KR101242049B1 - Uplink Scheduling Method Based On Maximum PF Selection And Apparatus For Performing The Same - Google Patents

Abstract

An uplink scheduling method and a scheduling apparatus are disclosed. In the uplink scheduling method, in a scheduling apparatus of a base station connected with a plurality of terminals through a wireless channel, N rows corresponding to positive integer-number terminals and KK are positive integer-number resource blocks (RB) Receiving a metric value of the N terminals for each RB in an N × K two-dimensional array consisting of K columns corresponding to the N columns; and based on the maximum metric value for each RB column of the K RB columns in the scheduling apparatus, the N Performing resource allocation in the x K two-dimensional array. Resource allocation path considerations are more flexible than scheduling techniques that extend existing resource allocation only to neighboring RBs, and have simpler and higher throughput with lower complexity than existing RME and IRME scheduling algorithms. Fairness can be guaranteed.

Description

Uplink Scheduling Method Based On Maximum PF Selection And Apparatus For Performing The Same}

The present invention relates to scheduling in a mobile communication base station, and more particularly, to an uplink scheduling method and a scheduling apparatus in a mobile communication base station.

3GPP (3rd Generation Partnership Project) is the next generation of mobile communication technology (Long Term Evolution) LTE (Long Term Evolution) has recently been addressed the issue of the processing of a vast amount of radio resources due to the increased demand for data services in the mobile terminal.

Orthogonal Frequency Domain Multiple Access (OFDMA) is used in the downlink of LTE, and Single Carrier Frequency Division Multiple Access (SC-FDMA) is used for the uplink. While OFDMA is strong in multipath fading and has high frequency efficiency and scalability, OFDMA has high peak-to-average power caused by the rapid change in RF power fluctuations within a single OFDM symbol during multicarrier modulation. ratio causes a problem in the power of the mobile terminal. SC-FDMA is used in the LTE uplink because of its single carrier feature and multiple access.

Resource allocation in the LTE frequency side is basically carried out by twelve carriers are bundled into one physical resource block (PRB). SC-FDMA and OFDMA are performed by Frequency Domain Packet Scheduling (FDPS), which allocates a resource block (RB), which is a part of system frequency resources, to each user in consideration of the state of a radio channel.

In the FDPS in the LTE uplink, the channel gain of each user in the frequency domain is used as the input of the scheduler. To this end, each user transmits a sounding reference signal (SRS) to a scheduling node-eNodeB, and the SRS signal is used for reporting Cchannel quality indicator (CQI) for uplink. The CQI information is updated at the base station at every transmission time interval (TTI). The FDPS determines the total number of users and RB values to be scheduled using the CQI information thus obtained and information on the number of users who need to perform retransmission received from the HARQ manager. The RB is composed of 12 subcarriers. Since the user position and channel state for each frequency are different, all RB values have different state values. A proportional fair (PF) algorithm may be used as a scheduling metric value of each user configured as a state value of the RB. The PF algorithm is required to prevent resource allocation from being concentrated on a specific user who has a good channel condition in scheduling, and to allocate a certain level of resources to all users to satisfy throughput and fairness.

SC-FDMA has a continuity constraint that requires localized frequency division multiple access (LFDMA) to maintain a particularly low PAPR, so that the RB to be allocated to each user must be contiguous. Resource allocation in the uplink uses channel-aware scheduling, which enables high multi-user diversity gain in conjunction with the above continuity constraint. Uplink scheduling is performed based on a scheduling metric that is generated by using information of consecutive radio channels ( ie , signal to interference and noise ratio, SINR) for each user's channel. For example, this is done based on a property fair (PF) metric to ensure a certain level of fairness.

As mentioned earlier, resource allocation in LTE uplink requires consecutive allocations in the frequency domain of RBs to be assigned to one user for each timeslot ( ie , sub-frame) in order to maintain a single carrier characteristic ("Proportional Fair Scheduling"). of Uplink Single-Carrier FDMA Systems, "IEEE Proc. of the PIMRC, Sep. 2006, J. Lim, HG Myung, and DJ Goodman.

Users with the highest PF metrics in the "Channel-Aware Scheduling Algorithms for SC-FDMA in LTE Uplink" (IEEE Proc. Of the PIMRC, Dec. 2008, L. Temino, G. Berardinelli, S. Fattasi, and P. Mogensen) Three techniques have been proposed to extend neighbors. The best evaluation result of the performance evaluation of the three proposed scheduling schemes is RME (Recursive Maximum Expansion).

1 is a conceptual diagram illustrating a resource allocation state when an RME algorithm is performed in an arrangement consisting of four UEs and seven RBs.

The RME algorithm selects a maximum PF metric from a multidimensional array of PF metrics input by a frequency domain scheduler, and then schedules the resource by extending resource allocation to neighboring RBs based on the maximum PF metric.

Specifically, referring to FIG. 1, the RME algorithm is performed by the following procedure. First, the UE and RB having the largest PF metric in the 2D array M are found and resources are allocated. Next, based on the assigned metric value, it is determined whether the values of the neighboring RBs on the right and left sides are the highest among all UEs of the RB and determine whether to assign them. This extension of resource allocation continues until another UE with a higher value of the RB appears than the metric to be compared. When the resource allocation is extended, the allocated UEs (rows) and RBs (columns) are deleted from the array. The scheduler selects the highest PF metric among the remaining arrays and repeats the above process until all UEs are allocated. If all UEs are allocated and all RBs are not allocated, the allocation of the remaining RBs is determined by the UE with the higher metric value of both UEs already allocated according to resource continuity constraints.

