CN100586220C - Method for packet scheduling in high speed downlink packet access network - Google Patents

Abstract

The invention discloses a packet scheduling method in high speed downlink packet access HSDPA network. The method includes: A. according to channel quality indication CQI information of priority queuing which is reported by presidial user equipment UE in current transmission time interval TTI, downlink average speed of PQ packet data transmitted to UE, data bulk of PQ packet data cache stored in RNC and scheduling priority index for PQ which is arranged by RNC, a Node B obtains PQ priority; B. the PQ priority sequence is different UE distribution radio resource, the Node B transmites packet data to UE through distributed radio resource. The method not only guaranteens system throughput and UE impartiality, but also guaranteens packet data QoS level and UE priority.

Description

The method of packet scheduling in high speed downlink packet access network

Technical field

The present invention relates to insert the method for in (HSDPA) network grouped data being carried out wireless resource scheduling in high speed downlink packet, particularly a kind of in the HSDPA network method of packet scheduling.

Background technology

The structure of Radio Access Network (RAN, Radio Access Network) as shown in Figure 1, RAN comprises one or several RNS (RNS).A RNS is made up of a radio network controller (RNC) and one or more Node B (NodeB).Interface between RNC and the CN is the Iu interface, and NodeB is connected by Iub interface with RNC.In RAN inside, interconnected between the RNC by the Iur interface, be used for distributing being attached thereto with control or the Radio Resource of relevant NodeB.NodeB will send subscriber equipment (UE) to or will be transferred to RNC by Iub from the packet that the Uu interface transmits by the Uu interface from the packet that Iub interface transmits, and NodeB also participates in a part of RRM simultaneously.

RAN can be applied in Wideband Code Division Multiple Access (WCDMA) (WCDMA) system or Time Division-Synchronous Code Division Multiple Access (TD-SCDMA) system.

At present, introduced high speed downlink packet access (HSDPA) technology in 3GPP Realse5 standard, this technology can be applied to WCDMA system and TD-SCDMA system, and the service of high-speed transfer downlink data is provided for UE.In the WCDMA system that uses the HSDPA technology, the peak rate of transmitting downlink data can reach 14.4 mbit/(Mbps) in theory; In the TD-SCDMA system that uses the HSDPA technology, adopt under the bandwidth condition identical with the WCDMA system, theoretic maximum downstream service rate is 16.8Mbps.The system that adopts the HSDPA technology sends to UE with data encapsulation in Packet Data Unit (PDU) when sending data for UE.

The function of packet scheduling is when to adjudicate, and distributes to which type of Radio Resource of which UE and communicates.Adopt the system of HSDPA technology that scheduler is moved on to NodeB from RNC, realized fast packet scheduling.Scheduler among the NodeB uses method for fast packet scheduling to determine which UE is the current transmission time interval (TTI) of high speed descending sharing channel (HS-DSCH) distributed to.Method for fast packet scheduling is equivalent to carry out diversity between many UE, although the current channel quality of HS-DSCH that uses of each UE has nothing in common with each other, but after scheduler schedules, it is (absolute or relative that the system of employing HSDPA technology always is in favourable channel condition, according to different fast dispatch algorithms and different) be the UE service, thereby realize adopting the high throughput of the system of HSDPA.

At present, three kinds of method for fast packet scheduling are arranged, below be described in detail respectively.

First kind: the Round Robin method

This dispatching method is very simple, promptly according to time sequencing the UE in the system that adopts HSDPA is carried out repeating query, makes UE obtain service successively.This dispatching method is not considered channel quality information.The rule of this dispatching method is: 1, each UE in the system of employing HSDPA is in the request queue, so that obtain service; 2, when carrying out wireless resource scheduling, the request in the non-NULL request queue is accepted service in the mode of repeating query; 3, before same UE accepts service once more in the request queue, the UE in other all non-NULL request queues must all serviced mistake once; Unless 4 only exist a non-NULL request queue, and this non-NULL request queue has only a request, otherwise a non-NULL request queue cannot be accepted twice service continuously.

Fairness and short time fairness when this dispatching algorithm can guarantee long between UE, and method is simple, but adopt the throughput of system of HSDPA low, and poor performance.

