Joint Transmission Method in Coordinated Multi-Point Transmission and Reception Systems

This work was supported by the National Science and Technology Major Project of the Ministry of Science and Technology of China under Grant No. 2009ZX03003.

    According to the requirements of Long Term Evolution Advanced (LTE-A), a cell’s average spectrum efficiency and edge spectrum efficiency, especially the latter, are attracting greater attention. This is both uplink and downlink of the LTE-A system are frequency division systems with the Orthogonal Frequency Division Multiplex (OFDM) as the basic
multi-access multiplexing mode. As opposed to the traditional wireless communication system based on Code Division Multiple Access (CDMA), the LTE-A system does not deal with gain and there is little interference because of the complete frequency division quadrature inside the cell, however, interference processing at the cell’s edge is rather complex.

    The technology of Coordinated Multi-Point (CoMP) transmission uses the transmitter antenna’s coordinated transmission of several cells to deliver high capacity and reliable transmission of the wireless link at the cell’s edge  thereby effectively solving the interference problem at the cell’s edge. There are two major types of CoMP joint transmission: the signal coherent one and signal non-coherent. The related transmission requires that signals transmitted by different nodes have coherence features at the reception end, so that the signals are able to obtain quite high combination gain. The
non-coherent transmission requires no coherence at the reception end but it obtains the diversity gain.

1 CoMP Transmission System
The multi-point transmission system is proposed with the main purpose solving the interference problem at the edge area of cells. It works on the principle of decreasing interference on edge users of cells by jointly scheduling several cells with rather strong edge interference, or by joint transmission, so that the reception power and service experience of a cell’s edge users can be improved. Figure 1 shows the basic principle of this transmission type. User Equipment 1 (UE1) at the cell edge is serviced by Cells 1, 2 and 3 at the same time. The interference from other cells to UE1 can be cleared in this way, and the signal quality of cell’s edge users can be enhanced.

2 Multi-Point Joint Transmission
Multi-point joint transmission is proposed to solve the serious interference problem by joint scheduling and joint transmission. There are two types of multi-point joint transmission:
non-coherent joint transmission and coherent joint transmission.

2.1 Non-Coherent Multi-Point Joint Transmission
The open-loop transmission solution is usually used for non-coherent transmission in a single cell. For example, References [1-3] give solutions of space time/frequency coding transmission diversity, and [4-6] propose transmission diversity solutions based on cyclic delay. Of the CoMP transmission systems, the simplest one adopts a transmission mode similar to the one used for Multimedia Broadcast and Multicast Services (MBMS), that is, different nodes send the same signals that are then combined in the space. In real cases, however, because of the randomness of space channels, effective combination of signals cannot be guaranteed. With the space
time/frequency coding scheme proposed by Alamouti, the greatest diversity gain can be obtained using two transmitter antennae. Beamforming technology, however, can have the power of transmitted signals stay at a certain angle, providing greater reception power at the reception end. The multi-path data obtained through space time/frequency packed encoding achieves combination gain at the reception end through a reception algorithm. Therefore, it’s advisable to separately map the multi-path data obtained through space time/frequency packed encoding to different nodes for transmission.

    Supposing the data sequence to be sent is [s 1, s 2, …, sn ], Space-Time Block Coding (STBC)/Space-Frequency Block Coding (SFBC) is performed first. The data that has undergone diversity processing is expressed as

and S 1 and S 2 are mapped to different nodes. The number of transmitter antennae of nodes performing CoMP transmission is supposed as Nti , where i stands for the ith transmitter node. To map data path Sk to Nti  antennae, the weighted value of beamforming is set as Bi , and Bi  is then used to process data Sk , where Bi  is the matrix of Nti  × 1. Once the signals sent from node i pass through the space channel, the expression measured at the UE is:

    where Hi  is the channel formed between Node i  and target users. Supposing the number of reception antennae of the target user is Nr , Hi  is the matrix of Nr  × Nti . For the sake of simplicity, we take the joint transmission performed by two nodes as an example to study the signal expression at the reception end.
If Node 1 sends the first channel of SFBC coded data and Node 2 sends the second channel of SFBC coded data, the signals received at the UE side from the two different nodes are respectively as follows:

    Then, the received signals of the reception end at two neighboring carriers are:

    As shown in Formula (2), the UE side can use Hei  to conduct the tests Alamouti proposed. Hei  is the coherent combined channels obtained through Bi  processing between several transmitter antennae and reception antennae, and therefore it is possible to achieve a high power gain. At the same time, because the nodes are located at different physical positions, even greater diversity gain can be obtained.