Specifically, referring to FIG. 1, the RME algorithm is performed by the following procedure. A resource is allocated by finding the UE and RB (50 of RB 6 of UE 2 in FIG. 1) having the highest PF metric values in the entire two-dimensional array M. FIG. Then, based on the assigned highest PF metric value (UE2 and RB 6), the metric values of neighboring (left and right) RBs have the highest value among the values of the corresponding RB column (including all UEs having the corresponding RB). If it has the largest value, assign it (as 42 is the largest in the left RB5 column, 42 is allocated and 40 in the right RB7 is not the largest value in the RB7 column), so it is extended to the neighboring RB. . This extension of resource allocation continues until another UE with a higher value of the RB appears than the metric to be compared. After the resource allocation is extended, the rows to which the allocated UE belongs and the columns to which the allocated RB belongs are deleted from the array. The scheduler selects the highest PF metric among the remaining arrays and repeats the above process until all UEs are allocated. If all UEs are allocated and all RBs are not allocated, the allocation of the remaining RBs is determined by the UE with the higher metric value of both UEs already allocated according to resource continuity constraints. Specifically, since 45 of UE1 and RB 7 are the highest values except for the excluded UE 2, RB 5, and RB 6, the RB extension process is performed for left and right RBs of RB 7 belonging to UE1. Is applied, 43 of RB 6 is not allocated. Next, of the remaining arrangements, 35 of UE3 and RB 4 are the largest metric values, and resource allocation, and if the RB extension is reapplied to the left and right RBs of RB 4 belonging to UE3, 28 of RB 3 is assigned to the excluded UE and RB. It allocates resources because it is the largest in the excluded RB 3 column, and since 25 of RB 5 belong to the excluded RB 5, it does not allocate resources. Then, in the remaining arrangements except the excluded UE and RB columns, 34 of UE 4 and RB 2 have the largest metric values, and then allocate resources, and apply RB extension to left and right RBs of RB 2 belonging to UE 4 again. In this case, since 30 of RB1 is the largest in the RB 1 column excluding the excluded UE and the RB, resource allocation is performed. Since 15 of RB 3 belongs to the excluded RB 3, it is not allocated. The existing RME algorithm assigns a metric value to each UE through such PF scheduling.

However, the conventional RME algorithm described above does not consider various resource allocation paths because the resource allocation is always extended by neighboring RBs.

Improved Recursive Maximum Expansion Scheduling Algorithms for Uplink Single Carrier FDMA System (IEEE Proc. Of the VTC Spring, Jun. 2010, F. Liu, X. She, L. Chen, and H. Otsuka) Expansion technique is proposed to compensate for the disadvantages in terms of throughput of the existing RME algorithm.

2 is a conceptual diagram illustrating a resource allocation state when an IRME algorithm is performed in an arrangement consisting of four UEs and seven RBs.

Referring to FIG. 2, the IRME is a technique first announced in 2010 and extends only to neighboring RBs, and has a ranking threshold ( Tr) than the RME algorithm that performs resource allocation. value) to consider resource allocation paths more flexibly. The Tr value refers to a range of resource extension performed to the neighboring RB by the value set as the input value. Specifically, the Tr value refers to the Tr-th value of the RB in the corresponding RB column. That is, whether the resource is allocated up to the Tr-th higher value among the metric values of the neighboring RB by the Tr value is considered.

IRME algorithm is performed as follows. Hereinafter, the case where the ranking threshold value Tr is 2 will be described as an example. First, similarly to the RME, the UE having the largest metric value and the RB (50 of RB 6 of the UE 2 in FIG. 2) in the entire two-dimensional array M are performed to perform resource allocation, and then the UE 2 having the highest metric value. And perform RB extension on the neighboring (right and left) RB metric values of RB 6 and continuously assign them until a higher value of the other UE of that RB appears. Column (column) is RB expansion to the procedure is complete, the UE is assigned a row (row) belonging to and assigned RB belongs is in the metric value and remain excluded from the entire array Tr Apply a value to find the value to assign. If Tr If the value is 1, it is performed in the same manner as the RME allocation process. Repeat the above process and give Tr A resource is allocated by finally determining a path having the highest sum of all metric values among resource assignable paths considered as the number of values. IRME consequently Tr The condition of the algorithm except for the value is the same as that of the RME. For example, the RB extension for neighbor (left and right) RB metric values of UE 2 and RB 6 is Tr If the value is 2, the first and second large values are selected from the left RB5 of the RB 6 to which UE 2 belongs, but the first to select the result of the case where the Global metric value (sum of the metrics of the entire matrix) is the highest as the final allocation path. And the second largest value to allocate the resource. That is, since 42 is the largest in column 5 of RB, 42 is allocated and 40 is the second largest value in column 7 of RB.

However, the above-described IRME has a disadvantage in that it is still more flexible than RME, has low fairness, and does not consider various resource allocation paths.

In general, the state of a channel (eg SINR) is closely related to time and frequency, but the association in the frequency domain is not as strong as that in the time domain (frequency selective fading distortion). This means that channel conditions are somewhat related in the overall frequency domain but not in one RB (small-scale variation).

The existing RME and IRME algorithms are extended according to the method of comparing the values of both neighboring RBs of the maximum PF metric value after searching the maximum PF metric value of the 2D array.

Existing RME and IRME algorithms attempt to allocate resources by extending scheduling to neighboring values based on selected metric values, and thus do not cope effectively with frequency selective fading.

In addition, the resource allocation scheme of the existing RME and IRME algorithm has a limitation in scheduling flexibility, and in the case of the IRME, the complexity increases as consideration of many allocation paths.

When scheduling is performed by each algorithm of FIGS. 1 and 2, Tr The IRME algorithm of value 2 considers one more path of resource allocation, which is better than RME. The RME algorithm is simple and provides better throughput (i.e., sum of metric values whose final allocation is determined) than static scheduling algorithms such as round robin.

However, the RME algorithm is an algorithm that depends only on the metric of the neighbor RB and does not provide flexible scheduling. IRME Algorithm Tr The value compensates for some of the disadvantages of the RME by considering the resource allocation path in various ways. But fixed Tr The value does not have agile and flexible response to the rapidly changing number of UEs and has the disadvantage that the scheduling complexity increases.

The first object of the present invention is to improve the throughput of the system while satisfying the continuity constraint that the RB should be allocated continuously in the frequency domain, and to ensure a lower complexity and an improved level of fairness than the existing scheduling algorithm. To provide.

In addition, a second object of the present invention is to provide a scheduling apparatus for performing the uplink scheduling method.

In addition, a third object of the present invention is to provide an apparatus including a processor for performing the uplink scheduling method.