Second kind: maximum signal interference ratio (SIR) dispatching method

This dispatching algorithm is to carry out on NodeB, acts on the inside of being administered, sub-district, guarantees that at any time always the best UE of SIR accepts service, and the throughput of system of this method is the upper bound of the throughput of system of employing HSDPA.The rule of this dispatching algorithm is: 1, each UE is in the request queue; 2, when carrying out wireless resource scheduling, the request in the non-NULL request queue is that priority is accepted service from high to low with the carrier/interface ratio (C/I) of UE; 3, the UE of any time acceptance service always has the highest C/I.

The advantage of this dispatching method is: when this dispatching method adopts Adaptive Modulation and Coding technology (AMC), adopt the wireless system of this dispatching method can obtain maximum system throughput.This dispatching method algorithm is simple, does not need special signaling and complicated calculating.But unfairness the UE of this dispatching algorithm in the system that adopts HSDPA, and this dispatching method can be subjected to the C/I influence of measurement error.

The third method: direct ratio equity dispatching method

This dispatching method is to carry out on NodeB, acts on the inside of being administered, sub-district.This dispatching method give to adopt each UE in the system of HSDPA to distribute a priority according to channel conditions, traffic carrying capacity size and service status, the UE that has limit priority in any time sub-district accepts service.The calculating of priority: ${\mathrm{prio}}_k=\frac{{\left[\frac{C}{I}\right]}_K\left(t\right)}{T_k\left(t\right)},k=\mathrm{1,2},...,N$

In the formula (C/I) _k(t) be that k UE is at t SIR constantly, T _k(t) for t being the throughput of system in the time window that ends up.

According to this dispatching method, the high UE of priority or has good instantaneous channel conditions for the average level of system; The very low and current channel condition of the throughput that obtained in the past be improved significantly.Since when UE receives service continuously, T _k(t) increase gradually, thereby the priority of this UE is reduced, so single UE can always not receive service.This dispatching method makes all UE that approximately uniform activation probability all be arranged, even its average channel quality is different fully, thereby guarantees fairness.If the UE instantaneous channel condition of cell edge is obviously improved, also can obtain the transfer of data of higher rate.But though this dispatching method emphasis has been looked after the UE of cell edge, the data in the whole sub-district remain non-homogeneous substep.

From above three kinds of grouping scheduling methods as can be seen: adopt the throughput of system of Round Robin method low, poor-performing; Though adopt the system of maximum SIR dispatching method that higher throughput is arranged, the UE unfairness; Adopt the system of direct ratio equity dispatching method realizing good compromise on the UE fairness and between the throughput, but the UE Determination of priority has just been considered the influence of the current use channel circumstance of UE, does not take into full account the factors such as priority of packet data service quality (QoS) requirement and different PQ.

Summary of the invention

In view of this, main purpose of the present invention be to provide a kind of in the HSDPA network method of packet scheduling, this method not only guarantees the fairness of throughput of system and UE, and guarantees the QoS level of grouped data and the priority of PQ.

According to above-mentioned purpose, technical scheme of the present invention is achieved in that

A kind of in the high-speed downlink packet access HSDPA network method of packet scheduling, this method comprises:

The channel quality indicator (CQI) information that A, node NodeB report at current transmission time interval TTI according to the user equipment (UE) of being administered, be the dispatching priority index that PQ is provided with to the data volume and the RNC of the grouped data of the PQ of the descending Mean Speed of the PQ grouped data of UE transmission, RNC institute buffer memory about the PQ of priority query, obtain the priority of PQ

Process is, the transmission block size that NodeB recommends in current TTI according to PQ multiply by the maximum timeslot number of distributing to high-speed downlink packet access HSDPA network shared channel in the TTI, multiply by each time slot allocation and give maximum code channel numbers of HSDPA network shared channel, the timeslot number of the HAPQ process that in a last TTI, is assigned with divided by PQ, the HSDPA network code channel number Channel_Code that in a last TTI, is assigned with divided by PQ again _NumThe value that obtains obtains the length of the maximum high speed medium cut-in control grouping data unit MAC-hs PDU that PQ can transmit in current TTI;

