CN102573092B - Improved proportional fairness scheduling algorithm based on multiuser eigenmode transmission (MET) precoding technology - Google Patents

Abstract

The invention discloses an improved proportional fair scheduling method based on multi-user eigenmode precoding technology in a multi-user multiple-input multiple-output (MU-MIMO) wireless communication system, including the following content: (1) The base station receives the user's service After requesting and channel state information, initialize the user average transmission rate, average request rate and service user set; (2) update the average request rate of each user, and select the user with the highest priority to form the first user of the service user set; (3 ) The base station uses a greedy algorithm to select users with the highest priority each time to form a service user set; (4) After selecting users that the system can serve at the same time, update the average transmission rate of each user for the next round of user scheduling. This method can improve the fairness of users with different channel conditions in the system on the basis of ensuring that the MET precoding technology can significantly improve the system capacity, and achieve a good trade-off between system capacity and user fairness. middle.

Description

Improved proportional fair scheduling method based on multi-user eigenmode precoding technology

technical field

The present invention relates to a multi-user scheduling technology and a multi-user eigenmode (MET) precoding technology in a multiple-input multiple-output (MIMO) communication system in wireless communication technology, and more specifically, relates to a multi-user eigenmode (MET) based An improved multi-user proportional fair scheduling method based on MET) precoding technology.

Background technique

Multiple Input Multiple Output (MIMO) technology is one of the core technologies of wireless communication B3G. By installing multiple transmitting and receiving antennas at the base station and the user end, it increases the degree of spatial freedom for the mobile communication system, and provides space multiplexing gain and space diversity gain to improve system throughput.

Precoding technology can well solve the problem of signal interference between users in a point-to-multipoint communication system. It realizes the parallel transmission of multiple signals by processing the wireless channel into multiple parallel sub-channels without interference with each other. , so as to obtain spatial multiplexing gain and increase the data transmission rate of the system.

In the wireless communication MIMO system, limited by the number of transmitting antennas, the sum of the receiving antennas of the users receiving service in each time slot cannot be greater than the number of transmitting antennas, and the number of users served at the same time is limited, and it is necessary to select the served users from all users , which is the multi-user resource scheduling problem. In order to achieve multi-user diversity in a multi-user MIMO system, it is necessary to solve the multi-user resource scheduling problem to solve how multiple users share limited wireless resources, so as to improve the spectrum efficiency of the system as much as possible, while taking into account the efficiency of each user in the system. service quality. The goal of multi-user scheduling is to provide a satisfactory system utilization (high efficiency) and variable QoS service (fairness) for multimedia services.

The traditional scheduling algorithm is to use the fading characteristics of the user channel to allow the base station to have the best channel state in each scheduling time slot, that is, to provide services to several users with the largest user channel capacity. Such a scheduling method can maximize the throughput of the system, but ignores the fairness among users. When the channel status of some users is in a poor state for a long time, that is, when the capacity of the user channel is small, these users will not be able to receive services for a long time.

In order to solve the problem of user fairness, the proportional fair scheduling algorithm is introduced into multi-user scheduling. It no longer uses the user's channel capacity as a measure of user priority, but uses the ratio of the user's instantaneous request rate to the instantaneous transmission rate, namely:

$\mathrm{<span\; class="notranslate">\; \&mu;</span>}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>}{{\mathrm{<span\; class="notranslate">\; T</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; )</span>$

T _k (t) represents the transmission rate of user k in the current time slot t. R _k (t) represents the instantaneous request rate of user k in the current time slot t. The proportional fairness algorithm can increase the scheduling priority of users who have not been scheduled for a long time, so that the fairness of the throughput of the users is improved, that is, the difference of the average throughput among different users is reduced.

Multi-user eigenmode (MET) precoding technology is a linear precoding technology based on block diagonalization. In the current point-to-multipoint broadcast channel, multi-user eigenmode transmission (MET) is in the linear precoding system. It has the best performance and can greatly improve the throughput of the system.

Based on this, we consider combining the proportional fair scheduling algorithm and MET precoding technology to achieve a compromise between system throughput and fairness. However, this combination, while improving system fairness, loses too much throughput, and cannot make good use of the superiority of MET precoding technology in improving system throughput. Therefore, it is necessary to provide an improved The multi-user proportional fairness method can achieve a better compromise between system throughput and fairness when it is combined with MET precoding technology.

