-
协作分集技术利用无线信道的广播特性,使多用户网络中的节点彼此共享天线构成虚拟天线阵,互相协作传输数据信息,从而获得分集增益,大幅提高无线通信系统的性能[1-3]。根据参与协作传输的中继节点数目不同,协作分集方法可分为多中继协作和单中继协作。以分布式空时码为代表的多中继协作方法,已经能使系统获得显著的分集增益[4-5]。但是,在这种多中继协同参与机制中,伴随中继节点数量的增加,频谱效率会有所下降。针对该问题,文献[6-7]分别提出了机会中继和选择协作,并且分析了其性能,证明了在高信噪比(signal-to-noise ratio, SNR)条件下这两种协作方法都能够取得全分集增益,验证了单中继协作方法的有效性。选择协作方法的基本原理是源节点发送数据后,在成功解码源节点信息的中继节点中,选择一个节点作为最优中继协助源节点转发数据。目前针对选择协作方法的研究,大多关注物理层本身的协作,较少考虑与其他通信机制进行联合优化。
在实际通信系统中,为了提高传输可靠性,链路层自动重传请求(ARQ)机制被广泛应用[8]。若将物理层协作分集和ARQ机制联合使用,能进一步提高系统的传输性能。文献[9]研究了ARQ机制在多输入多输出(multiple-input multiple-output, MIMO)系统中的性能,指出ARQ机制能够提供额外的分集增益[9]。不同于传统ARQ重传总是由源节点响应重传请求,在协作ARQ系统中,中继节点能够响应重传请求,且每次重传的链路质量最优[10-11]。文献[12-13]通过将ARQ协议与基于分布式空时码的多中继协作方法相结合,使系统的传输性能得以改善,获得了较高的系统分集增益。为了有效提高系统的频谱利用率,进一步考虑将ARQ协议与单中继协作方法相结合。文献[14]改进了机会中继在ARQ场景下的中继选择方式,通过利用源节点与目的节点间的信道状态信息(channel state information, CSI)进行最优中继节点选择,并且推导了系统的中断概率。文献[15]讨论了分布式选择协作与ARQ机制联合使用的方法,并指出在中继节点具备信号合并接收能力的情况下,系统能够获得最优分集复用时延(D-M-D)权衡。节点具备信号合并接收能力,即能够合并处理所有轮次中接收到的数据;不具备信号合并接收能力,即当前轮次解码失败时,直接丢弃接收数据。
所以,在中继节点不具备信号合并接收能力的情况下,要通过合理的协议设计,使系统获得比较好的性能增益。在文献[15]中,中继节点在第1轮重传开始前被选择为最优中继节点,那么它在后面所有重传轮次中都将作为最优中继参与协作传输。其他中继节点一直处于空闲状态,重传过程实质上是重复简单的重传模式,很大程度上降低了系统性能。基于此,本文提出一种在重传过程中系统解码集和最优中继选择都动态改变的ARQ选择协作方法,称之为ML-DR(memoryless-dynamicrelay),并分析了其D-M-D权衡性能和中断概率。
-
本文提出的ML-DR方法中单个数据包的传输过程如图 1所示,包含的主要步骤如下:
1) 源节点向目的节点直接发送数据,同时所有中继节点也能够接收该数据。如果目的节点正确解码数据,则返回一个确认(acknowledgement, ACK)帧,表明传输成功;否则,返回一个否决(negative acknowledgement, NACK)帧,表明传输失败,系统进入步骤2)。
2) 在步骤1) 中能够成功解码源节点数据的中继节点及源节点组成首轮重传的解码集D(s)。解码集中每个节点都能够接收步骤1) 中目的节点发送的NACK帧,并计算出各自与目的节点间的信道值。系统利用该值为每个节点设置一个定时器回退时间Ti,信道值越大回退时间越短,即最优中继与目的节点间的通信链路最好。在竞争过程中,节点持续监测信道,一旦发现其他节点开始发送数据,说明其他节点与目的节点间的链路更好,从而终止其自身的定时器并退出竞争;如果节点的定时器减为零后监测到信道没有被占用,则该节点成为当前轮次的最优中继。如果多个节点同时满足最优条件,则随机选取一个,最优中继节点向目的节点发送数据。
3) 目的节点接收最优中继节点发送的数据,与之前重传轮次中接收到的源节点或最优中继节点发送的数据进行合并处理。若正确解码数据,则返回一个ACK帧,表明传输成功;否则,返回一个NACK帧,表明传输失败,系统进入步骤4)。
4) 在步骤2) 中能够成功解码最优中继节点数据的中继节点及源节点组成第i轮(i=2, 3, …, L-1) 的解码集D(s)(i)。最优中继竞争和数据重传过程同步骤2) 和步骤3),且逐轮循环。
系统会在两种情况下结束当前数据的传输过程:1) 目的节点在当前轮次传输中正确解码数据,返回ACK帧;2) 目的节点在所有轮次传输中均未能正确解码数据,且ARQ传输达到最大重传轮数L。
ML-DR方法的优势主要有以下3点:
1) 利用ARQ机制的多轮反馈和重传机制,有效地提升了系统传输可靠性;
2) 在ARQ重传过程中保持系统解码集和最优中继选择动态变化,扩大了最优中继选取范围,充分利用了系统节点资源,且提高了系统性能;