NodeB obtains the function of the data to be transferred amount of PQ according to the data to be transferred amount of PQ in the RNC buffer area, comprising: when the data to be transferred amount of the PQ of RNC institute buffer memory during less than the minimum RNC buffer area of setting big or small, this function setup is 1; When the data to be transferred amount of the PQ of RNC institute buffer memory is not less than minimum RNC buffer area big or small of setting, this function setup is to deduct 1 after the size of data to be transferred amount divided by the minimum RNC buffer area of setting of PQ of institute's buffer memory, the difference that obtains multiply by α, add 1 value that obtains again, wherein α is by system's setting, and NodeB obtains when initialization;

The dispatching priority index Scheduling_Priority_Indicator that NodeB is provided with according to this PQ for UE obtains the function of PQ dispatching priority, and the function of described Scheduling_Priority_Indicator for the PQ setting is β ^{Dispatching priority index for this PQ setting}, wherein Scheduling_Priority_Indicator is distributed by the upper strata, is determined by the dispatching priority requirement of PQ; β is by system configuration, and NodeB obtains when initialization, is determined for the capability of influence of scheduling by the dispatching priority index;

The length of the maximum MAC-hs PDU that can transmit multiply by behind the PQ data to be transferred flow function of RNC institute buffer memory the average downstream rate divided by the PQ grouped data, the value that obtains multiply by the dispatching priority function of PQ again, obtain the priority of PQ, the average downstream rate of wherein said PQ grouped data is the average downstream rate that 1 difference that deducts filter factor multiply by PQ grouped data in the TTI, with, filter factor multiply by the instant downstream rate of current PQ grouped data, between and the value that obtains, described filter factor is determined by the fairness of serving between throughput of system and the PQ;

B, NodeB are the different UEs distributing radio resource according to the PQ priority orders that obtains, by institute's assigned radio resource to the UE transmitting grouped data.

The PQ priority orders that the described NodeB basis of step B obtains is that the process of different UEs distributing radio resource is:

B1, NodeB judge whether Radio Resource in the current TTI for empty, or do not have PQ in the dispatching priority formation, or the PQ of scheduling in this TTI reached in the TTI of setting the PQ numbers of scheduling at most, if finish described dispatching method; If not, execution in step B2 then;

B2, NodeB obtain the highest PQ of current priority from the PQ formation, NodeB judges whether the UE of this PQ correspondence has had PQ to be called in this TTI, if then change step B1 over to; Otherwise, change step B3 over to;

B3, NodeB judge whether available to HAPQ process and Radio Resource that should PQ, if, execution in step B4; Otherwise, change step B1 over to and continue to carry out;

Whether B4, NodeB judge has grouped data to need to retransmit among this PQ, if, then send the grouped data that needs re-transmission, the timeslot number that grouped data adopted of this re-transmission, code channel number, and these parameters that transmission block size all adopts when sending this grouped data last time are identical, change step B1 after retransmitting over to and continue to carry out, otherwise, change step B5 over to;

B5, NodeB judges whether that the buffer area of this PQ is for empty, or whether this PQ exceeds the transmission window of setting, if arbitrary condition satisfies, then stop to send the grouped data of this PQ, the buffer area of waiting for this PQ is not for empty, or the transmission window length of this PQ is during less than the maximum length of the transmission window of setting, be ready for sending the grouped data of this PQ again, otherwise, data to be transferred amount according to this PQ in MAC-hs PDU that is calculated by the CQI value and the current NodeB buffer area is the PQ Resources allocation, changes step B1 after complete over to and continues to carry out, and this Resources allocation comprises that this PQ sends institute's distributed time slot number in current TTI, code channel number, and transmission block size.

From such scheme as can be seen, after NodeB of the present invention is the priority of the dispatching priority Index for Calculation PQ that is provided with of PQ according to UE in the CQI information about PQ that current TTI reports, to the data volume of the grouped data of the PQ of the average downstream rate of the grouped data of this UE transmission PQ, RNC institute buffer memory and RNC, according to the PQ priority orders is corresponding UE distributing radio resource, is used for the transmitting grouped data to UE.Because method provided by the invention is to realize on the basis of the factors such as priority level of the fairness of the qos requirement (data volume of this PQ that sends to UE of RNC institute buffer memory) that combines channel quality status (CQI information) for the UE transmitting grouped data, UE, UE (average downstream rate) and PQ, so this method not only guarantees the fairness of throughput of system and UE, and the priority of the PQ of assurance packet data QoS level and UE.