    When there are many nodes involved in the coordination, the nodes can be sorted into differnet groups based on the abovementioned method. That is to say, to form the coordianted nodes into two groups, the nodes in the same group transmits the same channel of data that has undergone SFBC coding. To achieve greater diversity gain, if needed, perform SFBC+Frequency Switched Transmit Diversity (FSTD) encoding to data [s 1, s 2, …, sn ], that is, to map

 to several channels of data and allocate them to different nodes for transmission. For example, the following mapping can be made if there are 4 nodes participating in a joint transmission:

 
    Moreover, all channels of data are mapped, on different nodes and through beamforming weighted value, to several antennae of corresponding nodes for transmission.

2.2 Coherent Multi-Point Joint Transmission
The transmitter conducts pre-processing to some extent based on channel information, to simplify the detection algorithm of the receiver. This is precoding technology, which includes linear precoding and non-linear precoding[7-11]. Common linear precoding technology refers to precoding based on a channel’s Singular Value Decomposition (SVD), the precoding based on Zero Forcing (ZF) algorithm, and the precoding based on Minimum Mean-Squared Error (MMSE) algorithm. Common non-linear precoding technology refers to Tomlinson-Harashima Precoding (THP) or dirty paper coding. Current studies of standards and protocols in precoding technology are primarily focused on linear processing precoding.

    The precoding processing of multi-point joint transmission described in this article combines the precoding technology and beamforming technology for CoMP transmission. This method works on the principle that different nodes participating in the coordinated transmission separately transmit a channel of data obtained through precoding, then perform beamforming on the data, and finally send the data out. Figure 2 shows the method:  
Supposing the available number of layers for precoding is L, data to be sent on a certain carrier is [s 1, s 2, …, sn ]T, the precoding matrix is W, (that is, the P × L dimension precoding matrix, where P  is the number of transmitter ports or coordinating nodes), Bp =[b1, b2, …, bN ]T, (which is the beam vector used by the transmitter port to perform beamforming), and where [.]T means the transposition of matrix.

    This method supposes the number of coordinating nodes is P. Firstly, the beam vector Bp can be obtained from the channel Hp that corresponds to every node p, can be obtained in light of the channel mutual-benefit feature or with the method of Direction of Arrival (DOA) estimation. The equivalent channel after beamforming processing is He p =Hp·Bp, where Hp is the matrix of Nr ×Nt    , and He p  is the channel matrix of Nr × 1. Therefore, the equivalent channels of all nodes constitute He, the channel matrix of Nr × P. Furthermore, He can be used to obtain the precoding weighted value W  for the precoding processing. Corresponding to the coordinating node p, the Wp , W being in the pth line of the precoding weighted value matrix, is used for the precoding processing. The final received signal is as follows:

    In Formula (3),   is the channel of the nth node participating in the coordination as detected by the terminal, (being the Nr ×1 dimension matrix),   is the channel between the nth coordinating node and the terminal, (being the Nr ×N t , n dimension matrix), Bn is the direction weighted value’s vector, (being the Nt ,
n ×1 dimension vector), and W ’n, 1≤n≤PN  is the 1×L dimension matrix.

    In the above described processing, the precoding of every node breaks into the following two processes:

    (1) precoding processing, that is, every node uses a line in the precoding weighted value matrix to perform precoding processing, so as to guarantee the relevancy feature of signals of different nodes.

    (2) beamforming processing, to make sure signals of every node are set at a certain angle range that the terminal corresponds to, so as to obtain the power gain of beam forming.

3 Results of Simulation
References [12-14] describe some other transmission schemes, including Single Frequency Network (SFN)-based and Cyclic Delay Diversity (CDD)-based non-coherent transmission schemes, and joint pre-coding-based, independent pre-coding-based and SFN pre-coding coherent transmission schemes. The coherent and non-coherent transmission schemes are compared with the schemes described in this article, and the simulation results are shown in Figures 3 and 4.

    The following configuration parameters are used for the simulation tests:

• Every node is configured with two pairs of bi-polar antennae.
• Two cells are involved in the coordinated transmission.
• The channel model is urban macro cell Space Channel Model (SCM).
• The system has a bandwidth of 5 MHz, the ideal channel estimation is adopted.
• The spacing between antennae on both the BTS side and terminal side is 0.5 wavelength.
• The detection algorithm is ZF, and the beam weighted value is obtained with the method of Eigenvalue Based beamforming (EBB). Just one data stream is transmitted for non-coherent transmission and two data streams are transmitted for coherent transmission.

    According to the simulation results, when the ideal beam weighted value and the non-coherent mode are employed, very high gain can be achieved with the method proposed in this article; while in the coherent mode, the performance proves to be second only to the joint pre-coding scheme.

