WO2022015008A1 - Procédé et système pour déterminer une cellule cible pour le transfert intercellulaire d'eu - Google Patents

Procédé et système pour déterminer une cellule cible pour le transfert intercellulaire d'eu Download PDF

Info

Publication number
WO2022015008A1
WO2022015008A1 PCT/KR2021/008948 KR2021008948W WO2022015008A1 WO 2022015008 A1 WO2022015008 A1 WO 2022015008A1 KR 2021008948 W KR2021008948 W KR 2021008948W WO 2022015008 A1 WO2022015008 A1 WO 2022015008A1
Authority
WO
WIPO (PCT)
Prior art keywords
ues
location information
handover
location
mobility management
Prior art date
Application number
PCT/KR2021/008948
Other languages
English (en)
Inventor
Satya Kumar VANKAYALA
Ashvin Kaithara JOSEPH
Seungil Yoon
Pankaj Bhimrao THORAT
Original Assignee
Samsung Electronics Co., Ltd.
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Samsung Electronics Co., Ltd. filed Critical Samsung Electronics Co., Ltd.
Publication of WO2022015008A1 publication Critical patent/WO2022015008A1/fr

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/084—Backpropagation, e.g. using gradient descent
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W36/00—Hand-off or reselection arrangements
    • H04W36/0005—Control or signalling for completing the hand-off
    • H04W36/0083—Determination of parameters used for hand-off, e.g. generation or modification of neighbour cell lists
    • H04W36/00835—Determination of neighbour cell lists
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W36/00—Hand-off or reselection arrangements
    • H04W36/24—Reselection being triggered by specific parameters
    • H04W36/32—Reselection being triggered by specific parameters by location or mobility data, e.g. speed data
    • H04W36/322—Reselection being triggered by specific parameters by location or mobility data, e.g. speed data by location data
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/044—Recurrent networks, e.g. Hopfield networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/048—Activation functions