Description of drawings

Fig. 1 is the structural representation of prior art RAN;

Fig. 2 is the method flow diagram of the present invention's packet scheduling in the HSDPA network;

Fig. 3 is the method embodiment flow chart of the present invention's packet scheduling in the HSDPA network.

Embodiment

In order to make the purpose, technical solutions and advantages of the present invention clearer, below lift specific embodiment and, the present invention is described in more detail with reference to accompanying drawing.

In order not only to guarantee the fairness of throughput of system and UE, and (the PQ of priority query of assurance packet data QoS level and UE, Priority Queue) priority (this PQ comprises grouped data), the present invention defines the priority of PQ again, to be transferred to the priority of the PQ of each UE in the priority computing system of definition according to the present invention, and according to priority order queuing from high to low, give the PQ priority allocation Radio Resource of high priority then so that transmit the grouped data of this PQ to different PQ.

In the present invention, the computational methods of the PQ priority of each UE are as described below in the current TTI.

In grouping scheduling method of the present invention, at first determine the priority of each PQ, the PQ with limit priority obtains service earlier, and promptly the PQ that priority is the highest in next TTI is assigned with Radio Resource earlier, transmits grouped data.

The basic calculating formula of PQ priority is:

${\mathrm{Prio}}_k\left(t\right)=\frac{\mathrm{MAChs}\_\mathrm{Pdu}\_\mathrm{Size}\×f_1(\mathrm{User}\_\mathrm{Buffer}\_\mathrm{Size})}{{\mathrm{Rate}}_{\mathrm{out}}\left(t\right)}\×f_2(\mathrm{Scheduling}\_\mathrm{Priority}\_\mathrm{Indicator})$

Wherein, the length of the maximum high speed medium cut-in control grouping data unit (MAC-hs PDU) that MAChs_Pdu_Size can transmit under the channel circumstance of PQ in current TTI has for this reason reflected the channel quality of the current transmitting grouped data of UE.Indicated parameters R TBS and this CQI of channel quality indication (CQI) information according to current PQ indicates the timeslot number (TS of the automatic request retransmission of pairing mixing (HARQ) course allocation _Num) and code channel number (Channel_Code _Num), can calculate MAChs_Pdu_Size, see formula (1):

MAChs_Pdu_Size＝RTBS*TS _total*Channel_Code _total/(TS _num*Channel_Code _{n um}) (1)

In formula (1), TS _TotalAnd Channel_Code _TotalImplication be respectively: TS _TotalIt is the maximum timeslot number of distributing to the HSDPA shared channel in the TTI; Channel_Codetotal gives the code channel number of HSDPA shared channel for each time slot allocation.

User_Buffer_Size is the data to be transferred amount of this PQ in the RNC buffer area, has reflected propagation delay time and the peak transfer rate requirement of this PQ, sees formula (2) about the function of User_Buffer_Size:

$f_1(\mathrm{User}\_\mathrm{Buffer}\_\mathrm{Size})=\left\{\begin{array}{cc}1& \mathrm{User}\_\mathrm{Buffer}\_\mathrm{Size}<\mathrm{Min}\_\mathrm{buffer}\_\mathrm{Size}\\ 1+\mathrm{\&alpha;}\&times;(\mathrm{User}\_\mathrm{buffer}\_\mathrm{size}/\mathrm{Min}\_\mathrm{Buffer}\_\mathrm{Si}{\mathrm{ze}}^{-1})& \mathrm{else}\end{array}\right.---\left(2\right)$

In formula (2), Min_Buffer_Size is RNC buffer size minimum in the grouping scheduling method of the present invention, its objective is that the end of protection grouped data can access transmission timely; α has embodied the influence of RNC buffer area data volume to grouping scheduling method of the present invention, has represented the qos requirement (maximum rate of time delay, business etc.) of grouped data, and α is by system configuration, and NodeB obtains this parameter when initialization.When User_Buffer_Size＜Min_Buffer_Size, then functional value gets 1; When User_Buffer_Size＞=Min_Buffer_Size, then functional value is got User_Buffer_Size and is deducted 1 after divided by Min_Buffer_Size, and the difference that obtains multiply by α, adds 1 value that obtains again.