Contents of the invention

The purpose of the present invention is to provide a multi-user scheduling method, which can achieve a better compromise between system throughput and fairness when it is combined with the MET precoding method, and can not only improve the channel condition in a long-term poor state. The user's fairness will not cause too much loss of system capacity.

In order to achieve the above object, the present invention provides an improved multi-user proportional fair scheduling method, which includes the following steps: (1) After the base station receives the user's service request and channel state information, it initializes the user's average transmission rate, and the average request rate and service user set; (2) MET precoding is performed on all users, and the average request rate of each user is updated at the same time, and the priority of each user is calculated, and the user with the highest priority is selected as the first user in the service user set; (3) The base station uses a greedy algorithm to select the remaining users: each time a user is added to the service object set, according to the channel state information and the existing service object set, the MET precoding technology is used to calculate the channel capacity of each user after precoding, and also That is, the user's instantaneous request rate; according to the user's instantaneous request rate, the initial average transmission rate and the increment after the average request rate update, use the priority calculation formula to update and calculate the scheduling priority of each user, and select the user with the highest priority to join the service User set; since the average request rate of each user is only updated when the first user is selected, the increment of the average request rate used each time a service user is selected is updated when the first user is selected Incremental value, similarly, the average transmission rate of each user is only updated after all service users are selected, so the average transmission rate used by each service user is selected after all service users are selected last time. Value (4) If the number of service users has reached the maximum number of users that the system can serve at the same time, go to step (5), otherwise, continue to repeat step (3); (5) After the selection is complete, the system can serve After the number of users, the average transmission rate of each user is updated, and the updated average transmission rate will be used as the initial value of the average transmission rate for the next round of user service set selection.

In one embodiment of the present invention, in the step (2), the base station selects the user with the highest priority to form the first user in the service user set, and the priority calculation formula used is:

$\mathrm{<span\; class="notranslate">\; \&mu;</span>}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>}{\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; T</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>}}<span\; class="notranslate">\; \&CenterDot;</span>\frac{{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>}{{\mathrm{<span\; class="notranslate">\; e</span>}}^{<span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; -</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; )</span>}}$

in Represents the average transmission rate of user k in the past t _c time slots in the current round of scheduling, R _k (t) represents the instantaneous request rate of user k in the current time slot t, Indicates the average request rate of user k in the past t _c time slots in the current round of scheduling, Indicates the average request rate of user k in the past t _c time slots in the last round of scheduling.

In yet another embodiment of the present invention, in the priority calculation formula in the step (2), the The update is performed when the first user is selected, and the value after the first update is used when the remaining users are selected. The update formula is:

$\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; -</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}}<span\; class="notranslate">\; )</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; +</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}}{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>$

In yet another embodiment of the present invention, in the step (6), after selecting all users that the system can serve at the same time each time, the average transmission of user k in the past t _c time slots in the next round of scheduling rate It needs to be updated according to the following formula according to the status of user scheduling:

R _k (t) used for updating the average transmission rate is the channel capacity calculated according to the MET precoding when the last user is selected. To simplify the algorithm, the and The time window for updating is taken as the same.

In yet another embodiment of the present invention, the system is able to know the instantaneous channel information of each user, and needs to use this information to perform MET precoding, so as to maximize the system capacity. Compared with the prior art, the improved multi-user proportional fairness algorithm based on the multi-user eigenmode (MET) precoding technology of the present invention introduces the instantaneous request rate and the exponential function with the average request rate increment as the power exponent Ratio, improve the scheduling priority of users whose channel status is the best relative to their own average channel status and have relatively high capacity increments, thereby increasing users whose channel status is poor for a long time but have relatively high capacity increments relative to themselves The opportunity to be scheduled, while improving system fairness, better maintains the superiority of MET precoding technology in improving system capacity.

The present invention will become clearer through the following description in conjunction with the accompanying drawings, which are used to explain the embodiments of the present invention.

Description of drawings

FIG. 1 is a flow chart of the MET-based multi-user proportional fair method of the present invention.

Fig. 2 is a flowchart of the rate update strategy used in the present invention.

Detailed ways

Embodiments of the present invention will now be described with reference to the drawings, in which like reference numerals represent like elements.

