JP7602114B2 - Control device and control method - Google Patents

Description

The present invention relates to a control device and a control method.

In recent years, local 5G, which allocates 5G frequencies to entities other than telecommunications carriers in limited areas, has been attracting attention as a system that allows companies and other entities to independently build their own private wireless networks using 5G communications technology.

For example, a company may build a local 5G environment on its own premises or on premises entrusted to it by another owner. Hereinafter, the network operation unit in local 5G will be called a tenant. Various operations are possible in each tenant, such as a factory for manufacturing goods or a warehouse for logistics. In a 5G communication network, slices may be built in the core network according to the required QoS (Quality of Service) for each traffic, such as the importance and urgency of the data being transmitted and received, and these slices may be used to accommodate a variety of traffic.

Technologies relating to 5G are described, for example, in the following Patent Documents 1 to 6.

JP 2019-16875 A JP 2016-77017 A JP 2017-200172 A International Publication No. 2019/111317 Special table 2018-538751 publication JP 2019-021953 A

However, unlike public networks operated and managed by telecommunications carriers, local 5G, which is operated independently by individual carriers, may not be able to satisfy the required quality of each slice set in the core network on the wireless network side. For example, interference may occur between tenants or between cells in the same tenant. A common method for controlling interference between cells is to assign different frequencies (channels) between cells that are affected by mutual interference, and this method is adopted, for example, in wireless LANs (Local Area Networks). In this case, for example, in a certain base station device, when slices with low traffic volume but high required quality and slices with low required quality are mixed, it is assumed that different frequencies (channels) are assigned between the interfering base station device and other base station devices so that the quality of the slice with high required quality can be ensured. In this case, many frequency resources (number of channels) are required, and the utilization efficiency of wireless resources in the entire network may be significantly reduced. On the other hand, conversely, it is assumed that different frequencies (channels) are assigned between the interfering base station device and other base station devices so that the quality of the slice with low required quality can be ensured. In this case, radio resources that cause unacceptable interference may be allocated to a slice with high required quality, and the required quality of this slice may not be met.

Therefore, one disclosure provides a control device and a control method that suppresses a decrease in frequency utilization efficiency while satisfying the required quality of slices composed of traffic with the same required quality.

A control device in a communication system having a plurality of base station devices, a terminal device wirelessly connected to the base station devices, and a control device that controls the base station devices, the control device having an estimation unit that calculates the amount of required radio resources for each of a plurality of slices classified based on required quality, and estimates, for each of the plurality of base station devices, the amount of interference between the base station devices, the amount of interference between the base station devices and the terminal devices, and the amount of interference between the terminal devices in each of the plurality of slices, and an allocation unit that allocates, within the radio resources of a first base station device based on the estimation results, a first resource for a first slice, a second resource for a second slice having a required quality higher than that of the first slice, and a constrained resource whose use is restricted, and the allocation unit allocates resources other than the second resource and the constrained resource as the first resource, and allocates resources allocated in a second base station device that is affected by interference to the second slice as the constrained resource.

One disclosure can suppress a decrease in frequency utilization efficiency while satisfying the required quality of slices composed of traffic with the same required quality.

FIG. 1 is a diagram showing an example of the configuration of a communication system 10. FIG. 2 is a diagram illustrating an example of the configuration of the NMS server 300. As shown in FIG. FIG. 3 is a diagram showing an example of control between cells using the same channel. FIG. 4 is a diagram showing an example of wireless resource allocation. FIG. 5 is a diagram showing an example of wireless resource allocation. FIG. 6 is a diagram showing an example of a processing flowchart of the wireless resource allocation process S1000. FIG. 7 is a diagram showing an example of inter-cell interference measurement timing. FIG. 8 is a diagram showing an example of the position of a terminal device in the area worst value evaluation. FIG. 9 is a diagram illustrating an example of resource allocation. FIG. 10 is a diagram illustrating an example of resource allocation. FIG. 11 is a diagram illustrating an example of resource allocation. FIG. 12 is a diagram illustrating an example of a process flowchart of the process 1 in the second embodiment. FIG. 13 is a diagram illustrating an example of a process flowchart of the process 2 in the second embodiment.

[First embodiment]
A first embodiment will be described.

<Example of distribution system configuration>
1 is a diagram showing an example of the configuration of a communication system 10. The communication system 10 includes gNodeBs 200-1 to 3 (hereinafter, sometimes referred to as gNodeBs 200), an NMS (Network Management System) server 300, a core network server 500, and a core network 400. The communication system 10 is a system in which the NMS server 300 controls radio resources for each gNodeB 200 and a terminal device 100 (not shown).

The gNodeB 200 is a base station device that performs wireless communication with the terminal device 100. The gNodeB 200 constitutes a cell, which is an area where wireless communication with the terminal device 100 is possible. The radio resources of the gNodeB 200 are controlled by the NMS server 300 via the core network 400 to which it is connected. Note that the gNodeBs 200 may be base station devices in different tenants or may be base station devices in the same tenant. Note that although there are three gNodeBs 200 in FIG. 1, the number may be two or less, or four or more.

