水声双扩展信道空时Turbo通信系统

发布时间：2018-06-05 23:42

本文选题：信道不确实性 + 香农信道容量　；参考：《浙江大学》2014年博士论文

【摘要】：本文瞄准信道不确实性下的有效、可靠通信,以信息理论、信号处理和传播物理为支柱,构建通信问题中的最佳瞬态观察者,即逼近香农信道容量限的Turbo通信系统。结合传播物理,研究动态稀疏时延—多普勒双扩展信道的解卷估计问题,这属于欠定逆问题的求解。在此基础上,设计了双扩展信道估计与判决迭代均衡相协同的水声通信系统,提高了双扩展条件下的通信性能。由于空时阵可以通过复用与分集对信道容量、系统通信差错率等性能带来质的提升,因此从估计理论与信息理论相结合的角度,提出了作为最佳瞬变观察者的空-时-频广义旁瓣抵消器,实现了信道不确实条件下的有效、可靠通信。 Turbo通信系统是一个逼近香农信道容量限的最佳瞬态观察者。Turbo原理可以公式化为：采用迭代更新的外信息作为先验信息对符号和码进行序贯贝叶斯估计。Turbo原理实质上由：(1)状态—空间模型(2)新息过程(3)反馈三要素组成,因而是通信问题中的最佳瞬态观察者。为了研究Turbo通信系统的有效性,首先需要回答的是通信性能限的问题,即：在给定信道环境中最终能达到的可靠通信传输率是多少。香农信道容量定理表明,存在通信系统能达到信道容量限,且该容量限是紧致的,因此信道容量提供了一个标准来量化地衡量Turbo通信系统的性能。在本文所给出的例子中,Turbo通信系统仅与信道容量所预测的最小可靠通信信噪比相差1.4dB,体现了Turbo通信系统的有效性。 Turbo通信系统在双扩展信道下的具体运用,需要传播物理来推动。本文从波动方程出发,分析了稀疏时延—多普勒双扩展信道的物理成因,以波数积分法推导出波导环境中运动声源产生的声场物理表达式。由于深度分离格林函数仅在少数几个水平波数值处有较大的非零值,从而在物理源头就保证了信道在时延—多普勒维度上的稀疏性。时延—多普勒双扩展信道的解卷估计是一个欠定逆问题,常规的过定最小二乘方法失效,有无穷多个解。传播物理所导出的信道稀疏性这一物理先验,为约束优化方法开发利用并对解空间施以约束,使得欠定逆问题可解,其性能优于正向匹配求解逆问题的互模糊度函数方法,因为正问题方法没有利用该物理先验。由于海表面波和内波等因素的影响,时延—多普勒双扩展信道还具有动态特性,需要开发对信道的稀疏序贯估计理论。本文为了实现动态稀疏双扩展信道的序贯估计方法,将约束优化问题等价为求各约束集的交集问题。每个时刻的接收序列会产生一个新的约束集,并且以不断往约束集上投影的方法最终收敛于信道的真实解。在解决了动态稀疏时延—多普勒双扩展信道估计问题的基础上,设计了双扩展信道估计与判决迭代均衡相协同的水声通信系统。水声通信向来将信道的时延—多普勒双扩展及其变化视为其最大的挑战,传统基于多普勒频移补偿的相干通信技术并不能在时延—多普勒双扩展信道中实现可靠通信。第一个结合试图将双扩展信道视为机遇并由此获取处理增益。由此得到的新处理器可视作水声双扩展信道下的最佳瞬态观察者,在该信道中能够实现有效、可靠通信。从处理器结构上看,其均衡器部分在时—频两维上实现接收分集,可以可靠工作在时延—多普勒双扩展信道下；而均衡器与解码器之间的新息迭代又极大地降低系统的误符号率和误码率,使之逼近香农信道容量限,从而具有通信的有效性。在实现了第一个结合的基础上,本文进一步研究空时Turbo通信系统。首先要回答的问题是为什么要空时阵,为什么要与空时阵系统相结合。本文从多个角度回答了这一问题。首先推导了衰落信道下空时阵通信系统的通信差错概率为：P(error)～Gc·SNR-Gd (1)其中编码增益Gc提供的是乘性增益,而由空时阵通信系统所带来的分集增益Gd则体现在使得通信差错概率随信噪比上升而指数下降上,显然要提供比编码增益可观得多的性能改善。对本文所研究的动态时延—多普勒双扩展信道下的空时Turbo阵通信系统而言,推导得出Gd∝nrnt(M+1)L,即从发射、接收、多路径和频率四维分集来提高系统系能。因此从可靠性即对系统通信差错概率的控制上,发展空时Turbo阵通信系统将带来非常可观的好处。从香农信道容量的角度看来,空时阵通信系统可以通过复用使得信道容量随着有效信道数目呈线性比例增长；也可以通过分集使得衰落信道中断容量为一定时的中断概率指数下降。既然空时阵通信系统可以通过复用与分集对通信系统的有效性与可靠性带来质的提升,因此有必要实现双扩展信道下的空时Turbo通信系统。为了使最佳阵处理系统可靠地工作在双扩展信道下,将最佳阵处理器的等价形式—广义旁瓣抵消器推广至时延—多普勒双扩展信道,得到空—时—频广义旁瓣抵消器,通信系统通过空间—时延—频率分集与复用来极大地提高通信速率和可靠性。从估计理论与信息理论相结合的角度,提出双扩展信道下的空时Turbo阵通信系统。该处理器的特点可以归纳为：(1)将信道的时延—多普勒双扩展特性显式地考虑到处理器框架中,利用双扩展信道下的空—时—频广义旁瓣抵消器通过在空—时—频多维的相干分集来增强信号,抑制符号间干扰和同道干扰；由于同道干扰的问题得到了妥善的解决,因而可以通过信道的复用来极大提高通信系统的通信速率,保证了通信的有效性。(2)Turbo通信系统是逼近香农信道容量的最佳瞬态观察者。正是由于Turbo机制的引入,系统的稳定性通过嵌入在发送码字中的结构性知识来保证。(3)空—时—频广义旁瓣抵消器其阵驾驶向量是信道估计结果的函数,通过Turbo系统的迭代进行,信道估计的结果会随之改善,进一步提高阵处理的性能。综上所述,本文的创新点可以归纳为：(1)结合水声传播物理,分析研究了双扩展信道响应的稀疏性,提出了基于稀疏约束的双扩展信道序贯估计-跟踪方法,可有效应用于时变信道条件下的水声通信。(2)设计了双扩展信道估计与判决迭代均衡相协同的水声通信系统,提高了双扩展条件下的通信性能。(3)从估计理论与信息理论相结合的角度,提出了作为最佳瞬变观察者的空-时-频广义旁瓣抵消器,实现了信道不确实条件下的有效、可靠通信。本文主要章节按创新点分为三部分,第一部分详述了稀疏双扩展信道的物理成因,互模糊度函数信道估计方法与约束优化估计方法,动态时延—多普勒稀疏双扩展信道的序贯压缩重构方法,为后续章节打下信道估计前处理的基础；第二部分研究了通信问题中的最佳瞬态观察者,即Turbo通信系统,并通过与香农信道容量的比较分析了其有效性。得到了双扩展信道下的Turbo通信系统；第三部分研究了空时阵通信系统对通信有效性和可靠性的巨大提升,提出了能可靠工作在双扩展信道下的空—时—频广义旁瓣抵消器,得到本文最终的题目内容：水声双扩展信道下的空时Turbo阵通信系统。仿真和海试实验数据分析验证了系统的可靠性和有效性。
