informative when identifying high-level concepts can aid in distinguishing specific low-level features of the input.

The connections between the layers of a neural network can be fully connected or partially connected. FIG. 4A shows an example of a fully connected neural network 402 . In a fully connected neural network 402, a neuron in the first layer can send its output to every neuron in the second layer, so that every neuron in the second layer will receive input from The input to each neuron of the first layer. FIG. 4B shows an example of a locally connected neural network 404 . In a locally connected neural network 404, a neuron in a first layer may be connected to a limited number of neurons in a second layer. More generally, the locally connected layers of the locally connected neural network 404 may be configured such that each neuron in the layer will have the same or similar connectivity pattern, but with possibly different values (e.g., 410, 412 , 414 and 416) connection strength. Linear patterns of local connections can generate spatially distinct receptive fields in higher layers, because higher layer neurons in a given region can receive properties tuned to a restricted portion of the network's total input via training. enter.

An example of a locally connected neural network is a convolutional neural network. An example of a convolutional neural network 406 is shown in FIG. 4C . Convolutional neural network 406 may be configured such that connection strengths associated with inputs to each neuron of the second layer are shared (eg, 408 ). Convolutional neural networks may be well suited for problems where the spatial location of the input is meaningful.

One type of convolutional neural network is a deep convolutional network (DCN). FIG. 4D shows a detailed example of a DCN 400 designed to recognize visual features from an image 426 input from an image capture device 430 , such as a vehicle-mounted camera. The DCN 400 of the present example can be trained to recognize traffic signs and the digits provided on the traffic signs. Of course, the DCN 400 can be trained for other tasks, such as recognizing lane markings or recognizing traffic lights.

The DCN 400 may be trained using supervised learning. During training, DCN 400 may be given an image, such as image of a speed limit sign 426 , and then may compute a forward pass to produce output 422 . DCN 400 may include a feature extraction part and a classification part. Upon receiving the image 426 , the convolution layer 432 may apply a convolution kernel (not shown) to the image 426 to generate the first set of feature maps 418 . As an example, the convolution kernel used for the convolution layer 432 may be a 5x5 kernel that produces a 28x28 feature map. In this example, because four different feature maps were generated in the first set of feature maps 418 , four different convolution kernels are applied to the image 426 at the convolution layer 432 . A convolution kernel may also be called a filter or a convolution filter.

The first set of feature maps 418 may be subsampled by a max pooling layer (not shown) to produce the second set of feature maps 420 . The max pooling layer reduces the size of the first set of feature maps 418 . That is, the size of the second set of feature maps 420 (such as 14x14) is smaller than the size of the first set of feature maps 418 (such as 28x28). The reduced size provides similar information to subsequent layers while reducing memory consumption. The second set of feature maps 420 may be further convoluted through one or more subsequent convolutional layers (not shown) to generate one or more subsequent sets of feature maps (not shown).

In the example of FIG. 4D , the second set of feature maps 420 is convolved to generate the first feature vector 424 . Additionally, the first eigenvector 424 is further convolved to generate a second eigenvector 428 . Each feature of the second feature vector 428 may include a number corresponding to a possible feature of the image 426 such as "logo", "60" and "100". A softmax function (not shown) can convert the digits in the second feature vector 428 into probabilities. Thus, the output 422 of the DCN 400 is the probability that the image 426 includes one or more features.

In this example, the probability for "flag" and "60" in output 422 is higher than other items in output 422 (such as "30", "40", "50", "70", "80", "90 ” and “100”). Prior to training, the output 422 produced by the DCN 400 may be incorrect. Therefore, the error between output 422 and the target output can be calculated. The target output is the ground truth of the image 426 (eg, "sign" and "60"). The weights of the DCN 400 can then be adjusted so that the output 422 of the DCN 400 more closely aligns with the target output.

To adjust the weights, the learning algorithm can compute gradient vectors for the weights. The gradient can indicate the amount by which the error will increase or decrease if the weights are adjusted. At the top layer, the gradient may directly correspond to the value of the weights connecting the activation neurons in the penultimate layer and the neurons in the output layer. In lower layers, gradients can depend on the values of weights and the calculated error gradients of higher layers. The weights can then be adjusted to reduce the error. This way of adjusting weights can be called "backpropagation" because it involves a "backward pass" through the neural network.

In practice, the error gradient of the weights can be computed over a small number of instances such that the computed gradient approximates the true error gradient. This approach to approximation can be called stochastic gradient descent. Stochastic gradient descent can be repeated until the achievable error rate of the overall system has stopped decreasing or until the error rate has reached a target level. After learning, the DCN can be given a new image (eg, a speed limit sign of image 426 ), and a forward pass through the network can produce an output 422 that can be considered the DCN's inference or prediction.

Deep Belief Networks (DBNs) are probabilistic models that include multiple layers of hidden nodes. DBNs can be used to extract hierarchical representations of training data sets. A DBN can be obtained via superposition of layers of Restricted Boltzmann Machines (RBMs). RBM is a type of artificial neural network that learns a probability distribution over a set of inputs. Since RBMs can learn a probability distribution without information about the class each input should be classified into, RBMs are often used for unsupervised learning. Using a mixed unsupervised and supervised paradigm, the bottom RBM of a DBN can be trained in an unsupervised manner and can be used as a feature extractor, and the top RBM can be trained in a supervised manner (based on the input and target class) and can be used as a classifier.

Deep Convolutional Networks (DCNs) are networks of convolutional networks that are configured with additional pooling and normalization layers. DCN has achieved state-of-the-art performance on many tasks. DCNs can be trained using supervised learning, where both the input and output targets are known for many examples and used to modify the weights of the network through the use of gradient descent methods.

A DCN may be a feed-forward network. Additionally, as described above, connections from a neuron in a first layer of a DCN to a set of neurons in the next higher layer are shared across neurons in the first layer. DCN's feed-forward and shared connections can be used for fast processing. For example, the computational burden of a DCN may be much smaller than that of a similarly sized neural network that includes recurrent or feedback connections.

The processing of each layer of a convolutional network can be thought of as a spatially invariant template or base projection. If the input is first decomposed into channels, such as the red, green, and blue channels of a color image, a convolutional network trained on that input can be considered three-dimensional, with two spatial dimensions along the axis of the image and Retrieve the third dimension of color information. The output of the convolutional connection can be thought of as forming a feature map in subsequent layers, where each element of the feature map (e.g., 220) receives input from a sequence of neurons in a previous layer (e.g., feature map 218) and from multiple channels input for each of the channels. Values in the feature map can be further processed with non-linearity, such as correction, max(0,x). Values from neighboring neurons can be further pooled, which corresponds to downsampling and can provide additional local invariance and dimensionality reduction. Regularization (which corresponds to whitening) can also be applied via lateral inhibition between neurons in the feature map.