Definitions

  • the present disclosure relates to wireless communication system, and more specifically related to a method and a system for determining target cell for handover of a user equipment (UE).
  • UE user equipment
  • a user equipment can be under coverage of different technologies such as 2 nd generation (2G), 3 rd generation (3G), 4 th generation (4G), 5 th generation (5G), Wireless-fidelity (Wi-Fi), etc. at the same time.
  • 2G 2 nd generation
  • 3G 3 rd generation
  • 4G 4 th generation
  • 5G 5 th generation
  • Wi-Fi Wireless-fidelity
  • handover are performed based on Reference Signal Received Power (RSRP) threshold and hysteresis values.
  • RSRP Reference Signal Received Power
  • the UE can report up to a maximum of 8 cells and also reports only same frequency Ccells unless network provide sufficient gap for measuring inter frequency or Inter radio access technology (RAT) cells which effects user Experience.
  • RAT Inter radio access technology
  • dedicated preamble is assigned and network normally choose one among the reported cells from UE for the handover as gNB may not be aware of full network topology. If the network decides to choose any other cells for the handoff, then the UE measurements are required for those cells which create extra latency during time critical handover.
  • the network may not be able to choose a best cell for handover even if the best cell is available and suits requirement of the UE based on mobility (fast/slow), applications (Ultra Reliable Low Latency Communication (URLLC), Machine Type Communication (MTC), Massive Broad band, Voice), coverage, capability of the target cells, etc. in case the UE failed to report about the best cell.
  • the UE may be handed over to a wrong cell, subjected to early handover, too late handover, coverage Hole, etc. which effects user experience during the mobility scenario.
  • the principal objective of the embodiments herein is to provide a method and system for determining a target cell for handover of a UE using AI and ML based on a correlation between multiple UEs.
  • the correlation between the multiple UEs is derived using multiple network characteristics associated and location information associated with the UEs.
  • Another objective of the embodiment herein is to enable base stations to predict the target cell for the UE handover based on federated learning without having to receive a measurement configuration or a measurement report from the UE.
  • embodiments herein disclose a method for determining a target cell for handover of a User Equipment (UE) in a wireless network.
  • the method includes monitoring, by a mobility management platform of a Base Station (BS) in the wireless network, multiple network characteristics associated with multiple UEs in the wireless network.
  • the multiple UEs are located in multiple locations.
  • the method includes determining, by the mobility management platform, a correlation between the multiple UEs based on the multiple network characteristics associated with the multiple UEs and location information associated with each of the UEs.
  • BS Base Station
  • the method includes receiving, by the mobility management platform, location information from a UE of the multiple UEs and determining, by the mobility management platform, a measurement report corresponding to the location information received from the UE based on the correlation between the UEs.
  • the method also includes determining, by the mobility management platform, the target cell for the handover of the UE based on the location information received from the UE and the measurement report corresponding to the location information received from the UE.
  • the plurality of network characteristics associated with the plurality of UEs includes the location of the UE Location based on position reference signals (PRS) sent by the BS and a BS topology, UE measurements of neighboring cells and load information associated with the neighboring cells of the UE, a UE capability, a UE index, a Quality of Service (QoS) of the UE, a QoS Class Indicator (QCI) of the UE, operating frequency of a plurality of BSs, multi-in multi-out (MIMO) information associated with the UE, MIMO information associated with the BS, Modulation and coding scheme (MCS) information associated with the UE, resource blocks (RB) information associated with the UE, a queue information associated with the UE, mobility of the UE, a Doppler frequency, traffic information, load of a cell, route map information, signal map information, transmission mode, reception mode, a map indicating a number of Acknowledgments (ACKs) received from the UE, a
  • the LPP information is used by the mobility management platform for at least one of scheduling resources for downlink (DL), scheduling resources for uplink (UL), allocation of resource for handover from a neighboring cell, reserving bearer resources for high priority UEs.
  • DL downlink
  • UL uplink
  • resource for handover from a neighboring cell reserving bearer resources for high priority UEs.
  • learning by the mobility management platform, a pattern of at least one of RSRP, Reference Signal Receive Quality (RSRQ), channel quality indicator (CQI), channel state information (CSI) based on the location of the plurality of UEs over a specific period of time; and performing, by the mobility management platform, one of recommending the handover of the UE to the target cell based on the learning, predicting a handover threshold of the UE at a specific time based on the plurality of network characteristics associated with the plurality of UEs and reserving data resources and control resources for the handover threshold of the UE.
  • the handover threshold of the UE indicates a required number of handovers.
  • the mobility management platform recommends the handover based on a federated learning across the plurality of UEs without receiving one of a measurement configuration and the measurement report from the UE.
  • determining, by the mobility management platform, the correlation between the plurality of UEs based on the plurality of network characteristics associated with the plurality of UEs and location information associated with the plurality of UEs comprises applying one of a Machine Learning (ML) model and an Artificial intelligence (AI) model on the plurality of network characteristics associated with the plurality of UEs and the location information associated with the plurality of UEs.
  • ML Machine Learning
  • AI Artificial intelligence
  • the ML model corresponds to a Neural Network (NN) model
  • the NN model is trained by inputting the plurality of network characteristics associated with the plurality of UEs and the location information associated with the plurality of UEs to a plurality of input NN nodes of the NN model and determining an optimal weight of each input NN node based on a training method.
  • the method includes training the NN model based on the optimal weight of each input NN node of the plurality of input NN nodes, the plurality of network characteristics associated with the plurality of UEs and the location information associated with the plurality of UEs.
  • the mobility management platform is hosted on one of an O-RAN based RIC application using standard E2 Interface and a Proprietary Hardware.
  • determining, by the mobility management platform, the correlation between the plurality of UEs based on the plurality of network characteristics associated with the plurality of UEs and location information associated with the plurality of UEs includes generating, by the mobility management platform, a handover map indicating a coverage area associated with the BS using location information received from a plurality of UEs based a plurality of mapping parameters; wherein the plurality of mapping parameters are a function of function of time, BS IDs, topology of the heterogeneous network, operating frequency, UE/BS category/capability, bandwidth paths/BW, throughput requirements, QCI/QoS type, Number of BS antennas, external climate conditions, Doppler statistics, Channel statistics, mobility statistics, delay spread statistics, MIMO/non-MIMO capabilities, transmission type, time of the day, ACK/NACK statistics, SINR statistics as a function of the UE category and above mentioned parameters, transmit power of the BS and/or UE, ACK/NACK as a function of M
  • the antenna ports, number of antennas, position of the reference resources are selected dynamically based on at least one of the QOS/QCI, a UE category, a UE type.
  • the location based on the digital antenna array provides location information as a function of SNR, the number of antennas, number of symbols, repetitions, number of sources arriving at a receiver.
  • the location based on the antenna array provides location information as a function of pre-defined beam pattern, receiver SINR, channel environment, number of symbols.
  • the location is determined based on the hybrid beam formation one of an anolog formation technique and a digital beam formation technique is dynamically adopted.
  • a BS for determining a target cell for handover of a UE in a wireless network.
  • the BS includes a memory, a processor and a mobility management platform.
  • the mobility management platform is configured to monitor a plurality of network characteristics associated with a plurality of UEs in the wireless network and determine a correlation between the plurality of UEs based on the plurality of network characteristics associated with the plurality of UEs and location information associated with each of the plurality of UEs.
  • the plurality of UEs is located in a plurality of locations.
  • the mobility management platform is configured to receive a location information from at least one UE of the plurality of UEs and determine at least one measurement report corresponding to the location information received from the at least one UE based on the correlation between the plurality of UEs; and determine the target cell for the handover of the at least one UE based on the location information received from the at least one UE and the at least one measurement report corresponding to the location information received from the UE.
  • inventions herein disclose a UE for determining a target cell for handover of the UE in a wireless network.
  • the UE includes a memory, a processor, a communicator and a mobility management controller.
  • the mobility management controller is configured to send location information of the UE to the wireless network and receive handover command from the wireless network to handover the UE to the target cell.
  • the target cell is determined by the wireless network based on the location information sent by the UE and at least one measurement report corresponding to the location information sent by the UE.
  • According to various embodiments may provide a method and system for determining a target cell for handover of a UE using AI and ML based on a correlation between multiple UEs.
  • the correlation between the multiple UEs is derived using multiple network characteristics associated and location information associated with the UEs.
  • FIG. 1 illustrates a heterogeneous network comprising a user equipment (UE) which is under coverage of different technology at a given instant of time, according to a prior art;
  • UE user equipment
  • FIG. 2A is a block diagram of a BS for determining target cell for handover of the UE using ML and AI, according to an embodiment as disclosed herein;
  • FIG. 2B is the UE for determining the target cell for handover in the wireless network, according to an embodiment as disclosed herein;
  • FIG. 3 is a flow chart illustrating a method for determining the target cell for the handover of the UE using the ML and the AI, according to an embodiment as disclosed herein;
  • FIG. 4 illustrates an architecture diagram of Open RAN (ORAN) for determining the target cell for the handover of the UE , according to an embodiment as disclosed herein;
  • ORAN Open RAN
  • FIG. 5A illustrates an architecture diagram of a cloud radio access network (CRAN) or virtual RAN for determining the target cell for the handover of the UE, according to an embodiment as disclosed herein;
  • CRAN cloud radio access network
  • FIG. 5B illustrates various split option for the CRAN or the virtual RAN for determining the target cell for the handover of the UE, according to an embodiment as disclosed herein;
  • FIG. 6 illustrates a heterogeneous network where mobility management platform for determining the target cell for the handover of the UE is hosted on a Proprietary Hardware, according to an embodiment as disclosed herein;
  • FIG. 7A is a signalling diagram illustrating a scenario of handover performed by a gNB during a transition of the UE from an idle state to an active state, according to the prior art
  • FIG. 7B is a signalling diagram illustrating the scenario of the handover performed by the gNB during the transition of the UE from the idle state to the active state, according to an embodiment as disclosed herein;
  • FIG. 8A is a signalling diagram illustrating a scenario of the handover of the UE performed by the gNB for load balancing based on a cell load, according to the prior art
  • FIG. 8B is a signalling diagram illustrating a scenario of the handover of the UE performed by the gNB for the load balancing based on the cell load, according to an embodiment as disclosed herein;
  • FIG. 9A is a signalling diagram illustrating a scenario of configuration of static measurement objects in the UE in a connected mode, according to the prior art
  • FIG. 9B is a signalling diagram illustrating a scenario of configuration of dynamic measurement objects in the UE in the connected mode, according to an embodiment as disclosed herein;
  • FIG. 10 is a signalling diagram illustrating a call flow in the UE during mobility without a measurement report, according to an embodiment as disclosed herein;
  • FIG. 11A illustrates various inputs received by a ML model of mobility management platform for determining the target cell for the handover of the UE, according to an embodiment as disclosed herein;
  • FIG. 11B illustrates a neural network used in the ML model of the mobility management platform, according to an embodiment as disclosed herein;
  • FIG. 12 illustrates an example scenario of using an AI model in the mobility management platform of the BS, according to an embodiment as disclosed herein;