[Abstract]:In this paper, we aim at the effective and reliable communication under the inaccuracy of the channel, with information theory, signal processing and propagation physics as the backbone to construct the best transient observer in the communication problem, that is, the Turbo communication system approaching the Shannon channel capacity limit. This is a solution to the undetermined inverse problem. On this basis, a underwater acoustic communication system with double extended channel estimation and iterative equalization is designed to improve the communication performance under the condition of double expansion. Because the space time matrix can improve the performance of the channel capacity and the error rate of the system communication through multiplexing and diversity, the estimation of the performance of the system is improved. With the combination of theory and information theory, an empty time frequency generalized sidelobe canceller, which is the best transient observer, is proposed to achieve effective and reliable communication under the untrue channel condition.
The Turbo communication system is an optimal transient observer that approximating the Shannon channel capacity limit.Turbo principle can be formulated as: the sequential Bayesian estimation of symbols and codes by iterative and updated information as prior information is essentially composed of the (1) state space model (2) new interest process (3) feedback three elements. It is the best transient observer in the communication problem. In order to study the effectiveness of the Turbo communication system, the first thing to answer is the problem of the communication performance limit, that is, what is the reliable communication transmission rate that can eventually be achieved in a given channel environment. The Shannon channel capacity theorem shows that the communication system can reach the channel capacity limit and the capacity of the communication system. The limit is compact, so the channel capacity provides a standard to quantify the performance of the Turbo communication system. In the example presented in this paper, the Turbo communication system is only 1.4dB of the minimum reliable communication signal to noise ratio predicted by the channel capacity, which reflects the effectiveness of the Turbo communication system.