The performance of deep learning architectures can increase as more labeled data points become available or as computing power increases. Modern deep neural networks are programmatically trained using computing resources thousands of times larger than were available to the typical researcher 15 years ago. New architectures and training paradigms can further improve the performance of deep learning. Rectified linear units can reduce a training problem known as vanishing gradients. New training techniques can reduce overfitting and thus enable larger models to achieve better generalization. Encapsulation technology can abstract data in a given receptive field and further improve overall performance.

FIG. 5 is a block diagram illustrating a deep convolutional network 550 . The deep convolutional network 550 may include many different types of layers based on connectivity and weight sharing. As shown in FIG. 5, the deep convolution network 550 includes convolution blocks 554A, 554B. Each convolution block 554A, 554B may be configured with a convolution layer (CONV) 356 , a normalization layer (LNorm) 558 and a max pooling layer (max POOL) 560 .

The convolutional layer 556 may include one or more convolutional filters that may be applied to the input data to generate a feature map. Although only two convolution blocks 554A, 554B are shown, the present disclosure is not so limited and instead, any number of convolution blocks 554A, 554B may be included in the deep convolution network 550 according to design preferences. A normalization layer 558 may normalize the output of the convolution filter. For example, normalization layer 558 may provide whitening or lateral suppression. A max pooling layer 560 may provide spatially downsampled aggregation for local invariance and dimensionality reduction.

For example, parallel filter banks of deep convolutional networks can be loaded on CPU 302 or GPU 304 of SOC 300 to achieve high performance and low power consumption. In an alternate embodiment, the parallel filter banks may be loaded on the DSP 306 or the ISP 316 of the SOC 300 . In addition, deep convolutional network 550 may access other processing blocks that may exist on SOC 300 , such as sensor processor 314 and navigation module 320 dedicated to sensors and navigation, respectively.

The deep convolutional network 550 may also include one or more fully connected layers 562 (FC1 and FC2). The deep convolutional network 550 may also include a logistic regression (LR) layer 564 . Between each layer 556, 558, 560, 562, 564 of the deep convolutional network 550 are weights (not shown) to be updated. The output of each of the layers (e.g., 556, 558, 560, 562, 564) may be used as a subsequent one of the layers (e.g., 556, 558, 560, 562, 564) in the deep convolutional network 550 Layer input to learn hierarchical feature representations from input data 552 (eg, image, audio, video, sensor data, and/or other input data) supplied at a first convolution block in convolution block 554A. The output of the deep convolutional network 550 is a classification score 566 for the input data 552 . Classification score 566 may be a set of probabilities, where each probability is a probability of the input data including features from the feature set.

As noted above, Figures 3-5 are provided as examples. Other examples may differ from those described with respect to Figures 3-5.

As described above, intercellular interference may degrade a user's signal (eg, signal-to-interference-plus-noise ratio (SINR)). The user's signal degradation may be especially noticeable when the user is at the edge of the cell. Additionally, due to the introduction of massive multiple-input multiple-output (MIMO) antennas in next-generation base stations (e.g., gNBs), such intercellular interference can be highly directional and highly variable over time. Unfortunately, highly directional interference at the cell edge can reduce data rates and negatively impact user experience. In addition, high variability in interference makes it more difficult to perform link self-tuning (eg, predicting supportable modulation and coding schemes (MCS)). This is challenging for latency-sensitive applications with limited latency budgets. This limited delay budget may not be sufficient to recover packets via a hybrid automatic repeat request (HARQ) procedure when the selected modulation and coding scheme (MCS) is inaccurate.

6A and 6B illustrate communication networks according to aspects of the present disclosure, wherein spatial interference experienced by user equipment (UE) is based on downlink transmit beams from neighboring base stations to neighboring UEs. As shown in FIG. 6A, in a first interference scenario 600, a first UE 120-1 communicates with a first base station 110-1 via a downlink transmit beam i. Similarly, the second UE 120-2 communicates with the neighbor base station 110-2 via the downlink transmit beam j. In this example, the spatial inter-cell interference from downlink transmit beam j to downlink transmit beam i results in very little signal degradation (eg, SINR = 20 dB) at the first UE 120-1.

FIG. 6B shows a second interference scenario 650, in which the first UE 120-1 communicates with the first base station 110-1 via downlink transmit beam i. On the contrary, the second UE 120-2 communicates with the neighbor base station 110-2 via the downlink transmit beam k interfering with the downlink transmit beam i. In this example, spatial intercellular interference from downlink transmit beam k to downlink transmit beam i results in significant signal degradation (eg, SINR=5 dB) at the first UE 120-1, which reduces A user experience at UE 120-1.

FIG. 7 is a schematic diagram of a communication network 700 according to various aspects of the content of the present application. The communication network 700 shows signal strength measurements of downlink transmit beams from neighboring base stations to achieve spatial intercellular interference aware downlink coordination. According to various aspects of this case, the neural network is trained to infer the impact of any potential transmit beam of a neighboring base station (eg, gNB) on any potential victim UE. In this example, the victim user equipment (UE ₁ ) 120-1 communicates with the serving base station (gNB ₁ ) 110-1 via downlink transmit beam i. Unfortunately, the victim UE ₁ 120-1 experiences a disturbance of the downlink transmit beam k from the neighbor base station (gNB ₂ ) 110-2 used to communicate with the second user equipment (UE ₂ ) 120-2. Significant interference.

In this example, the spatial intercellular interference to the victim UE ₁ 120-1 is mitigated via coordination with the neighbor base station gNB ₂ 110-2. For example, neighbor base station _gNB2 110-2 may transmit on downlink beam j to avoid sending energy in the direction of downlink transmit beam k on resources serving victim _UE1 120-1. Aspects of this case coordinate with neighbor base station gNB ₂ 110-2 to transmit on downlink beam j on resources used to serve victim UE ₁ 120-1, thereby avoiding victimization Interference by UE ₁ 110-1.