Before explaining the improved proportional fair scheduling algorithm based on multi-user eigenmode (MET) precoding technology in this embodiment, the following concepts involved in the method are explained first:

Service user set: It is a set of users that the system will serve at a certain moment, and the users among them are the scheduled users. The process of user scheduling is a process of adding users to the service user set.

Scheduling priority: It is a measure of the user's priority in scheduling. The higher the scheduling priority, the higher the probability that the user will be scheduled in one scheduling. In actual implementation, the user with the highest scheduling priority is selected every time.

Greedy Algorithm: A solution method for a problem that does not pursue the global optimal solution, but only makes the best choice at present. In this method, a greedy algorithm is used for user selection, and each time a user is selected, only the user that can maximize the sum of user scheduling priorities of the currently serving users is selected.

Channel matrix: is a matrix in which the coefficients are the fading coefficients between the receiving antenna and the transmitting antenna. If the number of receiving antennas is N and the number of transmitting antennas is M, then the channel matrix is an N×M complex matrix. For a single user, the number of receiving antennas is the number of receiving antennas of the user; for multiple users, the channel matrix is the vertical connection of the channel matrices of each user, and the number of receiving antennas is the sum of the number of receiving antennas of all users.

MET precoding technology: as a block diagonal precoding technology, MET has three components, (1) block diagonalization (BD) technology based on linear precoding, (2) thorough user and feature flow Select, (3) for the power allocation to the selected eigenstreams. The block diagonalization technology in MET greatly reduces the complexity of transmitting precoding and receiving end. Multiuser Eigenmode Transmission (MET) currently has the best performance in linear precoding systems in point-to-multipoint broadcast channels.

The flow of the improved proportional fair method based on the multi-user feature mode selection precoding technology in this embodiment will be described in detail below. With reference to Figure 2, the improved proportional fair method includes the following steps:

Step S1, the user sends a service request to the base station, and at the same time feeds back its own channel state information to the base station, and the base station can obtain the channel matrix of each user according to the fed back channel state information;

Step S2, initialize the average transmission rate and the average request rate, the initial value of which can be taken as 1, that is: These two initial values are used in the first round of service user scheduling, and in the subsequent service user scheduling, the values updated during the process of selecting all users to serve each time are used as the initial values.

Step S3, initializing the service user set, making the service user set an empty set;

In step S4, the base station calculates the channel capacity of each user under the condition of no interference according to the adopted MET precoding method from the channel matrix of each user. The obtained result is taken as the instantaneous request rate R _k (t) of each user;

Step S5, using the obtained instantaneous request rate to update the average request rate of all users, And thus the increment of the average request rate can be obtained, and the updated value will be used as the initial value of the average request rate in the next round of service user scheduling;

Step S6, by the formula Calculate the priority of each user. The user with the highest priority is selected into the service user set, that is, the first user in the service user set is determined;

Then, according to the greedy selection algorithm, start preparing to select users from users who are not in the service user set to join the service user set. The next step is a tentative process, that is, tentatively add each candidate user to the service user concentration, examine the advantages and disadvantages of each solution, and finally choose the best solution, so as to determine the later joining service user concentration User. For the convenience of explanation, take the selection of the second user in the service user set as an example;

Step S7, try to add each candidate user to the set of service users. After joining, including the original users and new users, there are 2 users in the service user concentration at this time, and the channel capacity of each user is recalculated by using the channel matrix of these two users combined with the MET precoding technology;

Step S8, calculate the channel capacity of each user in the service user set, and use this new result as the instantaneous request rate R _k (t);

Step S9, according to the new instantaneous request rate obtained in step S8 and the average request rate increment and the initial average transmission rate updated when the user is selected for the first time, calculate the priority of each tentatively added user;

Step S10, select the user with the highest priority from all tentatively joined users to join the service user set, that is, determine the second user selected into the set;

Step S11, judging whether the number of users in the service user set has reached the maximum number of users that the system can serve simultaneously. If not, go back to step S7. At this time, instead of adding the second user to the service users, add the third user, and so on. If reached, then continue to step S12;

Step S12, using the instantaneous request rate obtained when the user joins the service user set tentatively for the last time, and updating the average transmission rate according to the status of user scheduling or not It is used as the initial value of the average transmission rate for the next round of service user set scheduling.

The present invention has been described above in conjunction with the best embodiments, but the present invention is not limited to the above-disclosed embodiments, but should cover various modifications and equivalent combinations made according to the essence of the present invention.

