JP4453010B2 - Base station, radio communication method, and base transmitter - Google Patents

Description

  The present invention relates to a mobile radio communication system including a radio communication base station and a terminal device, and more particularly to a technique for performing packet scheduling.

  FIG. 1 is a diagram for explaining an outline of a communication method adopted in, for example, cdma2000 1xEV-DO and W-CDMA HSDPA. The terminal 2-X (X is arbitrary) receives a pilot signal transmitted from the base station, measures its signal-to-interference noise power ratio (SINR), and transmits it to the base station 1. The base station 1 compares SINRs reported from all terminals that are communication candidates, selects a terminal having a high SINR, that is, maximizes the throughput, and performs assignment (scheduling) of packets to be transmitted. In this way, the efficiency of the wireless line is realized by transmitting a signal to the optimum terminal (user) at the timing for each frame.

  FIG. 2 is a sequence diagram showing this situation. In the figure, a plurality of terminals A and B, a base station, and a base station control station are described, and a downlink (that is, information transmission from the base station to the terminal) is taken as an example. Transmission information to be transmitted from the base station control station to the terminals A and B is sent in advance to the base station, and transmission information is queued in the base station buffer (waiting for transmission). From each terminal, propagation path information (that is, SINR or related information) indicating the state of the propagation path is transmitted to the base station. Upon receiving the propagation path information, the base station evaluates them according to the evaluation function, and performs scheduling for determining packet transmission to the most appropriate terminal. As a result of scheduling, a line is allocated to a terminal determined to be transmitted, and packet transmission is performed. As a result, the efficiency of the wireless line described above is realized.

  The reason why the line is made efficient by this technique will be briefly described with reference to FIG. FIG. 3 is a diagram showing temporal variation of propagation path information. In the figure, there are two terminals, each propagation path condition (propagation path information) is taken on the vertical axis, and time is shown on the horizontal axis. A curve 3-A indicates the propagation path information of the terminal A, and a curve 3-B indicates the propagation path information of the terminal B. In the figure, the higher the channel information value on the vertical axis (the better), the higher the line efficiency. Radio channel fading occurs due to multipath and the like, and as a result, the channel information of each terminal varies with time independently of each other. At a certain time, terminal A has higher line usage efficiency than terminal B. Therefore, it can be expected that the use efficiency of the line is improved by preferentially assigning terminal A. Also, at another time, terminal B has higher line usage efficiency than terminal A, and it can be expected to improve the line utilization efficiency by giving priority to terminal B at such time. As a result, as shown by a curve 3-C in FIG. 4, the line can be effectively used by allocating the line to a terminal (user) having a good propagation path at that time.

  Actually, for example, the SINR of a terminal close to a base station is high on average, and if a packet is allocated only with the above-mentioned propagation path information, only a specific terminal uses a line. A method using a function is used. This evaluation function is composed of instantaneous SINR / average SINR. By using the average SINR as the denominator, the average SINR disparity between terminals is corrected and the allocation opportunities are equalized, but the terminals that are in a favorable environment with fading peaks are given priority for packet allocation. Thus, the efficiency can be improved.

  FIG. 5 shows the configuration of a base station apparatus according to the prior art. A signal received by the antenna 100 is input to the reception RF circuit 102 via the deplexer 101. In the reception RF circuit 102, the signal having the RF frequency is down-converted using an amplifier, a mixer, and a filter, and converted into a baseband signal. The signal converted into the baseband signal is converted into a digital signal through an AD converter. The converted baseband digital signal is input to the reception baseband circuit 103. The reception baseband circuit 103 processes the baseband digital signal, estimates a propagation path from the pilot signal, and performs detection using the estimation result. Also, control information and received data are demodulated from the detection result, and information is extracted. The extracted control information includes the propagation path information described above. Information taken out by demodulation is taken into the signal processing unit 104 and processed according to its use.