According to various aspects of the content of this case, the neural network of the serving base station gNB ₁ 110-1 is trained to infer the impact of the downlink transmit beam of the neighboring base station gNB ₂ 110-2 on the victim UE ₁ 120-1. In these aspects of the content of this case, the serving base station gNB ₁ 110-1 coordinates with the neighboring base station gNB ₂ 110-2 to prohibit the neighboring base station gNB2 110-2 from serving the victim UE ₁ The communication of the downlink transmit beam k is performed on the time/frequency resource of 120-1. In some configurations, prediction of intercellular interference is performed using machine learning with a Neural Processing Engine (NPE), eg, as shown in FIG. 8 .

8 is a block diagram of a network including a neural processing engine configured to perform neural network-based prediction of distributions of spatial intercellular interference in accordance with aspects of the subject matter to achieve spatial intercellular Interference aware downlink coordination. In aspects of the subject matter, the database stores information about the impact of any potential transmit beam of a neighboring base station (eg, gNB) on any potential victim UE. In these aspects of the subject matter, the database stores UE channel state information (CSI) reference signal (CSI-RS) measurement reports, beam information and UE location to enable database inspection to determine any Potential impact on potential victim UE.

Figure 8 shows a network 800 comprising a UE 120 with a location block 830 and a base station 110 with a neural processing engine 810 for implementing a neural network. In this example, the location block 830 indicates the UE location 802 (X) to the neural processing engine 810 . The neighbor cell transmits a precoder 804(T) (eg, channel state information (CSI) beam index) also as input to the neural processing engine 810 from the neighbor cell. Based on the UE location 802 and the neighbor cell transmit precoder 804 , the neural processing engine 810 predicts an interference-based distribution 820 (F) for the current location of the UE 120 . For example, interference (eg, thermal interference) distributions can be predicted, or signal-to-interference-plus-noise ratio (SINR) distributions can be predicted.

In various aspects of the content of this case, the training of the neural network of the neural processing engine 810 may be based on UE channel state information (CSI) reference signal (CSI-RS) measurement reports and neighbor cell transmission precoder 804 and UE location 802. The UE CSI-RS measurement report provides the neighbor cell signal strength for each beam of the neighbor base station. For example, the signal strength measurement report may be based on a CSI report including, for example, SINR information or Reference Signal Received Power (RSRP) information. In some aspects of this disclosure, UE CSI-RS measurement reports may provide SINR information, where signal strength corresponds to serving cell and interference corresponds to each beam of neighboring base stations.

In some aspects of this disclosure, the base station 110 periodically sends the precoded CSI-RS signal to the served UE 120 . The served UE 120 can use the received CSI-RS signal to estimate the channel condition. The served UE 120 may also use the received CSI-RS signal to identify the single beam or beam combination that results in the strongest received signal quality (eg, the strongest beam) to account for data channel precoding. Additionally, CSI-RS signals can be measured from other cells. For example, UE can use CSI-RS signals from other cells to estimate the interference caused by other cells. The UE can use the measured CSI-RS signal to perform radio resource management. For example, the UE uses the measured CSI-RS signals of other cells to identify whether the neighbor cell is stronger than the current serving cell, which may trigger a handover. In fact, UE 120 reports to serving cell 110 the CSI-RS measurements performed by the UE.

As shown in FIG. 8 , the neural network of the neural processing engine 810 predicts an interference-based distribution 820 for a given UE location 802 based on the neighbor cell transmit precoder 804 and the UE location 802 . The neighbor cell transmit precoder 804 may be a precoder index within a known codebook, or may be a precoder weight. In aspects of this disclosure, UE location 802 may be represented as a combination of (x, y, z) coordinates of UE 120 from a positioning source. The positioning source may be, for example, a Global Navigation Satellite System (GNSS), a 5G NR location server, and the like. Alternatively, UE location 802 may be determined based on a set of metrics representing the location of UE 120 within the serving cell. For example, the set of metrics indicative of the location of UE 120 within a serving cell may include a serving cell reference signal received power (RSRP) signal, a serving cell precoder (e.g., a precoding matrix indicator (PMI)) indicating the direction of the strongest transmit beam , a serving cell channel quality indicator (CQI), and/or a path loss estimate for the channel between the serving cell and the UE 120 . Additionally, UE location 802 may be determined based on other UE sensor information. In some aspects, UE location 802 may be a geographic location.

In various aspects of the content of this case, the training of the neural network can be performed at various nodes. For example, the training of the neural network can be performed at the serving cell. In this example, the UE 120 sends the interference measurement report to the base station 110 of the serving cell together with information about the UE location 802 . The location estimation of the UE 120 can be performed at the serving cell base station 110 or transmitted to the serving cell base station 110 by a separate location server. Alternatively, UE 120 reports UE location indicator metrics to the serving cell to determine UE location 802 .

In some aspects of the subject matter, as shown in FIG. 7, the training of the neural network is performed at a neighbor cell, such as a neighbor base station gNB ₂ 110-2. In these aspects of the content of this case, the base station gNB ₁ 110-1 of the serving cell sends the UE location 802 and the interference measurement report to the neighboring base station _gNB2 110-2, so that the Perform training of the neural network. In other aspects of this case, the training of the neural network is performed at a centralized node. In these aspects of the content of this case, the base station gNB ₁ 110-1 of the serving cell sends the UE location 802, the serving cell identification (ID), the neighbor cell ID, and the interference measurement report to the centralized node for Perform neural network training.

Once the neural network is trained to infer the impact of the neighbor base station's transmit beam on the victim UE, the network device (e.g., gNB) coordinates with the neighbor base station. In these aspects of the subject matter, network device coordination may prevent communication of downlink transmit beams by neighbor base stations on time/frequency resources used to serve victim UEs. For example, as shown in Figure 7, UE ₁ 120-1 experiences significant interference from a downlink transmit beam k of a neighbor base station ( _gNB2 ) 110-2. In this example, spatial inter-cell interference is mitigated when neighbor base station gNB ₂ 110-2 avoids sending energy in the direction of downlink transmit beam k on resources used to serve UE ₁ 120-1.

According to various aspects of the subject matter, undesired neighbor cell beams are identified based on certain criteria based on predicted interference-based distributions. For example, an average or percentile exceeding a predetermined threshold may be a criterion. Additionally, the spatial interference characterization can be used to coordinate scheduling between nearby cells, such as neighboring base stations gNB ₂ 110-2, to prevent high interference events, eg, as shown in Figures 9-11.

