Physical cell identity optimization in open ran networks assisted by to artificial intelligence / machine learning techniques

Abstract

A PCI Optimization Module comprising: 1) a PCI Conflict Optimization Detection Module to detect the PCI conflicts and configured to apply the policy at a Non-RT-RIC and 2) a PCI Conflict Optimization Resolution Module ata Near-RT-RIC configured to execute an analyzing and determining method. The method comprises: detecting a PCI Conflict at a non-real-time (Non-RT) RIC; providing, by the Non-RT RIC, parameters and policies to a Near-RT RIC with an identity of cells involved in the PCI Conflict along with a new PCI to be applied to one of the identifed cells; at the Near-RT RIC: analyzing cell-level information and KPIs received from an E2 node and determining a candidate cell of the idenified cells whose PCI value should be changed; and determining an optimal value of a shutdown timer when the candidate cell can be shut down and the new PCI can be applied during a boot-up procedure.

Description

PHYSICAL CELL IDENTITY OPTIMIZATION IN OPEN RAN NETWORKS ASSISTED BY TO ARTIFICIAL INTELLIGENCE / MACHINE LEARNING TECHNIQUES BACKGROUND Field of the Disclosure

[0001] The present disclosure relates to systems and methods for radio accessnetworks. The present disclosure is related to the design of operation, administration and management of various network elements of 4G, 5G, and further in a Fifth Generation (5G) New Radio (NR) system. The present disclosure also related to Physical Cell Identity (PCI) Optimization in Open Radio Access Network (O-RAN) wireless networks and relates more particularly to Artificial Intelligence / Machine Learning (AI / ML) assisted techniques. Description of the Related Technology

[0002] Physical Cell Identity (PCI) of a cell which identifies a cell on a frequencylayer. It is derived by User Equipment UE from a Primary synchronization signal (PSS) and Secondary Synchronization Signal (SSS). PCI is used to scramble data in order to allow a UE to separate information from different cells (or eNodeBs), and PCI of each cell should be locally unique in a deployment area.

[0003] As number of cells increase in a network, PCI gets repeated over the footprintof the network (for a given frequency layer). This can result in reuse of some PCIs and can result in the following types of PCI Conflicts (for a given frequency band): 1) PCI Collision happens whenever a cell has a neighbor with identical PCI (in the same frequency band), and 2) PCI Confusion where a serving cell detects that two neighbor cells have the same PCI (in the same frequency band).

[0004] Network changes (such as increasing power of a cell) can also result in suchPCI Conflicts. PCI Conflicts can result in increased number of failed handover (and thus higher number of dropped calls) and increased channel interference.

[0005] Cell administration refers to permission to use cell resources and thisdecision is indicated through management services for that cell. When the administrative state of Distributed Unit DU in a 5G NR Cell is changed from unlocked to shutting down, DU should perform the cell shutdown operation (such as using a configured policy). Extant SON Self Organizing Networks based methods can be used to detect the PCI conflict based on the neighbor cell information available at the SON.

[0006] When there is a PCI collision between two neighboring cells, some of theexisting methods select a cell randomly (from the group of cells which is causing PCI collision) and update the PCI of that cell by selecting a PCI from the PCI pool. Choosing the cell randomly for PCI change can impact the user experience in the cell where PCI is updated. PCI change can only be applied after bringing down the cell and this can have impact on the existing sessions in that cell and degrade the KPIs (Key Performance Indicators) in that cell.

[0007] One other existing option is to wait for the maintenance period (typicallynon-busy hours) and apply the PCI change on that cell but this can impact synchronization for UEs in those cells and impact system KPIs. For example, PCI conflict can result in high call drops, high RRC failures, reduced intra-frequency handover success, degraded DL throughput, high block error rate, high DL Latency and the like.

[0008] PCI can be applied in a graceful manner by using the shutdown mechanismdescribed above. In this case, gNB-DU is configured to wait for the duration specified by the shutdown timer before releasing the existing sessions and the cell is not available for new users until that cell is brought up again after completion of the shutdown process. The shutdown timer value can range from few seconds to even hours, and choosing right value for the shutdown timer is important. SUMMARY

[0009] In an embodiment, described is a method comprising: detecting a PCI Conflictat a non-real-time (Non-RT) RIC; providing, by the Non-RT RIC, parameters and policies to a Near-RT RIC with an identity of cells involved in the PCI Conflict along with a new PCI to be applied to one of the identifed cells; and at the Near-RT RIC: analyzing cell-levelinformation and KPIs received from an E2 node and determining a candidate cell of the idenified cells whose PCI value should be changed; and determining an optimal value of a shutdown timer when the candidate cell can be shut down and the new PCI can be applied during a boot-up procedure. The method can further comprise providing a PCI Optimization Module further comprising: 1) a PCI Conflict Optimization Detection Module to detect the PCI conflicts and configured to apply the policy at the Non-RT-RIC and 2) a PCI Conflict Optimization Resolution Module at the Near-RT-RIC configured to execute the analyzing and determining steps.

[0010] In an embodiment, descibed is an apparatus comprisng: a PCI OptimizationModule configued to select a cell whose PCI should be changed and a corresponding shutdown timer for changing the PCI for the selected cell to resolve PCI Conflicts in an optimal manner. The PCI Optimization Module can further comprise: 1) a PCI Conflict Optimization Detection Module to detect PCI conflicts and configured to apply a policy at the Non-RT-RIC or 2) a PCI Conflict Optimization Resolution Module to help resolve PCI Conflicts in an optimal manner at the Near-RT-RIC, or both. The 1) PCI Conflict Optimization Detection Module at the Non-RT RIC can be configured to collect information from a gNodeB including: a PCI of each cell, and a list of neighbors of each cell along with neighbor PCIs from aNeighbor Relation Table (NRT). The NRT information can be collected from a gNB-CU-CP. The 1) PCI Conflict Optimization Detection Module cam be configured to collect a PCI List of available PCIs for each gNodeB. The 1) PCI Conflict Optimization can be configured to: analyze the collected information and detect a PCI Conflict; select a new PCI to be assigned from the PCI List a cell involved in the PCI Conflict; and send an identity of the cell involved in PCI Conflict and the new PCI to be assigned to the cell to the Near-RT RIC.

[0011] In an embodiment, the 2) PCI Conflict Optimization Resolution Module at theNear-RT-RIC can be configured to periodically collect information for each cell from the gNB-CU-CP comprising: a number of connected users in a cell; a number of emergency (VoNR) calls in a cell; UE measurement reports for from the cell; and handover related counters and metrics. The 2) PCI Conflict Optimization Resolution Module at the Near-RT-RIC can be configured to periodically collect information for each cell from the gNB-DU comprising: a number of active users in a cell; and a PRB utilization of the cell.

[0012] The 2) PCI Conflict Optimization Resolution Module at the Near-RT-RIC canbe configured to predict information comprising: a predicted number of connected users; a predicted number of active users; and a predicted cell load. The 2) PCI Conflict Optimization Resolution Module at the Near-RT-RIC is configured to use the predicted information to select a cell for PCI change between the two cells involved in a PCI Conflict and decide a value of the timer when that cell is to be shut down to change the PCI.