4 Conclusions
This article proposes a method of CoMP joint transmission in the CoMP transmission system. The method makes use of beamforming technology coupled with space time/frequency coding or pre-coding technology, performs space time/frequency coding or pre-coding to obtain channels of data, and then maps the data to different nodes for transmission, thus obtaining both the diversity gain and beamforming power gain. In addition, with beamforming technology, if every node performs the same forming processing on the pilot signals, the CoMP transmission and non-CoMP joint transmission are transparent to the terminal as the terminal’s processing method of received signal is concerned, and the processing at the terminal is accordingly simplified[15-17]. The simulation results show that the precoding and beamforming combined method leads to performance close to that of joint precoding, while the processing complexity of UE reception and the design complexity of codebook are much lower than those of joint precoding. In the non-coherent mode, the SFBC and beamforming technology combined performance is much better than the schemes based on SFN or CDD.

References
[1] ALAMOUTI S M. A simple transmitter diversity scheme for wireless communications [J]. IEEE Journal on Selected Areas in Communications, 1998, 16(10): 1451-1458.
[2] ALAMOUTI S M, TAROKH V, POON P. Trellis coded modulation and transmit diversity: Design criteria and performance evaluation [C]//Proceedings of the 1998 International Conference on Universal Personal Communications (ICUPC 98), Oct 5-9, 1998, Florence, Italy. Piscataway, NJ, USA: IEEE, 1998: 703-707.
[3] FOSCHINI G F. Layered space-time architecture for wireless communication in a fading environment when using multi-element antennas [J]. Bell Laboratories Technical Journal, 1996, 1(2): 41-59.
[4] DAMMAN A, KAISER S. Low complex standard conformable antenna diversity techniques for OFDM systems and its application to the DVB-T system [C]// Proceedings of the 4th International ITG Conference on Source and Channel Coding (ICSCC'02), Jan 28-30, 2002, Berlin, Germany. 2002: 253-259.
[5] DAMMAN A, LUSINA P, BOSSERT M. On the equivalence of space-time block coding with multipath propagation and/or cyclic delay diversity in OFDM [C]//Proceedings of IEEE European Wireless 2002: Vol 2, Feb 25-28, 2002, Florence, Italy. 2002: 847-851.
[6] HUEBNER A, BOSSERT M, SCHUEHLEIN F, et al. On cyclic delay diversity in OFDM based transmission schemes [C]//Proceedings of the 7th International OFDM Workshop (InOWo’02), Sep 10-11,2002, Hamburg, Germany. 2002: 1-5.
[7] COLLIN L, BERDER O, ROSTAING P, et al. Optimal minimum distance-based precoder for MIMO spatial multiplexing systems [J]. IEEE Transactions on Signal Processing, 2004, 52(3): 617-627.
[8] ZHOU S, GIANNAKIS G B. Optimal transmitter eigen-beamforming, and space-time block coding based on channel correlations [J]. IEEE Transactions on Information Theory, 2003, 49(7): 1673-1690.
[9] AKHTAR J, GESBERT D. A closed-form precoder for spatial multiplexing over correlated MIMO channels [C]//Proceedings of Global Telecommunications Conference (GLOBECOM '03), Vol 4, Dec 1-5,2003, San Francisco, CA, USA. Piscataway, NJ, USA: IEEE, 2003: 1847-1851.
[10] ROH J C, RAO B D. Multiple antenna channels with partial channel state information at the transmitter [J]. IEEE Transactions on Wireless Communications, 2004, 3(2): 677-688.
[11] SAMPATH H, STOICA P, PAULRAJ A. Generalized linear precoder and decoder design for MIMO channels using the weighted MMSE criterion [J]. IEEE Transactions on Wireless Communications, 2001, 49(12): 2198-2206.
[12] 3GPP R1-084144. Per-cell Precoding Methods for Downlink Joint Processing CoMP [S]. 2008.
[13] 3GPP R1-090142. Performance Evaluation of CoMP Solutions [S]. 2009.
[14] 3GPP R1-090193. Aspects of Joint Processing in Downlink CoMP [S]. 2009.
[15] Liu Zhimin, Yang Yi, Xu Yinqing. Key Technology and Experimental Research in Wireless Mesh Networks [J]. ZTE Communications, 2008, 6(2): 39-44.
[16] Yang Wendong, Cao Yueming, Xu Youyun. Wireless Cooperative Mesh Network: A New Architecture for Network Convergence [J]. ZTE Communications, 2008, 6(3): 20-23.
[17] Zhao Rui, Yu Fei, Yang Lvxi. Relay-Assisted Cooperative Communication Networks [J]. ZTE Communications, 2008, 6(3): 24-28.

[Abstract] In this paper, we propose a multi-point joint transmission method for Coordinated Multi-Point (CoMP) systems. This solution combines space time/frequency coding and precoding with beamforming technology. The multi-path data obtained by space time/frequency coding and precoding can be mapped by beamforming technology to multiple coordinated nodes for transmission. The signal receiver can also obtain bigger diversity gain and beamforming gain as well through the combination of space time/frequency coding and beamforming. Moreover, precoding and beamforming can help the signals transmitted from separate nodes in CoMP system combine at the receiver, which makes the signals at different layers more independent and obtains beamforming gain.