  • FIG. 13 is an example handover map indicating a coverage area associated with the BS, according to an embodiment as disclosed herein.
  • circuits may, for example, be embodied in one or more semiconductor chips, or on substrate supports such as printed circuit boards and the like.
  • circuits constituting a block may be implemented by dedicated hardware, or by a processor (e.g., one or more programmed microprocessors and associated circuitry), or by a combination of dedicated hardware to perform some functions of the block and a processor to perform other functions of the block.
  • a processor e.g., one or more programmed microprocessors and associated circuitry
  • Each block of the embodiments may be physically separated into two or more interacting and discrete blocks without departing from the scope of the invention.
  • the blocks of the embodiments may be physically combined into more complex blocks without departing from the scope of the invention
  • embodiments herein disclose a method for determining a target cell for handover of a User Equipment (UE) in a wireless network.
  • the method includes monitoring, by a mobility management platform of a Base Station (BS) in the wireless network, multiple network characteristics associated with multiple UEs in the wireless network.
  • the multiple UEs are located in multiple locations.
  • the method includes determining, by the mobility management platform, a correlation between the multiple UEs based on the multiple network characteristics associated with the multiple UEs and location information associated with each of the UEs.
  • BS Base Station
  • the method includes receiving, by the mobility management platform, location information from a UE of the multiple UEs and determining, by the mobility management platform, a measurement report corresponding to the location information received from the UE based on the correlation between the UEs.
  • the method also includes determining, by the mobility management platform, the target cell for the handover of the UE based on the location information received from the UE and the measurement report corresponding to the location information received from the UE.
  • a BS for determining a target cell for handover of a UE in a wireless network.
  • the BS includes a memory, a processor and a mobility management platform.
  • the mobility management platform is configured to monitor a plurality of network characteristics associated with a plurality of UEs in the wireless network and determine a correlation between the plurality of UEs based on the plurality of network characteristics associated with the plurality of UEs and location information associated with each of the plurality of UEs.
  • the plurality of UEs is located in a plurality of locations.
  • the mobility management platform is configured to receive a location information from at least one UE of the plurality of UEs and determine at least one measurement report corresponding to the location information received from the at least one UE based on the correlation between the plurality of UEs; and determine the target cell for the handover of the at least one UE based on the location information received from the at least one UE and the at least one measurement report corresponding to the location information received from the UE.
  • inventions herein disclose a UE for determining a target cell for handover of the UE in a wireless network.
  • the UE includes a memory, a processor, a communicator and a mobility management controller.
  • the mobility management controller is configured to send location information of the UE to the wireless network and receive handover command from the wireless network to handover the UE to the target cell.
  • the target cell is determined by the wireless network based on the location information sent by the UE and at least one measurement report corresponding to the location information sent by the UE.
  • the BS needs to handover the UE to a new cell the BS needs the UE to send the measurement report and decides the handover based on the RSRP.
  • the BS may not handover the UE to the best cell as the BS is not aware of complete network topology.
  • the proposed method includes the BS generating a correlation using ML/AI between the multiple UEs and the location of the UEs using the network characteristics associated with the UEs. Therefore, the BS is aware of the complete network topology and also the use of ML/AI helps in determini9ng the best cell for UE handoff.
  • the proposed method can perform the handover even without the requirement of the measurement report from the UE based on past learning. Also, the proposed method allows sharing of learning between multiple BS as a result the best cell is chosen for each UE and each time thereby enhancing the efficiency of the entire system.
  • the UE power efficiency can be improved and latency in the network reduces.
  • the proposed method minimizes the handover procedure and reduces signaling load with enhanced user experience by hand off to the suitable cell based on the mobility of the user [5G cells have lesser coverage that other technology, hence frequent handover is possible for fast moving user].
  • the proposed method can be implemented in distributed unit or centralized unit or cloud intelligence unit or RRH or Radio unit. Further, the proposed method can be used in a communication system.
  • the communication system can be, for example, but not limited to a 4G system, 5G system, 6G system, Wi-Fi system, and LAA system.
  • the proposed method can be used for various NN/ML/RL architecture based on various or sub-set of network parameters.
  • the BS can maintain a NN/ML/AI on per UE basis or a single NN/ML/AI for multiple UEs.
  • the system can also use sparse DNN, CNN, RNN, Sparse CNN, Sparse RNN, Sparse DNN and hybrid architectures.
  • the system can also use various activation functions, approximation of activation functions and/or linear approximation of activation functions. For example, one can use MDP based algorithms such as modified value iteration or policy iteration. The performance will be function of the activation function. One can approximate the activated functions to reduce the computational complexity.
  • the system can also intelligently remove the connections in the NN network if the weight of the link is negligible and again the system can retrain the NN to achieve the expected performance. In case, if performance does not meet the requirement, the system will go back earlier NN. Further, the system can use ML/AI to improve the performance of these algorithms. The person ordinary skill in the area can easily do slight modifications to the proposed solutions. These techniques can run in ML module or can run in hardware (HW).
  • FIGS. 2-13 where similar reference characters denote corresponding features consistently throughout the figures, there are shown preferred embodiments.
  • FIG. 2A is a block diagram of a BS (100) for determining target cell for handover of a UE (200) using ML and AI, according to an embodiment as disclosed herein.
  • the BS (100) comprises a memory (120), a processor (140), a transceiver (160), and a mobility management platform (180). Further, two or more components may be embodied in one single component, and/or one component may be configured using multiple sub-components to achieve the desired functionalities. Some components of the BS (100) may be configured using hardware elements, firmware elements and/or a combination thereof.
  • the memory (120) stores instructions to be executed by the processor (140).
  • the memory (120) may include non-volatile storage elements. Examples of such non-volatile storage elements may include magnetic hard discs, optical discs, floppy discs, flash memories, or forms of electrically programmable memories (EPROM) or electrically erasable and programmable (EEPROM) memories.
  • the memory (120) may, in some examples, be considered a non-transitory storage medium.
  • the term “non-transitory” may indicate that the storage medium is not embodied in a carrier wave or a propagated signal. However, the term “non-transitory” should not be interpreted that the memory (120) is non-movable.
  • a non-transitory storage medium may store data that can, over time, change (e.g., in Random Access Memory (RAM) or cache).
  • RAM Random Access Memory
  • the processor (140) is coupled with the memory (120), the transceiver (160) and the mobility management platform (180).
  • the processor (140) may include one or a plurality of processors.
  • the one or the plurality of processors may be a general-purpose processor, such as a central processing unit (CPU), an application processor (AP), or the like, a graphics-only processing unit such as a graphics processing unit (GPU), a visual processing unit (VPU), and/or an AI-dedicated processor such as a neural processing unit (NPU).
  • the processor (140) may include multiple cores and is configured to execute the instructions stored in the memory (120).
  • the mobility management platform (180) includes a UE characteristics management controller (182), a ML model (184), a handover management controller (186) and an AI model (188). According to various embodiments of the disclosure, at least part of elements, functions and/or operations of the mobility management platform (180) may be included and/or performed by the processor (140).
  • the UE characteristics management controller (182) is configured to receive and monitor multiple network characteristics associated with multiple UEs (200a-N) located in multiple locations in the wireless network (1000).
  • the multiple network characteristics associated with the multiple UEs (200a-N) may include but are not limited to a location of the UE Location based on position reference signals (PRS) sent by the BS (100) and a BS topology, UE measurements of neighboring cells and load information associated with the neighboring cells of the UE (200), a UE capability, a UE index, a Quality of Service (QoS) of the UE (200), a QoS Class Indicator (QCI) of the UE (200), operating frequency of a plurality of BSs (100a-N), multi-in multi-out (MIMO) information associated with the UE (200), MIMO information associated with the BS (100), Modulation and coding scheme (MCS) information associated with the UE (200), resource blocks (RB) information associated with the UE (200), a
  • the antenna ports, number of antennas, position of the reference resources are selected dynamically based on at least one of the QOS/QCI, a UE category, a UE type.
  • the location based on the digital antenna array provides location information as a function of SNR, the number of antennas, number of symbols, repetitions, number of sources arriving at a receiver.
  • the location based on the antenna array provides location information as a function of pre-defined beam pattern, receiver SINR, channel environment, number of symbols.
  • the location is determined based on the hybrid beam formation one of an anolog formation technique and a digital beam formation technique is dynamically adopted.
  • the ML model (184) is a neural network (NN) model.
  • the ML model (184) is configured to determine a correlation between the multiple UEs (200a-N) based on the multiple network characteristics associated with the multiple (200a-N) and location information associated with each of the plurality of UEs (200a-N). Further, the ML model (184) receives the location information from the UE (200) and fetches a measurement report corresponding to the UE (200) based on the correlation determined. Then, the ML model (184) determines the target cell for the handover of the UE (200) based on the location information and the measurement report corresponding to the location information received from the UE (200).
  • the ML model (184) is also configured to learn a pattern of one of RSRP, RSRQ, CQI, CSI based on the location of the multiple UEs (200a-N) over a specific period of time.
  • the learning is then utilized to recommend the handover of the UE (200) to the target cell based on the learning and also to predict a handover threshold of the UE (200) at a specific time based on the multiple network characteristics associated with the multiple UEs (200a-N). Therefore, the prediction of the handover helps the BS (100) to reserve data resources and control resources for the handover threshold of the UE (200).
  • the handover threshold of the UE (200) indicates a required number of handovers.
  • the ML model (184) is also configured to generate a handover map indicating a coverage area associated with the BS (100) using location information received from the multiple UEs (200a-N) based a plurality of mapping parameters.
  • the plurality of mapping parameters are a function of function of time, BS IDs, topology of the heterogeneous network, operating frequency, UE/BS category/capability, bandwidth paths/BW, throughput requirements, QCI/QoS type, Number of BS antennas, external climate conditions, Doppler statistics, Channel statistics, mobility statistics, delay spread statistics, MIMO/non-MIMO capabilities, transmission type, time of the day, ACK/NACK statistics, SINR statistics as a function of the UE category and above mentioned parameters, transmit power of the BS (100) and/or UE (200), ACK/NACK as a function of MCS and other mentioned parameters, shadowing, fading stats, distribution of a plurality of parameters, long term statistics and short term statistics of all the parameters.
  • the handover management controller (186) is configured to perform the handover by sending appropriate commands to the UE (200) based on the output of the ML model (184).
  • the appropriate commands includes RRC Reconfiguration Complete message.
  • the AI model (188) is used as a substitute in place of the ML model (184) for determining the correlation between the multiple UEs and to take the decision of the handover of the UE based on multiple network parameters associated with the multiple UEs (200a-N). The functions are further explained in FIG. 12.
  • the mobility management platform (180) is hosted on one of an O-RAN based RIC application using standard E2 Interface and a Proprietary Hardware which is explained in detail in FIG. 4 to FIG. 6. Therefore, the mobility management platform (180) enables estimating the appropriate target cell for the handover using the ML model (184)/AI model (188), which is in the RIC module (418) of the O-RAN systems based on the mentioned parameters, using the estimated appropriate BS ID value(s), the MAC layer will allocate the resources to the UE(s). There will be BS (100) guided RACH procedure for the new BSs.