The NMS server 300 is a control device that controls the entire network and controls the radio resources of the gNodeB 200, and is, for example, a server machine. The NMS server 300 controls the radio resources for each slice configured by the core network server 500, for example. The NMS server 300 also controls interference between the gNodeBs 200, for example, by controlling the radio resources of each gNodeB. The NMS server 300 also controls the core network 400, for example.

The core network server 500 is a server that provides various functions in the core network 400, and is, for example, a server machine. The core network server 500 constructs slices according to, for example, the required quality of various traffic. The required quality is, for example, an index such as a transmission rate, a delay amount, or a reliability. For example, URLLC (Ultra-Reliable and Low Latency Communications) does not require a high required transmission rate, but high reliability and low latency are required. Also, eMBB (enhanced Mobile Broadband) requires high-capacity communication, but the reliability and allowable delay amount are lower than those of URLLC.

The terminal device 100 (not shown) is a device that performs wireless communication via the gNodeB 200, and is, for example, a smartphone or a tablet terminal. The terminal device 100 may also be a device such as a manufacturing robot in a factory or a communication module attached to a production line. The terminal device 100 may also be a device such as a product management reader device or a communication module attached to an inventory management system in a logistics system.

The communication system 10 is a system in which the NMS server 300 manages the radio resources of the gNodeB 200, thereby optimizing the allocation of radio resources and interference between adjacent cells formed by each of the adjacent gNodeBs 200.

<Configuration example of NMS server 300>
2 is a diagram showing an example of the configuration of the NMS server 300. The NMS server 300 includes a CPU (Central Processing Unit) 310, a storage 320, a memory 330, and a communication circuit 340.

Storage 320 is an auxiliary storage device such as a flash memory, a hard disk drive (HDD), or a solid state drive (SSD) that stores programs and data. Storage 320 stores a wireless resource allocation program 321 and a base station device control program 322.

Memory 330 is an area into which programs stored in storage 320 are loaded. Memory 330 may also be used as an area in which programs store data.

The communication circuit 340 is a circuit that connects to and communicates with other devices. The communication circuit 340 is, for example, a network interface card or a communication port.

The CPU 310 is a processor that loads the programs stored in the storage 320 into the memory 330, executes the loaded programs, constructs each part, and realizes each process.

The CPU 310 executes the wireless resource allocation program 321 to construct an estimation unit and an allocation unit, and performs wireless resource allocation processing. The wireless resource allocation processing is processing for allocating wireless resources to the gNodeB 200 to which it is connected (controls). In the wireless resource allocation processing, the NMS server 300 performs a radio access network (RAN (Radio Access Network)) slice construction processing, a required wireless resource amount estimation processing, a TDD frame construction and setting processing, an inter-cell interference amount estimation processing, an interference relationship estimation processing, and a resource allocation processing.

The CPU 310 executes the RAN slice construction module 3211 of the wireless resource allocation program 321 to construct an estimation unit and an allocation unit, and performs a RAN slice construction process. The RAN slice construction process is a process of constructing a RAN slice by subdividing a slice constructed in the core network according to the RAN environment. Note that in the RAN slice construction process, the NMS server 300 may use the slice constructed in the core network as the RAN slice as it is.

The CPU 310 executes the required radio resource amount estimation module 3212 of the radio resource allocation program 321 to construct an estimation unit and an allocation unit, and performs the required radio resource amount estimation process. The required radio resource amount estimation process is a process for estimating (calculating) the amount of radio resources required for each RAN slice.

The CPU 310 executes the TDD frame configuration setting module 3213 of the wireless resource allocation program 321 to construct an estimation unit and an allocation unit, and to perform the TDD frame configuration setting process. The TDD frame configuration setting process is a process that determines the TDD frame configuration based on the required wireless resource amount, and sets the determined TDD frame.

The CPU 310 executes the inter-cell interference amount estimation module 3214 of the wireless resource allocation program 321 to construct an estimation unit and an allocation unit and perform inter-cell interference amount estimation processing. The inter-cell interference amount estimation processing is a process for estimating the amount of interference between cells for each RAN slice.

The CPU 310 executes the cell interference relationship estimation module 3215 of the wireless resource allocation program 321 to construct an estimation unit and an allocation unit and perform cell interference relationship estimation processing. The cell interference relationship estimation processing is a process for estimating the interference relationship between cells for each RAN slice.

The CPU 310 executes the resource allocation module 3216 of the wireless resource allocation program 321 to construct an estimation unit and an allocation unit and perform resource allocation processing. The resource allocation processing is a process of allocating wireless resources that can satisfy the required quality of the slices, starting from the RAN slice with the highest priority (highest required quality), based on the estimated interference relationship between cells.