[0013] In an embodiment, the PCI Optimization Module can at the gNB CU-CP andcomprise an AI / ML model trained to predict cell parameters. The AI / Ml cam comprise: LSTM or Transformer based deep neural network models to predict PRB utilization, a number of connected users and a number of active users for a given cell.

[0014] In an embodiment, described is a method comprising: detecting, by a PCIOptimization Module at a gNB-CU-CP, a PCI Conflict; and if the PCI Conflict is detected, it selecting a new PCI to be assigned to one of a plurality of cells that are involved in the PCI Conflict and selecting the cell to which this new PCI is to be assigned to resolve PCI Conflict in an optimal manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] FIG. 1a shows an example of a User Plane Stack.

[0016] FIG. 1b is a block diagram illustrating User Plane Stack protocols for a PDUsession.

[0017] FIG. 2 shows an example of a Control Plane Stack.

[0018] FIG. 3 shows an example of high-level NG-RAN including a gNB CU and DU.

[0019] FIG. 4 shows an example of a Separation of CU-CP and CU-UP in a 5G gNB.

[0020] FIG. 5a shows an example of an O-RAN architecture.

[0021] FIG. 5b illustrates a flow for an E2 node and Near-RT-RIC establishing an E2session.

[0022] FIG. 6 illustrates an overall structure of an LSTM network.

[0023] FIG. 7 illustrates an example of a PCI assignment for three cells.

[0024] FIG. 8 illustrates a PCI cell collision.

[0025] FIG. 9 illustrates a flow for an implementation.

[0026] FIG. 10 shows a PCI Conflict Optimization Module at a Near-RT RIC collectingcell information from gNB-CU-CP.

[0027] FIG. 11 shows a PCI Conflict Optimization Module at a Near-RT RIC collectingcell information from gNB-DU.

[0028] FIG. 12 illustrates a flow for an implementation.

[0029] FIG. 13 a flow for training and implementing and AI / ML model.

[0030] FIG. 14 illustrates an AI / ML model to predict certain parameters at a gNB-CU-UP.

[0031] FIG. 15 illustrates a logical flow for training and implementing and AI / MLmodel.

[0032] FIG. 16 illustrates a logical flow for training and implementing and AI / MLmodel.

[0033] FIG. 17 illustrates a flow for an implementation.

[0034] FIG. 18 illustrates a flow for an implementation.DETAILED DESCRIPTION OF THE DISCLOSURE

[0035] Reference is made to Third Generation Partnership Project (3GPP) and theInternet Engineering Task Force (IETF) and related standards bodies in accordance with embodiments of the present disclosure. The present disclosure employs abbreviations, terms and technology defined in accord with Third Generation Partnership Project (3GPP) and / or Internet Engineering Task Force (IETF) technology standards and papers, including the following standards and definitions.3GPP and IETF technical specifications (TS), standards (including proposed standards), technical reports (TR) and other papers are incorporated by reference in their entirety hereby, define the related terms and architecture reference models herein.

[0036] Abbreviations5GC: 5G Core Network 5G NR: 5G New Radio 5QI: 5G QoS Identifier ACK: Acknowledgement AI: Artificial Intelligence AI / ML (or AIML): Artificial Intelligence and Machine Learning AM: Acknowledged Mode APN: Access Point Name ARP: Allocation and Retention Priority BO: Buffer Occupancy BS: Base Station BSR: Buffer Status Report CNN: Convolution Neural Network CMS: Centralized Management SystemCP: Control Plane CSI: Channel State Information CU: Centralized Unit CU-CP: Centralized Unit – Control Plane CU-UP: Centralized Unit – User Plane DL: Downlink DDDS: DL Data Delivery Status DNN: Data Network Name DNN: Deep Neural Network DQN: Deep Q Network DRB: Data Radio Bearer DU: Distributed Unit eNB: evolved NodeB EPC: Evolved Packet Core EN-DC: E-UTRAN New Radio Dual Connectivity GBR: Guaranteed Bit Rate gNB: gNodeB GTP-U: GPRS Tunnelling Protocol – User Plane IP: Internet Protocol L1: Layer 1 L2: Layer 2 L3: Layer 3 L4S: Low Latency, Low Loss and Scalable Throughput LC: Logical ChannelLESS: Low Energy Scheduler Solution LSTM: Long Short-Term Memory MAC: Medium Access Control MDP: Markov Decision Process MIB: Master Information Block ML: Machine Learning MR-DC: Multi-RAT Dual Connectivity NACK: Negative Acknowledgement NAS: Non-Access Stratum NG-RAN: Next Generation Radio Access Network NR-U: New Radio – User Plane NSI: Network Slice Instance NSSI: Network Slice Subnet Instance NWDAF: Network Data Analytics Function O-RAN: Open Radio Access Network OAM: Operations, Administration Maintenance PCI: Physical Cell Identity PDB: Packet Delay Budget PDCP: Packet Data Convergence Protocol PDU: Protocol Data Unit PER: Packet Error Rate PF: Proportional Fair PHY: Physical Layer PRB: Physical Resource BlockQCI: QoS Class Identifier QFI: QoS Flow Identifier QoS : Quality of Service RAN: Radio Access Network RAT: Radio Access Technology RB: Resource Block RDI: Reflective QoS Flow to DRB Indication RIC: Radio Intelligent Controller RL: Reinforcement Learning RLC: Radio Link Control RLC-AM: RLC Acknowledged Mode RLC-UM: RLC Unacknowledged Mode RNN: Recurrent Neural Networks RQI: Reflective QoS Indication RRC: Radio Resource Control RRM: Radio Resource Management RTP: Real-Time Transport Protocol RTCP: Real-Time Transport Control Protocol RU: Radio Unit SCTP: Stream Control Transmission Protocol SD: Slice Differentiator SDAP: Service Data Adaptation Protocol SIB: System Information Block SLA: Service Level AgreementS-NSSAI: Single Network Slice Selection Assistance SST: Slice / Service Type TB: Transport Block TCP: Transmission Control Protocol TEID: Tunnel Endpoint Identifier UE: User Equipment UP: User Plane UL: Uplink UM: Unacknowledged Mode UPF: User Plane Function

[0037] In the following sections, overview of Next Generation Radio Access Network(NG-RAN) architecture and 5G New Radio (NR) stacks will be discussed.5G NR (New Radio) user and control plane functions with monolithic gNB (gNodeB) are shown in FIGS. 1a, 1b and 2. For the user plane (shown in FIG.1a, which is in accordance with 3GPP TS 38.300), PHY (physical), MAC (Medium Access Control), RLC (Radio Link Control), PDCP (Packet Data Convergence Protocol) and SDAP (Service Data Adaptation Protocol) sublayers originate in the UE 101 and are terminated in the gNB 102 on the network side.

[0038] As shown in FIG. 1b, which is a block diagram illustrating the user planeprotocols stacks for a PDU session, in accordance with 3GPP TS 23.501, PDU layer 9010 corresponds to the PDU carried between the UE 101 and the data network (DN) 9011 over the PDU session. As shown in FIG.1b, UE 101 is connected to the 5G access network (AN) 202, which AN 202 is in turn connected via the N3 interface to the Intermediate UPF (I- UPF) 203a portion of the UPF 203, which I-UPF 203a is in turn connected via the N9 interface to the PDU session anchor 203b portion of the UPF 203, and which PDU session anchor 203b is connected to the DN 9011. The PDU session can correspond to IPv4, IPv6, or both types of IP packets, when the PDU session is of type IPv4, IPv6 or IPv4v6, respectively. GTP-U shown in FIG.1b supports tunnelling user plane data over N3 and N9interfaces and provides encapsulation of end user PDUs for N3 and N9 interfaces.

