Wireless resource scheduling method based on priority
Technical Field
The invention relates to the field of wireless communication, in particular to a wireless resource scheduling method based on priority.
Background
The communication under the background of the smart grid has the biggest characteristic of having multiple types of services, and the difference of QoS (Quality of Service) requirements among different services is large, therefore, considering that the services in the smart grid are graded, namely, the services in the grid are graded according to the indexes such as time delay, throughput and packet loss rate, emergency services in the grid, such as field emergency command services, have very strict requirements on the Service Quality, and some periodic reporting services, such as intelligent meter reading, have very large tolerance on the Service Quality, because of the characteristics of wide coverage, high rate, large capacity and the like of L TE networks, at present, the domestic smart grid generally adopts L TE as a basic communication framework, but standard L TE does not have a corresponding solution mechanism aiming at the characteristics of the Service grade and the large difference of QoS requirements of the smart grid, if L TE is not optimized, the Service Quality of services with different priorities can not be differentiated, so that the high-priority Service cannot be guaranteed, therefore, a wireless resource scheduling algorithm based on priority is provided herein, namely, and the user performance of the system is guaranteed, thereby the resources are high in the system.
The priority ordering method in L TE scheduling determines when the system schedules users and how to allocate resources to the users, which affects the performance and efficiency of the whole system, so the method for determining user scheduling priority ordering in L TE multi-user dynamic scheduling is of great importance.
At present, L TE system scheduling has three mature sequencing methods, namely a maximum carrier-to-interference ratio scheduling algorithm (Max C/l), a Round Robin scheduling algorithm (RR, Round Robin) without a specific priority sequencing method, wherein the priority sequencing can obtain a system maximum throughput boundary according to the quality of a wireless channel of a user, but the fairness requirements of different users are not considered at all, each user is scheduled circularly, each user occupies distributable time slots and power with the same probability in each cycle, the optimal fairness among the users is realized, however, the method does not consider the wireless channel quality condition of different users, the throughput and the spectral efficiency of the system are low, and a Proportional fairness algorithm (PF, Proportional Fair) is used, the priority sequencing is comprehensively considered according to the scheduling frequency of the users and the quality of the channel, one compromise is taken between the system throughput and the fairness, the fairness among the users is represented by using an average data transmission rate, the fairness has better long-term fairness is provided, but the short-term scheduling method cannot guarantee the fairness among the users, and the fairness requirements of QoS are not considered.
A plurality of improved methods for priority ordering in a scheduling algorithm are provided in domestic and foreign papers, for example, a VT-M L WDF scheduling algorithm provided by Iturralde M in the article "Performance study of multimedia services using virtual token for resource allocation in L TE networks", on the basis of PF algorithm, QoS factors of users with different priorities are multiplied in a priority calculation formula, the QoS factor of a user with high priority is larger, so that the Performance of the user with high real-time requirement is improved, besides, the length of a queue to be sent in a buffer area is considered when the scheduling priority is calculated, the priority can be improved for the user with large data volume, and the system throughput is improved to a certain extent, but the user waiting time is not considered in the method.
For another example, Y.P, L i provides DP-VT-M L WDF algorithm in article "a delay priority scheduling algorithm for downlink-time traffic in L TE networks", which introduces a delay factor based on VT-M L WDF algorithm and can reflect the urgency of delay, and the longer the waiting time, the larger the scheduling priority factor of the user, thereby avoiding the packet loss due to timeout.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provides a wireless resource scheduling method based on priority, which divides users into three priorities of high, medium and low; on the basis of preferentially distributing resources to high-priority users, priority ranking of middle-priority users and low-priority users is realized by calculating priority factors, and further reasonable utilization of Resource Blocks (RB) is realized.
The purpose of the invention is realized by the following technical scheme: a wireless resource scheduling method based on priority comprises the following steps:
s1, dividing users into a high priority, a medium priority and a low priority;
s2, calculating scheduling priority factors of all users, and realizing scheduling priority ordering of the users with middle and low priorities according to the size of the scheduling priority factors;
s3, in each TTI, detecting whether the high-priority user has data to be sent; if yes, preferentially distributing Resource Blocks (RB) for the high-priority users, and then entering the step S4; if not, go directly to step S4;
and S4, if the residual resource blocks RB exist, distributing the residual RB to the users with the middle and low priorities according to the sequencing of the scheduling priority until all the resource blocks RB are distributed.