  User data transmitted by the terminal is transmitted to the network 106 via the network interface unit 105. Although not shown in the figure, the base station control station, which is a higher station, is connected to the network 106 and is responsible for collecting information and transmitting it to the core network. A part of the control information is transmitted to the base station control station and used for radio channel management such as handover. Furthermore, control information relating to propagation path information is stored in the signal processing unit 104 and used for packet allocation of a packet scheduler built in the processing unit. Information transmitted to the terminal is transmitted from the network 106.

  Transmission information from the network 106 is taken into the signal processing unit 104 via the network interface unit 105 and temporarily saved in the storage unit 107 for queuing. The packet scheduler built in the signal processing unit 104 assigns packets using proportional fairness as described above. Based on the assignment result, information to be transmitted is extracted from the storage means 107 and sent to the transmission baseband circuit 108. The scheduler built in the signal processing unit 104 selects a set of modulation scheme and coding rate from the SINR and notifies the transmission baseband circuit 108 of the selected set. The transmission baseband circuit 108 performs baseband modulation of transmission information according to the procedure. The transmission signal subjected to baseband modulation is sent to the transmission RF circuit 109, up-converted to a radio frequency through a mixer or filter, and then amplified through a power amplifier or the like, and through a deplexer 101. Output from the antenna.

  In conventional packet switching systems, schedulers have been studied with the highest priority on efficiency. However, it has become important to support services that guarantee bandwidth such as voice, and the importance of schedulers that support QoS guaranteed type sessions as well as these best effort types has increased. For example, in “Frequency domain type scheduling characteristics considering traffic in downlink OFCDM broadband wireless access” by Oto et al., IEICE 2004 Society Conference B-5-101, announced in September 2004 In the conventional scheduler method, a scheduling evaluation function of Formula 1 in which a correction term for a QoS guaranteed type session is added to proportional fairness is proposed. FIG. 6 is a flowchart of the scheduler according to the prior art. The scheduler evaluates the evaluation function for all users at once. In the evaluation function, the evaluation function P is defined in consideration of the five elements of instantaneous SINR, average SINR, retransmission or new assignment flag, time until delay deadline, and priority.

  However, since the evaluation function reflects the evaluation of the instantaneous SINR even in the QoS guaranteed type session, the SINR can be allocated in the future even though the bandwidth guarantee is considered. Allocation is implemented early because As a result, there is a problem that the degree of freedom of allocation is lowered and the efficiency of the entire system is lowered. To cope with this problem, a scheduling method that takes into account the risk management that predicts the future based on the current line state and makes assignment impossible is necessary, but the prior art does not have such a scheduling method.

  In addition, there is a need for a method for evaluating QoS guaranteed communication and best effort communication with the same index. However, the proportional fairness described above, that is, the best effort evaluation function, Whereas it is an evaluation method that competes for how good the state of the propagation path at the time is, QoS guaranteed type evaluation is an evaluation method that competes for the degree of urgency or dissatisfaction of each terminal. There is a fundamental reason that the comparison itself is difficult and the parameters become complicated.

Oto et al., “Frequency Domain Scheduling Characteristics Considering Traffic in Downlink OFCDM Broadband Wireless Access”, IEICE 2004 Society Conference B-5-101, published September 2004

  As already explained, in the prior art, the evaluation of the instantaneous SINR is reflected even in the QoS guaranteed type session. Therefore, it is sufficient to allocate in the future to guarantee the bandwidth. However, since the SINR is good, the assignment is performed at an early stage. As a result, there is a problem that the degree of freedom of assignment is lowered and the efficiency of the entire system is lowered. Also, a method for evaluating QoS guaranteed communication and best effort communication with the same index is necessary. However, the proportional fairness described above, that is, the best effort evaluation function, QoS-guaranteed evaluation is an evaluation method that competes for the degree of urgency or dissatisfaction of each terminal, and compares each with the same formula. There is a fundamental reason why it is difficult to do so and the parameters become complicated.

