In traditional cellular systems, each cell transmits to its own user equipment (UE) and in the process creates interference to UEs in adjacent cells. This is illustrated in
a) Coordinated Scheduling and/or Beamforming (CS/CB)
b) Joint Processing/Transmission (JP)
The first step in any CoMP scheme, such as CS/CB or JP, is to decide which BS will cooperate for a given UE. In “R1-092232, Summary of email discussions CoMP v2”, 3GPP TSG RAN WG1 meeting #57″, the following terms are introduced in this context.
The layout of cells is shown in
In a static cooperating set formation, the clusters of cells are formed apriori to the UE signaling and reporting of link gain values to different cells. The formation of clusters is based on geometry.
The problem with static cooperating set formation is that, for a given UE, the cells that have the strongest link gain to it after its serving cell (i.e. the cell that serves it in case of no CoMP), may not lie in its cooperating set. These cells would then cause strong interference. This is possible due to the random nature of link gain of which the random shadowing component is the biggest factor. Thus, in another way of forming the cooperating set of a given UE, the UE first determines the cells with the strongest link gains (by measuring the Reference Signal Received Power or RSRP of all cells and finding out which ones lie within a threshold of the serving cell) and then reports this to its serving cell. The serving cell then decides to form a cooperating set with these cells. Such approaches have been considered in “R1-093410, “Coordinated Beamforming Algorithms Based on Spatial Covariance Feedback and its Gain over Single-point SU/MU Beamforming”, 3GPP TSG RAN WG1 meeting #58”.
Frequency partitioning is another important method that increases the overall system performance. One simple illustrative example could be that the total available spectrum is partitioned into disjoint bands to transmit to different classes of UEs. This mitigates the problem of interference and allows each cell to perform SU MIMO algorithms in these disjoint bands. The other alternative would have to transmit to the different UEs in the total band but use MU-MIMO beamforming to mitigate interference. The former approach of frequency partitioning can be easier to implement. Often different frequency bands are used for transmitting to uses in the cell edge as they are the most prone to interference. Such approaches have been considered in “R1-093279, “Downlink CoMP based on cooperative precoding”, 3GPP TSG RAN WG1 meeting #58”.
Coordinated Multipoint Transmission/Reception (CoMP) is considered to be an important feature for Release-10 (Rel-10) LTE-Advanced technology as this enhances the system data rate and reliability. CoMP involves multiple cells cooperating together to transmit to a single UE. In the future it is envisaged that there will be networks which will have a mixture of Release 8 (Rel-8) LTE UEs that do not support CoMP and Rel-10 UEs that have the option of using the CoMP mode, especially during the period of transition when the market moves from the current LTE stage to the future LTE-A stage.
If a UE is using the CoMP mode, the transmission set or the cells that cooperate to transmit to it, needs to be decided. Instead of forming fixed transmission sets based on cell geometry, it is better to form these dynamically based on the UE link gains to the various cells (R1-093410, “Coordinated Beamforming Algorithms Based on Spatial Covariance Feedback and its Gain over Single-point SU/MU Beamforming”, 3GPP TSG RAN WG1 meeting #58).
At each Transmission Time Interval (TTI), the BS of a cell has to decide to which UE to transmit to amongst all UEs for which it is the serving cell. One approach would be to transmit to that UE to which it can support the highest data rate, which is the UE which has the best link gain to the BS. However as link gains are strongly dependant on distance, this means that most of the times the nearby UEs are scheduled. Thus this scheduling method is said to be unfair to the faraway UEs. To remedy this situation, the proportional fair (PF) metric has been introduced and is widely used in many wireless systems such as CDMA/HDR. In this method the BS calculates the instantaneous rate to a UE if that UE is scheduled and divides it by the average rate received by the UE till that TTI. The UE with the highest value of this metric is chosen for transmission. Thus in PF scheduling both instantaneous rate (strength of link gain) and average rate (long term fairness) is considered.
Thus, there is a need for efficient portioning of BS resources, such as time, bandwidth and also assigning BSs to cooperating sets of different UEs so that there is seamless operation of both Rel-8 and Rel-10 UEs.
In the embodiments of the present invention, proposed is a method in which a Rel-10 (CoMP enabled) UE chooses the BSs to be in its cooperating set and a BS partitions its bandwidth to serve its own UEs and UEs from other cells that have requested it to be in its cooperating set.
In the beginning of each scheduling instant, each BS decides which UE to schedule to using a proportional fair (PF) metric. If the UE supports Rel-10 (i.e. CoMP enabled UE), it opts for entering the CoMP mode. It reports, to its serving cell, the cell-id of the cell to which it has the highest received signal power (as measured by the Reference Signal Received Power, RSRP) after the serving cell. The serving cell contacts this new cell and informs it of this request.
If a cell receives multiple such CoMP requests, it uses another PF metric to choose a single UE to transmit to. The PF metric looks at the extra rate increase of a UE if it was scheduled by this cell and the average rate of that UE till that TTI.
Each cell divides a portion of its available bandwidth to serve its own UE for which it is the serving cell and the rest is used to serve the UEs from other cells who hand in a CoMP request. This frequency portioning is the simplest scheme and ensures that within the allocated band, each cell carries out Single User (SU) beamforming as in Rel-8.
In the embodiment of the present invention, an algorithm to be followed by the UEs and the BSs is provided so that each BS can determine whether it has to be a part of a cooperating set and if so then how much of its resources (e.g. power, bandwidth) to allocate in this cooperative transmission. One of the embodiments of this invention includes a system in which LTE and LTE-A enabled UEs can co-exist assuming that BSs are capable of running both protocols.