The CPU 310 executes the base station device control program 322 to perform base station device control processing. The base station device control processing is a process of notifying each gNodeB 200 of the determined radio resources and having them execute the processing.

<Co-channel interference control>
Fig. 3 is a diagram showing an example of control between cells using the same channel. As shown in Fig. 3(1), for example, if a 200 MHz band that is a band available in the 4.5 GHz band is divided into four channels with a minimum allocation unit of 50 MHz.

Figure 3 (2) is a diagram showing an example of allocation of 15 cells and channels used by each cell. As shown in Figure 3 (2), channels are assigned to each cell so that adjacent cells do not use the same channel. However, for example, cells C1 and C2, which are not adjacent to each other but are assigned the same channel 1, may interfere with each other. Therefore, the NMS server 300 performs radio resource control between cells taking into account various QoS flows (or slices) in the above-mentioned minimum allocation unit.

Figure 4 is a diagram showing an example of wireless resource allocation. Figure 4 shows an environment in which two types of traffic with different required qualities coexist. One is eMBB (enhanced Mobile Broadband) communication, which requires high speed and large capacity. The other traffic is URLLC (Ultra-Reliable and Low Latency Communications). In a factory, for example, URLLC slices are used for real-time control of manufacturing equipment, and eMBB slices are used for line monitoring with cameras. In this case, the top priority is to perform real-time control of the manufacturing equipment with high reliability, and in communication, the priority of the URLLC slice is set higher than that of the eMBB slice. Hereinafter, eMBB is considered to have a relatively lower priority than the URLLC slice.

As shown in FIG. 4, the NMS server 300 assigns the same frequency band to cells C1 and C2 as radio resources for eMBB. The NMS server 300 then divides the frequencies of the channels other than the radio resources for eMBB into two, and sets the frequency band that cell C1 uses with high priority as a restricted frequency band in C2, and sets the frequency band that cell C2 uses with high priority as a restricted frequency band in cell C1. The restriction in the restricted frequency band indicates, for example, a state in which use is prohibited. The restriction also includes restrictions to suppress interference with the other cell, such as restrictions on transmission power and beam direction.

As shown in FIG. 4, when cells C1 and C2 use high-priority radio resources, the other cell is restricted from using those radio resources, allowing high-priority communication to be carried out on resources with minimal mutual interference.

Figure 5 is a diagram showing an example of wireless resource allocation. As shown in Figure 5, the NMS server 300 allocates the same frequency band to cells C1 and C2 as wireless resources for eMBB. The NMS server 300 then allocates the frequencies of channels other than the wireless resources for eMBB as frequency bands to be used with high priority in cell C2, and as frequency bands with restrictions in cell C1. For example, depending on the environment in which the cell is used, there are cases where high-priority data communication is not required. For example, when high-priority data communication is not required in cell C1, the allocation method shown in Figure 5 may be used. By performing the allocation shown in Figure 5, it may be possible to obtain the same effect as in Figure 4.

<Radio Resource Allocation Processing>
A description will now be given of the wireless resource allocation process in the NMS server 300. When allocating wireless resources, the NMS server 300 executes a wireless resource allocation process S1000.

Figure 6 is a diagram showing an example of a processing flowchart of the wireless resource allocation process S1000. In the wireless resource allocation process S1000, the NMS server 300 constructs RAN slices in each gNodeB 200 (S1100), estimates the amount of required wireless resources for each RAN slice (S1200), sets the TDD frame configuration (S1300), estimates the amount of inter-cell interference (S1400), estimates the interference relationship between cells for each RAN slice (S1500), and allocates resources to the RAN slices (S1600). In the wireless resource allocation process S1000, the NMS server 300 repeats processes 1 to 6. The NMS server 300 performs the wireless resource allocation process S1000, for example, periodically. Each process is described below.

<Process 1: Construction of RAN slice>
In process 1, the NMS server 300 constructs a RAN (Radio Access Network) slice. In process 1, the NMS server 300 constructs a RAN slice by further subdividing a slice constructed in the core network based on QoS based on the characteristics of the RAN. Since the slice is subdivided based on the RAN characteristics, when radio resources are assigned to the RAN slice, it is possible to assign more appropriate radio resources according to the RAN characteristics.

For example, the NMS server 300 constructs a RAN slice based on the positional relationship between the terminal device 100 and the gNodeB 200 in the cell. The positional relationship includes, for example, the distance between the gNodeB 200 and the terminal device 100 during communication. The NMS server 300 subdivides slices of the same QoS into, for example, a RAN slice consisting of traffic of the terminal device 100 located at the cell edge (near the edge of the cell) and a RAN slice consisting of traffic of the other terminal devices 100. This allows the NMS server 300 to allocate radio resources equivalent to the allocation of radio resources in FFR (Fractional Frequency Reuse). FFR is a technology that separates the frequency band of the radio resources to be allocated between, for example, a terminal device located at the cell edge far from the base station device and a terminal device located near the center of the cell close to the base station device, and partially repeats the frequency.