In the uplink scheduling method according to an aspect of the present invention for achieving the first object of the present invention, the NN corresponds to a positive integer-number of terminals in the scheduling apparatus of the base station connected through a wireless channel with a plurality of terminals N rows and KK are N x K two-dimensional arrays of K columns corresponding to positive integer-resource blocks (RBs), and receiving metric values of the N terminals for each RB; And performing resource allocation in the N × K two-dimensional array based on a maximum metric value for each RB column of K RB columns in a scheduling apparatus.

In the scheduling apparatus, the resource allocation in the N × K two-dimensional array based on the maximum metric value for each RB column of the K RB columns may be considered by considering all the maximum metric values for each RB column belonging to the K RB columns. This can be done until resource allocation is complete for all RBs.

In the scheduling apparatus, resource allocation in the N × K two-dimensional array based on the maximum metric value for each RB column of the K RB columns may include selecting a maximum metric value for each RB column of the K RB columns and selecting a first metric value. Comprising a set, and selecting the maximum metric values included in the first set in descending order to determine whether to satisfy the resource continuity constraint.

 In the scheduling apparatus, performing resource allocation in the N × K two-dimensional array based on a maximum metric value for each RB column of K RB columns may be performed in succession with another metric value in which the selected maximum metric value has already been allocated to the terminal. If the resource allocation is not possible because the resource continuity constraint is not satisfied because the resource is not located in the RB, the resource allocation is performed by checking the metric values of the other UEs in the RB column including the selected maximum metric value in descending order. It may further comprise a step.

Performing resource allocation in the N × K two-dimensional array based on the maximum metric value for each RB column of the K RB columns in the scheduling apparatus includes all of the RB columns having the maximum metric value selected during the resource allocation. If the metric value is not assignable, the method may further include suspending allocation of the corresponding RB column and performing resource allocation by selecting a next higher metric value.

The metric value may be a PF (Proportional Fair) metric value calculated based on channel information and HARQ information of each UE for each time slot. The uplink scheduling method may be applied for continuous allocation of RBs in an LTE mobile communication system using SC-FDMA in uplink.

According to an aspect of the present invention for achieving a second object of the present invention, a scheduling apparatus of a base station connected with a plurality of terminals through a wireless channel, where NN denotes N rows and KK corresponding to positive integer-terminals. The maximum metric for each RB column of the K RB columns by receiving the metric values of the N terminals for each RB in an N x K two-dimensional array consisting of K columns corresponding to positive integer-Resource blocks (RBs). And a scheduler that performs resource allocation in the N × K two-dimensional array based on the value.

An apparatus including a processor for use in a mobile communication base station connected via a wireless channel with a plurality of terminals according to an aspect of the present invention for achieving the third object of the present invention, wherein the processor is NN is a positive integer N rows and KKs corresponding to N UEs are N × K two-dimensional arrays of K columns corresponding to positive integer-Resource Blocks (RBs). It is configured to perform resource allocation in the N × K two-dimensional array based on the maximum metric value for each RB column of K RB columns.

According to the uplink scheduling method as described above and the scheduling apparatus for performing the same, it is necessary to consider the maximum metric value for each RB column first while complying with the continuity constraint of continuously allocating resources (RBs) to all UEs in the frequency domain. Scheduling can be effectively performed even with a channel change. That is, unlike the ranking threshold used in the existing IRME algorithm, scheduling is performed to consider the input parameter values more variously.

Specifically, the uplink scheduling method selects a maximum metric value for each RB column among PF metrics for each UE and RB inputted to the 2D array M to determine whether to allocate resources according to resource continuity constraints starting from a large metric value. By confirming, the resource allocation path consideration can be performed more flexibly than the scheduling method of extending the existing resource allocation to the neighbor RB only. In addition, it has a lower complexity than the existing RME and IRME scheduling algorithm, it can have a simple and high throughput, and can guarantee an improved level of fairness.

In addition, scheduling is performed in consideration of the maximum PF metric value for each RB column, thereby enabling more flexible scheduling even in a sudden change in the number of users.

1 is a conceptual diagram illustrating a resource allocation state when an existing RME algorithm is performed in an arrangement consisting of four UEs and seven RBs.
FIG. 2 is a conceptual diagram illustrating a resource allocation state when an existing IRME algorithm is performed in an arrangement consisting of four UEs and seven RBs.
3 shows a two-dimensional array M composed of PF metric values generated based on each user UE and RB in each time slot.
4 illustrates a mobile communication system including a base station and a terminal to which a scheduling algorithm is applied according to an embodiment of the present invention.
5 is a conceptual diagram illustrating a resource allocation state when a scheduling algorithm according to an embodiment of the present invention is performed in an arrangement consisting of four user UEs and seven RBs, and FIG. 6 is a scheduling diagram according to an embodiment of the present invention. A flowchart for explaining the algorithm.
7 is a flowchart illustrating an operation of implementing a scheduling algorithm in pseudo code according to an embodiment of the present invention.
8 is a graph showing a change in fairness according to the number of users for the conventional RME and IRME technique and the MSCC scheduling method according to an embodiment of the present invention.
9 is a graph showing a change in system throughput according to the number of users for the existing RME and IRME techniques and the MSCC scheduling method according to an embodiment of the present invention.

While the invention is susceptible to various modifications and alternative forms, specific embodiments thereof are shown by way of example in the drawings and will herein be described in detail.

It should be understood, however, that the invention is not intended to be limited to the particular embodiments, but includes all modifications, equivalents, and alternatives falling within the spirit and scope of the invention.

The terms first, second, etc. may be used to describe various components, but the components should not be limited by the terms. The terms are used only for the purpose of distinguishing one component from another. For example, without departing from the scope of the present invention, the first component may be referred to as a second component, and similarly, the second component may also be referred to as a first component. And / or < / RTI > includes any combination of a plurality of related listed items or any of a plurality of related listed items.

When a component is referred to as being "connected" or "connected" to another component, it may be directly connected to or connected to that other component, but it may be understood that other components may be present in between. Should be. On the other hand, when an element is referred to as being "directly connected" or "directly connected" to another element, it should be understood that there are no other elements in between.