Further, the step S2 includes the following sub-steps:
s21, for any user i, calculating the signal to interference plus noise ratio (SINR) of the user i on each carrier according to the received signali,k(t); wherein i is 1,2,3, N; n represents the total number of users; specifically, the step S21 includes the following sub-steps:
calculating the channel gains of all carriers of the user i in the t TTI:
wherein, pli,k(t),ξi,k(t),pathi,k(t) respectively representing the path loss, the shadow fading gain and the multipath fading gain of the user i for the kth subcarrier at the tth TTI; gaini,k(t) represents the channel gain of the kth subcarrier of user i in the t-th TTI;
and calculating the signal-to-interference-and-noise ratio of the user i in each subcarrier according to the channel gain:
wherein, PkRepresenting the transmit power of the base station on the k sub-carrier, N0Representing the noise power; i represents inter-cell interference power; SINRi,k(t) represents the SINR of the kth subcarrier in the tth TTI for user i.
S22, converting the signal-to-interference-and-noise ratios of all carriers in each resource block RB occupied by the user i into the effective signal-to-noise ratio of the resource block RB:
wherein SNR isi,j(t) represents the signal-to-noise ratio of user i on jth RB at tth TTI; m represents one RBThe number of subcarriers of (a); sigma represents a factor which varies with different coding modulation modes;
s23, calculating the maximum transmission rate r which can be reached by the user i on the jth RB in the tth TTIi,j(t):
ri,j(t)=log2(1+SINRi,j(t));
S24, calculating the average transmission rate of the user i in the t TTI
Specifically, the step S24 includes the following sub-steps:
obtaining the average transmission rate of the user i in the t-1 TTI
According to the maximum transmission rate of the user i in the t TTI and the average transmission rate of the t-1 TTI, calculating the average transmission rate of the user i in the t TTI
In the formula, tcRepresenting the throughput window length.
S25, reading a packet loss rate threshold and a time delay threshold according to QoS requirements of users with different priorities, and calculating a QoS factor ai:
iIs the packet loss rate, τ, that user i can tolerateiIs a waiting delay threshold that user i can tolerate;
s26, obtaining the length Q of a buffer queue to be sent in the t TTI of the user ii(t);
S27, calculating a delay factor DP according to the waiting delay of a first packet of a queue to be sent; specifically, the step S27 includes the following sub-steps:
firstly, defining a guard interval PI of a medium priority user:
PI=0.3*τi,
in the formula, τiIs a waiting delay threshold that user i can tolerate;
second, calculate the delay factor DP:
if the condition is satisfied: user i belongs to a medium priority user, and Dhol,iAnd if the PI is larger than or equal to PI:
if the condition is not satisfied: user i belongs to a medium priority user, and Dhol,iAnd if the PI is larger than or equal to PI:
in the formula, Dhol,iIs the head of line latency for user i.
S28, obtaining r according to the steps S21-S27
i,j(t),
a
i,Q
i(t) DP, calculating a user scheduling priority factor K:
s29, repeating the steps S21-S28 for each medium-priority user and each low-priority user to obtain the scheduling priority factors of all the medium-priority users and all the low-priority users, and sorting the obtained scheduling priority factors in size.
Further, the step S3 includes the following sub-steps:
s31, detecting whether the high-priority user has data to be sent:
(1) if the high-priority user has data to be sent, the process goes to step S32;
(2) if the high-priority user does not have data to be sent, directly entering step S4;
s32, estimating the resources occupied by the high-priority users, preferentially distributing the required resources to the high-priority users according to the estimation result, realizing the resource scheduling of the high-priority users, and then entering the step S4, specifically, estimating the resources occupied by the high-priority users in the following way:
A. for the high-priority user i, calculating the maximum transmission rate which can be reached by the user i on the jth RB in the tth TTI according to the steps S21-S23;
B. according to the maximum rate value obtained by calculation, resource blocks RB are sequentially selected from high to low until the sum of the rate values of the selected resource blocks RB is larger than the data volume of the high-priority user i;
and repeating the steps A-B for each high-priority user until corresponding resource blocks RB are selected for all the high-priority users, wherein the sum of all the selected resource blocks RB is the estimated value of the resources occupied by the high-priority users.