[0039] For the control plane (shown in FIG. 2, which is in accordance with 3GPP TS38.300), RRC (Radio Resource Control), PDCP, RLC, MAC and PHY sublayers originate in the UE 101 and are terminated in the gNB 102 on the network side, and NAS (Non-Access Stratum) originate in the UE 101 and is terminated in the AMF (Access Mobility Function) 103 on the network side.

[0040] NG-Radio Access Network (NG-RAN) architecture from 3GPP TS 38.401 isshown in FIGS.3-4. As shown in FIG.3, the NG-RAN 301 comprises of a set of gNBs 302 connected to the 5GC 303 through the NG interface. Each gNB comprises gNB-CU 304 and one or more gNB-DU 305 (see FIG.3). As shown in FIG.4 (which illustrates separation of CU-CP (CU-Control Plane) and CU-UP (CU-User Plane)), E1 is the interface between gNB- CU-CP (CU-Control Plane) 304a and gNB-CU-UP (CU-User Plane) 304b, F1-C is the interface between gNB-CU-CP 304a and gNB-DU 305, and F1-U is the interface between gNB-CU-UP 304b and gNB-DU 305. As shown in FIG.4, gNB 302 can comprise a gNB-CU-CP 304a, multiple gNB-CU-UPs (or gNB-CU-UP instances) 304b and multiple gNB-DUs (or gNB-DU instances) 305. One gNB-DU 305 is connected to only one gNB-CU-CP 304a, and one gNB- CU-UP 304b is connected to only one gNB-CU-CP 304a.

[0041] Open Radio Access Network (O-RAN) is based on disaggregated componentswhich are connected through open and standardized interfaces based on 3GPP NG-RAN. An overview of O-RAN with disaggregated RAN CU (Centralized Unit), DU (Distributed Unit), and RU (Radio Unit), near-real-time Radio Intelligent Controller (Near-RT RIC 132) and non-real-time RIC is illustrated in FIG.5a.

[0042] As shown in FIG. 5a, the CU (shown split as O-CU-CP 304a and O-CU-UP304b) and the DU (shown as O-DU 305) are connected using the F1 interface (with F1-C for control plane and F1-U for user plane traffic) over a mid-haul (MH) path. One DU can host multiple cells (for example, one DU can host 24 cells) and each cell can support many users. For example, one cell can support 800 Radio Resource Control (RRC)-connected users and out of these 800, there can be 250 Active users (i.e., users that have data to send at a givenpoint in time).

[0043] A cell site can comprise multiple sectors, and each sector can supportmultiple cells. For example, one site can comprise three sectors and each sector can support eight cells (with each cell being on a different frequency band in a given sector). One CU-CP (CU-Control Plane) can support multiple DUs and thus multiple cells. For example, a CU-CP can support 500 cells and around 100,000 User Equipments (UEs). Each UE can support multiple Data Radio Bearers (DRBs) and there can be multiple instances of CU-UP (CU-User Plane) to serve these DRBs. For example, each UE can support 4 DRBs, and 400,000 DRBs (corresponding to 100,000 UEs) can be served by five CU-UP instances (and one CU-CP instance).

[0044] The DU can be located in a private data center, or it can be located at a cell-site. The CU can also be in a private data center or even hosted on a public cloud system. The DU and CU, which are typically located at different physical locations, can be tens of kilometers apart. The CU communicates with a 5G core system, which can also be hosted in the same public cloud system (or can be hosted by a different cloud provider). A RU (Radio Unit) (shown as O-RU 306 in FIG.5a) is located at a cell-site and communicates with the DU via a front-haul (FH) interface.

[0045] The E2 nodes (CU and DU) are connected to the near-real-time RIC 132 usingthe E2 interface. The E2 interface is used to send data (such as user and / or cell KPMs) from the RAN, and deploy control actions and policies to the RAN at near-real-time RIC 132. The applications or services at the near-real-time RIC 132 that deploys the control actions and policies to the RAN are called xApps. During the E2 setup procedures, the E2 node advertises the metrics it can expose, and an xApp in the Near-RT RIC 132 can send a subscription message specifying key performance metrics which are of interest. The near- real-time RIC 132 is connected to the non-real-time RIC 133 (which is shown as part of Service Management and Orchestration (SMO) Framework 130 in FIG.5a) using the A1 interface. The applications that are hosted at Non-RT-RIC are called rApps. Also shown in FIG.5a are O-eNB 307 (which is shown as being connected to the near-real-time RIC 132 and the SMO Framework 130) and O-Cloud 307 (which is shown as being connected to theSMO Framework 130).

[0046] As in FIG. 5b, E2 node (which is DU or CU) and Near-RT RIC 132 establish E2session using E2 SETUP REQUEST and E2 SETUP RESPONSE. Near-RT RIC 132 can subscribe to certain parameters from the E2 node (on behalf the xApp running at Near-RT RIC 132) using the RIC SUBSCRIPTION REQUEST and E2 node acknowledges this message by sending RIC SUBSCRIPTION RESPONSE to the Near-RT RIC 132. As part of this, xApp running at the Near-RT RIC 132 also provides the event triggers to E2 node, so that it can ask E2 node to REPORT subscribed parameters periodically to the xApp or to REPORT these subscribed parameters based on certain events to the xApp. E2 node communicates subscribed parameters to Near-RT RIC 132 (and the xApp) using RIC INDICATION as shown in FIG.5b. After analysing received parameters from the E2 nodes (and based on network operator policies), Near-RT RIC 132 can send RIC CONTROL REQUEST to take an action at the E2 node (such as influence mobility decision). E2 node acknowledges this message by sending RIC CONTROL ACKNOWLEDGE to Near-RT RIC 132 while E2 node takes action as asked by the Near-RT RIC 132.

[0047] In this section, basic structures and characteristics of neural networks will bediscussed. Recurrent neural networks (RNNs) are a family of neural networks that are suited for handling sequential data. RNNs use a hidden state associated with each time-step and the output at each time step is computed using the input and the previous hidden state. RNN architectures suffer from some limitations. First, RNNs fail to store information for a long period of time and thus do not handle the situations well in which a reference to certain information stored quite a long time ago is required to predict the current output. Second, there is no fine control to select which part of the context needs to be carried forward and or can be “forgotten”. Also, in RNNs, gradients in early layers are computed as the product of terms from later layers. With this, gradients in early layers can grow exponentially large (for example, if the terms in later layers are large enough) or can exponentially decrease (for example, if terms in later layers are small).

[0048] Long Short-Term Memory (LSTM) networks, which are an extension ofrecurrent neural networks (RNNs), have been introduced to handle situations where RNNsdo not work well. The basic difference between the architectures of RNNs and LSTMs is that the hidden layer of LSTM is a gated unit or a gated cell, which comprises of four layers that interact with one another in a way to produce the output of that cell along with the cell state. The output and the cell state are then passed onto the next hidden layer. Unlike RNNs which have only a single neural net layer of tanh (hyperbolic tangent), LSTMs comprise three logistic sigmoid gates and one tanh layer. It should be noted that output of sigmoid activation function is in the range (0 to 1) for any real value as input, while the output of tanh is in the range (-1 to 1).