The NMS server 300 may also construct a RAN slice for each terminal device 100. This enables allocation of radio resources according to the RAN characteristics of each terminal device 100.

The NMS server 300 may also collect location information and wireless conditions of the terminal device 100 in order to construct the RAN slice described above.

Furthermore, the NMS server 300 may use the slice constructed in the core network as the RAN slice without subdividing it. For example, if a sufficiently subdivided slice has been constructed on the core network side, the NMS server 300 constructs a RAN slice in a unique correspondence with the slice constructed in the core network.

<Process 2: Estimation of required radio resource amount for each RAN slice>
In process 2, the NMS server 300 estimates (calculates) the amount of required radio resources for each RAN slice in each gNodeB 200.

The NMS server 300 calculates the required packet size for each slice. The NMS server 300 calculates based on, for example, the Guaranteed Flow Bit Rate (GFBR), Maximum Flow Bit Rate (MFBR), Packet Delay Budget (PDB), Averaging Window, and Transmission Time Interval (TTI) length (Reference: 3GPP TS.23.501) that are set (specified) when constructing the slice on the core network side. For example, the calculation is performed using the following formula (1).

Packet size = MFBR (or GFBR) x TTI length (Equation 1)
(TTI<PDB, 1 slot/TTI)
Note that, for example, equation (1) does not take retransmission into consideration. Using equation (1), the upstream/downstream packet sizes are calculated.

Next, the NMS server 300 calculates the amount of radio resources required for packet transmission. The NMS server 300 determines the MCS (Modulation and Coding Scheme) based on a propagation environment map created in advance. The propagation environment map is generated, for example, from past measurement results or simulation results. In determining the MCS, it may be assumed that there is no interference, or that there is an acceptable level of interference.

Based on the determined MCS, the NMS server 300 calculates the number of upstream/downstream resource blocks required for transmitting the estimated required packet size. Note that the NMS server 300 may calculate the number of resource blocks by adding a predetermined margin, since calculating the number of resource blocks using an MCS that assumes no interference may result in insufficient capacity in the actual environment.

The NMS server 300 calculates the number of resource blocks required for each gNodeB 200. For example, the NMS server 300 calculates the total number of resource blocks required by the terminal devices 100 that are wirelessly connected to the gNodeB 200 as the number of resource blocks required by that gNodeB 200.

<Process 3: Setting the TDD frame configuration>
In process 3, the NMS server 300 sets the configuration of a TDD (Time Division Duplex) frame based on the required resource amount estimated (calculated) in process 2.

For example, when the frame configuration is the same between gNodeBs 200 in the same tenant, the NMS server 300 determines the TDD frame configuration according to the average load in the tenant and the load balance of critical cells (e.g., the cell with the highest communication volume, the cell transmitting the highest priority traffic, etc.).

Note that if the TDD frame configuration is fixed (the TDD frame configuration cannot be changed), the NMS server 300 does not need to perform process 3.

<Process 4: Estimation of Inter-Cell Interference Amount>
In process 4, the NMS server 300 estimates the amount of interference between two cells (between the gNodeBs 200, between the gNodeBs and the terminal devices, and between the terminal devices) for each slot. The slots follow the TDD frame configuration determined in process 3. The NMS server 300 estimates and measures the interference between the gNodeBs 200 and the terminal devices of the other cell and the gNodeBs 200 and the terminal devices of its own cell. The NMS server 300 estimates the amount of interference using an environmental model. The environmental model is a model of the layout relationship and propagation environment of the cells (gNodeBs 200, terminal devices) and tenants, and is created from layout information of the site and buildings. The environmental model includes, for example, the position of the terminal device, the antenna beam pattern applied to transmission and reception, and the transmission power. In addition, process 4 is repeatedly executed along with the repeated execution of the wireless resource allocation process. At this time, the environmental model used is also updated successively. This allows the latest environmental model to be used at all times, and an estimation result closer to the actual environment can be obtained.

When estimating the amount of interference, the NMS server 300 evaluates the impact of the amount of interference based on the differences in frame configuration for each base station device. The NMS server 300 also evaluates the impact of the amount of interference based on the offset amount of frame timing (differences in frame configuration, differences in transmission timing, etc.). This allows the NMS server 300 to evaluate the amount of interference taking into account cases where, for example, frame transmission timing and frame length differ for each base station device and the transmission timing of uplink and downlink signals overlap.

Figure 7 is a diagram showing an example of the timing for measuring inter-cell interference. Figure 7 shows an example of the uplink (UL) and downlink (DL) transmission intervals in one slot for each of cell 1 and cell 2. Note that Figure 7 shows a case where frame synchronization is performed in cell 1 and cell 2, but the concept of the measurement interval is the same even if frame synchronization is not achieved.