The terminology used herein is for the purpose of describing particular example embodiments only and is not intended to be limiting of the present invention. Singular expressions include plural expressions unless the context clearly indicates otherwise. In this application, the terms "comprise" or "have" are intended to indicate that there is a feature, number, step, operation, component, part, or combination thereof described in the specification, and one or more other features. It is to be understood that the present invention does not exclude the possibility of the presence or the addition of numbers, steps, operations, components, components, or a combination thereof.

Unless defined otherwise, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with the meaning in the context of the relevant art and are to be interpreted in an ideal or overly formal sense unless explicitly defined in the present application Do not.

Hereinafter, with reference to the accompanying drawings, it will be described in detail a preferred embodiment of the present invention. In order to facilitate the understanding of the present invention, the same reference numerals are used for the same constituent elements in the drawings and redundant explanations for the same constituent elements are omitted.

The terminal may be a mobile station (MS), a user equipment (UE), a user terminal (UT), a wireless terminal, an access terminal (AT), a terminal, a subscriber unit, a subscriber station (SS) ), A wireless device, a wireless communication device, a Wireless Transmit / Receive Unit (WTRU), mobile node, mobile or other terms. Various embodiments of the terminal may be used in various applications such as cellular telephones, smart phones with wireless communication capabilities, personal digital assistants (PDAs) with wireless communication capabilities, wireless modems, portable computers with wireless communication capabilities, Devices, gaming devices with wireless communication capabilities, music storage and playback appliances with wireless communication capabilities, Internet appliances capable of wireless Internet access and browsing, as well as portable units or terminals incorporating combinations of such functions However, the present invention is not limited thereto.

A base station generally refers to a fixed point for communicating with a terminal, and includes a base station, a Node-B, an eNode-B, a base transceiver system (BTS), and an access point. access point).

3 shows a two-dimensional array M composed of PF metric values generated based on each user UE and RB in each time slot.

는 사용자 i와 RB j의 PF 매트릭값을 의미하기 때문에 M은

로 나타낼 수 있다 (즉,

).M has the size of [ixj].

Is the PF metric of user i and RB j, so M is

Can be represented as

).

As a result, the FDPS scheduler is performed to satisfy the following Equation 1 with the input metric value of FIG. 3 every time slot.

는 0 또는 1값을 가지는 지시자(indicator)로서 해당

값에 대한 최종적인 자원 할당 여부를 나타낸다 (즉,

).

Is an indicator with zero or one value

Indicates whether or not the final resource is allocated for a value (ie

).

4 illustrates a mobile communication system including a base station and a terminal to which a scheduling algorithm is applied according to an embodiment of the present invention.

4, a mobile communication system according to an embodiment of the present invention includes a plurality of terminals 200 and a base station 100.

The base station 100 is connected to a core network through a wired network 400, and transmits and receives data to and from a plurality of terminals 200 and a wireless channel environment 300. The plurality of terminals includes UE 1 20-1, UE 2 20-2,..., UE N 20 -N (N is a positive integer). The base station 100 includes a scheduling device 120 that includes a scheduler 122.

The scheduling apparatus 120 of the base station 100 is based on the scheduling information provided from the terminal 200 in the scheduler 122 (Maximum PF Selection with Contiguity for the uplink channel of the corresponding terminal according to an embodiment of the present invention) Constraint) Performs scheduling. The scheduling information may include a PF (Proportional Fair) metric value calculated based on channel information and HARQ information of each terminal. The MSCC scheduling method according to an embodiment of the present invention will be described later.

The scheduling apparatus 120 may further include a storage 120 that stores the scheduling information.

Application of the MSCC scheduling method according to an embodiment of the present invention is not limited to LTE uplink SC-FDMA, hereinafter, for example, scheduling performed by the scheduler 122 for FDPS in LTE uplink SC-FDMA. Describe the algorithm. Hereinafter, a case of having a two-dimensional array consisting of four user UEs and seven RBs from RB1 to RB7 will be described as an example.

5 is a conceptual diagram illustrating a resource allocation state when a scheduling algorithm according to an embodiment of the present invention is performed in an arrangement consisting of four user UEs and seven RBs, and FIG. 6 is a scheduling diagram according to an embodiment of the present invention. A flowchart for explaining the algorithm.

The scheduling algorithm according to an embodiment of the present invention finds the maximum metric value for each RB column and checks whether the resource continuity constraint is satisfied while allocating the resource by first considering the maximum metric value for each RB column.

Referring to FIG. 6, first, metric values of all UEs are input to each RB column in a two-dimensional array M (step 601), and a set S is configured by selecting a maximum metric value for each RB column (step 620). The maximum metric is found in set S and it is checked whether the resource continuity constraint is met (step 630).

If the maximum metric selected cannot be allocated because it is not located in a contiguous RB with other metric values already assigned to the terminal and cannot be allocated because the continuity constraint of the resource is not satisfied, the other UE in the RB column containing the selected maximum metric is not included. Resource allocation is performed by checking metric values one by one in descending order (step 640).

If all the metric values in the RB column having the maximum metric value selected during the resource allocation are not available for resource allocation, the resource allocation of the RB column containing the currently selected maximum metric value is suspended and the next higher metric value is selected to allocate the resource allocation. Perform (step 650).

If it is determined that resource allocation is completed to all RBs (step 660), the resource allocation process is terminated when resource allocation is completed to all RBs.

Referring to FIG. 5, a scheduling algorithm according to an embodiment of the present invention will be described using an arrangement consisting of four user UEs and seven RBs as an example.

1) Construct the set S by selecting the highest value for each RB column in a given two-dimensional array, and arrange the values contained in the set S in descending order (50-45-42-38-33-34-30).

2) Resource allocations are checked in a high order in a state in which the values included in the set S are arranged in descending order (50-45-42).

3) If the maximum metric value for each selected RB column does not meet the resource continuity constraint because it is not located in a contiguous RB with other metrics already assigned to the UE, the maximum metric value for the selected RB column includes the maximum metric value for each selected RB column. The next higher metric value is checked for assignment (38 instead of 40, 30 instead of 33).

