KR20070023811A - Scheduling Method and Device in Wireless Network - Google Patents

Abstract

Techniques for scheduling flows and links for transmission are described. Each link is a directed source-destination pair and carries one or more flows. Each flow may be associated with throughput, delay, feedback (eg, acknowledgment (ACK)) and / or other requirements. The serving interval is determined for each flow based on the requirements for the flow. The serving interval is determined for each link based on the serving interval for all of the flows transmitted on the link. If system resources are available, each link is scheduled for transmission at least once in each serving interval to ensure that the requirements for all flows sent on the link are met. The link is also scheduled in a manner that facilitates closed loop rate control. The link is also scheduled such that ACKs for one or more layers in the protocol stack are sent at a sufficiently fast rate.

Transmission scheduling, serving intervals, delay requirements, data units

Description

METHOD AND APPARATUS FOR SCHEDULING IN A WIRELESS NETWORK}

I. Claims of Priority under 35 U.S.C. §119

This patent application is filed on June 2, 2004, entitled " Scheduling Method and Apparatus in a Wireless Network, " Provisional Application No. 60 / 576,721, assigned to the assignee of this patent application and expressly incorporated herein by reference. Insist on the priority of the call.

background

I. Field

TECHNICAL FIELD The present invention generally relates to communications and, more particularly, to transmission scheduling techniques in wireless networks.

II. background

Wireless networks are widely deployed to provide various communication services such as voice, data packets, and the like. These networks may support communication for multiple users by sharing the available system resources. Examples of such networks are wireless local area network (WLAN), wireless personal area network (WPAN), wireless wide area network (WWAN), code division multiple access (CDMA) network, time division multiple access (TDMA) network, frequency division multiple access (FDMA) network and the like. The terms "network" and "system" are often used interchangeably.

The wireless network may include any number of access points and any number of user terminals. An access point is typically a gateway or bridge between a backbone, which may be a wireless network and a wired network. A user terminal is a device that can communicate with an access point and / or another user terminal. Each user terminal may actively communicate or idle with the access point at any given time. Active user terminals may have different data requirements and capacities, and idle user terminals may likewise have different capacities. The wireless network may provide a specific transmission scheme, support one or more transmission schemes, and so forth. The key challenge will then select and schedule the user terminal for transmission and allocate available system resources to the selected user terminal as efficiently as possible based on their requirements and capacity. This task attempts more in high throughput wireless networks where scheduling has a greater impact on the overall performance of the wireless network.

Accordingly, there is a need in the art for techniques for scheduling transmissions efficiently in a wireless network.

summary

Techniques for scheduling "flow" and "link" for transmission are described herein. Each link is for a specific source station and a specific destination station. The station may be an access point or a user terminal. Each link carries one or more flows. Each flow carries data from the protocol stack to the higher layer and may be associated with certain requirements, such as throughput and delay requirements. Each flow and / or each link may also be associated with some feedback requirements. For example, each flow may require an acknowledgment (ACK) for the data sent for the flow. The serving interval is determined for each flow based on the requirements for the flow. The serving interval often indicates how a flow should be served to satisfy all of its requirements. The serving interval is then determined for each link based on the serving interval for all of the flows transmitted on the link. If system resources are available, each link is scheduled for transmission at least once in each serving interval to ensure that the requirements for all flows sent on the link are met.

The link is also scheduled in a manner to facilitate closed loop rate control. If the overhead channel is not available to send feedback information (eg, pilot, rate, etc.), then the reverse transmission may then be scheduled before each data transmission to provide feedback information. The link is also scheduled such that ACKs for one or more layers in the protocol stack are sent at a sufficiently fast rate so that data transmissions are not limited or stalled while waiting for an ACK.

Various aspects and embodiments of the invention are described in further detail below.

Brief description of the drawings

1 shows a wireless network.

2 illustrates an example protocol stack.

3 shows an exemplary transmission structure.

4 shows a process for determining a serving interval for a link.

5 shows data and reverse transmission between station A and station B. FIG.

6 shows a process for selecting and scheduling a link for transmission.

7 shows a process for scheduling a link in each frame.

8 shows a process for determining a TXOP duration for reverse transmission.

9 shows a process for determining a TXOP duration for data transmission.

10 shows a process for determining the rate used for transmission.

11 shows a process for determining the amount of time required for transmission.

12 is a block diagram of an access point and two user terminals.

13 is a block diagram of a CSI processor at an access point.

14 is a block diagram of a scheduler.

details

The word "exemplary" is used herein to mean "acting as an example, illustration, or illustration." Certain embodiments described herein as "exemplary" are not necessarily to be construed as preferred or preferred over other embodiments.

The scheduling technique described herein may be used for various wireless networks such as WLAN, WPAN, and the like. These techniques include single-input single-output (SISO) networks with single-antenna transmitting and receiving stations, single-input multi-output (SIMO) networks with single-antenna transmitting stations and multi-antenna receiving stations, multi-antennas Multi-input single-output (MISO) network with transmitting station and single-antenna receiving station, multi-antenna multiple-input multiple-output (MIMO) network with transmitting station and receiving station, or single-antenna and multi-antenna It may also be used in a hybrid wireless network with a combination of stations. These techniques include (1) time division duplex (TDD) networks in which data is transmitted on the downlink and uplink on a single frequency band at different time intervals and (2) data on the downlink and uplink on different frequency bands. It may also be used for transmitted frequency division duplex (FDD) networks. For clarity, some aspects of the scheduling technique are described below for a wireless TDD MIMO network.

1 illustrates a wireless network having one or more access points 110 and a multi-user terminal 120. Only one access point is shown in FIG. 1 for simplicity. An access point is generally a fixed station that communicates with a user terminal and may also be referred to as a base station or some other terminology. The user terminal may be fixed or mobile and may also be referred to as a mobile station, wireless device, user equipment (UE), or other terminology. Each access point may support communication for any number of user terminals. Each user terminal may communicate with one or more access points. The user terminal may also communicate peer to peer with other user terminals. For a centralized network architecture, system controller 130 combines the access points and provides coordinates and control for these access points. In the following description, "station" may refer to an access point or a user terminal.

2 illustrates an example protocol stack 200 that may be used for the wireless network 100. Protocol stack 200 includes transmission control protocol (TCP) / user datagram protocol (UDP) layer 210, internet protocol (IP) layer 220, media access control (MAC) layer 230, physical (PHY) Layer 240. Protocol stack 200 may also include other interlayers and / or sublayers. For example, a point-to-point protocol (PPP) layer, a radio link protocol (RLP) layer, and the like may exist between the IP layer and the MAC layer. TCP and UDP are two transport layer protocols. UDP is often used for real-time applications that provide transport services without a reliability mechanism and retransmissions are not necessarily applicable or less applicable. TCP provides reliable transport services and has error detection and error recovery mechanisms. The TCP / UDP layer supports higher layer applications and provides TCP packet / segment and / or UDP datagrams. The IP layer encapsulates TCP packets and / or UDP datagrams and provides IP packets. TCP, UDP, and IP functions are well known. The MAC layer encapsulates an IP packet and provides a MAC Service Data Unit (SDU). The MAC layer also performs other functions such as scheduling transmissions for the downlink and uplink, QoS coordination, and the like. The physical layer provides a mechanism for transmitting data over the air and performs various functions such as framing, encoding, and modulation.

There may not be a limited relationship between the TCP packet / UDP datagram, the IP packet, and the MAC SDU. Thus, each data unit in a given layer may carry one, partial, or multiple data units in another layer. However, for simplicity, the following description assumes a one-to-one correspondence between a TCP packet / UDP datagram, an IP packet, and a MAC SDU. For clarity, processing by TCP / UDP, IP, and physical layer will not be described below unless appropriate herein. The MAC layer receives the flow of packets from the higher layer. Each flow may be associated with some quality of service (QoS) requirements that may not be limited by a particular minimum rate and / or a particular maximum delay. The terms "rate" and "data rate" are synonymous in the following description.

The wireless network 100 may use an automatic retransmission request (ARQ) scheme at the MAC layer. For the ARQ scheme, the source station transmits each MAC SDU to the destination station one or more times until the destination station correctly decodes the MAC SDU or reaches the maximum number of transmissions for the MAC SDU. The destination station sends back an acknowledgment (ACK) for each correctly decoded MAC SDU. Wireless network 100 is a MAC message Block Ack It may support a block ACK mechanism that causes the source station to transmit a burst of MAC SDUs prior to requesting the status of these MAC SDUs by request . The destination station then blocks another MAC message The last block by sending Ack backward Ack It will communicate the status of all MAC SDUs received from the request . The ARQ scheme also utilizes an ARQ window that specifies the maximum number of MAC SDUs that may be sent without receiving an ACK for these MAC SDUs.

TCP supports other ACK schemes, for example, optional ACK and TCP Reno. Each ACK scheme sends an ACK for the TCP packet in a different manner, and the size of the ACK feedback for TCP depends on the selected ACK scheme used. TCP also uses a TCP window that specifies the maximum number of TCP packets that may be sent without receiving an ACK for these packets. TCP interferes with the transmission of TCP packets outside the TCP window. In addition, the TCP window may be shrunk due to inadequate ACK feedback on the transmitted TCP packet.