In addition, in process 4, in addition to the amount of interference in the same channel, the amount of interference in adjacent channels may also be estimated. When estimating the amount of interference in adjacent channels, for example, the leakage power from the adjacent channels is taken into consideration. Also, for example, when there is symbol synchronization between cells and the subcarrier spacing is the same, it is determined that the influence of adjacent carriers is small (limited), and it is not necessary to estimate adjacent carrier interference.

Section (a) is a section where the downlink communication of cell 1 and the downlink communication of cell 2 overlap. This is a section where the degree of interference caused by radio waves transmitted by gNodeB 200 of adjacent cell 2 is measured in a terminal device in cell 1 that receives downlink signals (radio waves) in cell 1. Also, this is a section where the degree of interference caused by radio waves transmitted by gNodeB 200 of adjacent cell 1 is measured in a terminal device in cell 2 that receives downlink signals in cell 2.

Section (b) is a section where the uplink communication of cell 1 and the uplink communication of cell 2 overlap. This is a section where the degree of interference caused by radio waves transmitted by a terminal device communicating in adjacent cell 2, when received by gNodeB 200 of cell 1, is measured in cell 1. Also, this is a section where the degree of interference caused by radio waves transmitted by a terminal device communicating in adjacent cell 1, when received by gNodeB 200 of cell 2, is measured in cell 2.

Section (c) is a section in cell 1 where downlink communication in cell 1 overlaps with uplink communication in cell 2. This is a section in which the degree of interference caused by radio waves transmitted by a terminal device communicating in adjacent cell 2 is measured in a terminal device receiving a downlink signal in cell 1. Section (d) is a section in cell 2 where downlink communication in cell 1 overlaps with uplink communication in cell 2. This is a section in which the degree of interference caused by radio waves transmitted by gNodeB 200 in adjacent cell 1 is measured in reception by gNodeB in cell 2. Note that if the TDD frame configuration is the same between the above two cells, there are no sections (c) and (d), and therefore there is no need to estimate the amount of interference.

The above estimation of the amount of inter-cell interference may be performed using the worst value in the area. This is, for example, a process of estimating or measuring inter-cell interference assuming that a terminal device is located at the cell edge of each cell, or that a terminal device is located at the cell edge of each cell.

Figure 8 shows an example of the location of a terminal device in the area worst value evaluation.

Pattern 1 is a diagram showing an example of a measurement pattern at the timing when gNodeB 200-1 transmits a signal to terminal device 100-1. gNodeB 200-1 is in communication with terminal device 100-1 and transmits a downlink signal S1 to terminal device 100-1. Terminal device 100-1 is, for example, at the cell edge of gNodeB 200-1 and is located in the vicinity of adjacent gNodeB 200-2 and terminal device 100-2 communicating with gNodeB 200-2. As a result, terminal device 100-1 is located in a poor environment where the reception strength of the signal from gNodeB 200-1 of its own cell is weak and the reception strength of the signal from gNodeB 200-2 of another cell and terminal device 100-2 communicating with another cell is strong.

At the timing when terminal device 100-1 receives signal S1, interference at timing (a) is measured (estimated) by measuring the downlink signal transmitted by gNodeB 200-2. Also, at the timing when terminal device 100-1 receives signal S1, interference at timing (c) is measured (estimated) by measuring the uplink signal transmitted by terminal device 100-2.

Pattern 2 is a diagram showing an example of a measurement pattern at the timing when the terminal device 100-1 transmits a signal to the gNodeB 200-1. The terminal device 100-1 transmits an uplink signal S2 to the gNodeB 200-1. The terminal device 100-1 is located, for example, at the cell edge of the gNodeB 200-1. Furthermore, the terminal device 100-2 of the adjacent cell is located near the gNodeB 200-1. As a result, the gNodeB 200-1 is located in a poor environment where the reception strength of the signal from the terminal device 100-1 of its own cell is weak and the reception strength of the signal from the terminal device 100-2 of the other cell is strong. Note that the positions of the gNodeB 200-1 and the gNodeB 200-2 are assumed to be fixed.

The interference at timing (b) is measured (estimated) by measuring the uplink signal transmitted by terminal device 100-2 at the timing when gNodeB 200-1 receives signal S2. Also, the interference at timing (d) is measured (estimated) by measuring the downlink signal transmitted by gNodeB 200-2 at the timing when gNodeB 200-1 receives signal S2.

<Process 5: Estimation of interference relationship for each RAN slice>
In process 5, the NMS server 300 estimates the inter-cell interference relationship for each RAN slice. The NMS server 300 identifies interfering cells for each cell and each RAN slice, and determines cells that have a large influence.