[0049] Gates are provided in LSTM in order to limit the information that is passedthrough the cell, i.e., the gates determine which part of the information will be needed by the next cell and which part is to be discarded. The output is usually in the range of 0 to 1, where “0” means “reject all”, and “1” means “include all”. Information is retained by the cells, and the memory manipulations are done by the gates. Three gates are provided in LSTM: 1) forget gate; 2) input gate; and 3) output gate.

[0050] FIG. 6 illustrates the overall structure of an LSTM network 1201. The overallLSTM network can be represented by the following expressions: f= (W . (h , x ) + b )i = (W . (h , x ) + b )o = (W . (h , x ) + b )c = tanh (W . (h , x ) + b )c = f c + i ch = o tanh (c )where, frepresents the forget gate;i represents the input gate;o represents the output gate;is the sigmoid activation function; tanh: hyperbolic tangent activation function; h: output of the previous LSTM block (at time stamp t-1);x : input at the current time stamp;W : weight matrix corresponding to the input gate;W : weight matrix corresponding to the forget gate;W : weight matrix corresponding to the output gate;W : weight matrix corresponding to the cell state;b : bias corresponding to the input gate;b : bias corresponding to the forget gate;b : bias corresponding to the output gate;b : bias corresponding to the cell state;c : represents candidate for cell state (memory) at time stamp t;c : cell state (memory) at time stamp t;h : output of the LSTM block (at time stamp t); and: element wise multiplication.

[0051] The information that is no longer useful in the cell state is removed with theforget gate 1201a as shown in FIG.6. Two inputs, xt (input at the particular time t) and ht-1 (previous cell output from time t-1), are fed to the gate and multiplied with weight matrices followed by the addition of bias. The resultant is passed through sigmoid activation function 1202a, which produces a binary output. For a particular cell state, i) if the output is 0, the piece of information is forgotten, and ii) if output is 1, the information is retained for future use. As noted above in connection with FIG.6, the forget gate is represented by the expression: f= (W . (h , x ) + b )

[0052] The addition of useful information to the cell state is done by the input gate1201b as shown in FIG.6. First, the information is regulated using the sigmoid function 1202b, which filters the values to be remembered using inputs ht-1 and xt. Then, a vector, , is created using the tanh function 1203a which is used to compute the new cell state(at time t), , as given below. The input gate and its associated functionality can berepresented by the following expressions:i = (W . (h , x ) + b )c = tanh (W . (h , x ) + b )

[0053] The task of extracting useful information from the current cell state to bepresented as output is done by the output gate 1201d, which is also shown in FIG.6. First, a vector is generated by applying the tanh function 1203b on the cell state ct. Then, the information is regulated using the sigmoid function 1202c and filtered for the values to be remembered using inputs ht-1and xt. Lastly, the values of the vector and the filtered (regulated) values are multiplied to be sent as an output of this cell and input to the next cell. The output gate 1201d and its associated functionality can be represented by the following expressions: o= (W . (h , x ) + b )h = o tanh (c )

[0054] In this section, described is the Physical Cell Identity (PCI) of a cell whichidentifies a cell on a frequency layer. It is derived by the UE from Primary synchronization signal (PSS) and Secondary Synchronization Signal (SSS) and is computed as: PCI = PSS + 3 * SSS. Here, PSS and SSS are integer numbers. For 5G NR, PSS can be equal to 0 or 1 or 2, and SSS can take a value between 0 and 335 (which will result in PCI being in the range 0 to 1007 for a particular 5G NR frequency layer for this case). For LTE, PSS can be equal to 0 or 1 or 2, and SSS can take a value between 0 and 167 (which will result in PCI being in the range 0 to 503 for a particular LTE frequency layer for this case). PCI is used to scramble data in order to allow a UE to separate information from different cells (or eNodeBs), and PCI of each cell should be locally unique in a deployment area. FIG.7 gives an example of PCI assignment for three cells.

[0055] As number of cells increase in a network, PCI gets repeated over the footprintof the network (for a given frequency layer). This can result in reuse of some PCIs and can result in the following types of PCI Conflicts (for a given frequency band): 1) PCI Collisionhappens whenever a cell has a neighbor with identical PCI (in the same frequency band), and 2) PCI Confusion where a serving cell detects that two neighbor cells have the same PCI (in the same frequency band).

[0056] Network changes (such as increasing power of a cell) can also result in suchPCI Conflicts. PCI Conflicts can result in increased number of failed handover (and thus higher number of dropped calls) and increased channel interference.

[0057] PCI Collision: Ideally, the physical separation between cells using the samePCI should be sufficiently high so that a UE never simultaneously receives the same PCI from more than one cell. It is not always possible to achieve this in real network deployments (for all the time periods when those cells are active) and this results in PCI Collision. FIG.8 shows PCI Collision between Cell P and Cell Z. A UE in the overlapping area of cells P and Z cannot perform signal synchronization and demodulation effectively.

[0058] PCI Confusion: Similarly, the physical separation between cells using thesame PCI should be sufficiently high to avoid neighbor ambiguity at a base station (or at a cell). It should be possible to link a specific PCI with a specific neighboring cell without any confusion. If a cell has multiple neighbors using the same PCI, it results in PCI Confusion. In FIG.8, if the UE is in Cell C and reports PCI = 42 to Cell C for handover then Cell C is in confusion whether the UE is referring to Cell P or Cell Z. This is PCI confusion from the perspective of Cell C and Cell C is referred as confused cell. This can result in increased handover failures (and increased number of call drops).

[0059] In addition to above, a UE should not simultaneously receive multiple PCIswith equal “PCI mode 3” or “PCI mode 4” or “PCI mode 30” values. For example, violation of “PCI mode 3” rule can result in negative impact on the Channel Quality Indication (CQI) values reported by UE and this can lead to lower throughput in LTE networks. There can be DMRS to DMRS interference if neighboring cells have equal “PCI mode 4” values in 5G NR networks.

[0060] In this section, a cell shutdown mechanism is described. Cell administrationrefers to permission to use cell resources and this decision is indicated throughmanagement services for that cell. This is represented by the parameter called, administrative state of that cell and it can take possible values such as “locked”, “unlocked” or “shuttingDown”.

[0061] When the administrative state of DU in a 5G NR Cell (denoted as NRCellDU) ischanged from unlocked to shuttingDown, DU should perform the cell shutdown operation (for example, using a configured policy). For example, in timer-based policy, gNB-DU will start a timer (such as for a duration specified by a timer called “shutdown” timer) and once this timer expires, DU will inform a Central Management System (CMS) 140 that shutdown process is completed. CMS 140 will then trigger cell-lock towards gNB-DU and state of the cell will be changed to “locked” state. As part of the lock procedure handling, the ongoing sessions (if any) can be released by gNB-DU in that cell or handed over to other cells. During the boot-up procedure for this cell, PCI of the cell can be changed to resolve PCI Conflicts.

[0062] Extant SON (Self Organizing Networks) or RIC, such as non-real-time (orNon-RT) RIC, based methods can be used to detect the PCI conflict (such as between Cell P and Cell Z in FIG.8) based on the neighbor cell information available at the SON or RIC. SON or RIC based methods can resolve this by assigning a suitable PCI to one of the cells (such as by assigning a suitable PCI to either Cell P or Cell Z in FIG.7).