  • the mobility management platform (180) conveys estimated handover values to a central cloud based MAC layer using E2 interface. Also, the proposed method reduces the request for measurement frequency after training phase and handover frequency after training phase. Maintaining the NN model on per UE(s) basis. Learning can be function of QCI/QoS and other parameters
  • the mobility management platform (180) enables maintaining a NN for all the users using federated learning algorithm.
  • the learning can be function of QCI/QoS and other mentioned parameters at the ML model (184).
  • the mobility management platform (180) allocates the resources without asking for the UE (200) using the estimated measurement/handover probability values. Periodicity of handover/measurements is more than the usual systems. Proposed solution can be implemented on cloud based systems as well as UE side. For example trigger the dual/multi connectivity.
  • the BS (100) can do force handovers without the Measurement reports based on the past/other UEs
  • This approach shall be extended to predict and calculate CQI, PMI, RI accordingly before giving the grant.
  • the embodiments disclosed herein can be implemented using at least one hardware device and performing network management functions to control the elements.
  • FIG. 2B is the UE (200) for determining the target cell for handover in the wireless network (1000), according to an embodiment as disclosed herein.
  • the UE (200) is for example but not limited to a mobile phone, a laptop, a smart phone, Personal Digital Assistant (PDA), a tablet, a wearable device, or the like.
  • the UE (100) includes a memory (220), a processor (240), a communicator (260) and a mobility management controller (280).
  • the memory (220) can include non-volatile storage elements. Examples of such non-volatile storage elements may include magnetic hard discs, optical discs, floppy discs, flash memories, or forms of electrically programmable memories (EPROM) or electrically erasable and programmable (EEPROM) memories.
  • the memory (220) may, in some examples, be considered a non-transitory storage medium.
  • the term "non-transitory” may indicate that the storage medium is not embodied in a carrier wave or a propagated signal. However, the term “non-transitory” should not be interpreted that the memory (220) is non-movable.
  • the memory (220) is configured to store larger amounts of information.
  • a non-transitory storage medium may store data that can, over time, change (e.g., in Random Access Memory (RAM) or cache).
  • RAM Random Access Memory
  • the processor (240) is coupled with the memory (220), the communicator (260) and the mobility management controller (280).
  • the processor (240) may include one or a plurality of processors.
  • the one or the plurality of processors may be a general-purpose processor, such as a central processing unit (CPU), an application processor (AP), or the like, a graphics-only processing unit such as a graphics processing unit (GPU), a visual processing unit (VPU), and/or an AI-dedicated processor such as a neural processing unit (NPU).
  • the processor (240) may include multiple cores and is configured to execute the instructions stored in the memory (220).
  • the communicator (260) includes an electronic circuit specific to a standard that enables wired or wireless communication.
  • the communicator (260) is configured to communicate internally between internal hardware components of the UE (200) and with external devices via one or more networks.
  • the mobility management controller (280) is implemented by processing circuitry such as logic gates, integrated circuits, microprocessors, microcontrollers, memory circuits, passive electronic components, active electronic components, optical components, hardwired circuits, or the like, and may optionally be driven by firmware.
  • the circuits may, for example, be embodied in one or more semiconductors.
  • the mobility management controller (280) is configured to send the location information of the UE (200) to the wireless network (1000). Further, the mobility management controller (280) receives a handover command from the wireless network (1000) to handover the UE (200) to the target cell.
  • the target cell is determined by the wireless network (1000) based on the location information sent by the UE (200) and the at least one measurement report corresponding to the location information previously sent by the UE (200) which is determined by the wireless network (1000).
  • the measurement report corresponding to the location information sent by the UE (200) is determined using the correlation between the plurality of UEs (200a-N), wherein the correlation between a plurality of UEs (200a-N) is determined based on a plurality of network characteristics associated with the plurality of UEs (200a-N) and location information associated with each of the plurality of UEs (200a-N).
  • the plurality of network characteristics associated with the plurality of UEs (200a-N) includes but is not limited to: the location of the UE Location based on position reference signals (PRS) sent by the BS (100) and a BS topology, UE measurements of neighbouring cells and load information associated with the neighbouring cells of the UE (200), a UE capability, a UE index, a Quality of Service (QoS) of the UE (200), a QoS Class Indicator (QCI) of the UE (200), operating frequency of a plurality of BSs (100a-N), multi-in multi-out (MIMO) information associated with the UE (200), MIMO information associated with the BS (100), Modulation and coding scheme (MCS) information associated with the UE (200), resource blocks (RB) information associated with the UE (200), a queue information associated with the UE (200), mobility of the UE (200), a Doppler frequency, traffic information, load of a cell, route map information, signal
  • FIG. 2B shows the hardware elements of the UE (100) but it is to be understood that other embodiments are not limited thereon. In other embodiments, the UE (100) may include less or more number of elements. Further, the labels or names of the elements are used only for illustrative purpose and does not limit the scope of the invention. One or more components can be combined together to perform same or substantially similar function.
  • FIG. 3 is a flow chart (300) illustrating a method for determining the target cell for the handover of the UE (200) using the ML and the AI, according to an embodiment as disclosed herein.
  • the method includes the BS (100) monitoring the multiple network characteristics associated with the multiple UEs (200a-N) in the wireless network.
  • the mobility management platform (180) is configured to monitor the multiple network characteristics associated with the multiple UEs (200a-N) in the wireless network.
  • the method includes the BS (100) determining the correlation between the multiple UEs (200a-N) based on the multiple network characteristics associated with the multiple UEs (200a-N) and the location information associated with each of the UEs (200a-N).
  • the mobility management platform (180) is configured to determine the correlation between the multiple UEs (200a-N) based on the multiple network characteristics associated with the multiple UEs (200a-N) and the location information associated with each of the UEs (200a-N).
  • the method includes the BS (100) receiving the location information from the UE (200).
  • the mobility management platform (180) is configured to receive the location information from the UE (200).
  • the method includes the BS (100) determining the measurement report corresponding to the location information received from the UE (200) based on the correlation between the UEs (200a-N).
  • the mobility management platform (180) is configured to determine the measurement report corresponding to the location information received from the UE (200) based on the correlation between the UEs (200a-N).
  • the method includes the BS (100) determining the target cell for the handover of the UE (200) based on the location information received from the UE (200) and the measurement report corresponding to the location information received from the UE (200).
  • the mobility management platform (180) is configured to determine the target cell for the handover of the UE (200) based on the location information received from the UE (200) and the measurement report corresponding to the location information received from the UE (200).
  • FIG. 4 illustrates an architecture diagram of ORAN for determining the target cell for the handover of the UE (200), according to an embodiment as disclosed herein.
  • FIG. 4 A conventional architecture diagram of the ORAN is shown in the FIG. 4, therefore explanation of the conventional functions of the conventional components (402-428, and 432-444) in the ORAN are omitted for the sake of brevity of the specification, and the explanation is focused on use of ML system (430) using RL based NN (184a further explained in FIG. 11B) inside the RAN Intelligent Controller (RIC) non-Real Time (non-RT) unit (418) of the ORAN.
  • RIC RAN Intelligent Controller
  • non-RT non-Real Time
  • the ML system (430) intelligently determines the correlation between the multiple UEs (200a-N) based on the multiple network characteristics associated with the multiple UEs (200a-N) and location information associated with the multiple UEs (200a-N).
  • the correlation between the multiple UEs (200a-N) is used to intelligently determine the target cell for the handover of the UE (200), in specific cases even without receiving the measurement reports from the UE (200).
  • the mobility management platform (180) can deploy as container based micro service solution in the RIC (418).
  • the RIC (418) can subscribe all Nodes to send the details of the node including handover report and is also able to communicate to all nodes regarding handover recommendation via E2AP and E2SM protocol. All the nodes can push the Neighbour information's of cells, Load Of each Cell, Cells Capabilities to the mobility management platform (180) through the E2 RIC Indication Message and the mobility management platform (180) can build the data base used for algorithm inputs. Further, the network nodes can request the mobility management platform (180) at any time and get the details about the target cell for mobility.
  • the mobility management platform (180) determines less active users and functionalities like a typical Handover Prediction in terms of location pattern recognition for the less active UEs are performed by the non-Real Time (non-RT) unit (418). For instance, a daily routine can be deterministic or fixed over weekdays for particular UEs and the training for the particular UEs is performed by the non-Real Time (non-RT) unit (418) for reducing a processing burden in the near-RT RIC. As a result, the processing capability (CPU, periodicity) for decision making can be reduced at the near-RT RIC.
  • the proposed method can be implemented on the VRAN, the Cloud RAN, a centralized RAN, and any cloud based RAN.
  • a centralized controller and a ML/Artificial Intelligence module is included in all the RAN architectures, where each RAN architecture have different interfaces. Further, the ML/NN/AI modules estimated values will be conveyed to MAC/PHY/L3 modules.
  • RL based local NN (184a) improved the efficiency of performing handover of the UE (200) by enabling the selection of the best RAT. Further, performing handover without receiving the measurement report from the UE (200) enables to reduce unnecessary signalling in the wireless network and also efficient use of the network resources.
  • FIG. 5A illustrates an architecture diagram of a CRAN or virtual RAN for determining the target cell for the handover of the UE, according to an embodiment as disclosed herein.
  • FIG. 5A A conventional architecture diagram of a centralized RAN is shown in notation (a) of the FIG. 5A.
  • the remote radio head (RRH) units (502a and 502b) are connected to baseband units (504) using a front haul.
  • the Baseband units (504) are linked to a core network (506) using a backhaul.
  • the RL based local NN (as shown in the FIG. 11B) is implemented in the Baseband units (504) to determine the correlation between the multiple UEs based on the multiple network characteristics associated with the multiple of UEs (200a-N) and the location information associated with each of the multiple of UEs (200a-N) to determine the target cell for the handover of the UE (200).
  • the estimated handover/measurement report values are given to a medium access control (MAC) layer of the CRAN and the MAC layer can give the grant without asking the UE (200).
  • MAC medium access control
  • a conventional architecture diagram of the VRAN includes the centralized RAN with a network function virtualization (NFV) of the baseband units (504) in a cloud RAN (508) is shown in notation (b) of the FIG. 5A.
  • the Cloud RAN (508) is linked to the core network using the backhaul.
  • the RRH units (502a and 502b) are connected to the Cloud RAN (508) using the front haul.
  • the core functions of the core network are collocated with the Cloud RAN (508).
  • the RL based local NN (as shown in the FIG.
  • the 11B is implemented in the Cloud RAN (508) to determine the correlation between the multiple UEs based on the multiple network characteristics associated with the multiple of UEs (200a-N) and the location information associated with each of the multiple of UEs (200a-N) to determine the target cell for the handover of the UE (200).
  • the estimated handover/measurement report values are given to the MAC layer of the VRAN and the MAC layer can give the grant without asking the UE (200).
  • FIG. 5B illustrates various split option for the CRAN or the virtual RAN for determining the target cell for the handover of the UE, according to an embodiment as disclosed herein.
  • the estimated handover/measurement report values are conveyed to the MAC layer.
  • the estimated handover/measurement report values are conveyed via X2 interface if the MAC scheduling module is in RRH. In case the MAC scheduling module is not in the RRH then, existing mechanism is used to convey the estimated handover/MR values to the MAC layer.
  • the proposed the method remains the same.