4) If all the metric values in the RB column that have the metric value selected during resource allocation are not assignable, the allocation of the corresponding RB column is suspended and the next higher metric value is selected to perform resource allocation. That is, if all of the metric values from UE 1 to UE 4 in the RB 4 column are not allocated, the allocation of the RB 4 column is suspended, the allocation of the next metric value of the RB 2 column first, and then the RB 4 is checked again.

7 is a flowchart illustrating an operation of implementing a scheduling algorithm in pseudo code according to an embodiment of the present invention. Table 1 shows a scheduling algorithm according to an embodiment of the present invention as a pseudo code. All notations used in FIG. 7 are used the same as the pseudo code of Table 1.

Like the basic FDPS system, the scheduler 122 receives PF metric values of all UEs for each RB column and performs scheduling until all RBs are allocated.

The resource allocation procedure of the MSCC according to an embodiment of the present invention is a step A (steps 703, 705, 707, 709, 711) and step B (steps 721, 723, 725, 727) after the initialization process (step 701). , 729, 731).

로 정의되며 n과 k는 각각 해당 메트릭값을 가리키는 UE와 RB를 의미한다. 집합 A는 자원 할당 여부를 나타내는 지시자(indicator)

들의 집합으로 S와 동일한 크기의 배열에 모든 값은 0으로 초기화된다. In the initialization process, the scheduler 122 generates a two-dimensional array ( ie , set U) for the UE and the RB using the input metric value, and finds the maximum metric value for each RB column (step 701). The set U means a collection of all scheduling metric values based on these N UEs and K RBs. The set S consists of the maximum metric values for each of these RBs. The selected metric value

N and k mean UE and RB indicating corresponding metric values, respectively. Set A is an indicator indicating whether to allocate resources

All values are initialized to zero in an array of the same size as S.

Step  A

이 자원의 연속성 제약 만족 여부를 확인한다(단계 703, 705, 707, 709, 711). Step A finds the highest metric value in set S and selects the selected metric value.

It is checked whether this resource satisfies the continuity constraint (steps 703, 705, 707, 709, and 711).

을 가지고 있는 UE에게 자원 할당이 처음이거나 할당받기로 결정된 자원과 연속적인 RB에 위치해야 하는 것을 의미한다. The continuity constraint of the resource is the selected metric value.

This means that the resource allocation for the UE having the first or the resource determined to be allocated must be located in consecutive RBs.

이 자원의 연속성 제약 만족하여 자원 할당이 결정되면, 스케줄러(122)는 최종 할당된 PF 매트릭 값들을 나타내는 집합 A의 지시자(indocator)

를 1로 변경한다(

←1, 단계 711). 변경된 지시자의 값은 해당 UE에게 그 RB가 할당되는 것을 의미한다. 또한, 집합 A에 선택된 매트릭 값이 포함되면, 스케줄러(122)는 집합 U에서 할당된 RB의 모든 메트릭 값들 (i.e., 집합 R)을 제외한다(U ← U - R, 단계 711). 집합 R은 선택된 매트릭 값

을 가지고 있는 RB의 모든 값들의 구성을 의미하며 선택된 값이 할당될 수 없게 된 경우 사용된다. 집합 U는 각 RB 열별 최대 매트릭 값들이 구성된 집합 S에서 선택된 메트릭 값이 제외되고 남은 값들의 모임을 의미한다. Selected metric value

If the resource allocation is determined by satisfying the continuity constraint of this resource, the scheduler 122 is an indicator of set A indicating the last assigned PF metric values.

Is changed to 1 (

← 1, step 711). The value of the changed indicator means that the RB is allocated to the corresponding UE. In addition, if the selected metric value is included in the set A, the scheduler 122 excludes all the metric values ( ie , set R) of the RBs allocated in the set U (U ← U-R, step 711). Set R is the selected metric

This is the configuration of all the values of the RB that have a value of. The set U means a collection of remaining values after the selected metric value is excluded from the set S having the maximum metric values for each RB column.

Step  B

이 자원의 연속성 제약으로 자원 할당을 할 수 없게 된 경우, 선택된 메트릭 값

을 포함하고 있는 RB 열의 다른 UE의 값(i.e.,

)들을 내림차순으로 하나씩 확인하여 자원 할당을 시도한다(단계 721, 723, 725, 727, 729, 731). As a result of the determination of step 709, the selected metric value

Selected metric value if this resource continuity constraint prevents resource allocation.

The value of the other UE in the RB column that contains (ie,

) And try to allocate the resources one by one in descending order (steps 721, 723, 725, 727, 729, and 731).

If the resource allocation cannot be determined even after comparing all the metric values of the corresponding RB column, the resource allocation of the RB column containing the currently selected metric value is suspended and the next higher metric value is selected from the set of maximum metric values for each RB column. ( Ie , X ← X + 1, steps 720 to 727).

The resource allocation of the reserved RB column is checked again whenever the allocation of the ongoing RB is determined. Since the scheduling method according to an embodiment of the present invention is performed until the set S is empty and all RBs are allocated according to the set A (steps 741 and 743), the reserved RB column may eventually be assigned some metric value. There is nothing else.

Referring to FIG. 5 again, the two-dimensional array of FIG. 5 is a result when the example used while describing the conventional RME and IRME techniques is applied to the MSCC scheduling method according to an embodiment of the present invention. As described in the flowchart of FIG. 7, the maximum metric values of each RB column are found in a given two-dimensional array M, and the continuity of resources is compared one by one from the highest metric value. The dotted circle in FIG. 5 is selected as the maximum metric value in the corresponding RB column, but it is excluded because another metric value of the corresponding RB column is allocated because the resource continuity constraint is not satisfied. As a result of performing the flowchart of FIG. 7, it can be seen that the sum of the metric values determined to be allocated to all RBs is 272, which is larger than the sum obtained when the RME 264 and the IRME 266 are performed.

The reason why the MSCC scheduling method according to an embodiment of the present invention may have a higher metric sum than the conventional RME and IRME schemes is extended to the neighbor of the corresponding RB after selecting one maximum metric value as in the conventional RME and IRME schemes. This is because the metric values are not compared to perform resource allocation, but the maximum metric values are preferentially assigned to each RB column. Therefore, the scheduling method according to an embodiment of the present invention can perform resource allocation more dynamically and flexibly at every execution.