For clarity, in the following description, a TCP ACK refers to an ACK for a TCP packet and a block ACK refers to an ACK for a MAC SDU. Each block ACK will transport the state of up to one and transmitted from the Block Ack N _SDU is equal to 64 or some other value may N MAC _SDU that SDU. The ARQ window for the MAC layer is denoted by W _ARQ and the TCP window for _TCP is denoted by W _TCP .

3 illustrates an example transmission structure 300 that may be used for the wireless network 100. Each access point in a wireless network maintains a separate timeline for all transmissions covered by that access point. The transmission timeline for one access point is described below. This access point periodically transmits a beacon on the downlink. This beacon carries the preamble and access point identifier (AD ID) used by the user terminal to detect and identify the access point. Target beacon transmit time (TBTT) refers to the time interval between the start of two consecutive beacons. TBTT may be fixed or variable depending on how the network is operated

The time interval between the beacons may include any combination of controlled access period (CAP), scheduled access period (SCAP), and contention period (CP) in which enhanced distributed channel access (EDCA) is used. . The CAP, SCAP and CP may be sent in any order. Each CAP covers a period of time used by the access point for network management. Each SCAP covers a period of time during which transmissions on the downlink and uplink are scheduled. Each CP covers a period of time for which transmissions are not scheduled. Beacons, CAPs, and SCAPs represent contention free periods in which only one station (which may be an access point or user terminal) transmits on a wireless channel at any given time. CP represents a contention period in which one or more stations transmit simultaneously on a wireless channel.

Each SCAP includes a SCHED frame and a scheduled access period. Each SCAP may span a fixed or variable time duration. The SCHED frame carries a schedule of all transmission opportunities (TXOPs) for the accompanying scheduled access periods. Each TXOP is a scheduled transmission from a concrete source (sending) station to a concrete destination (receiving) station. The scheduling information for each TXOP carries the source and destination stations, the start time and duration of the TXOP, and other suitable information as possible. The scheduled access period may include any number of TXOPs (to some limit), and each TXOP may be for any pair of source and destination stations. The SCHED frame may include a pilot that may be used by the user terminal to perform channel estimation and rate control on the uplink. Scheduled access periods may also include other types of transmissions.

In an embodiment, the transmission timeline is divided into MAC frames, or simply, "frames." Each frame has a predetermined time duration, for example approximately 2 milliseconds (ms). In an embodiment, each SCAP spans one frame.

3 shows an exemplary transmission structure. In general, the scheduling technique described herein may be used with any transmission structure having a period in which transmissions are scheduled.

The wireless network 100 allows a transmitting station to select an appropriate transmission mode and causes the receiving station to send back channel state information (CSI) to the transmitting station so that one or more rates are used for data transmission to the receiving station. May be employed. This CSI may be in the form of a steered or non-steered MIMO pilot (described below), an SNR estimate, an initial rate selected by the receiving station, or the like. In an embodiment, the wireless network 100 does not use an overhead channel dedicated for transmission of CSI. Thus, each station transmits CSI when TXOP is assigned.

The scheduler schedules transmissions on the downlink and uplink to achieve high throughput and robust performance. The scheduler may be co-located at the access point or may reside at some other network entity (eg, system controller 130 in FIG. 1). The scheduler may be responsible for the following tasks.

Assign TXOP to the station as efficiently as possible to ensure that the Qos (e.g. throughput and delay) requirements of the flow are met;

Assign TXOP as appropriate to ensure that the rate control mechanism is updated at a sufficient rate for good performance;

Allocate TXOP for transmission of higher layer ACKs and MAC ACKs so that data transmissions are not limited or staggered by ACK feedback.

TXOP is also allocated in a manner that facilitates the block ACK mechanism used by the ARQ scheme. The operation of the scheduler to perform three tasks is described below. The following description assumes that the scheduler has access to information available at the access point in the execution of the scheduling.

The following terms are used in the description below. The link is a directed source-destination pair (A, B) with a concrete source station A and a concrete destination station B. Links to source-destination pairs (A, B) are also referred to as links (A, B). For data transmission from station A to station B, link (A, B) is the direction of traffic data and link (B, A) is the direction of block ACK for the MAC layer and TCP ACK for TCP. Thus, two links for opposite directions are used for data transmission from station A to station B. If station B has traffic data sent to station A, then another link to the source-destination pair (B, A) is recorded by the scheduler to carry traffic data from station B to station A.

The flow is a higher layer (eg TCP or UDP) data stream sent over the link. The link may carry one or multiple flows for the same source-destination pair. If a higher layer protocol requires bi-directional flow, then two flows on opposite links are recorded by the scheduler. For example, TCP uses a bi-directional flow where one flow is for TCP packets and the other is for TCP ACK. The scheduler may handle two flows for TCP packets and TCP ACK as separate flows in scheduling.

The call control entity, which resides above the MAC layer, provides an authorization control algorithm and determines which flow to authorize for service. The call control entity may also provide a mechanism for regulating the flow, for example, regulating the rate supported for each flow. Various designs may be used for authorization control algorithms and regulatory mechanisms, as known in the art.

The scheduler schedules the link by TXOP such that the QoS requirements for flows sent on these links are met (if any). The scheduler may dynamically allocate time to the link based on the following characteristics of each flow.

Delay requirements of the flow;

Throughput requirements of the flow; And

Transmission duration required in real time by the source station

Details of the scheduling are described below. Most of the description below is for a bi-directional flow for TCP that is more complex in terms of scheduling than a single flow for UDP. The scheduler carries scheduled TXOPs to the source and destination stations on the SCHED frame.

One. serving  interval

In an embodiment, each link is associated with a serving interval that indicates how often that link should be scheduled by TXOP. The scheduler attempts to schedule each link by one or more TXOPs in each serving interval for that link. The serving interval for each link may be, for example, the delay requirement of the flow (s) transmitted on the link, the throughput requirement of the flow (s), the ARC scheme selected for use, the rate control mechanism, etc., or a combination thereof. It may be determined based on such various criteria. The calculation of the serving interval for the example links (A, B) is described below. Table 1 shows a list of variables used in the exemplary flow (F).

Table 1

   symbol                                   Explanation d _flow Delay requirement for flow F in seconds R _flow Rate requirement for flow F in units of bits / second (bps) S _flow MAC SDU payload size for flow (F) in units of bits / MAC SDU W _ARQ ARQ window size applicable to flow (F) in number of MAC SDUs W _TCP TCP window size applicable to flow (F) in the number of TCP packets     α Fraction of ARQ Window Covered by Block ACK Transmitted in Each Serving Interval N _max Maximum number of transmissions for MAC SDU by ARQ method T _ARQ Time interval for block ACK to transmit for flow (F) T _delay Time interval between transmissions of MAC SDU by ARQ scheme T _flow Serving Interval for Flow (F)

The serving interval for flow F is selected to achieve both throughput and delay requirements for the flow. The throughput requirement for flow F may be partially achieved by ensuring that the block ACK for the flow is sent at a rate fast enough so that the flow is not starved to the wait of these ACKs. In general, more MAC SDUs are sent at higher rates, and block ACKs are sent more frequently. The delay requirement for flow F may be partially satisfied by ensuring that each MAC SDU may be transmitted up to the product of N _tx within the delay d _flow specified for a flow where N _tx ≦ N _max .

Destination Station B is Block Ack Receives a request and sends a block ACK whenever a TXOP is allocated. The number of MAC SDUs covered by the block ACK may be given as α · W _ARQ , which is a fraction of the ARQ window. Station B is scheduled by TXOP for block ACK transmission at a rate fast enough so that station A is not starred to the wait of the block ACK. In an embodiment, the time interval T _ARQ for sending a block ACK for flow (F) is:

식(1)

Formula (1)

Is calculated as

As shown in equation (1), the block ACK interval is set to be equal to the time duration required to transmit a certain fraction of the ARQ window, which is the? W _ARQ MAC SDU. This time duration is equivalent to multiplying the αW _ARQ MAC SDU by the MAC SDU payload size (S _flow ) and divided by the rate for R _flow . R _flow is the rate that the scheduler guarantees or tries to achieve for the flow and not the instantaneous rate used for each TXOP scheduled for the flow. Equation (1) assumes that the MAC SDU payload size is fixed to flow (F). In general, a fixed or variable MAC SDU payload size may be used for flow (F). Equation (1) may then be modified to indicate an expected amount of time for T _ARQ to transmit the αW _ARQ MAC SDU. The block ACK interval is selected in equation (1) so that transmission of the block ACK for the α fraction of the ARQ window in each service interval ensures that the flow meets its rate requirements. The smaller fraction α corresponds to a shorter block ACK interval corresponding to a smaller probability of flow defined by the block ACK feedback. In an embodiment, the fraction is selected as α = 1/4 such that the block ACK is sent to one quarter of the MAC SDU in the ARQ window. Block ACK may also be sent more or less frequently.