The NMS server 300 determines the interfering cells for each cell and RAN slice in the following priority order (tier):

Tier 1: The effective downlink interference and uplink interference are estimated from the amounts of co-channel interference (a) to (d) from one other cell calculated in process 4, and a cell in which even one of the estimated amounts exceeds the allowable value set for each slice is determined to be an interfering cell in that slice. The effective downlink interference is the same as the amount of interference (a) if both cells have the same TDD frame configuration, and if they have different TDD frame configurations, the calculation is performed based on the amount of interference in sections (a) and (c) according to the ratio of sections (a) and (c). The effective uplink interference is the same as the amount of interference (b) if both cells have the same TDD frame configuration, and if they have different TDD frame configurations, the calculation is performed based on the amount of interference in sections (b) and (d) according to the ratio of sections (b) and (d). A similar determination is made for adjacent channel interference (interference due to radio waves in adjacent frequency bands, not the same frequency), and if there is interference, the cell is stored as having adjacent channel interference.

Tier 2: If the sum of the same channel interference from two cells, excluding the Tier 1 interfering cell, exceeds the allowable value, the cell with the higher interference level is determined to be the interfering cell. A similar determination is made for adjacent channel interference.

Tier 3: If the sum of the same channel interference from three cells, excluding the interfering cells of Tiers 1 and 2, exceeds the allowable value, the cell with the highest interference level among these three cells is determined to be the interfering cell. A similar determination is made for adjacent channel interference.

Note that similar processing may be performed by providing priorities for Tier 4 and above and increasing the number of cells.

<Process 6: Resource Allocation to RAN Slice>
In process 6, interference control (allocation of radio resources) is performed for each interfering cell determined in process 5.

For example, when two cells are interfering with each other (when each determines that the other cell is an interfering cell), the NMS server 300 assigns resources to the interfered cell that do not overlap with the resources assigned to the interfering cell, and sets the resources assigned to the interfered cell as constrained resources in the interfered cell, thereby suppressing interference between the cells and making it possible to keep the degree of interference below a predetermined value. Constrained resources are, for example, resources that are constrained during scheduling (for example, scheduling is prohibited).

The NMS server 300 regards the interfering cell determined in Tier 1 as the highest priority interfering cell and performs an allocation method that can best suppress interference. On the other hand, for the interfering cell determined in Tier 2 or lower, the NMS server 300 uses resources other than those allocated in the higher tier to suppress interference. Resource allocation is explained below.

Figure 9 is a diagram showing an example of resource allocation. For example, cell 1 and cell 2 are frame synchronized, the subcarrier spacing (SCS (Sub Carrier Spacing)) is 60 kHz, the slot length is 0.25 ms, and one subframe consists of four slots. Note that Figure 9 shows an example in which cell 1 and cell 2 use the same slot format.

The NMS server 300 allocates the BWP (Bandwidth part) 1 of the cell 1 as a resource for a high-priority RAN slice (high demand slice: for example, URLLC). The NMS server 300 then allocates the BWP 2 of the cell 1 as an unused restricted resource.

On the other hand, the NMS server 300 assigns BWP1 of cell 2 as an unused constrained resource.

Then, the NMS server 300 assigns BWP2 of cell 2 for the high priority RAN slice.

Furthermore, the NMS server 300 allocates BWP3 of cells 1 and 2 for a low priority RAN slice (e.g., eMBB).

In FIG. 9, cell 1 uses BWP1 for the high-priority RAN slice, and cell 2 does not use BWP1 (or it is a constrained resource; the same applies below). Also, in FIG. 9, cell 2 uses BWP2 for the high-priority RAN slice, and cell 1 does not use BWP2. In this way, by allocating a specific frequency band as a resource for the high-priority RAN slice for each cell, interference with high-priority RAN slices between cells can be suppressed.

Figure 10 is a diagram showing an example of resource allocation. In Figure 10, the BWP assigned to each slice is the same as in Figure 9, but the slot formats of cells 1 and 2 are different. Different slot formats may be used between cells as long as adjacent channel interference can be appropriately controlled.

Figure 11 is a diagram showing an example of resource allocation. As shown in Figure 11, the NMS server 300 suppresses interference by using BWP1 as a resource for a high-priority RAN slice in cell 1 and as an unused resource in cell 2. In this way, if one cell (cell 2 in this case) does not use a high-priority RAN slice, it is not necessary to allocate resources for a high-priority RAN slice to the other cell. In addition, when resources for RAN slices are allocated by separating frequency bands in this way, the slot length may differ for each cell, as shown in Figure 11.

[Second embodiment]
In the second embodiment, process 1 (construction of RAN slices) and process 2 (estimation of required radio resources for each RAN slice) in the radio resource allocation process S1000 will be described.

The NMS server 300 must prepare appropriate RAN slices for traffic with different wireless transmission path quality and required quality for each user, and allocate appropriate wireless resources to each RAN slice. Indicators of required quality include, for example, error rates such as BLER (BLock Error Rate), allowable delay time, and location ratios that indicate the proportion of space that satisfies requirements in a given space.

In order to assign an appropriate RAN slice to a user with high quality requirements, it is necessary to prepare a high-quality RAN slice. However, if radio resources with more margin than necessary are assigned to this RAN slice, the resource utilization efficiency may decrease.