Further, the step S4 includes: judging whether remaining resource blocks exist:
if so, allocating the rest RB resource blocks to the users with middle and low priorities according to the scheduling priority sequence obtained in the step S2 until all the RB resource blocks are allocated, and scheduling the resources in the next TTI;
if not, directly carrying out resource scheduling in the next TTI.
The invention has the beneficial effects that: the invention divides users into three priorities of high, middle and low; on the basis of preferentially distributing resources to high-priority users, priority ranking of middle-priority and low-priority users is realized by calculating priority factors, and further reasonable utilization of Resource Blocks (RB) is realized; in the process of calculating the priority factors of the middle-priority and low-priority users, a protection interval is introduced for calculating the time delay factors of the middle-priority users, and the time delay factors of the middle-priority users are exponentially increased in the time close to a packet loss threshold, so that packet loss is prevented; in the calculation process of the priority factor, factors such as user QoS requirements, the length of a queue to be sent, waiting time delay, overtime urgency and the like are comprehensively considered, and the transmission performance is improved.
Drawings
FIG. 1 is a flow chart of the present invention;
FIG. 2 is a flow chart of scheduling priorities for medium and low priority users;
FIG. 3 is a comparison graph of the high priority user throughput curves of the scheduling method and the PF, VT-M L WDF, DP-VT-M L WDF algorithms in the present invention;
FIG. 4 is a comparison graph of high-priority user delay curves of the scheduling method and three algorithms of PF, VT-M L WDF and DP-VT-M L WDF in the present invention;
FIG. 5 is a comparison graph of the scheduling method and the packet loss rate curves of the high-priority users of PF, VT-M L WDF, and DP-VT-M L WDF in the present invention;
FIG. 6 is a comparison graph of the system throughput curves of medium priority users of the scheduling method and three algorithms PF, VT-M L WDF, DP-VT-M L WDF in the present invention;
FIG. 7 is a comparison graph of the low priority user system throughput curves of the scheduling method and the PF, VT-M L WDF, DP-VT-M L WDF algorithms in the present invention;
FIG. 8 is a comparison graph of high priority user fairness curves for the scheduling method of the present invention and three algorithms PF, VT-M L WDF, DP-VT-M L WDF;
FIG. 9 is a comparison graph of medium priority user fairness curves of the scheduling method and three algorithms PF, VT-M L WDF, DP-VT-M L WDF in the present invention;
FIG. 10 is a comparison graph of low priority user fairness curves for the scheduling method of the present invention and the PF, VT-M L WDF, DP-VT-M L WDF three algorithms;
FIG. 11 is a schematic diagram of a scheduling model.
Detailed Description
The technical solutions of the present invention are further described in detail below with reference to the accompanying drawings, but the scope of the present invention is not limited to the following.
As shown in fig. 1, a method for scheduling radio resources based on priority includes the following steps:
s1, dividing users into a high priority, a medium priority and a low priority;
s2, calculating scheduling priority factors of all users, and realizing scheduling priority ordering of the users with middle and low priorities according to the size of the scheduling priority factors;
s3, in each TTI, detecting whether the high-priority user has data to be sent; if yes, preferentially distributing Resource Blocks (RB) for the high-priority users, and then entering the step S4; if not, go directly to step S4;
and S4, if the residual resource blocks RB exist, distributing the residual RB to the users with the middle and low priorities according to the sequencing of the scheduling priority until all the resource blocks RB are distributed.