In an embodiment the retransmission delay T _delay for flow F is:

식 (2)

Formula (2)

Is calculated as

The retransmission delay is selected to ensure that the N _tx ARQ cycle can be completed within the delay requirements of the flow. Each ARQ cycle covers one retransmission of a given MAC SDU. The plus one (+1) argument in the denominator of equation (2) describes the MAC SDU feedback for the N _tx th transmission of the MAC SDU.

In an embodiment, the serving interval T _flow for flow (F) is:

T _flow = min (T _ARQ , T _delay ) Equation (3)

Is selected as.

As shown in equation (3), the serving interval for flow (F) is determined to be the smaller of the block ACK interval and the retransmission delay. This ensures that the serving interval will satisfy both flow delay and feedback requirements for the ARQ scheme and may ensure that rate requirements for the flow can be achieved. The serving interval is typically governed by the block ACK interval for high rate flows and the retransmission delay for delay sensitive flows.

Specific embodiments for determining the serving interval for the flow have been described above. The serving interval for each flow may also be determined in other ways and / or using other criteria. For example, the serving interval for the flow may be selected to achieve the packet error rate (PER) required for the flow. The value N _tx for the flow may be selected based on the PER and delay requirements for the flow. The PER achieved by the ARQ scheme for flow is determined by (1) the number of transmissions for each MAC SDU for the flow and (2) the PER achieved by the physical layer for each transmission of the MAC SDU. N _tx may be chosen such that the PER achieved for the flow is less than or equal to the PER required for the flow. Lower PER may be achieved by more transmissions (ie, larger N _tx ) corresponding to shorter retransmission delays that alternately correspond to shorter serving intervals for the flow. The serving interval for each flow may also be selected by considering user priority, data requirements, other QoS requirements, and the like.

As noted above, multiple flows may be sent on the link. In such a case, the serving interval may be determined for each flow, for example as described above. The serving interval for the link may then be set equal to the shortest serving interval for all flows transmitted on the link. Each link is then scheduled based on its serving interval.

Bi-directional flow may be used for data transmission in the case where two flows on opposite links are recorded by the scheduler. The serving interval may be selected for each link that meets the requirements of data transmission. By way of example, for a TCP transmission, the first flow may be recorded on the first link carrying the TCP packet, and the second flow may be recorded on the second link in the opposite direction of carrying the TCP ACK. The serving interval for the first link carrying the TCP data flow may be determined as described above based on the rate, packet size, and delay requirements for the TCP transmission (if any). However, the rate for the TCP ACK flow may not be known as it depends on the TCP ACK scheme selected for use. In addition, the TCP ACK flow may not have a specific delay requirement, or the delay requirement for the TCP ACK flow may be in accordance with the TCP data flow.

The serving interval for the TCP ACK flow may be determined as described below. In the following description, N _{data / ACK} is the number of TXOPs scheduled for a TCP data flow between two consecutive TXOPs scheduled for a TCP ACK flow where N _{data / ACK} ≥ 1. Therefore, the serving interval for the TCP ACK flow is N _{data / ACK} multiplied by the serving interval for the TCP data flow. For simplicity, the following description assumes that TCP packets are not fragmented such that each TCP packet is sent in one MAC SDU. Also for simplicity, no additional delay due to retransmission by the ARQ scheme is considered.

At each time a TCP data flow is scheduled, an [alpha] W _ARQ MAC SDU (or TCP packet) may be sent for that flow. The number of TCP packets transmitted during the serving interval for the link for the TCP ACK flow may then be expressed as α · W _ARQ · N _{data / ACK} . To ensure that the TCP ACK reaches the TCP sender before the TCP sender's TCP window is consumed, the number of TCP packets transmitted during the TCP ACK serving interval is equal to α W _ARQ N _{data / ACK} <W _TCP . It is limited as such. The TCP sender is not limited if the following limitations are satisfied (e.g., no waiting for TCP ACK is idle).

식 (4)

Formula (4)

Equation (4) assumes that one TCP packet is sent to each MAC SDU. Equation (4) may be modified to account for the fragmentation of the TCP packet.

The serving interval for the TCP ACK flow is then:

T _ACK = N _{data / ACKT data} , (5)

Wherein T _data is a serving interval for the TCP data flow, and this serving interval may be calculated as described above as equations (1) to (3),

T _ACK is the serving interval for the TCP ACK flow.

The serving interval for the TCP ACK flow is chosen so that the TCP sender's TCP window is not consumed.

The above description suggests that the TCP ACK flow may be treated as a data flow having a rate of 1 / N _{data / ACK} multiplied by the rate of the TCP data flow. The calculation of T _ARQ for a TCP ACK flow ensures that the serving interval for this flow will not be greater than N _{data / ACK} multiplied by the serving interval for the TCP data flow. The MAC layer may handle TCP data flow and TCP ACK flow as two separate data flows transmitted in opposite directions. The only difference between the TCP data flow and the TCP ACK flow is the frequency at which the two flows are scheduled and this frequency is determined by their respective serving intervals.

4 shows a process 400 for determining a serving interval for a given link. Initially, all flows transmitted on the link are identified (block 412). If so, the requirements for each flow are determined (block 414). These requirements may include throughput, delay, feedback, and / or other requirements. For each requirement of each flow, a time interval in which the flow must be served to meet that requirement is determined (block 416). For example, the block ACK interval and the retransmission interval may be determined for each flow, as described above. The serving interval for each flow is then determined, for example, as the shortest time interval for all the requirements of that flow, and this serving interval ensures that all requirements can be met (block 418). The serving interval for the link is then determined, for example, as the shortest time interval for all flows transmitted on the link, ensuring that this serving interval can be met for all flows (block 420). ).

2. Closed-loop Rate  Control

The serving interval for each link determines how often or often the link should be scheduled. The scheduler attempts to assign one or more TXOPs to that link in each service interval for each link. The scheduler may also assign TXOP to the link in a manner to facilitate efficient operation by the physical layer.

For data transmission from station A to station B, sender station A typically needs the latest channel state information (CSI) for destination station B to achieve high throughput. This CSI may be in the form of a channel response estimate for the radio channel from station A to station B, a rate used for transmission, and the like. Station A may obtain CSI for station B based on a reverse transmission sent by station B to station A. For a TDD system, this reverse transmission may include (1) a pilot that causes station A to estimate the response of the radio channel from station A to station B, and (2) the rate (s) used for transmission to station B. have.

The scheduler may schedule a TXOP for reverse transmission from station B to station A following a TXOP for data transmission from station A to station B. The TXOP for link (B, A) may be selected long enough to cause station B to transmit a block ACK. Transmission of a block ACK on link (B, A) before each TXOP for link (A, B) enables efficient operation of the ARQ scheme. Station B will send a block ACK (if any) to the station A in the scheduled TXOP for the link (B, A). Otherwise, station B may transmit a null MAC SDU (or null packet) which is a MAC SDU that has no payload at all or has dummy data for the payload. Station B may also transmit a pilot in the reverse transmission to cause station A to estimate the channel response between stations A and B. The reverse transmission may also carry one or more rates selected by station B in the data rate vector feedback (DRVF) field. Reverse transmission from station B causes station A to update its channel response estimate (based on the pilot) for station B and obtain the rate (s) selected by station B (from the DRVF field).

5 shows data and reverse transmission by stations A and B for data transmission from station A to station B. FIG. This data transmission may be for TCP data, TCP ACK, UDP data, or some other type of data. The scheduler may schedule reverse transmission on link (B, A) prior to each data transmission on link (A, B) to improve system performance. Station B may send a pilot and block ACK (if any) on the reverse transmission. Reverse transmission on link (B, A) may be scheduled one or more frames prior to data transmission on link (A, B). The scheduling of reverse transmissions and data transmissions in other frames performs all of the required processing, e.g., station A with sufficient time to estimate the pilot based channel response, calculate the steering vector used for spatial processing, and the like. To provide. Scheduling of reverse and data transmissions in other frames may also simplify the scheduler. 5 also shows station A scheduled by one TXOP in each serving interval.

3. Scheduling

The scheduler executes scheduling at each scheduling interval to select and schedule the link for transmission. The scheduling interval may be any time duration. In an embodiment, scheduling is performed in each frame to schedule the link for transmission in the same frame. In an embodiment, the link carrying the data for the higher layer is recorded by the scheduler, and the link carrying the block ACK is scheduled based on the recorded link. Table 2 lists the variables that are kept for each link recorded by the scheduler.

Table 2

   parameter                              Explanation    Serving interval The time interval that the link should be scheduled by TXOP Last Served Time Frames on which link was last scheduled by TXOP   Status flags If present, indicates the type of transmission to be scheduled for the link.

The last served time for each link may be initialized to the frame in which the link is recorded by the scheduler. The status flag for each link may be initialized to "none" when the link is written by the scheduler.