Fig. 9 is a diagram showing, for example, UE 120 (120-1, ..., 120-N), base station 110-1 of serving cell 900, and neighbor cell 950 (950-1, ..., 950-N) is a timing diagram of an example procedure executed by base stations 110-2, . . . , 110-N for spatial inter-cell interference aware downlink coordination at serving cell 900.

According to various aspects of the content of this case, the base station 110-1 of the serving cell 900 predicts that the UE 120 (120-1, ..., 120-N) in the serving cell 900 experiences ..., 950-N) for potential downlink interference of different potential downlink transmit beams. For example, the base station 110-1 of the serving cell 900 identifies a subset of the serving UEs 120 for which the potential negative impact caused by the inter-cell downlink interference from the neighbor cell 950 exceeds a predetermined UE interference threshold. That is, neighbor cells 950 may include potentially interfering neighbor cells. This identification of the subset of potential victims of the UE may also include additional criteria, such as the type of transmission the UE is receiving. For example, UEs receiving delay sensitive traffic and/or high reliability traffic may be selected as part of the subset of potential victims of UEs.

At time t0, the base station 110 - 1 of the serving cell 900 sends a request message to the potentially interfering neighbor cell 950 for each victim subset of the UE 120 . The request message may include a proposal indicating the requested list of beam indices for potential interfering neighbor cells 950 to avoid. Additionally, the request message may indicate the requested time/frequency resource (eg, time slot/resource block (RB)) for which protection is requested. For example, the request message may indicate the number of resources to be protected. The base station 110-1 may determine how much bandwidth is needed by the vulnerable UE, for example, a quarter of the allocated resources. In some implementations, a predefined set of time/frequency resources is configured for each cell such that the signaling of the request message at time t0 may simply represent the index of the proposed set of resources. Alternatively, the choice of time/frequency resources is left to the neighbor cell 950 to make the decision.

At time t1, the base station 110-2-N of the aggressor neighbor cell 950 replies with a response message, which is received by the base station 110-1 of the serving cell 900 at time t1. The response message may be an acceptance of the offer indicated via the request message. For example, the response message may indicate agreement to limit transmitted energy in the identified beam direction for the indicated time/frequency resource. Alternatively, the response message may include a potential proposal or an alternative proposal for the time/frequency resource to be protected. For example, when the offer (eg, service offer) in the request message from the serving cell 900 is not accepted, the response message received at time t1 may include an alternative offer for a different set of resources.

When the victim UE 120 changes location, the channel condition changes and/or the traffic demand changes, the serving cell 900 can periodically repeat the spatial inter-cell downlink interference prediction. For example, UE 120 may be closer to serving base station 110-1, or UE may enter idle mode. At time t2, the serving cell 900 may send an updated request message to the neighbor cell 950 based on the updated prediction for the new condition. In addition, at time t3, the neighbor cell 950 can similarly respond to the updated request message with the updated response message. Once the interference threat to the victim UE 120 expires, at time t4, the serving cell 900 may send a cancel request message to stop resource protection. For example, when the victim UE 120 enters the idle mode, the serving cell 900 may send a cancel request message at time t4.

Fig. 10 is a diagram showing the execution by UE 120 (120-1, ..., 120-N), base station 110-1 of serving cell 900, and base station 110-2 of neighbor cell 950 according to various aspects of the content of this application. A timing diagram of an example procedure for spatial inter-cell interference-aware downlink coordination at neighbor cells 950.

In aspects of the present case, prediction of downlink interference between spatial cells occurs at neighbor cells 950 (which may be referred to as aggressor neighbor cells 950). In these aspects of the present disclosure, base station 110-1 of serving cell 900 identifies potentially vulnerable UEs based on their location, traffic type, or other selection criteria. Once identified, the base station 110-1 of the serving cell 900 sends a request message to the aggressor neighbor cell 950 at time t0. The request message may indicate a subset of: (1) location of vulnerable UEs; (2) UE interference tolerance threshold; and/or (3) resource requirements.

The request message sent by the base station 110-1 of the serving cell 900 at time t0 may omit the resource requirement, which is optional. When the request message omits the resource requirement, the protected resource is determined by the base station 110-2 of the aggressor neighbor cell 950. Also, instead of geographic location, the UE's location may be represented via a certain set of metrics representing the UE's location. Furthermore, there may be a predefined set of time/frequency resources configured for each cell. In this configuration, the signaling of the request message sent at time t0 may simply represent the index of the proposed set of resources that satisfy the resource requirements of the UE.

In these aspects of the present case, in response to the request message at time t0, the base station 110-2 of the aggressor neighbor cell 950 predicts whether the interference from the transmit beam of the aggressor neighbor cell 950 would be harmful to the victim UE Cause harmful effects exceeding the UE interference threshold. In response to the predicted deleterious effect, at time t1, the base station 110-2 of the aggressor neighbor cell 950 sends a response message to the serving cell 900 indicating that the requested time/frequency resource for the victim UE is granted at Limits the energy transmitted in the prohibited beam directions. Potentially, the response message at time t1 includes a proposal or an alternative proposal (if the resource requirement is not accepted) for time/frequency resources (such as a set of resources to be protected) for the protected UE.

In some implementations, the base station 110 - 1 of the serving cell 900 periodically evaluates the location and vulnerability of the protected UE 120 and protected resources when the UE 120 changes location, channel conditions change, or traffic demand changes. In response to any of these changed conditions, at time t2, serving cell 900 may send an updated request message with the new conditions. After predicting whether downlink interference is detrimental to the updated set of vulnerable UEs 120, at time t3, the base station 110-2 of the aggressor neighbor cell 950 responds to the updated request message by sending an updated response message to respond. In this example, once the interference threat to the set of vulnerable UEs 120 has expired, at time t4, the base station 110-1 of the serving cell 900 may send a message to the aggressor neighbor cell 950 to terminate resource protection cancellation message for .

Fig. 11 is a diagram showing, for example, UE 120 (120-1, ..., 120-N), base station 110-1 of serving cell 900, central node 960, and aggressor neighbor cell 950 according to various aspects of the present application. A timing diagram of an example procedure executed by base station 110-2 for spatial inter-cell interference aware downlink coordination at the central node.