As shown in fig. 2, the step S2 includes the following sub-steps:
s21, for any user i, calculating the signal to interference plus noise ratio (SINR) of the user i on each carrier according to the received signali,k(t); wherein i is 1,2,3, N; n represents the total number of users; specifically, the step S21 includes the following sub-steps:
calculating the channel gains of all carriers of the user i in the t TTI:
wherein, pli,k(t),ξi,k(t),pathi,k(t) respectively representing the path loss, the shadow fading gain and the multipath fading gain of the user i for the kth subcarrier at the tth TTI; gaini,k(t) represents the channel gain of the kth subcarrier of user i in the t-th TTI;
and calculating the signal-to-interference-and-noise ratio of the user i in each subcarrier according to the channel gain:
wherein, PkIndicating the transmit power of the base station on the k sub-carrier,N0Representing the noise power; i represents inter-cell interference power; SINRi,k(t) represents the SINR of the kth subcarrier in the tth TTI of the user i; in the embodiment of the present application, the total power of the base station is 43dBm, each subcarrier is allocated with the same power, and the transmission power of the k-th subcarrier is the total power divided by the number of subcarriers.
S22, converting the signal-to-interference-and-noise ratios of all carriers in each resource block RB occupied by the user i into the effective signal-to-noise ratio of the resource block RB:
wherein SNR isi,j(t) represents the signal-to-noise ratio of user i on jth RB at tth TTI; m represents the number of subcarriers in one RB, if a frame structure of a common prefix is used, the value of M is 7, and if a frame structure of an extended prefix is used, the value of M is 6; sigma represents a factor which varies with different coding modulation modes, and the value of sigma can be inquired according to the following table:
s23, calculating the maximum transmission rate r which can be reached by the user i on the jth RB in the tth TTIi,j(t):
ri,j(t)=log2(1+SINRi,j(t));
S24, calculating the average transmission rate (also called average throughput) of the user i in the t TTI
Specifically, the step S24 includes the following sub-steps:
obtaining the average transmission rate (also called average throughput) of user i in the t-1 th TTI
According to the maximum transmission rate of the user i in the t TTI and the average of the t-1 TTIThe transmission rate is calculated as the average transmission rate of the user i in the t TTI
In the formula, tcRepresenting the throughput window length.
S25, reading a packet loss rate threshold and a time delay threshold according to QoS requirements of users with different priorities, and calculating a QoS factor ai:
iIs the packet loss rate, τ, that user i can tolerateiIs a waiting delay threshold that user i can tolerate; if the waiting time of the data packet of the user in the buffer queue exceeds the delay threshold, the data packet is discarded. In the application, the packet loss rate only considers the overtime packet loss condition.
S26, obtaining the length Q of a buffer queue to be sent in the t TTI of the user ii(t);
S27, calculating a delay factor DP according to the waiting delay of a first packet of a queue to be sent; specifically, the step S27 includes the following sub-steps:
firstly, defining a guard interval PI of a medium priority user:
PI=0.3*τi,
in the formula, τiIs a waiting delay threshold that user i can tolerate;
second, calculate the delay factor DP:
if the condition is satisfied: user i belongs to a medium priority user, and Dhol,iAnd if the PI is larger than or equal to PI:
if the condition is not satisfied: user i belongs to medium priorityUsers, and Dhol,iAnd if the PI is larger than or equal to PI:
in the formula, Dhol,iThe queue head waiting time of the user i is adopted, and the DP reflects the urgency of time delay, because in the DP, the denominator is the difference value between the time delay threshold and the current waiting time, the urgency of time delay can be reflected, and the longer the waiting time is, the time delay factor can be rapidly increased, so that the overtime data packet of the user is prevented from being discarded; in addition, the introduction of the guard interval for the medium priority user enables the DP factor of the medium priority user to be exponentially increased in the time close to the packet loss threshold, thereby preventing packet loss and ensuring the performance of the medium priority user as much as possible.
S28, obtaining r according to the steps S21-S27
i,j(t),
a
i,Q
i(t), DP, calculating a user scheduling priority factor K:
s29, repeating the steps S21-S28 for each medium-priority user and each low-priority user to obtain the scheduling priority factors of all the medium-priority users and all the low-priority users, and sorting the obtained scheduling priority factors in size.