3) 在高SNR条件下,能使系统在不增加中继节点接收机复杂度的情况下获得全分集增益。
-
本文提出方法的系统中断概率由定理1给出。
定理 1 ML-DR方法的系统中断概率为:
$$ {P_{{\rm{out}}}} = \sum\limits_{\mathit{\Omega}} {{\rho ^{ - \left[{{k_L}\left( {1-\frac{r}{L}} \right) + \sum\limits_{i = 1}^L {(1-r)(N + 1-{k_i})} } \right]}}} $$ (1) 式中,L是ARQ最大重传轮数;r是源节点到目的节点直接传输复用增益;Ω为解码集k1, k2, …, kL的取值空间,表示为:
$$ {\mathit{\Omega}} = \{ ({k_1}, {k_2}, \cdots, {k_L})|1 \le {k_1}, {k_2}, \cdots, {k_L} \le N + 1\} $$ (2) 证明:设第i轮的解码集D(s)(i)成员数为ki,且随着重传轮数i的不同,ki是动态变化的。此外,源节点自动加入解码集,参与协作竞争。故ki的最大值为N+1,最小值为1。在ARQ系统中,系统中断概率由达到最大重传轮数后,目的节点仍未能成功解码数据的中断事件决定。中断事件的出现,表示源节点与目的节点间的互信息量无法支持传输速率R。根据全概率公式,系统的中断概率为:
$$ \begin{array}{l} \;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;{P_{{\rm{out}}}} = \Pr \{ {O_L}\} = \\ \sum\limits_{\mathit{\Omega}} {\Pr \{ {I_D}^{(L)} \le R|{k_1}, {k_2}, \cdots, {k_L}\} } \Pr \{ {k_1}, {k_2}, \cdots, {k_L}\} \end{array} $$ (3) 式中,ID(L)为达到最大重传轮数后源节点与目的节点间的互信息量。
计算过程如下:
1) 解码集条件下的中断概率
经过L轮重传后,源节点与目的节点之间的互信息量ID(L)为:
$$ {I_D}^{(L)} = \sum\limits_{i = 1}^L {\log (1 + \rho (|{\alpha _{s, d}}{|^2} + |\alpha _{b, d}^{(i)}{|^2}))} $$ (4) 式中,$ |\alpha _{b, d}^{(i)}{|^2} = \max \{ |{\alpha _{s, d}}{|^2}, |\alpha _{{r_1}, d}^{(i)}{|^2}, \cdots, |\alpha _{{r_{{k_i} - 1}}, d}^{(i)}{|^2}\} $。
设$ {\mu _{s, d}} $、$ \mu _{b, d}^{(i)} $分别为$ |{\alpha _{s, d}}{|^2} $和$ |\alpha _{b, d}^{(i)}{|^2} $的指数阶数,将式(4) 改写为:
$$ \begin{array}{l} {I_D}^{(L)} \buildrel\textstyle.\over= \sum\limits_{i = 1}^L {\log } {\rho ^{(1 - \min ({\mu _{s, d}}, \mu _{b, d}^{(i)}))}} \buildrel\textstyle.\over= \\ \sum\limits_{i = 1}^L {\log } {\rho ^{(1 - \mu _{b, d}^{(i)})}} = \log {\rho ^{\left( {L - \sum\limits_{i = 1}^L {\mu _{b, d}^{(i)}} } \right)}} \end{array} $$ (5) 式中,
$$ \mu _{b, d}^{(i)} = \min \{ {\mu _{s, d}}, {\mu _{{r_1}, d}}, \cdots, {\mu _{{r_{_{{k_i} - 1}}}}}_{, d}\} $$ (6) 通过观察可知:
$$ \mu _{b, d}^{(1)} \ge \mu _{b, d}^{(2)} \ge \cdots \ge \mu _{b, d}^{(L)} $$ (7) 根据指数阶数的性质[6],由式(4)~式(7) 可得:
$$ \begin{array}{l} \;\;\;\;\;\;\;\;\;\;\Pr \{ I_D^{(L)} < R|{k_1}, {k_2}, \cdots, {k_L}\} \buildrel\textstyle.\over= \\ \Pr \left\{ {\left. {\log {\rho ^{\left( {L - \sum\limits_{i = 1}^L {\mu _{b, d}^{(i)}} } \right)}} < r\log \rho |{k_1}, {k_2}, \cdots, {k_L}} \right\}} \right. = \\ \;\;\;\;\;\;\Pr \left\{ {\left. {\sum\limits_{i = 1}^L {\mu _{b, d}^{(i)}} > L - r|{k_1}, {k_2}, \cdots, {k_L}} \right\}} \right. \ge \\ \;\;\;\;\;\;\;\Pr \left\{ {\left. {\sum\limits_{i = 1}^L {\mu _{b, d}^{(L)}} > L - r|{k_1}, {k_2}, \cdots, {k_L}} \right\}} \right. = \\ \;\;\;\;\;\;\;\;\;\;\;\Pr \{ L\mu _{b, d}^{(L)} > L - r|{k_1}, {k_2}, \cdots, {k_L}\} = \\ \Pr \left\{ {\left. {{\mu _{s, d}} > 1 - \frac{r}{L}, \cdots, {\mu _{{r_{{k_L} - 1}}, d}} > 1 - \frac{r}{L}|{k_1}, {k_2}, \cdots, {k_L}} \right\}} \right. \buildrel\textstyle.\over= \\ \;\;\;\;\;\;\;\;\;\;{\rho ^{ - \left( {1 - \frac{r}{L}} \right)}}{\rho ^{ - \left( {1 - \frac{r}{L}} \right)}} \cdots {\rho ^{ - \left( {1 - \frac{r}{L}} \right)}} = {\rho ^{ - {k_L}\left( {1 - \frac{r}{L}} \right)}} \end{array} $$ (8) 2) 解码集概率
形成第i轮解码集D(s)(i)前,所有中继节点均能收到第i-1轮中最优中继节点发送的数据。此时,中继节点ri与最优中继节点间的互信息量为:
$$ I_R^{(i)} = \log (1 + \rho |\alpha _{b, r}^{(i - 1)}{|^2}) $$ (9) 该节点无法加入解码集时,有:
$$ \begin{array}{l} \;\;\;\Pr \{ {r_i} \notin D{(s)^{(i)}}\} = \Pr \{ I_R^{(i)} < R\} = \\ \Pr \{ \log (1 + \rho |\alpha _{b, r}^{(i - 1)}{|^2}) < r\log \rho \} = \\ \;\;\;\;\;\Pr \{ |\alpha _{b, r}^{(i - 1)}{|^2} < r - 1\} \buildrel\textstyle.\over= {\rho ^{ - (1 - r)}} \end{array} $$ (10) 进而得到:
$$ \begin{array}{l} \Pr \{ {k_i}\} = \left( {\begin{array}{*{20}{c}} {N + 1 - {k_{i - 1}}}\\ {N + 1 - {k_i}} \end{array}} \right)\prod\limits_{r \notin D{{(s)}^{(i)}}} {{\rho ^{ - (1 - r)}}} \times \\ \prod\limits_{r \in D{{(s)}^{(i)}}} {(1 - {\rho ^{ - (1 - r)}})} \buildrel\textstyle.\over= {\rho ^{ - (1 - r)(N + 1 - {k_i})}} \end{array} $$ (11) $$ \Pr \left\{ {\left. {{k_1}, {k_2}, \cdots, {k_L}} \right\}} \right. = \prod\limits_{i = 1}^L {\Pr \{ {k_i}\} } \buildrel\textstyle.\over= {\rho ^{ - \sum\limits_{i = 1}^L {(1 - r)(N + 1 - {k_i})} }} $$ (12) 将式(8) 和式(12) 带入式(3) 可得:
$$ \begin{array}{l} {P_{{\rm{out}}}} = \Pr \{ {{\rm O}_L}\} \le \sum\limits_{\mathit{\Omega}} {\left( {{\rho ^{ - {k_L}\left( {1 - \frac{r}{L}} \right)}}{\rho ^{ - \sum\limits_{i = 1}^L {(1 - r) \cdot (N + 1 - {k_i})} }}} \right)} = \\ \;\;\;\;\;\;\;\;\;\;\;\;\sum\limits_{\mathit{\Omega}} {{\rho ^{ - \left[{{k_L}\left( {1-\frac{r}{L}} \right) + \sum\limits_{i = 1}^L {(1-r)(N + 1-{k_i})} } \right]}}} \end{array} $$ (13) 证明完毕。