Pseudo code ( Pseudo Code )

의 자원의 연속성을 만족 여부를 파악한다. 만약 선택된 메트릭 값

이 연속성 제약을 만족하지 못하면, 변수 y를 이용하여 선택된 매트릭 값

을 포함하고 있는 RB 열의 다른 값이 큰 순서대로 반복 확인된다(i.e., step B). 선택된 RB 열의 모든 매트릭 값이 연속성을 만족하지 못하는 경우, 변수 x를 증가시켜 집합 S의 다음 메트릭 값을 우선 수행하게 된다 (i.e., 변수 y가 step B에서 N값을 초과했을 경우). 선택된 새 매트릭 값의 자원 할당이 완료되면 스케줄러(122)는 보류한 이전 메트릭 값을 새롭게 바뀐 배열 상태에서 자원할당이 가능한지 다시 확인한다.The pseudo code procedure of the MSCC according to an embodiment of the present invention is shown in Table 1. The pseudo code has the same procedure and notation as the flowchart of FIG. Full scheduling is performed until all RB column maximum metric values are considered and resource allocation of the RB is completed. This means that iterations are repeated until the metrics set in set S are completely lost (step A). Scheduler 122 selects the selected metric value.

Determine whether the resource continuity is met. If the selected metric value

If this continuity constraint is not met, the metric selected using the variable y

The other values in the RB column containing are repeated in large order (ie, step B). If all the metric values in the selected RB column do not satisfy the continuity, the variable x is incremented to perform the next metric value of the set S first (ie, if the variable y exceeds the N value at step B). When resource allocation of the selected new metric value is completed, the scheduler 122 checks again whether the resource allocation is possible in the newly changed arrangement state of the previously held metric value.

, all indicators in the set A, be 0

4: while (S ≠ 0 ) do // step A .
5:

← x-th highest metric in the set S

6: Let the set R be the set of all scheduling metric s in RBk
7: if (mn ,k is the first RB assigned to user n) or (

is adjacent to RB assigned to user n) then
8:

←1, U ← U - R, S ← S - {

}

9: x ←1

10: Else

11: while (y ≤N) do // step B .

12: x ←x + 1
13:

← y-th highest metric in the set R
14; if (

is adjacent to RB assigned to user

) then
15:

←1, U ← U - R, S ← S - {

}

16; x ←1 , y ← 2

17; break

18: else

19: y ←y + 1
1: x ← 1, y ← 2

2: Let the set S be the highest metric of each RB
3: Let

, all indicators in the set A , be 0

4: while (S ≠ 0) do // step A.
5:

← x- th highest metric in the set S

6: Let the set R be the set of all scheduling metric s in RBk
7: if ( mn , k is the first RB assigned to user n ) or (

is adjacent to RB assigned to user n) then
8:

← 1, U ← U-R, S ← S-{

}

9: x ← 1

10: Else

11: while (Y ≤N) do // step B .

12: x ← x + 1
13:

← y -th highest metric in the set R
14; if (

is adjacent to RB assigned to user

) then
15:

← 1, U ← U-R, S ← S-{

}

16; x ← 1, y ← 2

17; break

18: else

19: y ← y + 1

According to the MSCC scheduling method according to an embodiment of the present invention, resource allocation is performed in a two-dimensional array configured based on scheduling metric values calculated based on channel information and HARQ information of each UE. Unlike the existing RME and IRME schemes, the MSCC scheduling method according to an embodiment of the present invention extends to neighboring RBs for resource continuity and performs scheduling based on the maximum metric of each RB column without performing scheduling. do. Therefore, the MSCC scheduling method according to an embodiment of the present invention has an advantage of more efficiently coping with a sudden drop in channel gain that may occur in every UE. In the MSCC scheduling method according to an embodiment of the present invention, since the scheduling can be performed dynamically for every input array value without having a static input value ( ie , ranking threshold) as in the conventional IRME method, the number of UEs and the channel environment are abrupt. Dynamically and flexibly respond to change. The MSCC scheduling method according to an embodiment of the present invention is performed on the basis of the aligned maximum metric value in terms of complexity required for scheduling, which is lower than the conventional RME and IRME techniques.

Performance evaluation ( Performance Evaluation )

Hereinafter, the results of the performance evaluation of the MSCC scheduling method according to an embodiment of the present invention through the complexity analysis and simulation analysis.

First, the complexity analysis part analyzes the algorithm complexity of the existing RME, IRME, and MSCC scheduling method according to an embodiment of the present invention in terms of time.

For each algorithm technique, when the UE is larger than the RB in consideration of the worst case, the complexity calculated in consideration of the complexity when the RB is larger than the UE is shown. Simulation Analysis defines simulation parameters necessary to perform scheduling, describes related formulas in detail, and shows and analyzes simulation results.

Complexity analysis ( Complexity Analysis)

로 표현될 수 있으며 등차수열 및 등비수열 정리를 통해

로 정리된다. 반대로, K > N 의 경우, NK에서 (N-(N -1))·(K-(N-1))까지 비용이 필요하게 되며

로 나타난다. 정리된 복잡도는

로 나타나게 되므로, 두 경우를 종합하여 도출되는 최종 복잡도는

이다.
First, we analyze the complexity of RME. The RME attempts to allocate resources to neighboring RBs located in both directions based on the maximum metric value in a given two-dimensional array. In this case, it is assumed that extended resource allocation is not attempted to the neighboring RB considering the worst case of each scheduling. In the first case, the number of UEs is larger than the number of RBs (ie, N> K). Since the RME performs scheduling until the K value is 1, the complexity required to find the initial scheduling metric is that the time cost of (N- (K-1)) (K- (K-1)) in NK need. To sum up this complexity

Can be expressed as

It is cleaned up by. Conversely, for K> N, costs are needed from NK to (N- (N-1)) · (K- (N-1))

Appears. Summarized complexity

The final complexity derived from the two cases is

to be.