  • the RIC module will have the information about L1 PHY upper to RRC via E2/Proprietary Interface. The same is implied to other split options. Therefore, the design of the mobility management platform (180) is modular and the scalable. Irrespective of the spilt, the mobility management platform (180) requires information from the PHY, MAC, RLC, and RRC for deciding the best cell for handover/attach.
  • FIG. 6 illustrates a heterogeneous network where mobility management platform (180) for determining the target cell for the handover of the UE (200) is hosted on the Proprietary Hardware, according to an embodiment as disclosed herein.
  • the heterogeneous network includes different RAN nodes such as for example wireless access points, base station, routers, relay gateways, etc. which are connected to the mobility management platform (180).
  • the mobility management platform (180) can connect to the network node through the proprietary interface.
  • the proprietary interface includes for example but not limited to all the nodes in the wireless network sends the information associated with the neighbour cells, the load of each cell, cells capabilities to the mobility management platform (180).
  • the mobility management platform (180) uses the information received from the nodes in the wireless network to build a data base used for algorithm inputs. Further, once the database is created by the mobility management platform (180), the network nodes can request the mobility management platform (180) at any time and get the details about the target cell during call procedures.
  • the mobility management platform (180) [RIC/xAPP] is introduced for processing the handover decision during attach, mobility or after mobility and recommend the new target cell/Node incase the current cell is not the favorable cell with respect to providing the best user experience.
  • the mobility xAPP subscribe every RAN Node to get the necessary information from the node through the E2AP RIC Subscription Message.
  • the mobility will subscribe for PRB Usage, Cell Topology, Cell Load, Cell Capability through E2SM protocol which needs to be newly added in the E2SM container.
  • the RAN node sends all the subscription details to the RIC/ Mobility XAPP through the E2AP IRIC Indication message and the RIC/ Mobility XAPP store all the information and creates a network topology.
  • the RAN contacts the RIC/ Mobility XAPP E2AP Mobility Into Request [which is a new message] which contains [UE Capability, UE position, direction and service request and request for the recommended cell for the UE (200).
  • the RIC/ Mobility XAPP E2AP Mobility Into Response [which is a new message] and provides the recommended RAT/Cell for the UE (200).
  • the current RAT Node performs the handover to the recommended RAT/Cell if the measurement results are satisfactory.
  • the RAN contacts the RIC/ Mobility XAPP E2AP Mobility Into Request is performed during mobility or after mobility based on time critical scenario.
  • FIG. 7A is a signalling diagram illustrating a scenario of handover performed by the serving gNB (100a) during a transition of the UE (200) from an idle state to an active state, according to the prior art.
  • the UE (200) sends a RRC connection request to the serving gNB (100a) and at step 704a the serving gNB (100a) responds by sending the RRC connection setup.
  • the UE (200) sends the RRC connection complete message on completion of the RRC connection (as indicated by step708a).
  • the serving gNB (100a) sends an Initial UE message to the AMF controller (500) and at step 712a the AMF controller (500) responds with an initial Context setup request.
  • the serving gNB (100a) On receiving the initial Context setup request, at step 714a the serving gNB (100a) sends a RRC Reconfiguration Request message to the UE (200). At step 716a the UE (200) completes the RRC Reconfiguration procedure and sends the RRC Reconf Complete message to the serving gNB (100a). At step 718a, the serving gNB (100a) sends the Initial Context setup response message to the AMF controller (500) and the UE (200) transitions from the idle state to the active state (as indicated by the step 720a).
  • the access node (AMF controller (500)) establishes the call in the Primary cell where the UE (200) sends the RRC Connection Request.
  • the access node AMF controller (500)
  • the access node establishes the call in the Primary cell where the UE (200) sends the RRC Connection Request.
  • the access node AMF controller (500)
  • the access node establishes the call in the Primary cell where the UE (200) sends the RRC Connection Request.
  • the access node AMF controller (500)
  • the access node establishes the call in the Primary cell where the UE (200) sends the RRC Connection Request.
  • the other Intra RAT/ Inter RAT Cells may provide better user experience.
  • the possibility of the availability of the other Intra RAT/ Inter RAT Cells during the handover is not considered in the conventional methods.
  • FIG. 7B is a signalling diagram illustrating the scenario of the handover performed by the serving gNB (100a) during the transition of the UE (200) from the idle state to the active state, according to an embodiment as disclosed herein.
  • the steps 702b to 706b are same as steps 702a to 706a in the FIG. 7A and repeated description is omitted.
  • all the nodes periodically push information for example which includes but may not be limited to information of a neighbour cell, load of each cell, cells capabilities, etc. to the mobility management platform (180), as indicated by step 700b.
  • the mobility management platform (180) builds a master database based on the inputs provided by the nodes.
  • the serving gNB (100a) sends the Initial UE message to the AMF controller (500) and at step 712b the AMF controller (500) responds with the initial context setup request.
  • the serving gNB (100a) communicates to the mobility management platform (180) the recommended cell for the UE (200).
  • the serving gNB (100a) sends the information to the mobility management platform (180) to take an optimal decision of the target cell.
  • the mobility management platform (180) recommends the target cell based on the inputs from the UE (200) and the serving gNB (100a) and the ML model (184) of the mobility management platform (180).
  • the details sent by the serving gNB (100a) includes but are not limited to:
  • the serving gNB (100a) sends the RRC Reconfiguration Request message to the UE (200).
  • the serving gNB (100a) commands the UE (200) to measure the candidate cells power and provide the same to the mobility management platform (180).
  • the serving gNB (100a) receives the measurement report from the UE (200) and at step 730b, the serving gNB (100a) performs the intra/inter RAT handover of the UE (200) based on the measurement report.
  • the mobility management platform (180) recommends the other cells for performing the handover and commands the UE (200) to measure and perform the handover to that cells if applicable.
  • the Intra RAT/ Inter RAT Cells which provide the best user experience is selected and the handover of the UE (200) is performed.
  • FIG. 8A is a signalling diagram illustrating a scenario of the handover of the UE (200) performed by the serving gNB (100a) for load balancing based on a cell load, according to the prior art.
  • the UE (200) is RRC connected to the serving gNB (100a).
  • the serving gNB (100a) takes decision on load balancing /shedding based on CPU/Cell Load and performs the following:
  • the serving gNB (100a) i identifies the UE (200) for off load
  • the serving gNB (100a) finds the collocated cells for Handover Over based on the configuration data
  • the serving gNB (100a) commands the UE (200) to measure those cells and report back.
  • the UE (200) measures those target cells and reports back based on the RSR/RSRQ by sending the measurement report to the serving gNB (100a).
  • the serving gNB (100a) performs the handover of the UE (200) to the best cell based on the measurement report sent by the UE (200).
  • Intra RAT/ Inter RAT Cells which may be available for the UE (200) which may provide better user experience based on location, service and mobility parameters.
  • the possibility of the availability of the other Intra RAT/ Inter RAT Cells during the handover is not considered in the conventional methods.
  • FIG. 8B is a signalling diagram illustrating a scenario of the handover of the UE (200) performed by the serving gNB (100a) for the load balancing based on the cell load, according to an embodiment as disclosed herein.
  • the nodes periodically push information for example which includes but may not be limited to information of a neighbour cell, load of each cell, cells capabilities, etc. to the mobility management platform (180), as indicated by step 800b.
  • the serving gNB (100a) decides to perform the load balancing /shedding based on the CPU/Cell load
  • the serving gNB (100a) identifies the UE (200) for off-loading.
  • the mobility management platform (180) builds the complete tropology information, neighbors of each cells, its capability, load coordinates.
  • the serving gNB (100a) sends the off-loading UE details, the UE capability, the measurement report, the UE context data, the coordinates to the mobility management platform (180) and at step 810b, the mobility management platform (180) updates the serving gNB (100a) with the target system/cell information for each UE based on requirements i.e., the serving gNB (100a) is updated about the best cell for performing the off-loading.
  • the serving gNB (100a) then commands the UE (200) to measure the cells recommended by the mobility management platform (180) and report back to the serving gNB (100a).
  • the UE measures those target cells and sends the measurement reporting on the RSRP/RSRQ to the serving gNB (100a).
  • the serving gNB (100a) performs the handover of the UE (200) to the best cell based on the measurement report sent by the UE (200). Therefore, unlike to the conventional methods of handover, in the proposed method the mobility management platform (180) recommends the best cell for the handover of the UE (200) based on the inputs from the serving gNB (100a) and the handover is performed based on the measurement report sent by the UE (200) for the cells recommended by the mobility management platform (180). Hence, a double check mechanism is used which confirms the best cell for the UE (200) before performing the handover.
  • FIG. 9A is a signalling diagram illustrating a scenario of configuration of static measurement objects in the UE in a connected mode, according to the prior art.
  • the UE (200) sends the RRC connection request to the serving gNB (100a) and at step 904a the serving gNB (100a) responds by sending the RRC connection setup.
  • the UE (200) sends the RRC connection complete message on completion of the RRC connection (as indicated by step 908a).
  • the serving gNB (100a) sends the initial UE message to the AMF controller (500) and at step 912a the AMF controller (500) responds with the initial Context setup request.
  • the serving gNB (100a) sends the RRC Reconfiguration Request message to the UE (200) to configure the measurement objects in the UE (200) for mobility purpose.
  • the UE (200) completes the RRC Reconfiguration procedure and sends the RRC Reconf Complete message to the serving gNB (100a).
  • the serving gNB (100a) sends the initial context setup response message to the AMF controller (500) and the UE (200) transitions from the idle state to the active state (as indicated by the step 920a).
  • the access node (AMF controller (500)) establishes the call in the Primary cell where the UE (200) sends the RRC Connection Request.
  • the Measurement Objects configured generally are static based on the Primary cell to which the UE (200) connects. Therefore, a possibility of reconfiguration of the Measurement Objects dynamically which can enhance the performance of the UE (200) remains unexplored.
  • FIG. 9B is a signalling diagram illustrating a scenario of configuration of dynamic measurement objects in the UE in the connected mode, according to an embodiment as disclosed herein.
  • the steps 902b to 906b are same as steps 902a to 906a in the FIG. 9A and repeated description is omitted.
  • all the nodes periodically push information for example which includes but may not be limited to information of a neighbor cell, load of each cell, cells capabilities, etc. to the mobility management platform (180), as indicated by step 900b.
  • the mobility management platform (180) builds a master database based on the inputs provided by the nodes.
  • the serving gNB (100a) sends the Initial UE message to the AMF controller (500) and at step 912b the AMF controller (500) responds with the initial context setup request.
  • the serving gNB (100a) communicates to the mobility management platform (180) the recommended cell for the UE (200).
  • the serving gNB (100a) sends the information to the mobility management platform (180) to take an optimal decision of the target cell.
  • the mobility management platform (180) recommends the target cell based on the inputs from the UE (200) and the serving gNB (100a) and the ML model (184) of the mobility management platform (180).
  • the access node (AMF controller (500)) will contact with the Central Node (serving gNB (100a)) with required details, position, direction, services and request for the best cells for measurements.
  • the serving gNB (100a) sends the RRC Reconfiguration Request message to the UE (200).
  • the serving gNB (100a) will configure those frequencies recommended by the mobility management platform (180) in the measurement Objects and send in the RRC Reconfiguration Message.
  • the serving gNB (100a) commands the UE (200) to measure the candidate cells power and provide the same to the mobility management platform (180).
  • the serving gNB (100a) receives the measurement report from the UE (200) and at step 930b, the serving gNB (100a) performs the intra/inter RAT handover of the UE (200) based on the measurement report. Therefore, unlike to the conventional method of performing the handover, in the proposed method the serving gNB (100a) can contact the mobility management platform (180) during every Reconfiguration or periodically and check the validity of the measurement Object and is able to reconfigure dynamically the frequency sets. As a result, the Intra RAT/ Inter RAT Cells which provide the best user experience is selected and the handover of the UE (200) is performed.