6 shows a process 600 for selecting and scheduling a link for transmission. The scheduler executes process 600 in each frame. Initially, the scheduler identifies all links that are not scheduled for transmission in the serving interval of all links (block 612). This identifies (1) all links with a status flag set to "none" and (2) subtracting the last served time for the link in the upcoming frame n + 1, where each link is equal to or greater than the serving interval for the link. It can also be achieved by selecting. As noted above, the scheduler prioritizes each data transmission by reverse transmission or channel probe. Thus, the scheduler determines whether each link is scheduled for data transmission in the next frame n + 1 so that each link can schedule reverse transmission in the current frame n. The scheduler sets a status flag for each of these links “backwards” to indicate that reverse transmissions should be scheduled for the link. The scheduler also identifies links that should be served in the current frame n, such as (1) links that are not properly served in previous frame n-1 and (2) those by backward transmission in previous frame n-1.

The scheduler then classifies the identified links based on their priority and / or other criteria (block 614). The scheduler may assign higher priority to the link carrying the flow in real time and lower priority to the link carrying the flow with the best effort. The scheduler may also assign higher priority to links with higher ARQ cycles and lower priority to links carrying the first transmission of MAC SDUs. In general, the scheduler may determine the priority of the identified link based on the type of flow transmitted on the link, the Qos class, the actual or potential delay experienced by the flow, the consideration being evaluated, the station's priority, and the like. . The scheduler places the sorted links in the sorted list with the highest priority link at the top of the list and the lowest priority link at the bottom of the list (block 616). The scheduler then serves as many links as possible in the sorted list for transmission in the current frame (block 618).

7 shows an embodiment of block 618 in FIG. 6 for scheduling of a link for transmission. The scheduler selects the link at the top of the sorted list represented by links (A, B) (block 710). The scheduler then determines whether the link (A, B) needs to be scheduled for reverse transmission based on the link's status flag (block 712). If the status flag for link (A, B) is set to "reverse" and the response is 'yes' for block 712, then the scheduler then determines the amount of time (or duration) of the TXOP for reverse transmission as described below. ) And assign this TXOP to links B and A (block 716). The TXOP will be in a duration sufficient to allow station B to send appropriate channel state information and block ACK to station A as much as possible. The scheduler then sets the status flag of link (A, B) to "data" to indicate that data transmission should be scheduled for this link in the next frame (block 718). The scheduler also subtracts the TXOP duration for the link (B, A) from the time available for transmission in the current frame (block 720) and then proceeds to block 740.

In block 712, if the currently selected link (A, B) does not require reverse transmission, then the scheduler needs to have the link (A, B) scheduled for data transmission based on the link's status flag. Determine if there is a block (722). If the status flag of the link (A, B) is set to "data" (by the scheduler in the previous frame n-1) and the response is 'yes' to the block 722, then the scheduler will be described later, Determine the duration of the TXOP for links (A, B) (block 724). The scheduler then determines whether there is sufficient time in the current frame for the TXOP for links (A, B) (block 726). If the answer is yes, then the scheduler assigns TXOP to link (A, B) (block 728) and updates the last served time for link (A, B) with the current frame (block 730). The scheduler also sets the status flag of link (A, B) to "none" to indicate that no transmission needs to be scheduled for this link unless triggered by that serving interval (block 732). The scheduler also subtracts the TXOP duration for the link (A, B) from the time available for transmission in the current frame (block 734) and then proceeds to block 740.

In block 726, if the available time in the current frame is less than the TXOP for links (A, B), then the scheduler assigns the remaining time in the current frame to the TXOP for links (A, B). (Block 736). The scheduler will be selected again in the next frame without updating the last served time or status flag of the link (A, B) since the link is not fully served. The scheduler then ends scheduling for the current frame since there is no more time left to allocate to any other link.

In block 740, the scheduler removes the links (A, B) from the sorted list. If the sorted list is not empty and there is time available in the current frame, then as determined at block 742, the scheduler then returns to block 710 to send a TXOP for the next link in the sorted list. Schedule. Otherwise, if all links in the sorted list have been scheduled or if there is no more time remaining in the current frame, the scheduler then ends the scheduling.

6 and 7 illustrate that the scheduler selects each link to schedule a TXOP when the serving interval for the link is approaching. The scheduler may also include other links in which the serving interval is closed and expires in the sorted list so that these links may be scheduled before the expiration of their serving interval if time is available in the current frame.

8 shows an embodiment of block 714 in FIG. 7 that determines the duration of a TXOP for reverse transmission on links (B, A) from station B to station A. FIG. The scheduler initially determines the rate (s) for transmission from station B to station A, as described below (block 810). The scheduler then determines, if any, feedback sent from station B to station A (block 812). This feedback may include pilot, block ACK, rate feedback (DRVF), additional time requirements, some other type of feedback, or some combination thereof. The scheduler calculates the amount of time needed to transmit feedback at a predetermined rate (s) called “data time” (block 814). This data time includes the time required to transmit (if any) header data in the MAC SDU payload. Such header data may include all or part of a physical layer convergence protocol (PLCP) header defined by IEEE 802.11a.

The scheduler also determines the amount of time needed to transmit overhead for reverse transmission, called " overhead time " (block 816). The overhead may include a preamble, a PLCP header, a MIMO pilot, or the like or some combination thereof. The preamble is any type of pilot used for signal detection and possibly other purposes. The MIMO pilot may be discussed below and considered as part of the feedback in block 814 or as part of the overhead in block 816. The overhead may depend on whether station B is an access point or user terminal, the number of antennas in station B, and possibly other factors. For example, the access point may pre-transmit the preamble at the start of each SCAP, in which case the overhead may only include a PLCP header and a MIMO pilot. The duration of the overhead also depends on the system design. As a specific example, if station B is an access point, then the overhead may include two OFDM symbols for the PLCP header and four OFDM symbols for the MIMO pilot, where each OFDM symbol may have a duration of 4 μsec. . If station B is a user terminal, then the overhead may include four OFDM symbols for the preamble, two OFDM symbols for the PLCP header, and four OFDM symbols for the MIMO pilot. The scheduler then calculates the TXOP duration for link (B, A) as the sum of the data time and the overhead time (block 818). The scheduler allocates links (B, A) with this TXOP duration in block 716 in FIG.

9 illustrates an embodiment of block 724 in FIG. 7 that determines the duration of a TXOP for data transmission on links A and B from station A to station B. FIG. As noted above, the scheduler is present at an access point or network entity having access to appropriate information at the access point. Other types of information may be available to the scheduler depending on whether the access point is source station A or destination station B. The calculation of the TXOP duration may be different from other types of information available to the scheduler, as described below.

If source station A is an access point, the scheduler then has the information available at source station A, as determined at block 910. This information may include (1) the amount of data sent to destination station B for data transmission to station B and (2) one or more initial rates selected by station B. The scheduler determines the payload (eg, for all flows) that is sent on link (A, B) to station B. The scheduler also determines one or more final rates for transmission from station A to station B, as described below (block 914). The scheduler then calculates the amount of time (or data time) needed to transmit the payload at the selected rate (block 916). The scheduler also determines the amount of time (or overhead time) required to transmit overhead for data transmission, as described above with respect to FIG. 8 (block 918). The scheduler then calculates the TXOP duration for the link (A, B) as the sum of the data time and the overhead time (block 920).

If source station A is a user terminal, then the scheduler has information available at destination station B, as determined at block 910. This information may include the amount of time required by source station A, which may be transmitted in the duration required field in reverse transmission from station A in the previous frame. The scheduler determines the amount of time required by station A (block 922). If the duration required is greater than zero, as determined at block 924, the scheduler then sets the TXOP duration for links A and B at the desired duration (block 926). Otherwise, if the required duration, which may be due to station A not requiring time or station B receiving a request from a faulty station A, is zero, then the scheduler may be responsible for a short reverse transmission from station A to station B. Set the TXOP duration for links (A, B) to the amount of time needed. This reverse transmission allows station A to request time on links A and B when the buffer at station A is augmented. This reverse transmission also allows periodic update of channel state information so that accurate rate selection and rate control can be achieved when data transmission is initiated. The scheduler may only allocate a small amount of time for reverse transmission, which may be sufficient to send only the MAC header and pilot.

The scheduler uses the TXOP duration calculated at block 920, 926, or 928 and assigns this TXOP duration at block 728 of FIG. 7 to link (A, B).

10 shows a process 1000 for determining one or more rates used for transmission from station A to station B. As shown in FIG. Process 1000 may be used at block 812 in FIG. 8 (although stations A and B were swapped) and at block 914 in FIG. 9.

In an embodiment, for data transmission from the access point to the user terminal, the user terminal selects one or more initial rates based on the pilot received from the access point and transmits the initial rate (s) to the access point via reverse transmission. do. The scheduler then selects one or more final rates based on the initial rate (s) received from the user terminal. For data transmission from the user terminal to the access point, the access point receives the pilot from the user terminal and selects one or more final rates used for the data transmission. The access point then sends the final rate (s) to the user terminal via reverse transmission. Other types of information are available to the scheduler depending on whether the access point is source station A or destination station B. The final rate (s) may be determined in other ways depending on the information available.