Rate _OutThe average transmission rate that has reflected this PQ, concrete computing formula is seen formula (3):

R _out(t+1)＝(1-1/T _c)×R _out(t)+1/T _c×R _out(t+1) (3)

In formula (3), 1/T _cBe the filter filterable agent, it is chosen and has determined the fairness of serving between throughput of system and the PQ; T _cFor NodeB sends the length of the sliding time window of grouped data, reflected that a UE is not to receiving the suffertibility of packet data transmission, long T _cTo allow wait for that the long period improves up to the channel quality of this UE, this helps the raising of throughput of system, but may bring additional delay and be unfavorable for the fairness of system, can so think T _cThe more approaching polling dispatching method of the prior art of short more then this grouping scheduling method, T _cThe approaching more maximum SIR dispatching method of the prior art of long more this grouping scheduling method.In the present invention, 1/Tc is by system configuration, and NodeB obtains this parameter when initialization.

Dispatching priority index (Scheduling_Priority_Indicator) has embodied the dispatching priority requirement of PQ, and its computing formula is seen formula (4)

f ₂(Scheduling_Priority_Indicator)＝β ^{Schedule_Priority_Indicator} (4)

Wherein the β factor has reflected the capability of influence of Scheduling_Priority_Indicator for scheduling, and by system configuration, NodeB obtains this parameter when initialization.

Fig. 2 is the method flow diagram of the present invention's packet scheduling in the HSDPA network, and its concrete steps are:

Step 200, NodeB receive CQI information and acknowledge message/non-affirmation (ACK/NACK) message that each UE of being administered reports in current TTI.

Step 201, NodeB calculate MAChs_Pdu_Size according to the CQI information that each UE reports according to formula (1), calculate Rate according to formula (3) again _Out

MAChs_Pdu_Size and Rate that step 202, NodeB calculate according to step 201 _Out, and the function of the Scheduling_Priority_Indicator that is provided with for this PQ about the function of the data to be transferred amount User_Buffer_Size of PQ in the RNC buffer area with about RNC calculates the PQ priority of each UE.

Step 203, NodeB are the UE distributing radio resource according to the PQ priority orders that obtains UE, promptly in next TTI, distribute time-slot code channel number and transmission block size for transmitting grouped data, the high UE of PQ priority is distributing radio resource at first, distributing radio resource (if Radio Resource is less, then the UE that PQ priority is low can't obtain Radio Resource in next TTI) behind the low UE of PQ priority.

Step 204, NodeB are according to the PQ that transmits UE for UE institute assigned radio resource respectively.

In the present invention, NodeB obtains the grouping scheduling method parameters needed when initialization, as 1/T _c, Min_Buffer_Size, the α factor, the β factor, Max_Allocate_PQ_Num and maximum retransmission (N _{Retransmission}) etc.These parameters all are that the network performance index that the network based needs of HSDPA reach is set.

Carry out one time packet scheduling in the every TTI of NodeB, the method for packet scheduling as shown in Figure 3:

Step 300, NodeB receive UE by the message that the shared information channel (HS-SICH) that is used for the HSDPA technology sends, and obtain parameters such as ACK/NACK information and CQI information; Carry out interacting message with RNC, obtain parameters such as User_Buffer_Size.

Step 301, NodeB upgrade the Rate of each PQ _Out, calculate Rate _OutCan adopt formula (3).

Step 302, NodeB calculate the priority of PQ in this TTI of each UE according to the ACK/NACK information, CQI information and the User_Buffer_Size that receive, and concrete computational process is:

At first, the ACK/NACK information that receives is if be NACK and N＜N _{Retransmission}, with the top of this PQ as for the packet scheduling formation, it is the highest promptly to retransmit priority.If be ACK, then read CQI the information modulation system and the RTBS of PQ recommendation for this reason, and empty the grouped data in the corresponding re-transmission priority query, upgrade the lower boundary TSN that NodeB sends window _{Down_edge}

Then, indicate the TS of corresponding HAPQ course allocation according to the CQI of this PQ _NumAnd Channel_Code _Num, adopt formula (1), (2), (3) and (4) to calculate the MAChs_Pdu_Size that can send in this TTI, thereby calculate the priority of each PQ;

At last, with PQ number (Allocate_PQ_Num) zero setting of being dispatched in this TTI.

Following steps are how NodeB is the UE Resources allocation.