In some implementations, prediction may be performed at a central node 960 , such as network controller 130 . For example, the base station 110-1 of the serving cell 900 identifies potentially vulnerable UEs based on their location, traffic type, or other selection criteria. Once identified, at time t1, the base station 110-1 of the serving cell 900 sends a message to the central node 960 regarding the identified vulnerable UE. The message may indicate all or a subset of: (1) the location of the vulnerable UE; (2) the interference threshold of the vulnerable UE; and (3) the traffic load. As noted above, UE location can be represented using a combination of (1) the UE's (x, y, z) coordinates (e.g., geographic location) from a positioning source and/or (2) an indication that the UE is in A set of metrics for a location within the serving cell 900 . The positioning source may be, for example, a Global Navigation Satellite System (GNSS) location server, a 5G NR location server, or other location servers. In addition, the set of metrics representing UE location may include reference signal received power (RSRP) of serving cell 900, serving cell precoder (eg, strongest transmit beam direction), serving cell channel quality indicator (CQI) and/or Path loss estimation between cell 900 and UE 120. Other UE sensor information may also be included in the vulnerable UE message at time t0.

In operation, the central node 960 predicts whether the degradation to the victim UE 120 by interference from the downlink transmit beam of the aggressor neighbor cell 950 exceeds a UE interference threshold. At time t1, when the interference of the downlink transmit beam of the neighbor cell exceeds the UE interference threshold, the central node 960 sends a coordination request message to the aggressor neighbor cell 950 . The request message may include an offer to protect the UE by limiting transmitted energy in the specified beam direction for the requested time/frequency resource indicated by the request message. In response, the base station 110 - 2 of the aggressor neighbor cell 950 sends a response message to the central node 960 at time t2 . The response message may instruct the base station 110-2 of the aggressor neighbor cell 950 to accept the offer. Otherwise, the response message may suggest alternatives to protect the UE. At time t3, the central node 960 sends a response message to the serving cell 900, which may indicate the set of protected resources for the protected UE.

In some implementations, the central node 960 periodically evaluates the location and vulnerability of protected UEs and protected resources. Periodic evaluation may determine whether a change is detected in the UE's location, the UE's channel condition, and/or the UE's traffic requirements. In response to the detected change, at time t4, the central node 960 may send an updated request message to the aggressor neighbor cell 950 based on the changed condition. In response, at time t5, the base station 110-2 of the aggressor neighbor cell 950 may respond to the updated request message with an updated response message indicating the updated protected resource. At time t6, the central node 960 sends a response message to the serving cell 900, which can notify the base station 110-1 of the updated set of protected resources.

Once the interference threat expires, at time t7, the serving cell 900 may send a cancel message to the central node 960, and trigger a terminate message at time t8 to terminate resource protection. Some implementations including a central node may be beneficial when multiple aggressor neighbor cells may cause interference to the victim UE.

12 is a flowchart illustrating an example procedure 1200 for neural network-based spatial intercellular interference learning, such as performed by a network device, in accordance with various aspects of the subject matter. The example procedure 1200 is an example of network enhancement for neural network based spatial intercellular interference aware downlink coordination.

As shown in FIG. 12, in some aspects, procedure 1200 includes predicting spatial inter-cell downlink interference experienced by a UE (block 1202). For example, a base station (eg, using controller/processor 240 and/or memory 242) can predict spatial downlink inter-cell downlink interference. Prediction can occur at the serving base station, the aggressor base station, or the central node. In some aspects, the prediction is based on the location of the UE, a UE interference tolerance threshold, and/or resource requirements for the UE.

In some aspects, the procedure 1200 also includes communicating with a second network device to reduce spatial inter-cell downlink interference in the direction of the UE by protecting resources across the selected set of resources (block 1204). For example, the base station (e.g., using antenna 234, DEMOD/MOD 232, TX MIMO processor 230, transmit processor 220, controller/processor 240, and/or memory 242) may communicate with a second network device, to reduce spatial inter-cell downlink interference in the direction of the UE. In some aspects, protected resources include barred beam indices, time/frequency resources to be protected, and/or number of resources to be protected. Instance aspect

Aspect 1: A method of wireless communication by a first network device, comprising: predicting spatial intercellular downlink interference experienced by a UE; and communicating with a second network device to span selected resources via protected Aggregate resources to reduce spatial inter-cell downlink interference in the direction of the UE.

Aspect 2: The method of Aspect 1, wherein the first network device includes a serving cell and the second network device includes a potentially interfering neighbor cell, the method also includes: selecting, by the first network device, the predicted The UE whose inter-cell downlink interference exceeds a predetermined threshold; and sending a request message to the second network device.

Aspect 3: The method of Aspect 1 or 2, wherein the request message indicates a beam causing too much interference.

Aspect 4: The method according to Aspect 1 or 2, wherein the request message indicates the time/frequency resource to be protected.

Aspect 5: The method according to Aspect 1 or 2, wherein the request message indicates the quantity of resources to be protected, and the quantity is determined based on the UE's traffic requirement.

Aspect 6: The method according to Aspect 1 or 2, wherein the request message indicates the index of the selected resource set within the list of predetermined resource sets.

Aspect 7: The method according to Aspect 1 or 2, also includes: receiving a response message, the response message indicating acceptance of the proposal indicated by the request message.

Aspect 8: The method according to Aspect 1 or 2 also includes: receiving a response message, the response message indicating an alternative proposal for a different resource set.

Aspect 9: The method according to Aspect 1 or 2, further comprising: updating the request in response to an updated prediction based on the updated UE location, the updated channel condition for the UE, and/or the updated throughput requirement for the UE message.

Aspect 10: The method of any of Aspects 1-9, wherein the first network device includes a neighbor cell, and the second network device includes a serving cell, the method further comprising: from the second network device receiving a request message indicating the location of the UE, a UE interference tolerance threshold and/or a resource requirement for the UE; performing prediction based on the request message; and sending a response message to the second network device.

Aspect 11: The method according to any of aspects 1-10, further comprising: receiving an update about UE location, UE interference tolerance threshold, and/or UE-specific resource requirements from the second network device.

Aspect 12: The method according to Aspect 1, wherein the first network device includes a central node, and the second network device includes a serving cell, the method also includes: receiving a first request message from the second network device, the first request The message indicates the location of the UE, the UE interference tolerance threshold and/or the resource requirement for the UE; predicts based on the first request message; sends a second request message requesting resource protection to the aggressor neighbor cell; and sends a second request message to the second network The device sends a response message indicating the protected resource.

Aspect 13: The method according to Aspect 12, further comprising: updating the prediction based on the updated UE location, the updated channel condition for the UE, and/or the updated resource requirement for the UE.