As determined at block 1010, if source station A is an access point, then the scheduler obtains an initial rate (s) selected by station B and transmitted to station A (block 1012). The scheduler determines the age of the initial rate (s) (block 1014). The scheduler then derives the final rate (s) based on the initial rate (s) and their age (block 1016). For example, the scheduler may discount or decrease the initial rate (s) based on the age of the initial rate (s), as described below.

As determined at block 1020, if destination station B is an access point, then the scheduler estimates an initial signal-to-noise ratio (SNR) for the wireless channel from station A to station B based on the pilot received from station A. Obtain (block 1022). The scheduler then determines the age of the initial SNR estimate (block 1024) and derives the adjusted SNR estimate based on the initial SNR estimate and their age (block 1026). The scheduler may discount the initial SNR estimate based on the age of these SNR estimates, as described below. The scheduler then selects the final rate (s) based on the adjusted SNR estimate (block 1028). The scheduler may also discount the rate or SNR estimate by the amount of time between the current frame and the future frame in which the final rate (s) will be used.

If station A and station B communicate peer-to-peer (via a direct link protocol), then if station A and station B are not an access point, then the access point does not relay traffic data between these stations. Nevertheless, the scheduler manages peer to peer communication. The access point may continue to receive transmissions sent by the user terminals within its range and read the data rate vectors DRVs transmitted by these user terminals. DRV describes the rate used for each data stream transmitted over the MIMO channel to allow the receiver to demodulate the transmission. The scheduler stores the DRV information as well as the time when the DRV information is observed. If the link (A, B) is selected for scheduling, then the scheduler obtains the rate (s) in the DRV information sent by station A to station B (block 1032). The scheduler also determines the age of the DRV (block 1034) and derives the final rate (s) based on the initial rate (s) and their age in the DRV, as selected by station A (block 1036). Thus, the access point snoops on the DRV sent to the peer to peer by the user terminal within its scope, and the scheduler uses these DRVs to schedule the communication to peer to peer.

10 shows a scheduler for calculating the final rate (s) used for data communication. Another entity may also perform rate calculations and provide a final rate to the scheduler.

11 shows a process 1100 for determining an amount of time for requesting transmission. For example, whenever a user has data to send, the user terminal may execute process 1100. The user terminal determines the payload (eg, for all flows) to send to the access point or other user terminal (block 1112). The user terminal, if any, increases the payload by the amount of data discarded in the last serving interval because the delay for discarded data is exceeded (block 1114). Even if the discarded data is not transmitted, increasing the payload will ensure that the user terminal requires sufficient time to serve future data within the delay requirements so that no data is discarded. The user terminal determines one or more rates for transmission to the destination station, for example using process 1000 in FIG. 10 (block 1116). The user terminal then calculates the amount of time (or data time) required to transmit the entire payload at the selected rate (s) (block 1118). The user terminal also determines the amount of time (or overhead time) required to transmit the overhead for data transmission, as described, for example, in FIG. 8 (block 1120). The user terminal then calculates the amount of time required as the sum of the data time and the overhead time (block 1122). The user terminal then transmits this requested time duration to the access point.

4. send mode  And Rate  Selection

The wireless network 100 may support multiple transmission modes for improved performance and greater flexibility. Table 3 lists some transmission modes and their brief descriptions.

Table 3

Transmission mode                             Explanation Steered Mode Multiple data flows are sent on multiple vertical spatial channels (or native modes) of MIMO channels Non-steered mode Multiple data flows are sent on multiple spatial channels on MIMO channels

The MIMO system may employ spatial spreading for the non-steered mode to improve performance. Spatial spreading, also called virtual random transmission steering (PRTS), allows data transmission to benefit from the obtained changes observed over an ensemble of effective channels and to not lock onto a single bad channel realization for extended periods of time. The source station performs spatial processing by another steering vector.

Each transmission mode has different capacities and requirements. Stinged mode may typically be used if better performance can be achieved and the source station has enough channel and steering information to transmit data on the vertical spatial channel. The non-steered mode does not require any channel information, but the performance may not be the same as the steered mode. Suitable transmission modes may be selected to be used depending on the available channel information, the capacity of the source and destination stations, system requirements, and the like. For clarity, the following description is for data transmission from source (sending) station A to destination (receiving) station B.

는 스테이션 B로부터 스테이션 A로의 링크 (B, A) 에 대한 채널 응답 행렬이고, 여기서

는 H _ab의 전치 행렬이다. 역수의 채널에 대해, 스테이션 A는 스테이션 B에 의해 스테이션 A로 전송되는 MIMO 파일럿에 기초하여 링크 (A, B)에 대한 MIMO 채널 응답을 추정할 수 있다. 스테이션 A는 그 후 H _ab의 고유 모드상에서 데이터를 송신하는데 이용되는 고유 벡터를 획득하기 위해 (예를 들 면, 특이값 분해(singular value decomposition)를 이용하여) H _ab를 대각 행렬로 할 수도 있다. 고유 스티어링은 스티어링된 모드에 대한 공간 프로세싱을 언급한다. 고유 벡터는 고유 모드상의 송신을 허용하는 스티어링 벡터이다.For the steered mode, station A is on the N _S unique mode of the MIMO channel formed by the N _T transmit antenna at station A and the N _R receive antenna at station B with N _S ≤min {N _T , N _R } Send the data. The MIMO channel may be characterized by the N _R xN _T channel response matrix H. For a wireless TDD MIMO network, the channel responses of the two opposing links may be assumed to be reciprocal of each other after the calibration procedure has been performed to calculate the difference in the frequency response of the transmitting and receiving RF chains. Thus, if H _ab is the channel response matrix for the link (A, B) from station A to station B, then H _ba =

Is the channel response matrix for the link (B, A) from station B to station A, where

Is the transpose of H _ab . For the reciprocal channel, station A may estimate the MIMO channel response for link (A, B) based on the MIMO pilot sent by station B to station A. Station A may then make H _ab a diagonal matrix (e.g., using singular value decomposition) to obtain an eigenvector used to transmit data on the eigen mode of H _ab . . Inherent steering refers to spatial processing for a steered mode. The eigenvector is a steering vector that allows transmission on eigenmodes.

For the non-steered mode, station A transmits data to station B on the N _S spatial channel of the MIMO channel. Station A may transmit from its N _T transmit antenna to the N _S data flow without any spatial processing. Station A may also perform spatial processing with a steering vector known as station B to achieve spatial spreading. If station B does not need to be aware of the spatial processing performed by station A, station A may also perform spatial spreading on both MIMO pilot and data transmitted to station B.

The MIMO pilot is a pilot that allows the receiving station to characterize the MIMO channel. The non-steered MIMO pilot is a pilot consisting of N pilot transmissions transmitted from N transmit antennas, where pilot transmissions from each transmit antenna are identifiable by the receiving station. The steered MIMO pilot is a pilot transmitted on the native mode of the MIMO channel. Steered MIMO pilots transmitted on links (A, B) may be generated based on eigenvectors for links (A, B), and these eigenvectors are steered MIMO pilots received on links (B, A). Or may be derived from a non-steered MIMO pilot.

For the steered mode, station A uses the eigenvectors to perform eigensteering. The frequency at which the eigenvectors change depends on the variability in the MIMO channel. Station A may also use different rates for each unique mode. For the non-steered mode, station A may use the same rate for all spatial channels. The rate supported by each unique mode or spatial channel is determined by the SNR achieved by that unique mode / spatial channel. The SNR for each link may be estimated at each time a steered or nonsteered MIMO pilot is received over that link. The receiving station may calculate a set of rates supported by the link and send this rate information to the transmitting station.

In one embodiment, the particular transmission mode used for data transmission is determined based on the age of the available channel information. The rate for the spatial channel is determined based on the SNR estimates for the spatial channel and their age. Each station may track when a MIMO pilot is transmitted to and received from another station. Each station may use this information to determine the age and quality of currently available channel information. Table 4 shows a list of variables used in the following description of transmission mode and rate selection by station A for data transmission from station A to station B.

Table 4

   symbol                             Explanation

(A→B,n)

(A → B, n) The last time station A transmits a non-steered MIMO pilot to station B as determined in frame n.

(A → B, n) The latest time station A transmits a MIMO pilot steered to station B as determined in frame n.

(A ← B, n) The time when station A received a recently non-steered MIMO pilot from station B as determined in frame n.

(A ← B, n) The time when station A received a recently steered MIMO pilot from station B as determined in frame n.

Processing delay for non-steered MIMO pilot to obtain channel information

Processing Delay for Steered MIMO Pilot to Acquire Channel Information d _snr Processing Delay for MIMO Pilot to Acquire SNR / Rate Information

Maximum age to allow use of channel information

Maximum age to allow use of SNR / rate information SNR (A ← B, n) A set of SNR derived from, for example, the initial rate (s) received from station B, obtained from station B, for example. t _snr (A ← B, n) The time obtained by station A by SNR (A ← B, n)

의 딜레이가 채널 정보를 획득하기 위해 비스티어링된 MIMO 파일럿을 프로세스하는데 초래되고, 또는 동등하게, 채널 정보는 비스티어링된 MIMO 파일럿을 수신하는

초 후에 이용가능하다. 따라서, 현재 채널 정보를 유도하는데 이용될 수 있는 최근의 비스티어링된 MIMO 파일럿은 최소

초보다 이전에 수신되었다. 이 최근의 비스티어링된 MIMO 파일럿은 Station A may determine the age of channel information available in current frame n (or “current channel information”) as follows. If the current channel information is derived from a non-steered MIMO pilot received from station B, then the age of this information is equal to the age of the non-steered MIMO pilot. But,

A delay of is caused to process the non-steered MIMO pilot to obtain channel information, or equivalently, the channel information is received to receive the non-steered MIMO pilot.