  QoS-guaranteed and best-effort evaluation functions have different evaluation objectives, and there are two types of scheduling, QoS-guaranteed and best-effort, for issues that would complicate parameters when compared with the same formula. It is solved by dividing into two. Specifically, it is a wireless communication method adopted in a wireless communication system composed of a terminal and a base station, and the communication between the base station and the terminal is a best effort type in which throughput varies depending on the state of the wireless line. When communication and QoS guaranteed type communication that defines connection conditions according to services coexist, a packet scheduler that determines the use of a radio channel for each communication first performs scheduling only with QoS guaranteed type communication, This is solved by a wireless communication system characterized by performing best-effort communication scheduling for an empty channel that has not been scheduled in QoS-guaranteed communication.

  Further, a problem in which the instantaneous SINR evaluation is reflected in the QoS guaranteed session is solved by a scheduling method that takes into account the risk management in which the future is predicted from the current line state and allocation becomes impossible. Specifically, packet scheduling for determining QoS guaranteed communication uses the average SIR and instantaneous SIR of the radio channel, the request SIR necessary to satisfy the required QoS, and the margin until the packet transmission deadline. This is solved by a wireless communication system characterized by determining a transmission packet.

  According to the present invention, the parameters of the evaluation function of the scheduler are not complicated even when QoS guaranteed type and best effort type communication coexist. In addition, since the instantaneous SINR evaluation is excessively reflected even in the QoS guaranteed type session, there is no problem that the efficiency is lowered, and the efficiency is improved.

  A first embodiment of the present invention will be described with reference to FIG. FIG. 7 is a flowchart of one embodiment according to the present invention. In FIG. 7, the major difference from FIG. 6 which is the flow chart of the prior art is that the QoS guarantee type communication and the best effort type packet scheduling are completely divided into two (step 301 and step 302). It is. In addition, in the QoS guaranteed communication, not only the instantaneous SINR and the average SINR but also new elements such as the required SINR are added.

  In order to perform scheduling collectively when QoS guaranteed type and best effort type communication coexist, it is necessary to evaluate QoS guaranteed type communication and best effort type communication using the same index. However, proportional fairness, that is, the best-effort evaluation function, is an evaluation method in which each terminal competes for how good the state of the propagation path at that time is. On the other hand, the QoS guarantee type evaluation is an evaluation using an evaluation function competing for the degree of urgency or dissatisfaction of each terminal, and the comparison between the two is different. Therefore, it is difficult to compare the two with the same evaluation function, and there is a fundamental reason that the parameters become complicated.

The present invention solves the problem by completely separating the scheduler process by QoS guaranteed communication and best effort communication. As a result, parameter management between the two types of communication can be reduced. Specifically, in the QoS guarantee type evaluation, the priority is set higher and scheduling is performed first (step 301). In the QoS guarantee type scheduling (step 301), the urgency level or dissatisfaction level of each communication is compared, and a line is allocated to a terminal having the highest urgency level. At this time, if the value of the evaluation function of the terminal (or transmission information) having the highest degree of urgency is equal to or less than the threshold value, it is not necessary to immediately assign a line to the QoS guaranteed terminal, and QoS guaranteed communication is performed. See off the assignment. The communication frame for which allocation has been postponed is then subjected to best-effort scheduling (step 302). The best-effort communication evaluation index is implemented by conventional proportional fairness. By dividing into two stages in this way, the simultaneous scheduling of QoS guaranteed type and best effort type communication, which are different evaluation functions, can be avoided, and the control of complicated parameters for legally comparing the two can be eliminated. Can do. Thus, the problem is solved.

  Also in QoS guaranteed communication, an evaluation function is constructed by incorporating a new element called required SINR. The requested SINR is an element that is periodically updated at the start of a QoS session (that is, at the start of the communication) or during the communication. From the packet allocation frequency to guarantee the expected bandwidth and the minimum transmission rate that must be achieved in one packet transmission (or N or less predetermined number of packet transmissions) at that packet allocation frequency Be indexed. The SINR required to satisfy the minimum rate is the required SINR. The relationship between SINR and rate has a predetermined table, and the requested rate is determined by referring to the table. When propagation path information reported from the terminal is a data rate as in cdma2000 1xEVDO, “SINR” in FIG. 7 may be read as “transmission rate”.