Aspect 14: An apparatus for wireless communication by a first network device, comprising: means for predicting spatial downlink inter-cell interference experienced by a UE; and communicating with a second network device , to reduce spatial downlink inter-cell interference in the direction of the UE by protecting resources across the selected set of resources.

Aspect 15: The apparatus of Aspect 14, wherein the first network device includes a serving cell and the second network device includes a potentially interfering neighbor cell, the apparatus further comprising: for selecting via the first network device for the A unit for the UE whose predicted spatial downlink inter-cell interference exceeds a predetermined threshold; and a unit for sending a request message to the second network device.

Aspect 16: The device according to Aspect 15, further comprising: means for receiving a response message indicating acceptance of the offer indicated through the request message.

Aspect 17: The device according to Aspect 15, further comprising: means for receiving a response message indicating an alternative proposal for a different set of resources.

Aspect 18: The apparatus according to Aspect 15, further comprising: updating the request in response to an updated prediction based on the updated UE location, the updated channel condition for the UE, and/or the updated throughput requirement for the UE The unit of the message.

Aspect 19: The apparatus according to Aspect 14, wherein the first network device includes a neighbor cell, and the second network device includes a serving cell, the device also includes: means for receiving a request message from the second network device, The request message indicates the location of the UE, the UE interference tolerance threshold and/or the resource requirement for the UE; a unit for predicting based on the request message; and a unit for sending a response message to the second network device.

Aspect 20: The apparatus according to Aspect 19, further comprising: means for receiving updates about UE location, UE interference tolerance threshold and/or UE-specific resource requirements from the second network device.

Aspect 21: The apparatus according to Aspect 14, wherein the first network device includes a central node, and the second network device includes a serving cell, the apparatus also includes: means for receiving the first request message from the second network device , the first request message indicates the position of the UE, the UE interference tolerance threshold and/or the resource requirement for the UE; a unit for predicting based on the first request message; a second request for resource protection sent to the aggressor neighbor cell A unit for requesting a message; and a unit for sending a response message indicating a protected resource to a second network device.

Aspect 22: The apparatus according to Aspect 21, further comprising means for updating the prediction based on the updated UE location, the updated channel condition for the UE, and/or the updated resource requirement for the UE.

Aspect 23: A first network device, comprising: a processor; a memory coupled to the processor; and operable instructions stored in the memory, the instructions cause the first network device to Predicting spatial intercellular downlink interference experienced by the UE; and communicating with a second network device to reduce spatial intercellular downlink in the direction of the UE by protecting resources across the selected set of resources road interference.

Aspect 24: The first network device according to Aspect 23, wherein the first network device includes a serving cell and the second network device includes a potentially interfering neighbor cell, the instructions also causing the first network device to: : selecting UEs for which the predicted inter-cell downlink interference exceeds a predetermined threshold; and sending a request message to the second network device.

Aspect 25: The first network device according to Aspect 23, wherein the first network device includes a neighbor cell, and the second network device includes a serving cell, and the instruction also causes the first network device to perform the following operations: The second network device receives the request message, the request message indicates the location of the UE, the UE interference tolerance threshold and/or the resource requirement for the UE; predicts based on the request message; and sends a response message to the second network device.

Aspect 26: The first network device according to Aspect 23, wherein the first network device includes a central node, and the second network device includes a serving cell, and the instructions also cause the first network device to perform the following operations: from the second The network device receives the first request message, the first request message indicates the location of the UE, the UE interference tolerance threshold and/or the resource requirement for the UE; predicts based on the first request message; sends a request for resource protection to the aggressor neighbor cell the second request message; and sending a response message indicating the protected resource to the second network device.

Aspect 27: The first network device according to Aspect 26, wherein the instructions also cause the first network device to: based on the updated UE location, the updated channel condition for the UE, and/or the updated resource for the UE needs to update the forecast.

Aspect 28: A non-transitory computer-readable medium having recorded thereon program code executed by a processor of a first network device and comprising: a method for predicting a spatial inter-cell downlink experienced by a UE and code for communicating with a second network device to reduce spatial inter-cell downlink interference in the direction of the UE by protecting resources across the selected set of resources.

Aspect 29: The non-transitory computer-readable medium of clause 28, wherein the first network device includes a neighbor cell and the second network device includes a serving cell, the non-transitory computer-readable medium also includes: Code for receiving a request message from a second network device, the request message indicating the location of the UE, UE interference tolerance threshold and/or resource requirements for the UE; code for making predictions based on the request message; Code for sending a response message from the second network device.

Aspect 30: The non-transitory computer-readable medium of clause 28, wherein the first network device comprises a central node and the second network device comprises a serving cell, the non-transitory computer-readable medium also comprises: for Code for receiving a first request message from a second network device, the first request message indicating the location of the UE, UE interference tolerance threshold and/or resource requirements for the UE; code for making prediction based on the first request message ; a program code for sending a second request message requesting resource protection to the aggressor neighbor cell; and a program code for sending a response message indicating the protected resource to a second network device.

The foregoing disclosure provides illustration and description, but is not intended to be exhaustive or to limit aspects to the precise forms disclosed. Modifications and variations are possible in light of the above disclosure or may be acquired from practice of the various aspects.

As used, the term "component" is intended to be interpreted broadly as hardware, firmware, and/or a combination of hardware and software. As used, a processor is implemented in hardware, firmware, and/or a combination of hardware and software.

Some aspects are described in conjunction with thresholds. As used, meeting a threshold may mean that a value is greater than the threshold, greater than or equal to the threshold, less than the threshold, less than or equal to the threshold, equal to the threshold, not equal to the threshold, etc., depending on the context.

It will be apparent that the described systems and/or methods can be implemented in different forms of hardware, firmware, and/or a combination of hardware and software. The actual specialized control hardware or software code used to implement these systems and/or methods is not a limitation of the various aspects. Thus, the operation and behavior of the systems and/or methods are described without reference to specific software code, it being understood that software and hardware can be designed to implement the systems and/or methods based at least in part on the description.

Even if specific combinations of features are described in the claims and/or disclosed in the specification, these combinations are not intended to limit the disclosure of each aspect. In fact, many of these features may be combined in ways not specifically recited in the claims and/or disclosed in the specification. Although each dependent claim item listed below may directly depend on only one claim item, the disclosure of the various aspects includes each dependent claim item in combination with every other claim item in the set of claims. A phrase referring to "at least one of" a list of items means any combination of those items, including individual members. For example, "at least one of a, b, or c" is intended to encompass a, b, c, a-b, a-c, b-c, and a-b-c, and any combination of multiples of the same element (e.g., a-a, a-a-a, a-a-b, a-a-c, a-b-b , a-c-c, b-b, b-b-b, b-b-c, c-c, and c-c-c, or any other ordering of a, b, and c).