Available in seconds. Thus, the latest non-steered MIMO pilot that can be used to derive current channel information is minimal.

Received earlier than seconds. This recent non-steered MIMO pilot

식 (6)

Formula (6)

Is sent in the latest frame k _u that satisfies.

Equation (6) determines the most recent frame k _u that a non-steered MIMO pilot received in a frame can be used to derive current channel information. The age of the current channel information derived from the non-steered MIMO pilot is then

식 (7)

Formula (7)

May be expressed as where Age ^u = -∞ if a non-steered MIMO pilot has not been received.

의 딜레이가 스테이션 B로부터 수신되는 스티어링된 MIMO 파일럿을 프로세스하는데 스테이션 A에 의해 초래되고,

의 딜레이가 스테이션 A에 의해 전송되는 대응하는 비스티어링된 MIMO 파일럿을 프로세스하는데 스테이션 B에 의해 초래된다. 따라서, 현재 채널 정보를 유도하는데 이용될 수 있는 최근의 비스티어링된 MIMO 파일럿은 최소

+

초보다 이전에 수신되었다. 이 최근의 비스티어링된 MIMO 파일럿은 If the current channel information is derived from a steered MIMO pilot received from station B, then the age of this information is in synchronism with the age of the corresponding non-steered MIMO pilot from which the steered MIMO pilot is derived.

Delay of is caused by station A in processing the steered MIMO pilot received from station B,

A delay of is caused by station B to process the corresponding non-steered MIMO pilot sent by station A. Thus, the latest non-steered MIMO pilot that can be used to derive current channel information is minimal.

+

Received earlier than seconds. This recent non-steered MIMO pilot

식 (8)

Formula (8)

Is transmitted in the latest frame k _S that satisfies.

Equation (8) determines the most recent frame k _S that a non-steered MIMO pilot sent in a frame can be used to derive current channel information. The age of the current channel information derived from the steered MIMO pilot is then

식 (9)

Formula (9)

It may be expressed as, where Age ^s = −∞ if no steered MIMO pilot is received.

Age of current channel information, Age _{ch _ inf} (n)

Age _{ch _ inf} (n) = min (Age ^u , Age ^s ) equation (10)

It may be expressed as.

The transmission mode may be selected based on the age of the current channel information,

식(11)

Formula (11)

Same as

The transmission mode may also be selected based on other suitable information, such as the time floating nature of the MIMO channel. For example, Age ^u and Age ^s may be a function of the channel type (eg, fast or slow fading), other age thresholds may be used for other channel types, and so on.

Station A may select the final rate (s) for data transmission to station B based on the age on which the MIMO pilot is used to derive these rate (s). The rate (s) supported by the link (A, B) from station A to station B depends on the received SNR at station B, which may be estimated based on the steered or non-steered MIMO pilot received from station A. . Station B may convert the received SNR at the initial rate (s) and then send these initial rate (s) back to station A. Station A may estimate the SNR received at station B based on the initial rate (s) obtained from station B. For example, station A may provide each initial rate to a reverse lookup table that may then provide the SNR required for the initial rate. The set of SNRs available to station A in the current frame n (or " current SNR information &quot;) is denoted SNR (A ← B, n) and is obtained at time t _snr (A ← B, n).

The delay of d _snr includes (1) processing the steered or non-steered MIMO pilot for station B, estimating the received SNR, and returning the initial rate (s) to station A and (2) to station A. Resulting in processing the initial rate (s) to obtain current SNR information. Thus, the latest MIMO pilot that can be used to obtain current SNR information is transmitted by the station at least d _snr seconds ago,

식 (12)

Formula (12)

May be identified as.

Equation (12) determines the most recent frame i for which the latest steered or non-steered MIMO pilot for the frame can be used to derive the current SNR information.

The age of the current SNR information has since

식 (13)

Formula (13)

Can be expressed as

) 을 초과하면, 그 후 SNR 정보는 너무 스테일 (stale) 한 것으로 간주되고 이용으로부터 폐기될 수도 있다. 이러한 경우에서 가장 강건한 송신 모드 및 가장 낮은 레이트 (예를 들면, 비스티어링된 모드에서의 가장 낮은 레이트) 는 스테이션 B로의 데이터 송신에 이용될 수도 있다. 현재 SNR 정보의 에이지가 SNR 에이지 임계값보다 작으면, 그 후 스테이션 A에 의해 획득되는 SNR은 SNR 정보의 에이지에 기초하여 조정될 수도 있고, 조정된 SNR은 그 후 최종 레이트를 선택하는데 이용될 수도 있다. SNR 조정은 다양한 방식으로 실행될 수도 있다.The final rate (s) may be selected based on the current SNR information, the age of the SNR information, and other possible information. For example, the age of the current SNR information is the SNR age threshold (or Age _{snr _inf} (n)>

), The SNR information is then considered too stale and may be discarded from use. In this case the most robust transmission mode and the lowest rate (eg, the lowest rate in the non-steered mode) may be used for data transmission to station B. If the age of the current SNR information is less than the SNR age threshold, then the SNR obtained by station A may be adjusted based on the age of the SNR information, and the adjusted SNR may then be used to select the final rate. . SNR adjustment may be performed in various ways.

If the steered mode is selected for use, then station A receives the initial rate for each unique mode, determines the SNR required for each unique mode based on the initial rate for that unique mode, It is also possible to adjust the SNR required for each unique mode based on the age. For example, the SNR back-off is based on the linear function of the age:

May be calculated as: where SNR _{adj _ rate} is the rate of adjustment to the SNR (eg, SNR _{adj _ rate} = 50 dB / sec).

The SNR adjusted for each unique mode is then:

식 (15)

Formula (15)

May be calculated as: where SNR _{req , m} (n) is the SNR (obtained from initial rate) required for eigen mode m;

는 스티어링된 모드에 대한 백오프이고 (예를 들면,

=0dB);

Is the backoff for the steered mode (e.g.,

= 0 dB);

(n)은 스티어링된 모드에 대한 고유 모드 m에 대해 조정된 SNR이다.

(n) is the SNR adjusted for the natural mode m for the steered mode.

Station A may provide the adjusted SNR for each unique mode in the lookup table which then provides the final rate for that unique mode. Station A may use the same lookup table that station B uses to obtain the initial rate for each unique mode or other lookup table.

If the non-steered mode is selected for use, then station A may then receive an initial rate for each unique mode and determine a single final rate for data transmission in the non-steered mode. Adjusted SNR is for each unique mode

식 (16)

Formula (16)

은 비스티어링된 모드 (예를 들면,

=3dB) 에 대한 백-오프이고;May be calculated as

Is a non-steered mode (e.g.,

= 3 dB) back-off;

(n)는 비스티어링된 모드에 대한 고유 모드 m에 대해 조정된 SNR이다.

(n) is the SNR adjusted for the eigenmodes m for the non-steered mode.

은 예를 들면, 모든 N_S 공간 채널에 걸쳐 분포된 전체 송신 전력, 각 데이터 패킷에 거치는 SNR에서의 변화에 기인한 성능에서의 손실 등과 같은 다양한 요소를 계산하는데 이용될 수도 있다.

,

및 SNR_{adj _ rate}는 컴퓨터 시뮬레이션, 실험에 의한 추정 등에 의해 결정될 수도 있다.

For example, all N _S It may be used to calculate various factors such as total transmission power distributed over the spatial channel, loss in performance due to changes in SNR across each data packet, and the like.

,

And SNR _{_ adj rate} may be determined by estimation based on computer simulation, lab.

(n)≥SNR_th이면, 그 후 고유 모드는 N_sch(n)로 카운트된다. 따라서 비스티어링된 모드에 이용되는 공간 채널의 수는 고유 모드의 수 이하이거나, N_sch(n)≤N_S이다. 비스티어링된 모드에 대한 평균 SNR, SNR_avg(n)은 The number of spatial channels used for data transmission in the current frame n, N _sch (n), may be determined by counting the number of " good " eigen modes by the SNR threshold, adjusted SNR greater than SNR _th . For each unique mode m,

If (n) ≧ SNR _th , then the eigenmode is counted as N _sch (n). Thus, the number of spatial channels used in the non-steered mode is equal to or less than the number of eigenmodes, or N _sch (n) ≦ N _S. Average SNR for non-steered mode, SNR _avg (n) is

식(17)

Formula (17)

May be calculated as

Station B selects an initial rate for each unique mode based on the assumption that all N _S unique modes are used for data transmission and equal transmit power is used for all unique modes. If fewer spatial channels than N _S are used in the non-steered mode, then higher transmit power may be used for each selected spatial channel. The first term on the right side of the equal sign in equation (17) describes the use of higher transmit power for each spatial channel when a spatial channel smaller than N _S is selected for use. The second term on the right side of the equal sign in equation (17) is the average SNR (in dB) for the N _sch (n) spatial channel selected for use in frame n.