The point is that the base station may send a packet at any time as long as the quality is higher than the required SINR. Even if the current quality is good as in the prior art, the packet is sent at a timing when the quality is excessive. It is not necessary, and it can save other communication opportunities that are lost.
From the above, in QoS guaranteed communication, Formula 2 is used as the evaluation function P.

  In this equation, 1-δ is a risk factor. The closer δ is to 1, the smaller the risk, indicating that future predictions are easier from the current line state. Δ in the denominator of the evaluation function P indicates the possibility of a future packet transmission opportunity. The addition number N indicates the number of frames until the allowable delay (the deadline when the corresponding packet must be transmitted), and indicates the future opportunity of the packet of the terminal to be evaluated. By taking δ as the future evaluation, the opportunity discount due to risk is implemented. In addition, the evaluation of the next frame is δ, while the evaluation of N frames ahead is δ raised to the Nth power, so that evaluation at a time point far from the present time can be legitimately performed. For example, when the number of frames up to the allowable delay is 2 and δ is a high value of 0.9, the denominator value is 1.9, and there are almost two transmission opportunities. On the other hand, the transmission opportunity is (1 + 0.5 + 0.25 + 0.125) = 1.875 in a terminal in which the number of frames until the allowable delay is 4 and δ is an unreliable value of 0.5. Therefore, compared with the former, it can be seen that although the latter has a longer number of frames until the allowable delay, there are few expected transmission opportunities and it is necessary to preferentially assign them.

  As described above, in this embodiment, the evaluation function is a value obtained by discounting the parameter ε indicating whether or not current transmission is possible by the number of transmission opportunities. The parameter ε is a flag for determining whether the instantaneous SINR satisfies the required SINR. When the instantaneous SINR is higher than the required SINR, ε is 1, indicating that communication is possible. Further, when the instantaneous SINR is lower than the required SINR, MIN≈0, and the possibility that this packet will be assigned approaches 0 as much as possible. Whether MIN is 0 or close to 0 depends on the concept of the system. In a system based on the concept that transmission should be performed even if the required SINR (required transmission rate) is not reached, a small nonzero value is desirable. On the other hand, if the requested rate is not reached, MIN = 0 should be set in a system based on the concept that transmission is not possible.

  When the packet approaches the K frame by the deadline time, if N ≦ K, it is necessary to preferentially assign this packet, so MAX (≧ 1) is set. As the value of the evaluation function P approaches the deadline (allowable delay), the transmission opportunity decreases and the priority increases. On the other hand, for communication showing a risk factor close to 1 and high, the value of δ is close to 0, and even if there are many opportunities for transmission in the future, it is not reliable, so the priority is set high. . Therefore, the control of Expression 2 can provide a scheduling method that takes into account the risk management that makes assignment impossible, so that the problem can be solved. As a result, MAX = 1 is set so that packets close to the deadline are preferentially transmitted without specially emphasizing the priority of the user nearing the deadline. The risk factor depends on the ratio of fading frequency, instantaneous SINR, and required SINR. For this reason, it is desirable to update the risk factor by an outer loop.

  FIG. 8 shows a flowchart for updating the risk factor by the outer loop. By executing this flow for each terminal, the risk factor can be updated according to each terminal. First, the deadline for delay is determined (step 310), and the risk factor is updated when the delay is closed. For terminals that have reached the update time, it is determined whether the delay target has been met (step 311). For terminals that cannot be satisfied, the risk factor value is updated (that is, the value of δ is decreased) (step 312). The update may be subtracted from the value of δ in a fixed step, or may be multiplied by a fixed coefficient. In any case, the lower limit is determined so that the minimum value of δ is 0 or more. In addition, a counter is provided to determine whether all 50 consecutive transmissions, for example, achieve the delay target (steps 313, 314, 315, and 317). If, for example, 50 consecutive transmissions result in all packets being transmitted without delay, the risk factor value is lowered and updated so as to increase the value of δ (step 316). Specifically, a fixed value may be added, or increase of δ may be realized by multiplying by a fixed coefficient. By such control, the value of the risk factor is changed according to the moving speed of the terminal. Therefore, assignment in consideration of future risks becomes possible, and the problem is solved.