No element, act, or instruction used should be construed as critical or essential unless expressly described as such. Also, as used, the articles "a" and "an" are intended to include one or more items and may be used interchangeably with "one or more". In addition, as used, the terms "collection" and "set" are intended to include one or more items (eg, related items, unrelated items, a combination of related items and unrelated items, etc.), and may be combined with "one or more" Used interchangeably. Where only one item is intended, the phrase "only one" or similar language is used. Furthermore, as used, the terms "has", "have", "having" and the like are intended to be open-ended terms. Additionally, the phrase "based on" is intended to mean "based at least in part on," unless expressly stated otherwise.

100: Internet 102a: Macrocytosis 102b: pico cells 102c: Femtocells 110:BS 110-1: The first base station 110-2: Adjacent base station 110a:BS 110b:BS 110c:BS 110d:BS 110-N: base station 120:UE 120-1:UE 120-2:UE 120a:UE 120b:UE 120c:UE 120d:UE 120e:UE 120-N:UE 130: Network controller 150: Neural Processing Engine 160: Neural Processing Engine 200: Design 212: Sources of information 220: send processor 230: Transmit (TX) multiple-input multiple-output (MIMO) processor 232a: Modulator (MOD) 232t: modulator (MOD) 234a: Antenna 234t: Antenna 236:MIMO detector 238: Receive processor 239: data slot 240: Controller/Processor 242: memory 244: Communication unit 246: Scheduler 252a: Antenna 252r: Antenna 254a: Demodulator (DEMOD) 254r: demodulator (DEMOD) 256:MIMO detector 258: Receive processor 260: data slot 262: Sources of information 264: send processor 266:TX MIMO processor 280: Controller/Processor 282: memory 290: Controller/Processor 292: memory 294: Communication unit 300: System on Chip (SOC) 302: Central Processing Unit (CPU) 304: Graphics Processing Unit (GPU) 306:Digital signal processor (DSP) 308: Neural Processing Unit (NPU) 310: connection block 312:Multimedia Processor 314: sensor processor 316: Image Signal Processor (ISP) 318: memory block 320:Navigation module 400:DCN 402: Neural Network 404: Neural Network 406:Convolutional Neural Network 408: Connection Strength 410: value 412: value 414: value 416: value 418: The first feature map set 420: The second feature map set 422: output 424: The first eigenvector 426: Image 428: Second eigenvector 430: image capture device 432:Convolution layer 550: Deep Convolutional Networks 552: Input data 554A: Convoluted block 554B: Convoluted block 556:Convolution layer 558:Regularization layer 560: Maximum pooling layer 562: Fully Connected Layers 564:Logistic regression (LR) layer 566: Classification Score 600: The first interference scene 650: The second interference scene 700: communication network 800: network 802: UE location 804: Precoder 810: Neural Processing Engine 820: Interference-based distribution 830: location block 900: Service cells 950: Adjacent cells 950-1: Adjacent Cells 950-N: Neighbor Cells 960: central node 1200: example program 1202: block 1204: block t0: time t1: time t2: time t3: time t4: time t5: time t6: time t7: time t8: time

So that the features of the present disclosure can be understood in detail, the specific description can be had by reference to various aspects, some of which are shown in the accompanying drawings. It is to be noted, however, that the drawings only illustrate certain aspects of the subject matter and are therefore not to be considered limiting of its scope, as the description may admit other equally valid aspects. The same reference numbers in different drawings may identify the same or similar elements.

FIG. 1 is a block diagram conceptually illustrating an example of a wireless communication network according to various aspects of the present disclosure.

FIG. 2 is a block diagram conceptually illustrating an example of communication between a base station and a user equipment (UE) in a wireless communication network according to various aspects of the disclosure.

3 illustrates an example implementation of designing a neural network using a system-on-chip (SOC) including a general-purpose processor in accordance with certain aspects of the subject matter.

4A, 4B, and 4C are schematic diagrams illustrating neural networks according to aspects of the present disclosure.

4D is a schematic diagram illustrating an exemplary deep convolutional network (DCN) according to aspects of the present disclosure.

5 is a block diagram illustrating an exemplary deep convolutional network (DCN) according to aspects of the present disclosure.

6A and 6B illustrate communication networks according to aspects of the present disclosure, wherein spatial interference experienced by UEs is based on downlink transmit beams from neighboring base stations to neighboring user equipments (UEs).

7 is a schematic diagram of various aspects of the communication network according to the content of the present application, the communication network shows the signal strength measurement of the downlink transmit beams from neighboring base stations to realize the spatial inter-cell interference aware downlink coordination.

8 is a block diagram of a network including a neural processing engine configured to implement spatial intercellular interference-aware downlink coordination in accordance with aspects of the present disclosure.

9 is a timing diagram illustrating an example procedure, eg, performed by a network, for spatial inter-cell interference-aware downlink coordination at a serving cell, in accordance with various aspects of the subject matter.

10 is a timing diagram illustrating an example procedure, eg, performed by a network, for spatial inter-cell interference-aware downlink coordination at neighbor cells, in accordance with aspects of the subject matter.

11 is a timing diagram illustrating an example procedure, eg, performed by a device, for spatial intercellular interference-aware downlink coordination at a central node, in accordance with aspects of the subject matter.

12 is a flowchart illustrating an example procedure for spatial intercellular interference-aware downlink coordination, such as performed by a network device, in accordance with aspects of the present disclosure.