[0063] When there is a PCI collision between two neighboring cells, some of theexisting methods select a cell randomly (from the group of cells which is causing PCI collision) and update the PCI of that cell by selecting a PCI from the PCI pool. Choosing the cell randomly for PCI change can impact the user experience in the cell where PCI is updated. PCI change can only be applied after bringing down the cell and this can have impact on the existing sessions in that cell and degrade the KPIs (Key Performance Indicators) in that cell.

[0064] One other existing option is to wait for the maintenance period (typicallynon-busy hours) and apply the PCI change on that cell, but this can impact synchronization for UEs in those cells and impact system KPIs. For example, PCI conflict can result in highcall drops, high RRC failures, reduced intra-frequency handover success, degraded DL throughput, high block error rate, high DL Latency and the like.

[0065] PCI can be applied in a graceful manner by using the shutdown mechanismdescribed above. In this case, gNB-DU is configured to wait for the duration specified by the shutdown timer before releasing the existing sessions and the cell is not available for new users until that cell is brought up again after completion of the shutdown process. The shutdown timer value can range from few seconds to even hour(s) and choosing right value for the shutdown timer is important.

[0066] This disclosure addresses the above two problems. It provides improvedmethods for selecting the candidate cell for PCI change to minimize impact on the existing sessions and to determine good values of shutdown timer dynamically for each cell.

[0067] IMPLEMENTATION (METHOD I)

[0068] Methods for selecting a cell whose PCI should be changed to resolve PCIConflicts are described here. These methods also provide the corresponding shutdown timer after which the cell can be shut down and new PCI can be applied as part of the boot- up or activation procedure. These methods can be placed at CU, RIC or OAM server as described below.

[0069] In the first method, PCI Conflict (Collision or Confusion) detection isperformed at the non-real-time (Non-RT) RIC 133. Non-RT RIC 133 provides parameters and policies to the Near-RT RIC 132 with identity of cells involved in PCI Conflict (such as PCI Collision or Confusion) along with a new PCI that is to be applied to one of these cells. As in FIG.9, the Near-RT RIC 132 analyzes cell-level information and KPIs received from E2 node (i.e. DU and CU) and determines the candidate cell whose PCI value should be changed. It also determines an optimal value of shutdown timer when the chosen cell can be shut down and the new PCI can be applied during the boot-up procedure.

[0070] The PCI Optimization module which is used to select the cell whose PCIshould be changed and the corresponding shutdown timer for changing PCI for thisselected cell, to resolve PCI Conflicts in an optimal manner comprises 1) a PCI Conflict Optimization Module (I) 901, to detect PCI conflicts and help apply various policies at the Non-RT-RIC and 2) a PCI Conflict Optimization Module (II) 902, to help resolve PCI Conflicts in an optimal manner at the Near-RT RIC 132 in this method.

[0071] The PCI Conflict Optimization Module (I) 901 at the Non-RT RIC 133 collectsinformation from gNodeB such as PCI of each cell and list of neighbors of each cell along with their PCIs from Neighbor Relation Table (NRT). The NRT related information is collected from gNB-CU-CP 304a. It also collects list of available PCIs for each gNodeB. The PCI Conflict detection module analyzes the above information and detects PCI Conflicts, if any. If there is any PCI Conflict, it selects a new PCI to be assigned from the PCI List to one of the cells which is involved in this PCI Conflict and sends the following information to the Near-RT RIC 132. Identity of cells involved in PCI Conflict, New PCI to be assigned to one of these cells.

[0072] As in FIG. 10, the PCI Conflict Optimization Module (II) 902 at the Near-RTRIC 132 collects following information from gNB-CU-CP 304a for each cell periodically: Number of connected users in a cell Number of emergency (VoNR) calls in a cell Measurements reports for UEs from that cell (such as RSRP, RSRQ) Handover related counters and metrics Other system KPIs

[0073] As in FIG. 11, the PCI Conflict Optimization Module (II) 902 at the Near-RTRIC 132 collects following information from gNB-DU 305 for each cell periodically: Number of active users in a cell PRB utilization of that cellOther system KPIs (such as related to throughput, block error rate, latency etc.)

[0074] The following are also predicted or captured at the PCI Conflict OptimizationModule (II) 902 at the Near-RT RIC 132: -Predicted number of connected userso Number of connected users for every time T (for each cell) arecommunicated from CU-CP to an AI / ML module (running at Near-RT RIC 132) to predict number of connected users. LSTMs or Transformer based deep neural network models are used to predict number of connected users. -Predicted number of active userso Number of active users for every time T (for each cell) arecommunicated from DU to an AI / ML module (running at Near-RT RIC 132) to predict number of active users for future. LSTMs or Transformer based deep neural network models are used to predict number of active users. -Predicted cell load (for example, in terms of PRB utilization)o PRB utilization for every time T (for each cell) is communicated fromDU to an AI / ML module (running at Near-RT RIC 132) to predict PRB utilization for future. LSTMs or Transformer based deep neural network models are used for this purpose. -Contextual informationo Contextual information, if available, is also provided to the RICmodule to improve accuracy of above predicted parameters.

[0075] Current load of cell c at time t is denoted as CL(c;t). Predicted cell load for cellc for time interval [t+1, t+m] is denoted as predictedCL(c; t+1, t+m). This method uses Wcl * CL(c;t) + Wpredict,cl * predictedCL(c; t+1, t+m) as one of the metric to help determine which cell to shut down. Here, Wcl and Wpredict,cl are the weights assigned to current cell load, CL(c;t), and predicted cell load, predictedCL(c;t+1,t+m), respectively.

[0076] The PCI Conflict Optimization Module (II) 902 at the Near-RT RIC 132 usesthe above information to select the cell for PCI change between the two cells involved in PCI Conflict and decides the value of the timer when that cell should be shut down to change the PCI.

[0077] For example, for two cells (c1, c2) which are involved in PCI Collision, ifcurrent load in cell c1 is less than the cell load in c2, and predicted load in cell c1 is (much) less than cell c2, the PCI Optimization Module (II) 902 can choose cell c1 for which a new PCI should be used.

[0078] As another example, if value of Wcl * CL(c1;t) + Wpredict,cl *predictedCL(c1; t+1, t+m) for cell c1 is greater than value of Wcl * CL(c2;t) + Wpredict,cl * predictedCL(c2; t+1, t+m) for cell c2, and if this difference is above a pre-configured threshold, this method decides to change PCI of cell c2.

[0079] For estimating value of the shutdown timer (to apply new PCI to the selectedcell), the PCI Conflict Optimization Module (II) 902 at the Near-RT RIC 132 analyzes measurement reports from the UEs, handover related performance measures for each cell, other system KPIs and other parameters discussed above and estimates the number of cell- edge UEs which can be involved in PCI conflicts in the neighboring cells. For example, consider Cell A , which is having Cell B (with PCI = 42) and Cell C (with PCI = 42) as neighbor cells. There can be cell-edge UEs in Cell A which are moving towards Cell B or Cell C. When a UE reports PCI = 42 in the measurement report, Cell A may not be able to correctly decide whether to handover this UE to Cell B or Cell C. In some cases, it can result in handover failures due to handover to wrong cell. The PCI Optimization Module (II) 902applies different policies in such cases.