Station A may provide an average SNR to the lookup table, which then provides the final rate for the non-steered mode. Station A may use the same lookup table, or another lookup table, that station B uses to obtain the initial rate for the non-steered mode. Alternatively, station A may receive a single initial rate for the non-steered mode from station B. In this case, station A determines the required SNR for the non-steered mode based on the initial rate, adjusts the required SNR based on the age of the SNR information, and enters the non-steered mode based on the adjusted SNR. The final rate may be determined.

는 채널 타입의 함수 (예를 들면, 빠른 또는 느린 페이딩) 일 수도 있다. 간략화를 위해, SNR 백-오프는 식 (14)에 도시된 바와 같이, 에이지의 선형 함수에 기초하여 계산되었다. 일반적으로, SNR 백-오프는 에이지의 어떤 선형 함수 또는 비선형 함수 및/또는 다른 파라미터일 수도 있다.The final rate (s) may also be determined based on other suitable information such as the time floating nature of the MIMO channel. For example, SNR back-off, SNR _{age _ bo} (n), and / or age threshold,

May be a function of the channel type (eg, fast or slow fading). For simplicity, the SNR back-off was calculated based on the linear function of the age, as shown in equation (14). In general, the SNR back-off may be any linear or nonlinear function and / or other parameter of the age.

5. system

12 shows a block diagram of an access point 110 and two user terminals 120x and 120y in the wireless network 100. Access point 110 is N _ap Antennas 1224a-1224ap are equipped. The user terminal 120x is equipped with a single antenna 1252x and the user terminal 120y is equipped with N _ut antennas 1252ya to 1252yu.

On the uplink, at each user terminal 120 scheduled for uplink transmission, the transmit (TX) data processor 1288 receives traffic data from the data source 1286 and receives data (eg, from the controller 1280). , Block ACK). TX data processor 1288 encodes, interleaves, modulates, and provides data symbols based on the final rate (s) selected for the user terminal. At each user terminal with multiple antennas, TX spatial processor 1290 performs spatial processing (if applicable) on the data symbols for the steered or non-steered mode and provides transmit symbols. The pilot symbols may be multiplexed in data symbols or transmit symbols as needed. Each transmitter unit (TMTR) 1254 processes each transmit symbol flow to generate an uplink signal (eg, convert to analog, amplify, filter, and frequency upconvert). Uplink signal (s) from transmitter unit (s) 1254 are transmitted from antenna (s) 1252 to an access point.

At access point 110, N _ap antennas 1224a-1224ap receive uplink signals from user terminals. Each antenna 1224 provides the received signal to a respective receiver unit (RCVR) 1222 that processes the received signal and provides the received symbol. Receive (RX) spatial processor 1240 performs receiver spatial processing on symbols received from all receiver units 1222 and provides detected symbols that are estimates of data symbols sent by the user terminal. RX data processor 1242 demodulates, deinterleaves, and decodes the detected symbols for each user terminal based on the final rate (s) used by each user terminal. Decoded data for each user terminal is stored in data sink 1244 and / or provided to controller 1230.

On the downlink, at the access point 110, the TX data processor 1210 receives traffic data from the data source 1208 for all user terminals scheduled for downlink transmission, and receives data from the controller 1230. For example, block ACK) is controlled and information is scheduled from the scheduler 1234. TX data processor 1210 encodes, interleaves, and modulates data for each user terminal based on the final rate (s) selected for the user terminal. TX spatial processor 1220 performs spatial processing on data symbols (if applicable), multiplexes in pilot symbols, and provides transmit symbols for each user terminal for the steered or non-steered mode. Each transmitter unit 1222 processes each transmit symbol stream and generates a downlink signal. N _ap N _ap downlink signals from the transmitter unit 1222 N _ap It is transmitted from the antenna 1224 to the user terminal.

At each user terminal 120, the antenna (s) 1252 receives a downlink signal from the access point 110. Each receiver unit 1254 processes the signal received from the associated antenna 1252 and provides the received symbol. At each user terminal with multiple antennas, RX spatial processor 1260 performs receiver spatial processing on the received symbols from all receiver units 1254 and provides the detected symbols. RX data processor 1270 demodulates, deinterleaves, and decodes the detected symbols and provides decoded data for the user terminal.

Controllers 1230, 1280x and 1280y command operation at access point 110 and user terminals 120x and 120y, respectively. The controller 1280 for each user terminal may send feedback information (eg, initial rate (s), required duration, etc.) to the access point. Memory units 1232, 1282x and 1282y store program codes and data used by controllers 1230, 1280x and 1280y, respectively. As described above, the scheduler 1234 executes scheduling for the access point and the user terminal. The scheduler 1234 may exist at an access point or at another network entity, as shown in FIG. 12.

13 shows a block diagram of an embodiment of a CSI processor 1228 at an access point 110. Channel estimator 1312 receives the pilot sent by each user terminal transmitting on the uplink and derives a channel response estimate for the user terminal. The transmission mode selector 1314 selects a steered or non-steered mode for each user terminal with multiple antennas, for example, based on the channel information and its age, as described above. SNR estimator 1316 estimates SNR for each user terminal based on the pilot received from the user terminal. Rate selector 1318 determines the final rate (s) for each user terminal based on the SNR estimate from SNR estimator 1316 or the initial rate (s) sent by the user terminal, as described above. The CSI processor 1278 for each user terminal may also be provided in a similar manner as the CSI processor 1228.

14 shows a block diagram of an embodiment of a scheduler 1234 at an access point 110. Calculation unit 1412 receives throughput, delay, and / or other requirements for flow on each link recorded by scheduler 1234 and calculates a serving interval for the link, as described above. The memory unit 1414 stores information for each registered link, such as the serving interval of each recorded link, the last served time, status flags, priority information, and the like. The link selector 1416 selects a link for transmission based on the serving interval of the link, and / or other criteria. Calculation unit 1418 calculates TXOP duration for each selected link based on (1) queue / buffer information and rate (s) for the link or (2) the required duration for the link. The link scheduler 1420 allocates the link selected by the TXOP calculated by unit 1418, updates the scheduled link, and provides scheduling information for the scheduled link.

The link selector 1416 and link scheduler 1420 may execute a process as shown in FIGS. 6 and 7. Calculation unit 1418 may execute a process as shown in FIGS. 8 and 9. The rate selector 1318 in FIG. 13 may execute a process as shown in FIGS. 10 and 11.

Those skilled in the art will appreciate that information and signals may be represented using any of a variety of other technologies and expertise. For example, data, instructions, commands, information, signals, bits, symbols, and chips that may be referred to throughout the above description may be applied to voltages, currents, electromagnetic waves, magnetic fields or particles, optical fields or particles, or any combination thereof. May appear.

 Those skilled in the art will also recognize that the various exemplary logical blocks, modules, circuits, and algorithm steps described in connection with the embodiments disclosed herein may be provided as electronic hardware, computer software, or a combination of both. To clearly illustrate this hardware and software compatibility, various illustrative components, blocks, modules, circuits, and steps have been described above generally in terms of their functionality. Whether such functionality is provided as hardware or as software depends on the particular application and design constraints imposed on the overall system. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.

The various illustrative logic blocks, modules, and circuits described in connection with the embodiments disclosed herein may be general purpose processors, digital signal processors (DSPs), application specific semiconductors (ASICs), field programmable gate arrays (FPGAs), or other programmable logic devices. May be provided or executed by discrete gate or transistor logic, discrete hardware components, or any combination designed to carry out the functions described herein. The multipurpose processor may be a microprocessor, but in the alternative, the processor may be any conventional processor, controller, microcontroller, or state machine. A processor may also be implemented as a combination of computing devices, eg, a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors associated with a DSP core, or some other such configuration.

The steps of a method or algorithm described in connection with the embodiments disclosed herein may be included directly in hardware, in a software module executed by a processor, or in a combination of the two. The software module may be in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disks, removable disks, CD-ROMs, or any other form of storage medium known in the art. An exemplary storage medium is coupled to the processor that can read information from and write information to the storage medium. In the alternative, the storage medium may be whole of a processor. The processor and the storage medium may reside in an ASIC. The ASIC may reside in a user terminal. In the alternative, the processor and the storage medium may reside as discrete components in a user terminal.

Items are included here for reference and to help locate them in any section. These items are not intended to limit the scope of the concepts described therein, and these concepts may have applicability in other sections throughout the entire specification.