For example, basically BS 0 has to take care of UE 0 as a matter of fairness. Note that if all the UEs were following CoMP, they would all have a set of BSs transmitting to them and the fairness issue for an individual UE is different. However, the illustrated example involves the network which comprises both kinds of UEs—those that have CoMP capability, i.e. UE n, and those that do not, i.e. UE 0.
UE n requests Cell 0 to be in its cooperating set and Cell 0 agrees. Cell 0 partitions its bandwidth into two parts and uses αW bandwidth for transmitting to UE 0 using SU transmission techniques. The remaining bandwidth (1−α)W is used to transmit to UE n. BS n transmits to UE n using the entire bandwidth W. Since UE 0 is non-CoMP enabled, it sees interference from BS n.
In practice, αW will be implemented in a discrete manner. For a 10 MHz bandwidth there are 50 Physical Resource Blocks (PRBs) per TTI with each PRB being of 180 KHz bandwidth. BS 0 could allocate N of them to UE 0, where N>25 and 50−N to UE n.
Cell 0, after frequency partitioning, follows LTE transmission to UE 0, the UE that was not Rel-10 enabled. For transmission to UE n, both cells can follow either CB/CS or JP CoMP techniques as aforementioned.
In STEP 701, at each TTI, each cell decides which UE to transmit to by running a PF algorithm. In this embodiment, Cell i decides the UE i to transmit to, via PF algorithm on all UEs for which it is the serving cell. In STEP 702, whether UE is CoMP-enabled is determined. In this embodiment, whether UE i is CoMP-enabled is determined. If the UE is not CoMP enabled, then the process proceeds to STEP 703. In STEP 703, whether UEs from other cells ask Cell i for cooperation is determined. If there is any request from UEs from other cells for cooperation, Cell i determines UE n to transmit to, using PF algorithm as illustrated in
If the UE is CoMP-enabled, UE i reports to Cell i, the id of the Cell that has the next strongest link gain to it (i.e. after Cell i itself) (STEP707). In this embodiment, Cell m is assumed to be the Cell that has the next strongest link gain to UE i. Cell i requests Cell m to be in UE i's cooperating set (STEP708). Then whether Cell m accepts the request from Cell i for cooperation is determined (STEP709). If Cell m accepts the request from Cell i after running PF algorithm as illustrated in
If there is no request to Cell i from UEs from other cells for cooperation, Cell i serves UE i using full bandwidth W via CoMP with Cell m (STEP711). If there is any request to Cell i from UEs from other cells for cooperation, Cell i determines UE n to transmit to using a PF algorithm as illustrated in
The advantage of the embodiment of the present invention is the backward compatibility of CoMP enabled Rel-10 with Non-CoMP enabled Rel-8 as both types of UEs co-exist. The algorithm can be implemented with minimum complexity and gives flexibility to all UEs and BSs involved in the cellular system. Another embodiment of this invention is a new dynamic algorithm for forming a CoMP cooperating set. The UE initiates this cooperating set formation by sending requests to BSs and the BS chose only one UE in order to minimize computational complexity.
This is a divisional application of U.S. application Ser. No. 12/687,405 filed Jan. 14, 2010. The entire disclosure of U.S. application Ser. No. 12/687,405 is incorporated herein in its entirety by reference.
Number | Name | Date | Kind |
---|---|---|---|
8219128 | Zangi et al. | Jul 2012 | B2 |
20080242308 | Gunnarsson et al. | Oct 2008 | A1 |
20080316955 | Yu | Dec 2008 | A1 |
20090010219 | Lee et al. | Jan 2009 | A1 |
20090201867 | Teo et al. | Aug 2009 | A1 |
20090219875 | Kwak et al. | Sep 2009 | A1 |
20090257390 | Ji et al. | Oct 2009 | A1 |
20090323614 | Wang et al. | Dec 2009 | A1 |
20100069064 | Hannu et al. | Mar 2010 | A1 |
20100165930 | Zangi | Jul 2010 | A1 |
20100173660 | Liu et al. | Jul 2010 | A1 |
20100195599 | Zhang et al. | Aug 2010 | A1 |
20100203887 | Kim | Aug 2010 | A1 |
20100261493 | Guey et al. | Oct 2010 | A1 |
20100304682 | Choi et al. | Dec 2010 | A1 |
20100323611 | Choudhury | Dec 2010 | A1 |
20110026483 | Rudrapatna et al. | Feb 2011 | A1 |
20110098054 | Gorokhov et al. | Apr 2011 | A1 |
20110170422 | Hu et al. | Jul 2011 | A1 |
20110237272 | Gorokhov et al. | Sep 2011 | A1 |
20110237282 | Geirhofer et al. | Sep 2011 | A1 |
20110255436 | Jian et al. | Oct 2011 | A1 |
20110269449 | Kazmi et al. | Nov 2011 | A1 |
20110269459 | Koo et al. | Nov 2011 | A1 |
20120076038 | Shan et al. | Mar 2012 | A1 |
20120142389 | Wei | Jun 2012 | A1 |
20120147773 | Kim et al. | Jun 2012 | A1 |
Entry |
---|
PCT International Search Report mailed on Sep. 19, 2011. |
Number | Date | Country | |
---|---|---|---|
20130137442 A1 | May 2013 | US |
Number | Date | Country | |
---|---|---|---|
Parent | 12687405 | Jan 2010 | US |
Child | 13718484 | US |