  • FIG. 10 is a signalling diagram illustrating a call flow in the UE (200) during mobility without a measurement report, according to an embodiment as disclosed herein.
  • step 1002 all the nodes periodically push information for example which includes but may not be limited to information of a neighbour cell, load of each cell, cells capabilities, etc. to the mobility management platform (180).
  • the proposed solution is to totally avoid the requirement that the UE (200) needs to get the measurement for mobility.
  • the UE (200) is configured with A1/A2 for measurement of the current cell power.
  • the UE (200) sends the A2 measurement report to the serving gNB (100a).
  • the mobility management platform (180) builds a master database based on the inputs provided by the nodes and at step 1008 the handover decision is taken.
  • the serving gNB (100a) can communicate to the mobility management platform (180) the recommended cell for the UE (200).
  • the mobility management platform (180) will recommend a target cell based on the inputs from the UE (200) and the serving gNB (100a) and the ML model (184).
  • the serving gNB (100a) will perform the handover to that recommended cell so that the UE (200) will have better performance.
  • FIG. 11A illustrates various inputs received by the ML model (184) of the mobility management platform (180) for determining the target cell for the handover of the UE, according to an embodiment as disclosed herein.
  • the ML model (184) of the mobility management platform (180) corresponds to a Neural Network (NN) model.
  • the NN model is trained by inputting the multiple network characteristics corresponding to the multiple UEs (200a-N) and the location information associated with the multiple UEs (200a-N) to multiple input NN nodes of the NN model.
  • the conventional inputs received by the ML model (184) include channel quality, the number of users, QCI/QoS/Bearer info, the resource availability and the user classes.
  • the ML model (184) also receives mobility pattern and location information, expected interference and traffic, expected QCI/power to be transmitted and predicted user classes which are determined by the NN model.
  • FIG. 11B illustrates a neural network used in the ML model (184) of the mobility management platform (180), according to an embodiment as disclosed herein.
  • the various inputs are provided to the NN model along with the old and new parameters. Further, an optimal weight of each input NN node of the NN model is to be determined based on a training method.
  • the trained NN is configured to receive the multiple network characteristics associated with the UE (200) and determine the correlation between the multiple UEs (200a-N). The correlation between the multiple UEs (200a-N) is used to take decisions on the handover of the UE (200a-N).
  • the BS (100) can maintain different NN/ML/AI modules for different UEs or Single NN/ML/AI for set of UEs.
  • the UEs classification can be done various parameters such as QCI or/and QoS or/and MCS or/and location other network or UE parameters.
  • Maintaining the NN model for each of the UEs may be required when the UEs are MIMO capable UEs.
  • the SINR distribution for the MIMO capable UEs are different and hence requires separate NN model for each of the UEs. Usually, average SINR of MIMO users is greater than average SINR of non-MIMO users.
  • the NN model is formed by considering the UE Category type, transmission type. Further, different applications have different packet loss tolerance and delay requirements.
  • the NN model will decide whether to perform early handover or not (whether it can afford to delay the handover or not). For example, browsing applications may have relaxed deadline on the packets, relaxed packet loss tolerance, etc. Then for such browsing applications the handover can be delayed and can trigger handover based on the Coverage area/SINR/CQI distribution map. For gaming applications, the deadline and the packet loss tolerance are stringent. In such scenarios, the NN model decides to perform the early handover.
  • the trained NN model is a reinforcement learning (RL) based NN (hereinafter may also interchangeably referred as "trained RL based local NN or trained local NN”.
  • the trained NN may comprise a deep NN (DNN), a convolution NN (CNN), recurrent NN (RNN), sparse NN (SNN), or an artificial NN (ANN).
  • DNN deep NN
  • CNN convolution NN
  • RNN recurrent NN
  • SNN sparse NN
  • ANN artificial NN
  • the trained ML model (184) comprises an input layer (1102), a first hidden layer (1104), a second hidden layer (1106), an output layer (1108).
  • the environment is the communication between the BS (100) and the multiple UEs (200a-N) for sending the measurement reports and receiving the handover decisions.
  • the first hidden layer (1104) and the second hidden layer (1106) of the trained NN comprises optimal weights assigned to multiple layers (including the input layer (1102), the first hidden layer (1104), the second hidden layer (1106) and the output layer (1108)) of the trained NN, a bias, and an activation function.
  • the different layers of the NN include plurality of NN nodes, therefore the layers of the NN may also be refereed as "NN nodes" of the trained NN.
  • the NN (506) is operable in two operating modes: a) training mode and b) testing mode.
  • the training mode the NN act as the RL agent and learns to predict the CQS information related to unknown frequency bands by interacting with the environment.
  • the NN For every interaction with the environment, the NN generates an output which is given to the error function controller to determine an error function of the output generated by the NN, and whether the generated output is less than a threshold error value or not.
  • the output of the error function controller is provided to the feedback function controller which is configured to provide feedback to a RL model in form of rewards based on performance of the NN.
  • the RL model receives the feedback and accordingly updates the weights assigned to each layer of the NN to achieve minimum error function.
  • the RL model further incorporates the rewards and accordingly updates its policy function and value function to enhance performance of the NN.
  • the NN is under training until an error function of the output of the NN is less than a threshold error value.
  • the threshold error value is predefined or user defined value.
  • the correlation information can be a function of time, BS IDs, topology of the heterogeneous network, operating frequency, UE/BS category/capability, bandwidth paths/BW, throughput requirements, QCI/QoS type, Number of BS antennas, external climate conditions, Doppler statistics, Channel statistics, mobility statistics, delay spread statistics, MIMO/non-MIMO capabilities, transmission type (for example, transmit or receive diversity modes etc.), time of the day, ACK/NACK statistics, SINR statistics as a function of the UE category and above mentioned parameters, transmit power of the BS and/or UE, ACK/NACK as a function of MCS and other mentioned parameters, shadowing, fading stats, distribution of all parameters, long/short term statistics of all the parameters.
  • the operating mode of the NN is switched to the testing mode.
  • the testing mode of the NN one or more inputs are provided to the NN and if the RL model determines that the error function of the output of the trained NN exceeds the threshold error value, it sends a negative feedback and accordingly the mode of operation of the NN may be switched back to training mode until the desired output (error function less than or equal to the threshold error value) is achieved.
  • a reason for the NN to predict the output with more error function could be change in environment of the multiple UEs (200a-N) which may not be known to the trained NN therefore the NN needs to learn and update according to the changed environment.
  • the training of the NN is conducted based on at least one of belief propagation, a back-propagation or an adaptive learning optimization (Adam).
  • the NN model can be trained online whenever there is a need. For example, if the accuracy is going down, the NN can start learning in the online or it can do federated learning. Learning can be triggered by RIC module.
  • the activation function used in the NN is a ReLu activation function.
  • other activation function can also be used such as unit step, signum, linear regression, piece-wise linear, hyperbolic tangent, logistic (sigmoid), rectifier softplus, etc.
  • the weight of each input node of the NN model is determined as:
  • the NN model is trained based on the optimal weight of each input NN node, the multiple network characteristics associated with the multiple UEs (200a-N) and the location information associated with the multiple UEs (200a-N).
  • the estimated aa of the handovers/bearer load/QCI/QoS info/Resource is obtained as the output of the NN model.
  • the location information can be estimated using LPP, LPPe protocols, SRS estimation, beam number, measurement information.
  • UE speeds can be estimated on per user basis.
  • the output of the NN model includes but not limited to recommendations of whether the handover is needed or not, to which neighboring BS (100) should the call be transferred, conveying the number of handover to the neighboring BSs (100a-N), estimating the QoS Stat details of the users, estimating the QCI Stat details of the users, estimating the GBR and the Non-GBR stat traffic details along with the bear load, interference stats, etc.
  • the network parameters can be updated either periodically or aperiodically or on request basis or a demand basic.
  • the learning can be translated into offline or rule based formulas. Even BS (100) can run the ML/AI techniques periodically or aperiodically. For example if the error is more than x% then re-learning can be done.
  • the BS (100) can request for other BSs NN/ML/AI.
  • the BS (100) can apply one UE learning on other UEs intelligently exploiting the characteristics such as location information, beam information, UE capabilities, UE transmission, UE type and other wireless parameters etc. even sub-set of parameters can be used to learn. Some of the parameters will be available on non-real time basis, i.e., with different periodicity. Each parameter can have different periodicity.
  • BS_1 can learn handover strategy (whether early handover is need or not) based on UE_1 location (x, y, z). Learning of UE_1 can be applied to other UEs which are entering into (x, y, z) zone or other zones in BS_1 category.
  • Each BS can learn from other BS learnings by updating the weights of NN (a central node can average the weights of NN model of all BSs, which are operating in same frequency/same type).
  • FIG. 12 illustrates an example scenario of using an AI model (188) in the mobility management platform (180) of the BS (100), according to an embodiment as disclosed herein.
  • the AI model (188) may be used in place of the ML model (184) for determining the correlation between the multiple UEs and to take the decision of the handover of the UE based on multiple network parameters associated with the multiple UEs (200a-N).
  • the AI model (188) uses an agent and an environment model. In the AI model (188) every action is rewarded and penalized accordingly.
  • the BS (100) can learn the parameters such as the number of required handovers using the AI model (188). Further, quasi queue based learning algorithms can be used to improve the performance of the AI model (188). Further, finite MDP models can also be used by the AI model (188) for learning. In another example, cost function generated using multiple parameters can also be used by the AI model (188) for learning.
  • the AI model (188) can be implemented using Reinforcement Learning, which in turn uses neural networks.
  • the AI model (188) can be implemented in a flexible manner. Further, various cost functions can be used in training the AI model (188) and the different cost functions can give the AI corresponding to different external factors. Therefore, multiple AI agents can be developed in the AI model (188) to react to changes in different external conditions.
  • FIG. 13 is an example handover map indicating the coverage area associated with the BS (100), according to an embodiment as disclosed herein.
  • the handover map is generated which indicates the coverage area associated with the serving BS (100).
  • the handover map is generated using the various parameters in table. 1.
  • the RSRP measurements are used to trigger the HO.
  • the RSRP of the UE (200) is periodically monitored and compared with the target BS to make the HO decision by the serving BS (100).
  • the conventional HO procedure is mainly threshold-based and involves multiple signalling between the UE (200), the serving BS (100a), and the target BS (100b), which leads to signalling overhead and incurs higher delay, thereby impacting the QoS of the UE (200).
  • the UE movement pattern is used by the serving BS (100) to learn to predict in advance whether a particular UE (200) needs the HO based on the QoS statistics of the UE and the usage pattern.
  • the machine learning framework is used for deciding on the HO mechanism.
  • the O-RAN architecture facilitates RAN intelligence and computing which is also used for deciding the HO.
  • the BS (100) can learn and make current HO decisions based on the earlier decisions taken for the UE (200) using the movement of the UE (200) and the QoS usage pattern. Further, the learning from a particular UE can also be applied to other UEs (200a-N) to make the HO decisions.
  • the learning problem is a classic example of federated learning which enables us to develop a model-driven federated-learning-based method for the HO.
  • the federated-learning based mechanism can make an inference on whether a particular UE needs the HO, in advance.
  • the proposed method benefits by substantially reducing the periodicity of the RSRP measurements which thereby enhances the UE power utilization and increases the spectral efficiency.
  • the learning from one BS (100) can also be applied to take the HO decision by other BSs (100a-N) with significantly reduced signalling.
  • the location prediction is used for taking the Handover decision without any measurements, based on the past decisions.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • General Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