Domestic deposit information (please note in order of depositor, date, and number) none Overseas storage information (please note in order of storage country, institution, date, and number) none

110-1: The first base station

110-2: Adjacent base station

110-3~110-N: base station/BS

120-1~120-N:UE

900: Service cells

950-1~950-N: Adjacent cells

Claims (30)

A method of wireless communication performed by a first network device, comprising the following steps: predicting spatial intercellular downlink interference experienced by a UE; and Communicating with a second network device to reduce the spatial inter-cell downlink interference in a direction of the UE by protecting resources across the selected set of resources. The method according to claim 1, wherein the first network device includes a serving cell, and the second network device includes a potentially interfering neighbor cell, the method also includes the following steps: selecting, by the first network device, the UE for which the predicted inter-cellular downlink interference exceeds a predetermined threshold; and Send a request message to the second network device. The method according to claim 2, wherein the request message indicates beams causing too much interference. The method according to claim 2, wherein the request message indicates the time/frequency resource to be protected. The method according to claim 2, wherein the request message indicates a quantity of resources to be protected, and the quantity is determined based on a transmission capacity requirement of the UE. The method according to claim 2, wherein the request message indicates an index of the selected resource set in a predetermined list of resource sets. The method according to claim 2 also includes the following steps: receiving a response message indicating acceptance of a proposal indicated by the request message. The method according to claim 2 also includes the step of: receiving a response message indicating an alternative proposal for a different set of resources. The method according to claim 2, further comprising the step of: updating in response to an updated prediction based on an updated UE location, an updated channel condition for the UE, and/or an updated throughput requirement for the UE The request message. The method according to claim 1, wherein the first network device includes a neighbor cell, and the second network device includes a serving cell, the method also includes the following steps: receiving a request message from the second network device, the request message indicating a location of the UE, a UE interference tolerance threshold and/or a resource requirement for the UE; make predictions based on the request information; and Send a response message to the second network device. The method according to claim 10 also includes the step of: receiving an update from the second network device about the location of the UE, the interference tolerance threshold of the UE and/or the resource requirement for the UE. The method according to claim 1, wherein the first network device includes a central node, and the second network device includes a serving cell, the method also includes the following steps: receiving a first request message from the second network device, the first request message indicating a location of the UE, a UE interference tolerance threshold and/or a resource requirement for the UE; predicting based on the first request message; sending a second request message requesting resource protection to an aggressor neighbor cell; and A response message indicating protected resources is sent to the second network device. The method according to claim 12, also comprising the step of updating the prediction based on an updated UE location, updated channel conditions for the UE and/or updated resource requirements for the UE. An apparatus for wireless communication by a first network device, comprising: means for predicting spatial downlink inter-cell interference experienced by a UE; and Means for communicating with a second network device to reduce the spatial downlink inter-cell interference in a direction of the UE by protecting resources across a selected set of resources. The device according to claim 14, wherein the first network device includes a serving cell, and the second network device includes a potentially interfering neighbor cell, the device also includes: means for selecting, via the first network device, the UE for which the predicted spatial downlink inter-cell interference exceeds a predetermined threshold; and A unit for sending a request message to the second network device. The device according to claim 15, further comprising: means for receiving a response message indicating acceptance of an offer indicated by the request message. The device according to claim 15, further comprising: means for receiving a response message indicating an alternative proposal for a different set of resources. The apparatus according to claim 15, further comprising: updating in response to an updated prediction based on an updated UE location, an updated channel condition for the UE, and/or an updated traffic demand for the UE The unit of the request message. The device according to claim 14, wherein the first network device includes a neighbor cell, and the second network device includes a serving cell, the device also includes: a unit for receiving a request message from the second network device, the request message indicating a location of the UE, a UE interference tolerance threshold and/or a resource requirement for the UE; means for making predictions based on the request message; and A unit for sending a response message to the second network device. The apparatus according to claim 19, further comprising: a unit for receiving updates about the UE location, the UE interference tolerance threshold and/or the resource requirement for the UE from the second network device. The device according to claim 14, wherein the first network device includes a central node, and the second network device includes a serving cell, the device also includes: A unit for receiving a first request message from the second network device, the first request message indicating a location of the UE, a UE interference tolerance threshold and/or a resource requirement for the UE; a unit for performing prediction based on the first request message; means for sending a second request message requesting resource protection to an aggressor neighbor cell; and means for sending a response message indicating protected resources to the second network device. The apparatus according to claim 21, further comprising: means for updating the prediction based on an updated UE location, updated channel conditions for the UE and/or updated resource requirements for the UE. A first network device, comprising: a processor; a memory coupled to the processor; and Operable instructions stored in the memory, the instructions cause the first network device to perform the following operations when executed by the processor: predicting spatial intercellular downlink interference experienced by a UE; and Communicating with a second network device to reduce the spatial inter-cell downlink interference in a direction of the UE by protecting resources across the selected set of resources. The first network device according to claim 23, wherein the first network device includes a serving cell and the second network device includes a potentially interfering neighbor cell, the instructions also cause the first network device to Do the following: selecting the UE for which the predicted inter-cellular downlink interference exceeds a predetermined threshold; and Send a request message to the second network device. According to the first network device of claim 23, wherein the first network device includes a neighbor cell, and the second network device includes a serving cell, the instructions also cause the first network device to perform the following operations : receiving a request message from the second network device, the request message indicating a location of the UE, a UE interference tolerance threshold and/or a resource requirement for the UE; make predictions based on the request information; and Send a response message to the second network device. According to claim 23 of the first network device, wherein the first network device includes a central node, and the second network device includes a serving cell, the instructions also cause the first network device to perform the following operations: receiving a first request message from the second network device, the first request message indicating a location of the UE, a UE interference tolerance threshold and/or a resource requirement for the UE; predicting based on the first request message; sending a second request message requesting resource protection to an aggressor neighbor cell; and A response message indicating protected resources is sent to the second network device. The first network device according to claim 26, wherein the instructions also cause the first network device to perform the following operations: based on an updated UE location, updated channel conditions for the UE and/or updated channel conditions for the UE A resource requirement for updating the forecast. A non-transitory computer readable medium having recorded thereon code for execution by a processor of a first network device and comprising: code for predicting spatial intercellular downlink interference experienced by a UE; and Code for communicating with a second network device to reduce the spatial inter-cell downlink interference in a direction of the UE by protecting resources across a selected set of resources. The non-transitory computer-readable medium according to claim 28, wherein the first network device includes a neighbor cell and the second network device includes a serving cell, the non-transitory computer-readable medium also includes: code for receiving a request message from the second network device, the request message indicating a location of the UE, a UE interference tolerance threshold and/or a resource requirement for the UE; code for making predictions based on the request message; and A program code for sending a response message to the second network device. The non-transitory computer-readable medium according to claim 28, wherein the first network device includes a central node, and the second network device includes a serving cell, the non-transitory computer-readable medium also includes: code for receiving a first request message from the second network device, the first request message indicating a location of the UE, a UE interference tolerance threshold and/or a resource requirement for the UE; a code for performing prediction based on the first request message; code for sending a second request message requesting resource protection to an aggressor neighbor cell; and Code for sending a response message indicating protected resources to the second network device.

Images