  FIG. 11 is a plot of the probability density of the Rayleigh distribution (variance 1). As an example, the case where the required SINR is smaller than the average SINR is shown. In this case, when the required SINR cannot be satisfied, it is 10% or less. Thus, when the average SINR is sufficiently higher than the required SINR, the risk factor is low. Therefore, the initial value of the risk factor is calculated from the ratio of the average SINR to the required SINR. Since the cumulative probability distribution of the Rayleigh distribution is 1-exp (-x ^ 2/2 / s ^ 2), the instantaneous SINR is less than or equal to the required SINR by setting x to the required SINR and s to the average SINR. Can be calculated. Setting this value as a risk factor completes the initial value setting.

  The change of the risk factor has been described by taking the method of using the outer loop as an example, but this is not restrictive. For example, in a system in which SINR is reported from a terminal every frame, the stability of SINR can be evaluated from several past SINR correlation calculations. The risk factor can be estimated by using this stability. Stability can be assessed using statistical cross-correlation. The autocorrelation is the mean square of the SINR (true value) of each frame <SINR (n) × SINR (n)>, and the correlation value of SINR <SINR (n−1) × SINR (n )>. The stability is obtained, for example, by (autocorrelation / time correlation).

  Further, in the method of performing division at a frequency using OFDM as shown in Non-Patent Reference 1, SINR estimation can be performed for each subcarrier, so that several samples can be obtained for the frequency domain at a time. Can be taken. Therefore, by averaging in the frequency domain, the average number in the time domain can be significantly reduced, and the risk factor can be calculated using the instantaneous SINR.

  The scheduler function described in this embodiment is realized on the signal processing unit 104. The request SINR is a parameter of the scheduler and may be stored in the storage unit 107 in a form linked to a terminal or a session. When scheduling the corresponding terminal or session, it is read from the storage means 107 and used for scheduling. The instantaneous SINR is input from the reception baseband circuit 103. Since the latest information is sent from the terminal for each frame, the value is used for calculation. In a TDD system in which the uplink and downlink are in the same frequency band, it is possible to estimate the propagation path using a specific signal (eg, pilot signal) transmitted from the terminal without sending the estimation result at the terminal. SINR estimation is possible. The average SINR is obtained by averaging the instantaneous SINR. The average may be included in the reception baseband circuit 103 as hardware, or the effect is not affected even if it is implemented by an arithmetic unit incorporated in the signal processing device 104. A flag indicating whether the packet is a retransmission or a new packet is an internal parameter of the scheduler. In packet transmission, an evaluation as to whether or not a packet has been received is received from the terminal as ACK information, and it is determined whether or not to retransmit based on the result. The packet that has received the ACK information from the terminal is dequeued by the signal processing device 104. The time until the delay deadline is determined from the allowable delay of the packet and the time when the packet is queued. The allowable delay is defined by classifying when a session is established. Even priorities are defined and classified when a session is established.

  A second embodiment of the present invention will be described with reference to FIGS. The present embodiment relates to the setting of the allowable delay (deadline time) of packet transmission described above. Packet arrival time varies depending on the network management unit and environmental conditions. Therefore, the deadline is also a Poisson distribution mechanism. As a result, for specific communications, it is inevitable that the deadline of QoS guaranteed communication will be the same for a plurality of users. FIG. 8 is a description of a conventional example, and shows a case in which there are four communications (sessions), and two of these deadlines are shared. As a result, the number of cases in which a plurality of assignments must be performed at the deadline time in the first embodiment increases, and the number of communications that cannot meet the deadline increases.

  For this reason, in setting the deadline time, the deadline time is advanced as shown in FIG. 9 to prevent the congestion of the deadline. With this function, the above problem does not occur and the problem is solved.

  The setting of the deadline time described in the present embodiment can be realized by the functional block of the conventional example of FIG. The deadline time is a scheduler parameter and may be stored in the storage unit 107. When a session start flow occurs, the deadline time is calculated in the program or hardware on the signal processing apparatus, and the result is stored in the storage means 107.