로 계산되어 정리된 결과는

가 된다. 반대로 K > N 의 경우, 복잡도는

가 필요하게 되어

로 정리된다. 두 경우를 모두 고려하여 계산된 최종 비용은

이다. The following is an IRME complexity analysis. The IRME technique performs similarly to the RME scheduling scheme, but requires a threshold, such as a ranking threshold value, to provide more flexible scheduling in terms of throughput. The threshold value is a static value that is statically provided to the scheduler 122 as a value input before performing scheduling. Therefore, IRME is performed. The worst case is a case where a ranking threshold is applied as many as the number of UEs so that a threshold value is performed every time all scheduling is performed. In this case, the complexity of N > K requires the same threshold as the UE and the complexity of the RME.

Calculated and summarized as

. Conversely, for K> N, the complexity

Is needed

It is cleaned up by. Taking into account both cases, the final cost is

to be.

Finally, the complexity of the MSCC scheduling method according to an embodiment of the present invention is analyzed. The MSCC scheduling method according to an embodiment of the present invention performs scheduling based on the maximum metric values found for each RB in the input two-dimensional array. Here, the worst case maximum metric values of each selected RB column in the algorithm execution of the MSCC scheduling method according to an embodiment of the present invention do not satisfy the resource continuity constraint and all the metric values of the RB column having the selected metric value are selected. Suppose you need to compare.

의 복잡도는 정리되어

로 나타난다. K > N 의 경우, 복잡도는

로 정리되어 두 경우를 종합하여 최종적으로 나타낼 수 있는 복잡도는

가 된다. 만약 UE나 RB 둘 중 한 값이 극도로 작은 환경에서 복잡도가 고려된다면 결과는 다를 수 있겠지만, 본 발명의 일실시예에 따른 MSCC 스케쥴링 방법의 복잡도는 현재 이동통신 환경에서 고려될 때 세 기법 중 항상 가장 낮은 값을 가진다. First, the need for the scheduler 122 to perform a binary search with all the metric values in order to find the highest metric values of each RB column is NKlog ₂ N. Each RB heat by maximum metric values are selected because the need to perform a descending order to the scheduling from the higher order is also required by the complexity of the Klog ₂ K. Finally, the complexity of all the metric values in the RB column to which the selected metric value belongs is NK. For N> K,

Complexity of the

Appears. For K> N, the complexity

The complexity that can be finally expressed by combining the two cases is

. If the complexity of the UE or RB value is considered to be extremely small, the results may be different. However, the complexity of the MSCC scheduling method according to an embodiment of the present invention is always considered in the current mobile communication environment. Has the lowest value.

Table 2 summarizes the complexity of each technique.

Simulation analysis Simulation Analysis )

System-level simulation of SC-FDMA is performed based on the 3GPP LTE system model to compare and analyze the fairness and throughput performance of the MSCC scheduling method and the existing RME and IRME algorithms according to an embodiment of the present invention. .

In this case, performance measurement is performed in a single cell environment in which there is no cell causing peripheral interference. Based on the Typical Urban channel model, Table 3 shows the parameters used to perform the simulation. Here, we use the MCS table rather than the new capacity to analyze more realistic results. In order to guarantee fairness above a certain level regardless of the technique, the scheduling metric uses a PF (Proportional Fair) method and is expressed as in Equation 2.

는 UE i가 t시간에 RBj를 할당받을 수 있는 데이터 레이트(data rate)를 의미한다.

는 UE i가 t시간까지 할당받은 RB의 누적 데이터 레이트(data rate)를 의미한다. here,

Denotes a data rate at which UE i can be allocated an RBj at time t.

Denotes a cumulative data rate of the RB allocated to UE i until time t.

By using PF as a scheduling metric, a UE that is provided unfairly relative to the corresponding time t can be compensated for the next t.

Fairness (Fairness) In order to perform the simulation, JFI (Jain's fairness index) was used, the equation is shown in Equation 3.

는 시간 간격(time interval)

동안 i의 실제 데이터 전송률을 나타내고, N은 전체 시스템에 있는 사용자의 수를 나타낸다.
here,

Is the time interval

I represents the actual data rate of i, and N represents the number of users in the overall system.

Simulation results ( Simulation Results )

In the simulation, each technique is analyzed in terms of fairness and system throughput, and the MSCC scheduling method according to an embodiment of the present invention is performed with higher performance in the overall section than the existing RME and IRME techniques, and the maximum fairness is 3% and 19% increase in throughput.

8 is a graph showing a change in fairness according to the number of users for the conventional RME and IRME technique and the MSCC scheduling method according to an embodiment of the present invention.

Referring to FIG. 8, it can be seen that the fairness decreases as the total number of users increases regardless of the technique. This is because, as the number of users increases, the number of RBs that can be allocated to each user decreases inversely, thereby decreasing the selection of resources.

All three schemes generally have similar fairness because the scheduling metric value is calculated in PF to guarantee a certain level of fairness. As shown in FIG. 8, since the IRME technique 820 is an improved technique in terms of throughput in the RME, the IRME technique 820 has lower fairness than the RME 830. The MSCC scheduling method 810 according to an embodiment of the present invention guarantees a fairness similar to that of the RME, and has a fairness of up to about 12% higher than the IRME and up to about 3% higher than the RME in the interval of 40 to 80 users. (fairness). Therefore, it is shown that the MSCC scheduling method according to the embodiment of the present invention guarantees fairness similar to the existing RME and IRME techniques.

9 is a graph showing a change in system throughput according to the number of users for the existing RME and IRME techniques and the MSCC scheduling method according to an embodiment of the present invention. All three techniques yield higher throughput as the number of users increases, resulting in saturation at some level. The reason for reaching saturation is that the throughput was calculated using the Modulation and Coding (MCS) table. As the number of users increases, the system throughput increases regardless of the technique because the resources can be allocated to users with better performance. This is defined as multi-user diversity gain. Referring to FIG. 9, the IRME 920 has a higher throughput in all sections than the RME 930 and shows a difference of up to 17% at 10 users and 20 users. This is because the scheduling threshold performs more flexible scheduling than the RME. The MSCC scheduling method 930 according to an embodiment of the present invention exhibits a higher throughput in all sections than the IRME 920 and shows a throughput of up to 12% greater than the IRME 920 in the section of 40 users. Show closest satisfaction.