-
本文提出方法的D-M-D权衡由定理2给出。
定理 2 ML-DR方法的D-M-D权衡为:
$$ d({r_e}, L) = (N + 1)\left( {1 - \frac{{{r_e}}}{L}} \right) $$ (14) 式中,N是中继节点个数;re是综合ARQ多轮传输后的有效复用增益。
证明:根据分集阶数的定义,由式(13) 可得:
$$ d(r, L) = \min \left\{ {\left. {{k_L}\left( {1 - \frac{r}{L}} \right) + \sum\limits_{i = 1}^L {(1 - r)(N + 1 - {k_i})} } \right\}} \right. $$ (15) 令:
$$ f({k_1}, {k_2}, \cdots, {k_L}) = {k_L}\left( {1 - \frac{r}{L}} \right) + \sum\limits_{i = 1}^L {(1 - r)(N + 1 - {k_i})} $$ (16) 通过计算可得,当$ 0 \le r < 1 $,$ {k_1} = {k_2} = \cdots = $N+1时$ f({k_1}, {k_2}, \cdots, {k_L}) $取得最小值,即:
$$ d(r, L) = (N + 1)\left( {1 - \frac{r}{L}} \right) $$ (17) 利用文献[15]中的结论可知,在高SNR下,$ {r_e} \buildrel\textstyle.\over= r $,$ d({r_e}, L) = d(r, L) $。证明完毕。
从定理1和定理2可以看出,在中继节点不具备信号合并接收能力的情况下,保持系统解码集和最优中继节点选择动态变化,同样能使系统取得最优D-M-D权衡及全分集增益。
An ARQ-Based Selection Cooperation Scheme with Dynamic Best Relays
-
摘要: 在以多轮重传增加时延为代价提高可靠性的自动重传请求(ARQ)协作网络中,对中继节点无信号合并接收能力时的跨层协作策略进行了研究,提出了一种在ARQ重传过程中系统解码集和最优中继选择动态改变的选择协作方法。在Rayleigh衰落信道环境中,该方法能够在不增加中继节点接收机复杂度的情况下取得最优分集-复用-时延(D-M-D)权衡。仿真结果表明,该方法的系统中断概率性能与中继节点具备信号合并接收能力方案的中断概率性能接近,能够在接收机复杂度和系统性能之间取得较好的权衡。Abstract: In cooperative networks with automatic repeat request (ARQ), the multiple-rounds retransmission is used to improve the reliability at the cost of additional delay. In this paper, a cross-layer cooperative strategy is investigated in ARQ cooperative networks for the case that relay nodes do not have ability of signal combining reception. A selection cooperation scheme is proposed, in which both the decoding set and the selection of best relay are changed dynamically during each round of retransmission. In Rayleigh fading channels, the proposed scheme can achieve optimal diversity-multiplexing-delay (D-M-D) tradeoff without increasing the receiver complexity of relay nodes. Simulation results show that the outage probability performance of the proposed scheme is close to that of the scheme using combining reception at relay nodes. It achieves better tradeoff between the receiver complexity and the system performance.