The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims (60)

Obtaining one or more requirements for one or more flows of data; And Scheduling the one or more flows for transmission based on the one or more requirements for the one or more flows. The method of claim 1, If yes, determining a serving interval for each of the one or more flows based on the requirements for the flow. The method of claim 2, Scheduling the one or more flows for the transmission, Scheduling each flow by one or more transmission opportunities (TXOPs) in each serving interval for the flow if system resources are available. The method of claim 1, Determining delay requirements for each flow; And Selecting a serving interval for each flow to satisfy the delay requirement for each flow. The method of claim 1, Determining delay requirements for each flow; Determining the number of transmissions allowed for any one data unit; And Selecting a serving interval for each flow based on the delay requirement for the flow and the number of transmissions allowed for any one data unit. The method of claim 1, Determining a throughput requirement for each flow; And Selecting a serving interval for each flow to satisfy a throughput requirement for the flow. The method of claim 1, Determining a feedback requirement for each flow; And Selecting a serving interval for each flow to satisfy a feedback requirement for the flow. The method of claim 1, Determining a rate for sending an acknowledgment for each flow; And Selecting a serving interval for each flow to satisfy an acknowledgment rate for the flow. A controller operative to obtain one or more requirements for one or more flows of data; And And a scheduler operative to schedule one or more flows for transmission based on the one or more requirements for the one or more flows. The method of claim 9, The scheduler is further operative to determine a serving interval for each of the one or more flows, if any, based on the requirements for the flow. The method of claim 10, The scheduler is further operative to schedule each flow by one or more transmission opportunities (TXOP) in each serving interval for the flow when system resources are available. The method of claim 9, And the wireless network supports multiple-input multiple-output (MIMO) transmission. Means for obtaining one or more requirements for one or more flows of data; And Means for scheduling the one or more flows for transmission based on the one or more requirements for the one or more flows. The method of claim 13, Means for determining a serving interval for each of the one or more flows, if any, based on the requirements for the flow. The method of claim 14, The means for scheduling the one or more flows for the transmission includes means for scheduling each flow by one or more transmission opportunities (TXOP) in each serving interval for the flow when system resources are available. Devices on the network. Identifying one or more links, each link carrying one or more flows of data; Obtaining one or more requirements for the one or more flows for each link; And Scheduling the one or more links for transmission based on the one or more requirements for the one or more flows for each link. The method of claim 16, Determining a serving interval for each link based on the one or more requirements for the one or more flows for the link. The method of claim 16, For each of the one or more links, If any, determining a serving interval for each flow for the link based on the requirements for the flow; And Determining a serving interval for the link based on one or more serving intervals determined for the one or more flows for the link. The method of claim 17, Determining a serving interval for each link, Determining a feedback requirement for the link; And Selecting a serving interval for the link to satisfy the feedback requirement for the link. The method of claim 17, Scheduling each link by one or more transmission opportunities (TXOPs) in each serving interval for the link if system resources are available. A controller identifying one or more links carrying one or more flows of data, the controller operative to obtain one or more requirements for the one or more flows for each link; And And a scheduler operative to schedule the one or more links for transmission based on the one or more requirements for the one or more flows for each link. The method of claim 21, The scheduler is further operative to determine a serving interval for each link based on the one or more requirements for the one or more flows for the link. The method of claim 22, The scheduler is further operative to schedule each link by one or more transmission opportunities (TXOP) in each serving interval for the link when system resources are available. Means for identifying one or more links carrying one or more flows of data; Means for obtaining one or more requirements for the one or more flows for each link; And Means for scheduling the one or more links for transmission based on one or more requirements for the one or more flows for each link. The method of claim 24, And means for determining a serving interval for each link based on the one or more requirements for the one or more flows for the link. The method of claim 25, Means for scheduling one or more links for the transmission comprises means for scheduling each link by one or more transmission opportunities (TXOP) in each serving interval for the link if system resources are available. Device. Identifying one or more links to schedule for data transmission based on requirements for the one or more links; Determining a transmission opportunity (TXOP) for each of the one or more links; And Scheduling each link by a TXOP determined for the link. The method of claim 27, Identifying one or more links to schedule for the data transmission, Identifying a link that is not scheduled for data transmission within a serving interval for the link. The method of claim 27, Identifying one or more links to schedule for the data transmission, Identifying a link that was not properly served in a previous scheduling interval. The method of claim 27, Determining TXOP for each of the one or more links, Determining an amount of data to transmit for the link, Determining one or more rates to use for the link, and Calculating a duration of the TXOP for the link based on one or more rates and amount of data to transmit for the link. The method of claim 27, Determining TXOP for each of the one or more links comprises: Determining a duration of a TXOP for the link based on the duration required for the link. The method of claim 27, Classifying the one or more links based on priority; And wherein the one or more links are scheduled in sorted order. The method of claim 27, Classifying the one or more links based on delay requirements for the one or more links, And wherein the one or more links are scheduled in sorted order. A selector operative to identify one or more links to schedule for data transmission based on requirements for the one or more links; A computing unit operative to determine a transmission opportunity (TXOP) for each of the one or more links; And And a scheduler operative to schedule each link by the TXOP determined for the link. The method of claim 34, wherein And the selector is operative to identify a link that is not scheduled for data transmission within a serving interval for the link. 36. The method of claim 35 wherein And the calculating unit is operative to determine the duration of the TXOP for each link based on the duration required for the link or one or more rates and buffer sizes for the link. Means for identifying one or more links to schedule for data transmission based on requirements for the one or more links; Means for determining a transmission opportunity (TXOP) for each of the one or more links; And Means for scheduling each link by the TXOP determined for the link. The method of claim 37, Means for identifying one or more links to schedule for the data transmission, Means for identifying a link that is not scheduled for data transmission within a serving interval for the link. The method of claim 38, Means for determining the duration of the TXOP for each link based on the duration required for the link or one or more rates and buffer sizes for the link. Identifying one or more links to be scheduled for data transmission in the first direction; Scheduling the one or more links for reverse transmission in a second direction opposite the first direction; And Scheduling the one or more links for data transmission in the first direction. The method of claim 40, And the one or more links are scheduled for reverse transmission in a first frame and for data transmission in a second frame after the first frame. 42. The method of claim 41 wherein And the second frame is subsequent to the first frame. The method of claim 40, Scheduling one or more links for the reverse transmission, For each link that has no data to transmit in the second direction, scheduling the link for reverse transmission of null packets having one or more rates used for data transmission in the first direction. A selector operative to identify one or more links to be scheduled for data transmission in the first direction; And A scheduler operative to schedule the one or more links for reverse transmission in a second direction opposite the first direction and to schedule the one or more links for data transmission in the first direction. Device. The method of claim 44, The reverse transmission on each link includes a pilot. The method of claim 44, The reverse transmission for each link includes one or more rate or one or more signal to noise ratio (SNR) estimates for the link. The method of claim 44, The reverse transmission for each link includes an acknowledgment for previous data transmission in the first direction for the link. Means for identifying one or more links to be scheduled for data transmission in the first direction; Means for scheduling the one or more links for reverse transmission in a second direction opposite the first direction; And Means for scheduling the one or more links for data transmission in the first direction. Identifying one or more data flows to be scheduled for data transmission; Determining a rate for transmission of an acknowledgment for each of the one or more data flows; And Scheduling reverse transmission for each data flow to achieve the acknowledgment rate for the data flow. The method of claim 49, Determining a serving interval for each data flow to achieve the acknowledgment rate for the flow. The method of claim 49, Identifying one or more acknowledgment (ACK) flows corresponding to the one or more data flows; Determining a transmission rate for each ACK flow; And Scheduling each ACK flow at a rate determined for the ACK flow. The method of claim 49, Determining the amount of time to schedule for reverse transmission for each data flow based on the estimated data rate, physical layer overhead, acknowledgment (ACK) block size, or any combination thereof for the data flow. Further comprising a data transmission scheduling method. A selector operative to identify one or more data flows to be scheduled for data transmission; A computing unit operative to determine a rate for sending an acknowledgment for each of the one or more data flows; And And a scheduler operative to schedule reverse transmission for each data flow to achieve the acknowledgment rate for the data flow. The method of claim 53 wherein And the acknowledgment is a medium access control (MAC) acknowledgment. The method of claim 54, wherein And the computing unit is further operative to determine a serving interval for each data flow to achieve the acknowledgment rate for the flow. The method of claim 53 wherein The selector is further operative to identify one or more ACK flows corresponding to the one or more data flows, The calculating unit is further operative to determine a transmission rate for each ACK flow, The scheduler is further operative to schedule each ACK flow at a rate determined for the ACK flow. The method of claim 56, wherein Wherein each ACK flow carries a Transmission Control Protocol (TCP) ACK for a corresponding data flow. Means for identifying one or more data flows to be scheduled for data transmission; Means for determining a rate for sending an acknowledgment for each of the one or more data flows; And Means for scheduling reverse transmission for each data flow to achieve the acknowledgment rate for the data flow. The method of claim 58, And means for determining a serving interval for each data flow to achieve an acknowledgment rate for the flow. The method of claim 58, Means for identifying one or more ACK flows corresponding to the one or more data flows; Means for determining a transmission rate for each ACK flow; And And means for scheduling each ACK flow at a rate determined for the ACK flow.

Images