[0080] For example, if there are many UEs which potentially need to be handed overto neighboring cells (which have same PCI) or if the number of handover attempt failures towards cells involved in PCI collision is on the higher side (such as above a pre-configured threshold), the PCI Conflict Optimization Module (II) 902 selects a shorter value of shutdown timer to change PCI of a cell (to resolve PCI conflict) reasonably quickly.

[0081] As another example, if number of active UEs and PRB utilization is high in acell (such as above a pre-configured threshold) and if this is the cell selected by the PCI Optimization Module (II) 902 to change the PCI (in the initial step), then the PCI Conflict Optimization Module (II) 902 chooses a somewhat higher value of the shutdown timer to gracefully terminate active sessions of the UEs in this cell before changing the PCI value of this cell.

[0082] It is possible that one of the AIML models does not provide very accuratevalue of a predicted parameter (for example, predicted cell load may not be very accurate during the initial phase of training). This method continues to compare predicted and the actual observed values. If it is found that the inaccuracy in prediction of a parameter is above a pre-configured threshold, this method reduces the weight used for this predicted parameter (in the PCI related decision making) and also provides feedback to AIML training module to help improve accuracy of the predicted parameter. For example, if it is found that the inaccuracy in cell load prediction (in terms of PRB utilization) is above a pre- configured threshold (such as during the initial phase), value of the corresponding weight, Wpredict,cl, is reduced. It is increased later on as accuracy of the model improves.

[0083] Also, if it is observed that some of the KPIs degrade more than expected aftercell selection and shutdown timer related decisions (such as degradation in handover failures above a pre-configured threshold), the weightage given to the predicted values, such as Wpredict,cl, can be reduced and increased at a later stage as the system KPIs stabilize and start improving.

[0084] The PCI Conflict Optimization Module (II) 902 at the Near-RT RIC 132 canprovide this information (i.e. the cell for which the PCI should be changed and the shutdown timer for that cell) to Non-RT-RIC which can communicate this to the CMS 140 which can further communicate this to gNB-DU 305 for that specific cell. Some of these steps are shown in FIG.12 and given below: -401a, 401b: The PCI Conflict Optimization Module (I) 901 keeps gettinginformation for each cell from the corresponding gNB-DU 305 at 401a and gNB-CU-CP 304a at 401b as described earlier. -402: The PCI Conflict Optimization Module (I) 901 identifies the cells that arecausing PCI Conflicts. It finds new PCIs that can be assigned to some of these cells and provides this information to the PCI Optimization Module (II) 902 at the Near-RT RIC 132. -403: The PCI Conflict Optimization Module (II) 902 module keeps gettingparameters such as number of active UEs, cell load (in terms of PRB utilization) and other system KPIs for each cell from the gNB-DU 305 via the E2 interface. It can get this information periodically or based on specific event triggers. -404: The PCI Conflict Optimization Module (II) 902 module keeps gettingparameters such as number of connected UEs, number of emergency call UEs, UE measurement reports, handover related counters for each cell from the gNB-CU-CP 304a and other system KPIs via the E2 interface. It can get this information periodically or based on specific event triggers. The PCI Conflict Optimization Module (II) 902 analyzes the information available with it and selects the cells for which PCIs should be changed. It also selects a shutdown timer for each cell for which PCI should be changed and PCI change can be applied after the expiry of this shutdown timer. -405: The PCI Conflict Optimization Module (II) 902 provides aboveinformation (i.e. selected cells for which PCIs need to be changed and theshutdown timers) to the PCI Conflict Optimization Module (I) 901. New PCI to be applied to these cells is also communicated. Note that RIC can be taking decisions over a geographical area where there are multiple gNodeBs deployed and more than one PCI Conflicts can be detected around the same time. Thus, RIC can provide identities of multiple cells to help resolve more than one PCI Conflicts at the same time. -406: The PCI Conflict Optimization Module (I) 901 communicates to CMS 140to shut down each such cell and provides the corresponding values of the shutdown timers. -407: CMS 140 triggers shut down for each such cell (as per the instructioncommunicated from the PCI Conflict Optimization Module (I) 901 from Non- RT-RIC). CMS sends a message to the corresponding gNB-DUs 305 for this purpose and provides value of shutdown timer too. -408 gNB-DU 305 starts shutting down that cell and performs actions such as1) Sets CellBarred to Barred state in the MIB for that cell, 2) gNB-DU 305 stops accepting new sessions for that cell, and 3) Shuts down the cell at the expiry of the shutdown timer for that cell. -409: gNB-DU 305 informs CMS 140 when cell shut down is completed.- 410: CMS 140 triggers cell lock for that cell.- 411: gNB-DU 305 indicated to CMS 140 that cell lock is completed.- 412: CMS 140 can boot-up or activate that cell now and apply the new PCIwhich it had received for that cell from the PCI Conflict Optimization Module (I) 901.

[0085] As an alternative solution, the PCI Optimization Module (II) 902 at the Near-RT RIC 132 can send RIC Control message via E2 interface to gNB-DU 305 to update the PCI of the chosen cell. It also provides value of the shutdown timer to the gNB-DU 305 whenthat chosen cell can be shut down and the new PCI can be applied during the boot-up procedure. This is shown in FIG.13, and these steps are given below: -501a, 501b The PCI Conflict Optimization Module (I) 901 keeps gettinginformation (such as PCI for each cell and its neighboring cells) for each cell from the corresponding gNB-CU-CP 304a as described earlier. -502: The PCI Conflict Optimization Module (I) 901 identifies the cells that arecausing PCI Conflicts. It finds new PCIs that can be assigned to some of these cells and provides this information to the PCI Optimization Module (II) 902 at the Near-RT RIC 132. -503: The PCI Conflict Optimization Module (II) 902 module keeps gettingparameters such as number of active UEs, cell load (in terms of PRB utilization) and other system KPIs for each cell from the gNB-DU 305 via the E2 interface. It can get this information periodically or based on specific event triggers. -504: The PCI Conflict Optimization Module (II) 902 module keeps gettingparameters such as number of connected UEs, number of emergency call UEs, UE measurement reports, handover related counters and other system KPIs for each cell from the gNB-CU-CP 304a and other system KPIs via the E2 interface. It can get this information periodically or based on specific event triggers. -505: The PCI Conflict Optimization Module (II) 902 analyzes the informationavailable with it and selects the cells for which PCIs should be changed. It also selects a shutdown timer for each cell for which PCI should be changed, and PCI change can be applied after the expiry of this shutdown timer. -506: The PCI Conflict Optimization Module (II) 902 provides aboveinformation (i.e. selected cells for which PCIs need to be changed and the shutdown timers) to the corresponding gNB-DUs 305. New PCI to be appliedto these cells is also communicated to the corresponding gNB-DUs 305. -507: gNB-DU 305 communicates to CMS 140 to shut down each such cell andprovides the corresponding values of the shutdown timers. -508: CMS 140 triggers shut down for each such cell (as per the instructioncommunicated from the PCI Conflict Optimization Module (I) 901 from Non- RT-RIC). CMS 140 sends a message to the corresponding gNB-DU 305 for this purpose and provides value of shutdown timer too. -509: gNB-DU 305 starts shutting down that cell and performs actions such as1) Sets CellBarred to Barred state in the MIB for that cell, 2) gNB-DU 305 stops accepting new sessions for that cell, and 3) shuts down the cell at the expiry of the shutdown timer for that cell. -510: gNB-DU 305 informs CMS 140 when cell shut down is completed.- 511: CMS 140 triggers cell lock for that cell.- 512: gNB-DU 305 indicated to CMS 140 that cell lock is completed.- 513: CMS 140 can boot-up that cell now and apply the new PCI which it hadreceived for that cell from the PCI Conflict Optimization Module (I) 901.