-
[1] LANEMAN J N, TSE D N C, WOMELL G W. Cooperative diversity in wireless networks:Efficient protocols and outage behavior[J]. IEEE Transactions on Information Theory, 2004, 50(12):3062-3080. doi: 10.1109/TIT.2004.838089 [2] XU Jie, DUAN Ling-jie, ZHANG Rui. Cost-aware green cellular networks with energy and communication cooperation[J]. IEEE Communications Magazine, 2015, 53(5):257-263. doi: 10.1109/MCOM.2015.7105673 [3] 肖海林, 王鹏, 聂在平, 等.基于遗传算法的多基站协作通信功率分配方案[J].电子科技大学学报, 2014, 43(1):26-30, 41. http://kns.cnki.net/KCMS/detail/detail.aspx?filename=dkdx201401005&dbname=CJFD&dbcode=CJFQ XIAO Hai-lin, WANG Peng, NIE Zai-ping, et al. Power allocation scheme based on genetic algorithm for multi-base station cooperative communication[J]. Journal of University of Electronic Science and Technology of China, 2014, 43(1):26-30, 41. http://kns.cnki.net/KCMS/detail/detail.aspx?filename=dkdx201401005&dbname=CJFD&dbcode=CJFQ [4] 田心记, 袁超伟, 王秋才, 等.基于相位旋转的坐标交织分布式空时码[J].电子与信息学报, 2011, 33(5):1131-1135. http://www.cnki.com.cn/Article/CJFDTOTAL-DZYX201105020.htm TIAN Xin-ji, YUAN Chao-wei, WANG Qiu-cai, et al. Co-ordinate interleaved distributed space-time code based on phase rotation[J]. Journal of Electronics & Information Technology, 2011, 33(5):1131-1135. http://www.cnki.com.cn/Article/CJFDTOTAL-DZYX201105020.htm [5] ABOU-RJEILY C. Distributed space-time codes for full-duplex IR-UWB amplify-and-forward cooperation[J]. IEEE Transactions on Wireless Communications, 2015, 14(4):2144-2155. doi: 10.1109/TWC.2014.2381264 [6] BLETSAS A, KHISTI A, REED D P, et al. A simple cooperative diversity method based on network path selection[J]. IEEE Journal on Selected Areas in communications, 2006, 24(3):659-672. doi: 10.1109/JSAC.2005.862417 [7] BERES E, ADVE R. Selection cooperation in multi-source cooperative networks[J]. IEEE Transactions on Wireless Communications, 2008, 7(1):118-127. doi: 10.1109/TWC.2008.060184 [8] LARSSON P, RASMUSSEN L K, SKOGLUND M. Throughput analysis of ARQ schemes in gaussian block fading channels[J]. IEEE Transactions on Communications, 2014, 62(7):2569-2588. doi: 10.1109/TCOMM.2014.2321383 [9] El GAMAL H, CAIRE G, DAMEN M O. The MIMO ARQ Channel:Diversity-multiplexing-delay tradeoff[J]. IEEE Transactions on Information Theory, 2006, 52(8):3601-3621. doi: 10.1109/TIT.2006.878173 [10] KAO J C. Performance analysis of relay-assisted network-coding ARQ with space-time cooperation in wireless relay networks[J]. IEEE Transactions on Wireless Communications, 2014, 13(8):4132-4145. doi: 10.1109/TWC.2014.2315793 [11] ABUZAINAB N, EPHREMIDES A. Energy efficiency of cooperative relaying over a wireless link[J]. IEEE Transactions on Wireless Communications, 2012, 11(6):2076-2083. doi: 10.1109/TWC.2012.404012.110242 [12] ZHANG Chao, ZHANG Jun, WEI Guo, et al. An efficient cooperative ARQ protocol for wireless relay networks[J]. Computer Communications, 2012, 36(1):105-112. doi: 10.1016/j.comcom.2012.08.002 [13] BAI Zhi-quan, YANG Xiang-lai, KWAK K. ARQ-based updatable relay selection for distributed EO-STBC cooperative system[J]. International Journal of Communication Systems, 2014, 27(1):116-125. doi: 10.1002/dac.v27.1 [14] LIU Kuang-hao, CHEN H H. Probabilistic relay selection for fast selection cooperation in half-duplex wireless networks[C]//Global Telecommunications Conference, 2009. Honolulu:IEEE, 2009:1-6. [15] WANG Heng, LI Min, LIN Jin-zhao, et al. Diversitymultiplexing-delay tradeoff in selection cooperation networks with ARQ[J]. IEEE Transactions on Communications, 2012, 60(6):1729-1740. doi: 10.1109/TCOMM.2012.040212.100459 [16] TANNIOUS R, NOSRATINIA A. Spectrally-efficient relay selection with limited feedback[J]. IEEE Journal on Selected Areas in Communications, 2008, 26(8):1419-1428. doi: 10.1109/JSAC.2008.081008