In the above, the FDPS scheme is proposed, for example, when applied to SC-FDMA required for low PAPR in LTE uplink. The MSCC scheduling method according to an embodiment of the present invention can perform scheduling with low complexity while satisfying the continuity constraint of resources of SC-FDMA while ensuring high system throughput and a certain level of fairness.

In the MSCC scheduling method according to an embodiment of the present invention, the performance of the MSCC scheduling method according to the embodiment of the present invention is compared with the existing RME and IRME algorithms using the PF metric as the scheduling metric. It was confirmed that it has a high throughput in all sections. In addition, it was confirmed that the time complexity of the MSCC scheduling method according to the embodiment of the present invention analyzed through Big O notation is always performed with a lower complexity than the existing two techniques based on the LTE mobile communication environment. The reason why the MSCC scheduling method according to the embodiment of the present invention is superior to the existing two methods is that scheduling is performed based on the maximum metric value of each RB rather than extending to neighboring RBs and performing resource allocation. Therefore, in a frequency selective fading environment where the gain may change rapidly even in the channel state of one UE, it is possible to provide more advanced scheduling than existing techniques.

Although described with reference to the embodiments above, those skilled in the art will understand that the present invention can be variously modified and changed without departing from the spirit and scope of the invention as set forth in the claims below. Could be.

100: base station 120: scheduling device
122: scheduler 124: storage unit

Claims (16)

In the uplink scheduling method performed in the scheduling apparatus of a base station connected through a plurality of terminals and a wireless channel,
NN is an N x K two-dimensional array consisting of N rows corresponding to positive integer-terminals and KK corresponding to K integer columns corresponding to positive integer-resource blocks (RBs). Receiving a metric value of the terminal; And
And performing resource allocation in the N × K two-dimensional array based on a maximum metric value for each RB column of K RB columns in the scheduling apparatus.
The uplink scheduling method is applied to uplink scheduling for continuous allocation of resource blocks (RBs) in an LTE mobile communication system using SC-FDMA in uplink. The method of claim 1, wherein the scheduling apparatus performs resource allocation in the N × K two-dimensional array based on a maximum metric value for each RB column of K RB columns.
Uplink scheduling method characterized in that until the resource allocation is completed for all the RB in consideration of all the maximum metric value for each RB column belonging to the K RB columns. 3. The method of claim 2, wherein the scheduling apparatus performs resource allocation in the N × K two-dimensional array based on a maximum metric value for each RB column of K RB columns.
Selecting a maximum metric value for each RB column of the K RB columns to form a first set; And
Selecting the maximum metric values included in the first set in descending order to determine whether to satisfy a continuity constraint of a resource. The method of claim 3, wherein the scheduling apparatus performs resource allocation in the N × K two-dimensional array based on a maximum metric value for each RB column of K RB columns.
RB including the selected maximum metric value when the selected maximum metric value is not located in the continuous RB with other metric values already allocated to the terminal and thus does not satisfy the resource continuity constraint so that resource allocation cannot be performed. And performing resource allocation by checking metric values of the other terminals in the column one by one in descending order. 5. The method of claim 4, wherein the scheduling apparatus performs resource allocation in the N × K two-dimensional array based on a maximum metric value for each RB column of K RB columns.
If all of the metric values in the RB column having the maximum metric value selected during the resource allocation are not assignable, then holding the allocation of the corresponding RB column and selecting the next higher metric value to perform resource allocation. Uplink scheduling method characterized by. The uplink scheduling method of claim 1, wherein the metric value is a PF (Proportional Fair) metric value calculated based on channel information and HARQ information of each UE for each time slot. delete In the scheduling apparatus of a base station connected through a wireless channel with a plurality of terminals,
NN is an N x K two-dimensional array consisting of N rows corresponding to positive integer-terminals and KK corresponding to K integer columns corresponding to positive integer-resource blocks (RBs). Including a scheduler for receiving the metric value of the terminal and performs resource allocation in the N x K two-dimensional array based on the maximum metric value for each RB column of the K RB columns,
The scheduler is a scheduling device, characterized in that applied for the continuous allocation of resource blocks (RB) in the LTE mobile communication system using SC-FDMA in the uplink. The method of claim 8, wherein the scheduler is
The scheduling apparatus, characterized in that until the resource allocation is completed for all the RB considering all the maximum metric value for each RB column belonging to the K RB columns. 10. The system of claim 9, wherein the scheduler is
Selecting a maximum metric value for each RB column of the K RB columns to form a first set,
And selecting maximum metric values included in the first set in descending order to determine whether a resource continuity constraint is satisfied. 11. The method of claim 10, wherein the scheduler is
RB including the selected maximum metric value when the selected maximum metric value is not located in the continuous RB with other metric values already allocated to the terminal and thus does not satisfy the resource continuity constraint so that resource allocation cannot be performed. The scheduling apparatus characterized in that the resource allocation is performed by checking the metric values of the other terminals in the column one by one in descending order. 12. The system of claim 11, wherein the scheduler is
If all of the metric values in the RB column having the maximum metric value selected during the resource allocation is not allocable, the scheduling apparatus is characterized in that the allocation of the RB column is suspended and the next higher metric value is selected to perform resource allocation. . The scheduling apparatus of claim 8, wherein the metric value is a PF (Proportional Fair) metric value calculated based on channel information and HARQ information of each terminal for each time slot. delete The method of claim 8, wherein the scheduling device
And a storage unit for storing scheduling information including a PF (Proportional Fair) metric value calculated based on channel information and HARQ information of each terminal. A communication device including a processor for use in a mobile communication base station connected via a wireless channel with a plurality of terminals,
The processor comprising:
NN is an N x K two-dimensional array consisting of N rows corresponding to positive integer-terminals and KK corresponding to K integer columns corresponding to positive integer-resource blocks (RBs). RB in LTE mobile communication system using SC-FDMA in uplink while receiving metric values of UE and performing resource allocation in the N × K 2D array based on the maximum metric value for each RB column of K RB columns And a processor configured to be adapted for successive allocation of s.

Images