[0086] Some of the examples in this method have been given for the scenarios whenthe PCI conflict can be resolved by changing PCI for one of the two (candidate) cells. This method is also applicable for the scenario where more than two neighboring cells of a cell can have a same PCI.

[0087] For example, consider a scenario where cell c1 (with PCI p1) has fourneighboring cells indicated as cell c2 (with PCI p2), cell c3 (with PCI p2), c4 (with PCI p4) and cell c5 (with PCI p5). In this case, cell c2, c3 and c5 have the same PCI (as each has PCI p2). With above method, the PCI Conflict Optimization Module (I) 901 at the Non-RT-RIC detects that the cells c2, c3 and c5 are neighboring cells of the cell c1 and these cells (i.e. c2,c3 and c5) have the same PCI (i.e. PCI p2). The PCI Conflict Optimization Module (II) 902 at the Non-RT-RIC analyzes the topology (and various measured parameters) and finds that the PCI Conflict can be resolved by changing PCIs of two out of these three cells. It provides identity of these cells (c2, c3 and c5) and two new PCIs ( p6 and p7) which will not cause PCI conflict in that cluster of cells and provides these to the Near-RT RIC 132. The PCI Conflict Optimization Module (II) 902 at the Near-RT RIC 132 evaluates relevant parameters (such as current and predicted cell load) for these three cells and decides about the cells to which new PCI should be applied and also estimates good values of shutdown timers for this purpose (for the two selected cells). For example, in this case, this method can decide to change PCI of cell c2 and c5 to avoid PCI Conflict. FIG.12 and FIG.13 show this method for the case where PCI of one of the cells is being changed. For the case, when this method decides to change PCI of more than one cell, the PCI Conflict Optimization Module (II) 902 communicates identity of these chosen cells (along with their shutdown timers) to the PCI Conflict Optimization Module (I) 901 at the Non-RT-RIC, and the Non-RT- RIC communicates these to the CMS 140 for the chosen cells in FIG.12. Similarly, the PCI Conflict Optimization Module (II) 902 at the Near-RT RIC 132 communicates identity of multiple cells (along with their shutdown timers) to the respective DUs.

[0088] With Method I above, the PCI Optimization Module (I) 901 and PCIOptimization Module (II) 902 can resolve multiple PCI Conflicts in an optimal manner.

[0089] IMPLEMENTATION II

[0090] In another method, a PCI Optimization module 903 is configured to helpselect the cell whose PCI should be changed (and the corresponding shutdown timer) to resolve PCI Conflicts in an optimal manner, is placed at the gNB-CU-CP 304a and this method can work without support for RIC in the system. As discussed earlier, each gNB-CU can control a large number of cells in a geographical area and inter-CU communication can take place via CU-CU interface (such as Xn in 5G networks). As shown in FIG.14, an AI / ML model 904 to predict certain parameters (for example, number of connected users, number of active users and PRB utilization per cell) is hosted at gNB-CU-CP 304b. Actual AI / ML training for this purpose can happen at gNB-CU-UP 304b itself or at another entity (such asan external analytics server).

[0091] F1 Application layer Protocol (F1AP) running over the F1-C interfacebetween DU and gNB-CU-CP 304a is used to communicate PRB utilization, number of active users and other system KPIs (such as KPIs related to throughput, block error rate, latency, and the like) for each cell from gNB-DU 305 to gNB-CU-CP 304a.

[0092] E1 Application layer Protocol (E1AP) running between gNB-CU-CP 304a andgNB-CU-CP 304b is enhanced to communicate PRB utilization, number of active users, number of connected users and number of emergency call users and other system KPIs from gNB-CU-CP 304a to gNB-CU-CP 304b.

[0093] The gNB-CU-CP 304b uses LSTM or Transformer based deep neural networkmodels to predict PRB utilization, number of connected users and number of active users for a given cell. E1AP is also enhanced to communicate predicted PRB utilization, predicted number of active users and predicted number of connected users for each cell from gNB- CU-CP 304b to gNB-CU-CP 304a.

[0094] The PCI Optimization module 903 which takes the final decision to resolvePCI Conflicts is hosted at the gNB-CU-CP 304a in this method and it gets access to following parameters for each cell: -Number of connected users in a cell (available at gNB-CU-CP 304a)- Number of emergency calls in a cell (available at gNB-CU-CP 304a)- Measurements reports for UEs from that cell, such as RSRP, RSRQ (availableat gNB-CU-CP 304a) -Handover related counters and metrics (available at gNB-CU-CP 304a)- Number of active users in a cell (received from gNB-DU 305 via F1AP)- PRB utilization of that cell (received from gNB-DU 305 via F1AP)- Other system KPIs (from gNB-DU 305 and gNB-CU)- Predicted number of connected users (received from gNB-CU-CP 304b viaE1AP) -Predicted number of active users (received from gNB-CU-CP 304b via E1AP)- Predicted PRB utilization of that cell (received from gNB-CU-CP 304b viaE1AP) -PCI for each cell- As CU-CP 304a hosts Neighbor Relation Table for its cluster of cells, it alsohas information about the list of neighbors of each cell along with their PCIs

[0095] The PCI optimization module 903 (running at gNB-CU-CP 304a) analyzesabove information and detects PCI Conflicts, if any. If any PCI Conflict is detected, it selects a new PCI to be assigned to one of the cells that are involved in the PCI Conflict and selects the cell to which this new PCI should be assigned to resolve PCI Conflict in an optimal manner. It also decides the value of shutdown timer when that cell should be shut down to change the PCI.

[0096] In the above method, the AI / ML Inference (or Prediction) model 904 can bedirectly downloaded at the gNB-CU-CP 304a (instead of gNB-CU-CP 304b) if feasible for a given hardware and software system. In that case, prediction of parameters (such as number of connected users, number of active users and PRB utilization in each cell) can be directly done at the gNB-CU-CP 304a, and the AI / ML model 904 can be downloaded at the gNB-CU-CP 304a as shown in FIG.15. Alternatively, AI / ML training can be done as part of Non-RT-RIC as shown in FIG.16 and the AI / ML Inference model 904 can be downloaded from Non-RT-RIC to gNB-CU-CP 304a.

[0097] For the case where there are multiple gNodeBs in a network, CU-CPs 304a inthat cluster can exchange relevant parameters (as discussed above) via XnAP and one of the CU-CP 304a can take a decision to resolve PCI Conflict in an optimal manner. One CU-CP304a can act as Master CU-CP 304a for this purpose. Note that there is one CU-CP 304a per gNodeB.

[0098] XnAP (Xn Application layer Protocol) over the Xn interface, supports DataCollection procedure between NG-RAN nodes (such as gNodeBs) as shown in FIG.17. This procedure is used by an NG-RAN node to request AI / ML information from another NG-RAN node. XnAP is enhanced to exchange PCI optimization related parameters between CU-CPs 304a. The Master CU-CP 304a chooses the cell for which PCI should be changed and informs the corresponding gNB-CU-CP 304b using the XnAP protocol.