Claims (6)

1. based on multi-user features pattern MET precoding technique modified model proportional fair dispatching method, comprise the steps:

(1) base station receives after user's service request and channel condition information, initialization user average transmission rate, average request speed and service-user collection;

(2) upgrade each user's average request speed, and the user who selects to have greatest priority forms first user of service-user collection; Select greatest priority user as the concentrated user of service-user, the priority computing formula using is:

$\mathrm{\&mu;}=\frac{R_k\left(t\right)}{T_k\left(t\right)}\&CenterDot;\frac{R_k\left(t\right)}{e^{-(\stackrel{\&OverBar;}{R_k\left(t\right)}-\stackrel{\&OverBar;}{R_k(t-1)})}}---\left(1\right)$

Wherein be illustrated in user k t in the past in epicycle scheduling _caverage transmission rate in individual time slot, R _k(t) represent the instantaneous request rate of user k in current time slots, be illustrated in user k t in the past in epicycle scheduling _caverage request speed in individual time slot, represent user k t in the past in last round of scheduling _caverage request speed in individual time slot;

After all users that selected each time system simultaneously to serve, whether dispatch state according to user and upgrade average transmission rate formula be:

Upgrade average transmission rate R used _k(t) be while selecting last user, the channel capacity of calculating according to MET precoding;

be illustrated in user k t in the past in epicycle scheduling _caverage request speed in individual time slot, it just upgrades in the time selecting first user, while selecting all the other users, all uses the value after upgrading for the first time, and more new formula is:

$\stackrel{\&OverBar;}{R_k(t+1)}=(1-\frac{1}{t_c})\stackrel{\&OverBar;}{R_k\left(t\right)}+\frac{1}{t_c}{R}_k\left(t\right)---\left(3\right)$

Upgrade average request speed R used _k(t) be while selecting first user, the each user's who calculates according to MET precoding channel capacity;

(3) user that base station utilizes the each selection of greedy algorithm to have greatest priority adds service-user collection; Concrete steps are: step S7, try by each user to be selected join service-user concentrate go; After adding, count the user that original user and Xin add in, service-user is concentrated and is now had 2 users, utilizes these two users' channel matrix in conjunction with MET precoding technique, recalculates the channel capacity that obtains each user; Step S8, calculation services user concentrates each user's channel capacity, and using this new result as instantaneous request rate R _k(t); Step S9, the average request rate increment of upgrading according to the new instantaneous request rate obtaining in step S8 with while selecting user for the first time and initial average transmission rate, calculate each exploratory user's who adds priority; Step S10 selects the user of priority maximum to join service-user and concentrates from all exploratory users that add, and has determined second user who is selected in set; Step S11, judges whether the concentrated number of users of service-user has reached the maximum number of user that system can be served simultaneously; If do not reached, get back to step S7, now, be no longer to concentrate and add the 2nd user to service-user, but add the 3rd user, the like;

(4), after the user who has selected system simultaneously to serve, upgrade each user's average transmission rate and dispatch for next round user.

2. as claimed in claim 1 based on multi-user features pattern MET precoding technique modified model proportional fair dispatching method, be further characterized in that: by right with the time window upgrading is taken as the same.

3. as claimed in claim 1 based on multi-user features pattern MET precoding technique modified model proportional fair dispatching method, be further characterized in that, system is to know each user's instantaneous channel information, and need to utilize this information to carry out MET precoding, makes largeizationr of power system capacity.

4. as claimed in claim 1ly it is characterized in that: described channel capacity is according to channel matrix based on multi-user features pattern MET precoding technique modified model proportional fair dispatching method, calculate with reference to the method for precoding adopting.

5. as claimed in claim 1 based on multi-user features pattern MET precoding technique modified model proportional fair dispatching method, be further characterized in that, also comprise: for the poor user of channel condition, at channel condition during in himself peak value, can obtain, than the higher scheduling probability of dispatching algorithm of selecting heap(ed) capacity user under simple MET precoding technique, can well improving the fairness of system.

6. as claimed in claim 1 based on multi-user features pattern MET precoding technique modified model proportional fair dispatching method, be further characterized in that, also comprise: in the time of the poor user of selective channel condition, it is larger that the user with larger capacity increment obtains the probability of scheduling, can ensure like this in improving system fairness, be unlikely to the power system capacity that loss is too large, can better utilize MET precoding technique in the superiority improving in power system capacity.