Des modes de réalisation de la présente invention concernent un procédé de détermination d'une cellule cible pour un transfert intercellulaire d'un UE. Le procédé comprend le suivi, par une plateforme de gestion de mobilité, de multiples caractéristiques de réseau associées à de multiples EU et la détermination, par la plateforme de gestion de mobilité, d'une corrélation entre les multiples EU sur la base des multiples caractéristiques de réseau associées aux multiples EU et des informations de localisation des multiples EU. Le procédé comprend la réception, par la plateforme de gestion de mobilité, des informations de localisation de l'EU parmi les multiples EU et la détermination, par la plateforme de gestion de mobilité, d'un rapport de mesure correspondant aux informations de localisation reçues de l'EU sur la base de la corrélation. Et le procédé comprend la détermination, par la plateforme de gestion de mobilité, de la cellule cible pour le transfert intercellulaire de l'EU sur la base des informations de localisation reçues de l'EU et du rapport de mesure.
PCT/KR2021/008948 2020-07-13 2021-07-13 Procédé et système pour déterminer une cellule cible pour le transfert intercellulaire d'eu WO2022015008A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
IN202041029710 2020-07-13
IN202041029710 2021-07-06

Publications (1)

Publication Number Publication Date
WO2022015008A1 true WO2022015008A1 (fr) 2022-01-20