[0099] Once the PCI Optimization Module 903 at the CU-CP 304b decides the cell forwhich PCI needs to be changed, it informs this to the CMS 140, along with the new PCI for that cell and the value of the shutdown timer (on expiry of which the new PCI can be applied). Some of these steps are shown in FIG.18 and given below: -601: gNB-CU-CP 304a provides identity of the cell to be shut down along withthe value of the shutdown timer (and the new PCI that is to be applied to that cell after the expiry of the shutdown timer) to the CMS 140. It can be a direct message or can be sent in the form of an event (notification). -602: CMS 140 triggers shut down for this cell by sending a message to thecorresponding gNB-DU 305 for this purpose and provides value of shutdown timer too. -603: gNB-DU 305 starts shutting down that cell and performs actions such as1) Sets CellBarred to Barred state in the MIB for that cell, 2) gNB-DU 305 stops accepting new sessions for that cell, and 3) shuts down the cell at the expiry of the shutdown timer for that cell. -604: gNB-DU 305 informs CMS 140 when cell shut down is completed.- 605: CMS 140 triggers cell lock for that cell.- 606: gNB-DU 305 indicated to CMS 140 that cell lock is completed.- 607: CMS 140 can boot-up that cell now and apply the new PCI, which it hadreceived for that cell from the PCI Conflict Optimization Module (I) 901.

[0100] Some of the examples in this method have been given for the scenario wherea PCI Conflict can be resolved by changing PCI of one of the two (candidate) cells. As with the previous method, this method is also applicable for the scenario where more than two neighboring cells of a cell can have same PCI. In such scenarios, the PCI Conflict Optimization Module (running at CU-CP 304a in FIG.14 and FIG.15) identifies multiple cells (and good values of shutdown timers for these cells) for which PCI needs to change to resolve PCI Conflicts. Identity of these cells (along with their shutdown timers to apply new PCIs) is communicated from CU-CP 304a to CMS 140 and the CMS 140 provides this to the corresponding DUs for these cells.

[0101] It will be understood that implementations and embodiments can beimplemented by computer program instructions. These program instructions can be provided to a processor to produce a machine, such that the instructions, which execute on the processor, create means for implementing the actions specified herein. The computer program instructions can be executed by a processor to cause a series of operational steps to be performed by the processor to produce a computer-implemented process such that the instructions, which execute on the processor to provide steps for implementing the actions specified. Moreover, some of the steps can also be performed across more than one processor, such as might arise in a multi-processor computer system or even a group of multiple computer systems. In addition, one or more blocks or combinations of blocks in the flowchart illustration can also be performed concurrently with other blocks or combinations of blocks, or even in a different sequence than illustrated without departing from the scope or spirit of the disclosure.

Claims

CLAIMS 1. A method comprising: detecting a Physical Cell Identity (PCI) Conflict at a non-real-time (Non-RT) RadioIntelligent Controller (RIC); providing, by the Non-RT RIC, parameters and policies to a Near-RT RIC with an identity of cells involved in the PCI Conflict along with a new PCI to be applied to one of the identifed cells; at the Near-RT RIC: analyzing cell-level information and KPIs received from an E2 node and determining a candidate cell of the idenified cells whose PCI value should be changed; and determining an optimal value of a shutdown timer when the candidate cell can be shut down and the new PCI can be applied during a boot-up procedure.

2. The method of claim 1, further comprising: providing a PCI Optimization Module further comprising: 1) a PCI Conflict Optimization Detection Module to detect the PCI conflicts and configured to apply the policy at the Non-RT-RIC and 2) a PCI Conflict Optimization Resolution Module at the Near- RT-RIC configured to execute the analyzing and determining steps.

3. An apparatus comprisng: a Physical Cell Identity (PCI) Optimization Module configued to select a cell whose PCI should be changed and a corresponding shutdown timer for changing the PCI for the selected cell to resolve PCI Conflicts in an optimal manner 4. The apparatus of claim 3, wherein the PCI Optimization Module further comprises: 1) a PCI Conflict Optimization Detection Module to detect PCI conflicts and configured toapply a policy at a a non-real-time (Non-RT) Radio Intelligent Controller (RIC) or 2) a PCIConflict Optimization Resolution Module to help resolve PCI Conflicts in an optimal manner at a Near-RT-RIC, or both.

5. The apparatus of claim 4, wherein the 1) PCI Conflict Optimization Detection Module is at the Non-RT RIC, and is configured to collect information from a gNodeB including: a PCI of each cell, and a list of neighbors of each cell along with neighbor PCIs from a Neighbor Relation Table (NRT).

6. The appartatus of claim 5, wherein the NRT information is collected from a gNodeB(gNB)- Centralized Unit – Control Plane (CU-CP).

7. The apparatus of claim 5, wherein the 1) PCI Conflict Optimization Detection Module is configured to collect a PCI List of available PCIs for each gNodeB.

8. The apparatus of claim 7, wherein the 1) PCI Conflict Optimization Detection Module is configured to: analyze the collected information and detect a PCI Conflict; select a new PCI to be assigned from the PCI List a cell involved in the PCI Conflict; and send an identity of the cell involved in PCI Conflict and the new PCI to be assigned to the cell to the Near-RT RIC.

9. The apparatus of claim 3, wherein the 2) PCI Conflict Optimization Resolution Module at the Near-RT-RIC is configured to periodically collect information for each cell from the gNB-CU-CP comprising: a number of connected users in a cell; a number of emergency (VoNR) calls in a cell;UE measurement reports for from the cell; and handover related counters and metrics.

10. The apparatus of claim 9, wherein the 2) PCI Conflict Optimization Resolution Module at the Near-RT-RIC is configured to periodically collect information for each cell from the gNB-Distributed Unit (DU) comprising: a number of active users in a cell; and a PRB utilization of the cell.

11. The apparatus of claim 3, wherein the 2) PCI Conflict Optimization Resolution Module at the Near-RT-RIC is configured to predict information comprising: a predicted number of connected users; a predicted number of active users; and a predicted cell load.

12. The apparatus of claim 11, wherein the 2) PCI Conflict Optimization Resolution Module at the Near-RT-RIC is configured to use the predicted information to select a cell for PCI change between the two cells involved in a PCI Conflict and decide a value of the timer when that cell is to be shut down to change the PCI.

13. The apparatus of claim 3, wherein the PCI Optimization Module is at the gNB CU-CP andcomprises an Artificial Intelligence / Machine Learning (AI / ML) model trained to predictcell parameters.

14. The appratus of claim 12, wherein the AI / Ml comprises: LSTM or Transformer based deep neural network models to predict PRB utilization, a number of connected users and a number of active users for a given cell.

15. A method comprising:detecting, by a Physical Cell Identity (PCI) Optimization Module at a gNodeB (gNB)- Centralized Unit – Control Plane (CU-CP), a PCI Conflict; and if the PCI Conflict is detected, it selecting a new PCI to be assigned to one of a plurality of cells that are involved in the PCI Conflict and selecting the cell to which this new PCI is to be assigned to resolve PCI Conflict in an optimal manner.

Images