In the second embodiment, we provide processes 1 and 2 that are applicable to users who require high quality and that suppress a decrease in resource utilization efficiency.

<Process 1: Construction of RAN slice>
In constructing a RAN slice, for example, there is a method of selecting a RAN slice by comparing the average wireless quality (e.g., average SINR (Signal to Interference plus Noise Ratio)) for each user with a threshold value determined for each RAN slice. With this method, although it occurs infrequently, if the wireless quality is significantly worse than the average value, the required quality of the user may not be ensured.

In constructing a RAN slice in the second embodiment, a probability distribution of wireless quality for each user is estimated, and by referring to the shape of this distribution, a RAN slice that can satisfy the extremely high required quality of a certain traffic (quality that does not tolerate quality degradation that occurs only occasionally) is constructed and selected. Wireless quality includes, for example, wireless quality between a base station device and a terminal device, and wireless quality between base station devices and between terminal devices.

Figure 12 is a diagram showing an example of a processing flowchart of process 1 in the second embodiment. The NMS server 300 determines the RAN slice candidate to be mapped based on the 5QI of each QoS flow (S1100-1). The 5QI includes indicators related to QoS characteristics such as priority level, packet delay budget, packet error rate, average window size, and maximum data burst amount.

The NMS server 300 instructs a RAN node (e.g., gNodeB, terminal device, etc.) to measure the UE's received SIR value and obtains the measurement results (S1100-2).

Then, the NMS server 300 estimates the distribution of SIR for each UE (terminal device) (S1100-3). As an estimation method, for example, if sufficient measured values are available, a histogram (probability distribution) is calculated. As an estimation method, for example, if sufficient measured values are not available, a histogram is calculated by applying Bayesian estimation.

Since it is difficult to estimate the tail of the distribution from a histogram with high accuracy, the NMS server 300 estimates the variance (variance value) of Bayesian inference with the likelihood distribution as a normal distribution, and estimates the Bayesian prediction interval. The distribution for each UE can be approximated by a single log-normal distribution and can be obtained analytically. Since the measurement values reported from the UEs have been subjected to L3 filtering, the NMS server 300 corrects the estimated variance value according to the filter coefficient. That is, the NMS server 300 corrects the estimated variance value according to the coefficient (filter coefficient) used in the smoothing process used on the measurement values. L3 filtering is, for example, filtering performed on the measurement results in a terminal device. The NMS server 300 can create a more accurate probability distribution by correcting the probability distribution taking into account the coefficient used in filtering.

The NMS server 300 determines the RAN slice for each QoS flow by comparing the UE's SIR distribution prediction interval with the required SIR threshold for each RAN slice (S1100-4) and ends the process.

<Process 2: Estimation of required radio resource amount for each RAN slice>
The NMS server 300 determines (estimates) the amount of radio resources to actually be allocated to each RAN slice so as to satisfy the required quality of the above-mentioned traffic.

The required packet size shown in formula (1) in the first embodiment is the size required for transmission by one user. Since packets from multiple users are multiplexed into a slice, the required slice transmission rate after multiplexing is determined in order to determine the required size of the slice.

One method for this is to calculate the required packet size for all users belonging to one slice, and then add up the calculated results to determine the slice required transmission rate. This method estimates the maximum rate for all users defined by the MFBR, and assumes that all users are multiplexed simultaneously, which may result in excessive allocation of radio resources.

Therefore, in the second embodiment, the NMS server 300 determines the minimum slice transmission rate required to satisfy the required quality of each traffic, taking into consideration the characteristics of each traffic (e.g., packet arrival interval, arrival packet size, maximum allowable delay, etc.) and the number of users to be multiplexed. The NMS server 300, for example, refers to the shape of the SINR probability density distribution, extracts the minimum SINR value that satisfies the extremely high required quality, and determines the amount of radio resources that satisfies the required delay characteristics of the accommodated traffic based on the minimum SINR value.

The NMS server 300 determines the amount of radio resources to be allocated to a RAN slice based on, for example, the minimum transmission rate of the RAN slice and attribute information of the traffic (UE) multiplexed into the RAN slice. The attribute information is, for example, the packet size. The attribute information also includes the data rate and the allowable delay time.

Figure 13 is a diagram showing an example of a processing flowchart of process 2 in the second embodiment. The NMS server 300 calculates the required transmission rate using SNC (Stochastic Network Calculus) based on the traffic information and required QoS (tolerable delay amount) of each RAN slice (S1200-1).

The NMS server 300 calculates the required number of RBs (initial value) for each RAN slice, assuming that the minimum MCS is MCS0 (initial value) (S1200-2). Note that process S1200-2 does not need to be executed if the required reliability is not high (lower than a predetermined value) or if sufficient information regarding the SIR/MCS distribution is available in advance.