Further, the step S3 includes the following sub-steps:
s31, detecting whether the high-priority user has data to be sent:
(1) if the high-priority user has data to be sent, the process goes to step S32;
(2) if the high-priority user does not have data to be sent, directly entering step S4;
s32, estimating the resources occupied by the high-priority users, preferentially distributing the required resources to the high-priority users according to the estimation result, realizing the resource scheduling of the high-priority users, and then entering the step S4, specifically, estimating the resources occupied by the high-priority users in the following way:
A. for the high-priority user i, calculating the maximum transmission rate which can be reached by the user i on the jth RB in the tth TTI according to the steps S21-S23;
B. according to the maximum rate value obtained by calculation, resource blocks RB are sequentially selected from high to low until the sum of the rate values of the selected resource blocks RB is larger than the data volume of the high-priority user i;
and repeating the steps A-B for each high-priority user until corresponding resource blocks RB are selected for all the high-priority users, wherein the sum of all the selected resource blocks RB is the estimated value of the resources occupied by the high-priority users.
Further, the step S4 includes: judging whether remaining resource blocks exist:
if so, allocating the rest RB resource blocks to the users with middle and low priorities according to the scheduling priority sequence obtained in the step S2 until all the RB resource blocks are allocated, and scheduling the resources in the next TTI;
if not, directly carrying out resource scheduling in the next TTI.
In the embodiment of the application, PF, VT-M L WDF, DP-VT-M L WDF and the scheduling method of the invention are compared on an L TE system-level platform, a high-priority user respectively counts four indexes of throughput, time delay, packet loss rate and fairness, and a medium-priority non-real-time user and a low-priority non-real-time user are counted to obtain two indexes of throughput and fairness, and according to the parameter requirements of a L TE system, simulation parameters of the invention are as follows:
parameter(s)
|
Numerical value
|
Type of emulation
|
tri_sector_tilted
|
System frequency
|
2GHz
|
Bandwidth of
|
3MHz
|
TTI duration
|
1ms
|
User movement model
|
still
|
Path loss model
|
TS36942
|
Channel model
|
Winner II+
|
Number of RBs
|
15
|
High, medium, low priority user percentage
|
20%40%40% |
PF, VT-M L WDF and DP-VT-M L WDF are used as comparison algorithms, and simulation comparison analysis is carried out on the method for scheduling the wireless resources provided by the invention as follows:
fig. 3, 4 and 5 show the throughput, delay and packet loss performance curves of high priority users; in fig. 3, we can notice that the average throughput between the four schedulers is almost the same when the number of users is less than 15; however, when the number of users exceeds 15, the performance of the scheduler proposed by the present invention is significantly better than that of other schedulers, and the throughput will rise linearly before the system capacity is reached; in fig. 4 and 5, it can be seen that the proposed scheduler has lower delay and packet loss rate than other schedulers; the performance improvement of the high-priority users is significant in an emergency repair scene of the smart grid.
Fig. 6 shows the average throughput for medium priority users. It can be seen that there is similar performance in the four algorithms. When the number of users exceeds 30, the throughput of the medium priority users is slightly higher than that of other algorithms because the guard interval is used for the medium priority users in the algorithm of the present invention, which effectively prevents the data packet from being discarded due to the timeout.
The throughput of the low priority user is shown in fig. 7. It can be seen that the proposed algorithm has some degradation in throughput performance compared to other algorithms. In the smart grid, low priority users, such as smart meter reading terminals, which have low QoS requirements, are acceptable.
The fairness indicators for high, medium, and low priority users are shown in fig. 8, 9, and 10. It can be seen from the figure that the fairness of the high priority users is greatly improved, and the fairness of the medium priority users is improved.
In the embodiment of the present application, according to the radio resource scheduling method of the present application, an obtained scheduler model is shown in fig. 11, and it can be known from the model that the present application supports resource preemption of a high priority user and introduces interval protection to a medium priority user in the process of resource allocation to medium and low priority users through an algorithm; therefore, reasonable utilization of Resource Blocks (RB) and packet loss protection of medium priority users are achieved.