Family

ID=79556156

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/KR2021/008948 WO2022015008A1 (fr) 2020-07-13 2021-07-13 Procédé et système pour déterminer une cellule cible pour le transfert intercellulaire d'eu

Country Status (1)

Country Link
WO (1) WO2022015008A1 (fr)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115483960A (zh) * 2022-08-23 2022-12-16 爱浦路网络技术(南京)有限公司 低轨卫星的跳波束调度方法、系统、装置及存储介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20150024754A1 (en) * 2011-05-06 2015-01-22 Telefonaktiebolaget L M Ericsson (Publ) Methods and Nodes Supporting Cell Change
US20150341828A1 (en) * 2009-02-09 2015-11-26 Qualcomm Incorporated Method and apparatus for maintaining location continuity for a ue following handover
US20160105831A1 (en) * 2013-05-10 2016-04-14 Telefonaktiebolaget L M Ericsson (Publ) Methods and Arrangements for Enabling Continuation of Ongoing Positioning Measurements at Handover
US20160330671A1 (en) * 2012-08-21 2016-11-10 Kyocera Corporation Handover management based on speeds of wireless communication devices
US9674740B2 (en) * 2012-08-03 2017-06-06 Telefonaktiebolaget Lm Ericsson (Publ) Method and arrangement for mobility procedures supporting user equipment groups

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20150341828A1 (en) * 2009-02-09 2015-11-26 Qualcomm Incorporated Method and apparatus for maintaining location continuity for a ue following handover
US20150024754A1 (en) * 2011-05-06 2015-01-22 Telefonaktiebolaget L M Ericsson (Publ) Methods and Nodes Supporting Cell Change
US9674740B2 (en) * 2012-08-03 2017-06-06 Telefonaktiebolaget Lm Ericsson (Publ) Method and arrangement for mobility procedures supporting user equipment groups
US20160330671A1 (en) * 2012-08-21 2016-11-10 Kyocera Corporation Handover management based on speeds of wireless communication devices
US20160105831A1 (en) * 2013-05-10 2016-04-14 Telefonaktiebolaget L M Ericsson (Publ) Methods and Arrangements for Enabling Continuation of Ongoing Positioning Measurements at Handover

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115483960A (zh) * 2022-08-23 2022-12-16 爱浦路网络技术(南京)有限公司 低轨卫星的跳波束调度方法、系统、装置及存储介质
CN115483960B (zh) * 2022-08-23 2023-08-29 爱浦路网络技术(南京)有限公司 低轨卫星的跳波束调度方法、系统、装置及存储介质

Similar Documents

Publication Publication Date Title
US12047803B2 (en) Measurement reporting method and related device
WO2021101347A1 (fr) Procédé et système de prédiction d'état de qualité de canal dans un réseau sans fil à l'aide d'un apprentissage automatique
WO2020226374A1 (fr) Procédé et appareil permettant de traiter un transfert intercellulaire dans un système de communication sans fil
WO2010087618A2 (fr) PROCÉDÉ POUR EFFECTUER UN TRANSFERT PAR UN ENSEMBLE CoMP
US20200413316A1 (en) Managing communication in a wireless communications network
WO2021029649A1 (fr) Procédé et appareil pour la gestion d'un transfert conditionnel dans un réseau de communication sans fil
Fehske et al. Small-cell self-organizing wireless networks
WO2011025224A2 (fr) Procédés et appareil prenant en charge l'atténuation de l'interférence coordonnée dans des réseaux à plusieurs niveaux
WO2020055172A1 (fr) Améliorations apportées et se rapportant à des réseaux de télécommunication
WO2020184929A1 (fr) Procédé et appareil de gestion de la mobilité d'un dispositif dans un réseau
US20210321306A1 (en) First network node, wireless device and methods performed thereby for handling a link switch
WO2015009130A1 (fr) Procédé et appareil de communication coopérative dans un système de communication sans fil
WO2020004688A1 (fr) Dispositif de communication v2x et procédé de transmission de données associé
EP3957127A1 (fr) Procédé et ue pour rapporter une mesure d'accessibilité dans un système de communications sans fil
EP4201156A1 (fr) Système à multiples points de réception/émission (trp) et son procédé
US20140003275A1 (en) Radio communication system and communication control method
WO2022015008A1 (fr) Procédé et système pour déterminer une cellule cible pour le transfert intercellulaire d'eu
WO2023101304A1 (fr) Procédé et appareil de réalisation d'une communication dans un système de communication sans fil
WO2021191498A1 (fr) Mesures proactives visant à réduire des effets de réduction de puissance
US9344928B2 (en) Neighbouring cell optimization method and mobile communication system
WO2023158053A1 (fr) Procédés et appareils pour économiser l'énergie d'un ue en utilisant des informations auxiliaires au niveau de l'ue
WO2022250377A1 (fr) Procédé et appareil pour optimiser les performances de mobilité d'un équipement utilisateur dans un système de communication sans fil
WO2023008956A1 (fr) Procédé et appareil pour fournir un service de localisation basé sur un signal de liaison montante
WO2021118322A1 (fr) Conception et architecture pour système de réseau multi-connectivité multi-radio
WO2019240304A1 (fr) Appareil de communication v2x et procédé de transmission multicanal correspondant

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21841461

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21841461

Country of ref document: EP

Kind code of ref document: A1