The NMS server 300 estimates the Bayesian prediction interval based on the latest SIR/MCS distribution of each RAN slice, and estimates the minimum MCS (MCSmin) for each RAN slice (S1200-3). Estimation methods include MCMC and particle filters, but the SIR distribution of the RAN slice is a mixture model, which imposes a large computational load. Therefore, the NMS server 300 performs estimation from the individual SIR distribution of the slice-configured UE.

The NMS server 300 calculates (data assimilates) the number of RBs required for each RAN slice (updated value) assuming the minimum MCS (MCSmin) estimated above (S1200-4), and ends the process.

In the second embodiment, the minimum possible transmission rate is estimated (calculated) from a probability distribution, which prevents the allocation of excessive wireless resources with an unnecessary margin, while ensuring the quality required by the user even when the wireless quality is significantly worse than the average.

[Other embodiments]
The allocation of radio resources shown in the first and second embodiments is merely an example. Therefore, for example, the slot configuration and the subframe configuration may be changed. Furthermore, the allocation patterns of radio resources shown in the first and second embodiments may be combined with each other.

Also, the allocation of uplink/downlink allocations may be changed. Also, in the wireless resource allocation process, some processes may not need to be performed. For example, if the TDD frame configuration is fixed, process 3 may not need to be performed.

10: communication system 100: terminal device 300: NMS server 310: CPU
320: Storage 321: Radio resource allocation program 3211: RAN slice construction module 3212: Required radio resource amount estimation module 3213: TDD frame configuration setting module 3214: Inter-cell interference amount estimation module 3215: Cell interference relationship estimation module 3216: Resource allocation module 322: Base station device control program 330: Memory 340: Communication circuit 400: Core network 500: Core network server

Claims (13)

A control device in a communication system having a plurality of base station devices, terminal devices wirelessly connected to the base station devices, and a control device that controls the base station devices,
an estimation unit that calculates a required radio resource amount for each of a plurality of slices classified based on a required quality, and estimates, for each of the plurality of base station devices, an amount of interference between the base station devices, an amount of interference between the base station device and the terminal device, and an amount of interference between the terminal devices;
an allocation unit that allocates, based on the result of the estimation, a first resource for a first slice, a second resource for a second slice having a required quality higher than that of the first slice, and a constrained resource having a usage constraint, within radio resources of a first base station device;
A control device in which the allocation unit allocates resources other than the second resource and the constrained resource as the first resource, and allocates a resource used as a third resource for the second slice of a second base station device as the constrained resource. The control device according to claim 1 , wherein the allocation of resources for the first slice in the first base station device allocates resources different from resources allocated in a second base station device that is affected by interference to the second slice. The control device according to claim 1 , wherein the plurality of slices are slices obtained by further classifying constituent slices formed by a core network according to a radio environment. The control device according to claim 1 , wherein the estimation unit estimates the amount of interference by estimating interference that occurs when the base station devices use the same channel. The control device according to claim 4 , wherein, in estimating the amount of interference, the estimation unit further estimates interference occurring when the base station devices use adjacent channels. The control device according to claim 1 , wherein, in estimating the amount of interference, the estimation unit estimates interference that occurs when the terminal device is located near an edge of a cell formed by the base station device. The control device according to claim 1 , wherein the estimation unit performs an influence evaluation of the amount of interference based on a difference in frame configuration for each of the base station devices or an offset amount of frame timing in estimating the amount of interference. The control device according to claim 1 , wherein the estimation unit constructs a slice so as to satisfy the required quality based on a probability distribution of wireless quality between the base station device and the terminal device. The control device according to claim 8 , wherein the estimation unit determines the amount of radio resources to be allocated to the slice based on a minimum possible transmission rate of the slice and attribute information of traffic multiplexed into the slice. The control device according to claim 9 , wherein the attribute information includes any one or a combination of a data rate, a packet size, and an allowable delay time. The estimation unit calculates a minimum possible transmission rate that satisfies the required quality in the slice as follows:
The control device according to claim 8 , wherein the estimation is based on a distribution of transmission rates in the slices. The control device according to claim 8 , wherein the estimation unit collects measurement values of wireless quality from at least one of the base station device and the terminal device, and corrects the estimated probability distribution according to a coefficient used for smoothing the measurement values. A control method in a control device of a communication system having a plurality of base station devices, terminal devices wirelessly connected to the base station devices, and a control device that controls the base station devices, comprising:
an estimation step of calculating a required radio resource amount for each of a plurality of slices classified based on a required quality, and estimating, for each of the plurality of base station devices, an amount of interference between the base station devices, an amount of interference between the base station device and the terminal device, and an amount of interference between the terminal devices;
and an allocation step of allocating, based on the result of the estimation, within radio resources of a first base station device, a first resource for a first slice, a second resource for a second slice having a required quality higher than that of the first slice, and a constrained resource having a usage constraint;
A control method in which, in the allocation step, resources other than the second resource and the constrained resource are allocated as the first resource, and a resource used as a third resource for the second slice of a second base station device is allocated as the constrained resource.

Images