Step 303, NodeB detect remaining shared channel resources in this TTI, the Allocate_PQ_Num situation of distributing in the PQ number in the dispatching priority formation and this TTI: if remaining shared channel resources is 0, be but that remaining time slot scheduling and code channel number are zero, or the PQ number in the priority query is 0, or Allocate_PQ_Num==Max_Allocate_PQ_Num, then change step 309 over to; Otherwise, change step 304 over to.

Max_Allocate_PQ_Num represents the PQ number that NodeB can dispatch at most in the TTI.

Step 304, NodeB select the highest PQ of priority query's medium priority, and this PQ is deleted in priority query.NodeB judges whether the UE of this PQ correspondence has had PQ to be scheduled in this TTI: if there has been PQ to be scheduled, then directly change step 303 over to; Otherwise NodeB judges that whether this UE has available HAPQ process and available HS-SCCH and HS-SICH, if having, changes step 305 over to; Otherwise, change step 303 over to.

Whether step 305, NodeB judge has grouped data to need to retransmit, if having, then to change step 308 among this PQ; Otherwise, change step 306 over to.

The buffer area that step 306, NodeB judge whether this PQ is for empty, or the grouped data among the PQ exceeds the transmission window, if arbitrary condition satisfies, then changes step 303 over to; Otherwise, change step 307 over to.

Step 307, NodeB are the PQ Resources allocation in the current TTI according to the MAChs_Pdu_Size that calculates and the size of NodeB buffer area, and detailed process is as follows:

At first, NodeB judges the remaining code channel number (Channel_Code of time slot in this TTI _Remain).

Secondly, NodeB judges whether to satisfy Allocate_PQ_Num=Max_Allocate_PQ_Num-1, if, then with remaining all Channel_Code _RemainAll distribute to this UE.Otherwise according to the data block size that retransmits or the time-slot code channel number of NodeB buffer size dispensed, formula specific as follows:

The data volume that the code channel of each bar SF=16 can send:

Pdu_Size_PerCode＝MAChs_Pdu_Size/TS _total/Channel_Code _total

Wherein, TS _TotalIt is the maximum timeslot number of distributing to the HSDPA shared channel in the TTI; Channel_Code _TotalGive the code channel number of HSDPA shared channel for each time slot allocation.

NodeB in current TTI for this PQ distributed time slot code channel number is:

TS _num＝TS _total

Channel_Code _num＝NodeB_Buffer_Size/Pdu_Size_PerCode/TS _num

Channel_Code _NumIf then do not round up for integer.

Simultaneously should satisfy Channel_Code _Num＜=Channel_Code _Remain

At last, the MAC-hs PDU size of distribution is calculated by following formula:

Mac-hs_Pdu_Length＝Pdu_Size_PerCode*TS _num*Channel_Code _num

The PDU length that the Mac-hs_Pdu_Length that calculates according to following formula chooses protocol compliant downwards sends.

Allocate_PQ_Num＝Allocate_PQ_Num+1

Put the number of transmissions N=1 of this PQ

Execute and change step 303 over to.

Step 308, NodeB retransmission packet data: use and when transmitting this PQ first identical time-slot code channel number and MAC-hs PDU size by eating dishes without rice or wine to UE transmission grouped data.Simultaneously, Allocate_PQ_Num=Allocate_PQ_Num+1, N=N+1.Change step 303 after complete over to.

Step 309, this grouping scheduling method finish.

The present invention is on the basis of the factors such as fairness of the PQ priority level of the qos requirement of the influence of comprehensive channel environment, UE, UE and UE, a kind of HSDPA grouping of network dispatching method has been proposed, to reach the QoS demand that guarantees UE, improve throughput of system as much as possible, make the fair as far as possible purpose of scheduling simultaneously.

The above only is preferred embodiment of the present invention, not in order to restriction the present invention, all any modifications of being made within the spirit and principles in the present invention, is equal to and replaces and improvement etc., all should be included within protection scope of the present invention.