  The present invention can be applied to packet scheduling in a mobile communication system, and is preferably applied to a system in which QoS guaranteed communication and best effort communication are mixed. Further, in the embodiment described above, scheduling for each carrier has been described in particular. However, for example, a system capable of frequency-domain channel division using OFDM can also be used for scheduling of the frequency channel.

The figure explaining the outline | summary of the packet communication system in a cellular communication system. The sequence diagram explaining the outline | summary of the packet communication system in the cellular by a prior art. The figure explaining the effect of the packet communication system in cellular. The figure explaining the effect of the packet communication system in the cellular by a prior art. The figure explaining the structure of a packet communication base station. The flowchart explaining the scheduler of the packet communication system in the cellular by a prior art. The flowchart explaining the scheduler of 1st Example of this invention. The flowchart explaining the update of a risk factor with the scheduler of 1st Example of this invention. The figure which shows the deadline time for every session of a prior art. The figure which shows the deadline time for every session of the 2nd Example of this invention. Cumulative probability density distribution diagram in Rayleigh distribution.

Explanation of symbols

1 base station, 2 (A) terminal A, 2 (B) terminal B, 2 (C) terminal C, 2 (D) terminal D, 3 (A) propagation path information of terminal A, 3 (B) terminal B Propagation path information, 3 (C) propagation path information after scheduling at terminals A and B,
DESCRIPTION OF SYMBOLS 100 Antenna, 101 Deplexer, 102 Reception RF circuit, 103 Reception baseband circuit, 104 Signal processing apparatus, 105 Network interface part, 106 Network, 107 Storage means, 108 Transmission baseband circuit, 109 Transmission RF circuit,
200-207 packet allowable delay (transmission deadline), 300 prior art scheduler steps, 301-317 scheduler steps of the present invention.

Claims (5)

A base station in a wireless communication system comprising a terminal and a base station,
A processing unit that schedules transmission packets, a baseband processing unit that performs baseband processing of transmission packets based on the scheduling, and a transmission unit that transmits a signal subjected to the baseband processing,
When the best effort type communication whose throughput varies according to the state of the radio line and the QoS guaranteed type communication that defines the connection condition according to the service coexist under the jurisdiction, the processing unit firstly guarantees the QoS. For each type of communication, calculate a value obtained by discounting the parameter indicating the current transmission availability by the possibility of a packet transmission opportunity that can be expected from the next time as an evaluation function,
Perform scheduling to determine the communication to be assigned to the channel by the evaluation function,
Next , a base station that performs best-effort communication scheduling for an empty channel that has not been assigned to the QoS-guaranteed communication .   In a wireless communication system composed of a terminal and a base station, a wireless communication method for performing transmission by performing scheduling of a transmission packet at the base station and performing a base panda process based on the scheduling,
  When the best-effort type communication in which the throughput varies according to the state of the radio line and the QoS guaranteed type communication that defines the connection condition according to the service coexist under the jurisdiction of the base station, first, the QoS guaranteed type For each communication, a value obtained by discounting the parameter indicating whether or not the current transmission is possible with the possibility of a packet transmission opportunity that can be expected from the next time is calculated as an evaluation function
  Perform scheduling to determine the communication to be assigned to the channel by the evaluation function,
  Next, a best-effort communication scheduling is performed for an empty channel that has not been assigned to the QoS-guaranteed communication.   The parameter indicating whether or not the current transmission is possible,
  A flag indicating whether the current instantaneous SINR of the line satisfies the required SINR as compared to the required SINR, which is the SINR of the line required to satisfy the QoS required for communication;
  3. The wireless communication method according to claim 2, wherein the value becomes smaller when the instantaneous SINR is lower than the required SINR, thereby reducing the possibility of allocating a line to the communication.   The parameter indicating whether or not the current transmission is possible increases when the number N of frames remaining until the packet transmission deadline is equal to or less than a predetermined number, so that a line is preferentially assigned to a communication approaching the packet transmission deadline. The wireless communication method according to claim 2, wherein:   5. The packet transmission deadline at the same time is set so as not to occur between a plurality of QoS guaranteed communications when setting the packet transmission deadline for each QoS guaranteed communication. Wireless communication method.

Images