Claims (2)

1, a kind of in the high-speed downlink packet access HSDPA network method of packet scheduling, it is characterized in that this method comprises:

The channel quality indicator (CQI) information that A, node NodeB report at current transmission time interval TTI according to the user equipment (UE) of being administered, be the dispatching priority index that PQ is provided with to the data volume and the RNC of the grouped data of the PQ of the descending Mean Speed of the PQ grouped data of UE transmission, RNC institute buffer memory about the PQ of priority query, obtain the priority of PQ

Process is, the transmission block size that NodeB recommends in current TTI according to PQ multiply by the maximum timeslot number of distributing to high-speed downlink packet access HSDPA network shared channel in the TTI, multiply by each time slot allocation and give maximum code channel numbers of HSDPA network shared channel, the timeslot number of the HAPQ process that in a last TTI, is assigned with divided by PQ, the HSDPA network code channel number Channel_Code that in a last TTI, is assigned with divided by PQ again _NumThe value that obtains obtains the length of the maximum high speed medium cut-in control grouping data unit MAC-hs PDU that PQ can transmit in current TTI;

NodeB obtains the function of the data to be transferred amount of PQ according to the data to be transferred amount of PQ in the RNC buffer area, comprising: when the data to be transferred amount of the PQ of RNC institute buffer memory during less than the minimum RNC buffer area of setting big or small, this function setup is 1; When the data to be transferred amount of the PQ of RNC institute buffer memory is not less than minimum RNC buffer area big or small of setting, this function setup is to deduct 1 after the size of data to be transferred amount divided by the minimum RNC buffer area of setting of PQ of institute's buffer memory, the difference that obtains multiply by α, add 1 value that obtains again, wherein α is by system's setting, and NodeB obtains when initialization;

The dispatching priority index Scheduling_Priority_Indicator that NodeB is provided with according to this PQ for UE obtains the function of PQ dispatching priority, and the function of described Scheduling_Priority_Indicator for the PQ setting is that β is this P ^{The dispatching priority index that Q is provided with}, wherein Scheduling_Priority_Indicator is distributed by the upper strata, is determined by the dispatching priority requirement of PQ; β is by system configuration, and NodeB obtains when initialization, is determined for the capability of influence of scheduling by the dispatching priority index;

The length of the maximum MAC-hs PDU that can transmit multiply by behind the PQ data to be transferred flow function of RNC institute buffer memory the average downstream rate divided by the PQ grouped data, the value that obtains multiply by the dispatching priority function of PQ again, obtain the priority of PQ, the average downstream rate of wherein said PQ grouped data is the average downstream rate that 1 difference that deducts filter factor multiply by PQ grouped data in the TTI, with, filter factor multiply by the instant downstream rate of current PQ grouped data, between and the value that obtains, described filter factor is determined by the fairness of serving between throughput of system and the PQ;

B, NodeB are the different UEs distributing radio resource according to the PQ priority orders that obtains, by institute's assigned radio resource to the UE transmitting grouped data.

2, the method for claim 1 is characterized in that, the PQ priority orders that the described NodeB basis of step B obtains is that the process of different UEs distributing radio resource is:

B1, NodeB judge whether Radio Resource in the current TTI for empty, or do not have PQ in the dispatching priority formation, or the PQ of scheduling in this TTI reached in the TTI of setting the PQ numbers of scheduling at most, if finish described dispatching method; If not, execution in step B2 then;

B2, NodeB obtain the highest PQ of current priority from the PQ formation, NodeB judges whether the UE of this PQ correspondence has had PQ to be called in this TTI, if then change step B1 over to; Otherwise, change step B3 over to;

B3, NodeB judge whether available to HAPQ process and Radio Resource that should PQ, if, execution in step B4; Otherwise, change step B1 over to and continue to carry out;

Whether B4, NodeB judge has grouped data to need to retransmit among this PQ, if, then send the grouped data that needs re-transmission, the timeslot number that grouped data adopted of this re-transmission, code channel number, and these parameters that transmission block size all adopts when sending this grouped data last time are identical, change step B1 after retransmitting over to and continue to carry out, otherwise, change step B5 over to;

B5, NodeB judges whether that the buffer area of this PQ is for empty, or whether this PQ exceeds the transmission window of setting, if arbitrary condition satisfies, then stop to send the grouped data of this PQ, the buffer area of waiting for this PQ is not for empty, or the transmission window length of this PQ is during less than the maximum length of the transmission window of setting, be ready for sending the grouped data of this PQ again, otherwise, data to be transferred amount according to this PQ in MAC-hs PDU that is calculated by the CQI value and the current NodeB buffer area is the PQ Resources allocation, changes step B1 after complete over to and continues to carry out, and this Resources allocation comprises that this PQ sends institute's distributed time slot number in current TTI, code channel number, and transmission block size.

Images