The specific application of the Turbo communication system in the double spread channel requires the propagation of physics. This paper analyzes the physical causes of the sparse delay Doppler double propagation channel from the wave equation, and derives the physical expression of the sound field generated by the moving sound source in the waveguide environment by the wave number integral method. Because the depth separation of the Green function is only less. Several horizontal wave values have large non zero values, thus ensuring the sparsity of the channel in the time delay - Doppler dimension at the physical source. The deconvolution estimation of the time delay - Doppler double expansion channel is an underdetermined inverse problem. The conventional overdetermined least squares method is invalid and there are infinitely many solutions. The channel sparsity derived from the propagation physics The physical prior is developed and utilized for constrained optimization methods and constrained to the solution space, which makes the undetermined inverse problem solvable, and its performance is better than the reciprocal ambiguity function method for the forward matching to solve the inverse problem, because the positive problem method does not use the physical prior. Due to the influence of the sea surface wave and internal wave, time delay Doppler double The extended channel also has dynamic characteristics and needs to develop the sparse sequential estimation theory for channels. In order to realize the sequential estimation of the dynamic sparse and double extended channel, the constrained optimization problem is equivalent to the intersection problem of each constraint set. The receiving sequence at each time will produce a new constraint set and to the constraint set. The projection method finally converges to the true solution of the channel.
On the basis of solving the problem of dynamic sparse delay Doppler double propagation channel estimation, a underwater acoustic communication system with dual extended channel estimation and iterative equalization is designed. Underwater acoustic communication has always considered the time delay of the channel - Doppler double expansion and its change as its biggest challenge, and the traditional phase based on Doppler frequency shift compensation The dry communication technology can not achieve reliable communication in the time delay Doppler double spread channel. The first combination tries to see the dual extended channel as an opportunity and thus obtain the processing gain. The resulting new processor can be considered as the best transient observer in the underwater acoustic dual propagation channel, which can be effective and reliable communication in this channel. In view of the structure, the equalizer part realizes the reception diversity on the time frequency two dimension. It can work reliably under the delay Doppler double expansion channel, while the new interest iteration between the equalizer and the decoder greatly reduces the system's false symbol rate and the bit error rate, making it close to the Shannon channel capacity limit and thus has the effectiveness of the communication.
On the basis of the first combination, this paper further studies space-time Turbo communication systems. The first question to answer is why the space-time matrix is needed and why it should be combined with the space-time matrix system. This paper answers this problem from many angles. Firstly, the error probability of the space time matrix communication system under the fading channel is P (ER ROR) ~ Gc. SNR-Gd (1) the encoding gain Gc provides multiplicative gain, and the diversity gain Gd brought by the space-time matrix communication system is reflected in the exponential decline of the probability of communication error with the signal to noise ratio, obviously providing a much greater performance improvement than the coding gain. The dynamic delay of this paper has been studied by Doppler double. For the space-time Turbo array communication system under the extended channel, the Gd nrnt (M+1) L is derived from the transmission, reception, multipath and frequency four dimension diversity to improve the system power. Therefore, the development of the space-time Turbo matrix communication system will bring a very considerable benefit from the reliability and the control of the error probability of the system communication. From the Shannon channel capacity The space time matrix communication system can increase the capacity of the channel with the number of effective channels in a linear proportion, and can also reduce the interruption probability index when the interruption capacity of the fading channel is certain by diversity. Since the space time matrix communication system can be effective and reliable by the reuse and diversity of the communication system. Because of the qualitative improvement, it is necessary to realize the space-time Turbo communication system under dual extended channels.
In order to make the best array processing system work reliably in the double spread channel, the equivalent form of the optimal array processor is extended to the delay Doppler double extension channel, and the space-time frequency generalized sidelobe canceller is obtained. The communication system improves the communication speed greatly through the spatial time delay frequency diversity and multiplexing. Rate and reliability. A space-time Turbo matrix communication system with double extended channels is proposed from the combination of estimation theory and information theory. The characteristics of this processor can be summed up as follows: (1) the time delay and Doppler double expansion characteristics of the channel are explicitly considered in the processor framework and the space-time frequency broad sidelobe under the dual extended channel is used. The canceller enhances the signal through the coherent diversity of the space-time frequency multidimensional diversity to suppress the inter symbol interference and the channel interference. Because the problem of the channel interference is properly solved, the communication rate of the communication system can be greatly improved through the multiplexing of the channel, and the effectiveness of the communication is guaranteed. (2) Turbo communication system is an approximation of the incense. The best transient observer of the capacity of the agricultural channel is due to the introduction of the Turbo mechanism. The stability of the system is guaranteed by the structural knowledge embedded in the transmission code. (3) the driving vector of the space-time generalized sidelobe is a function of the channel estimation result. The result of the channel estimation will follow the iteration of the Turbo system. The improvement will further improve the performance of the array processing.
To sum up, the innovative points of this paper can be summed up as follows: (1) the sparsity of dual propagation channel response is analyzed and studied in combination with underwater acoustic propagation physics. A double extended channel sequential estimation and tracking method based on sparse constraints is proposed, which can be effectively applied to underwater acoustic communication under the condition of time-varying channel. (2) a double extended channel estimation and decision iteration is designed. The underwater acoustic communication system with equilibrium phase synergy improves the communication performance under the condition of double expansion. (3) an empty time frequency generalized sidelobe canceller, which is the best transient observer, is proposed from the angle of combination of the estimation theory and information theory, and the effective and reliable communication under the untrue channel condition is realized.
The main section of this paper is divided into three parts. In the first part, the physical origin of the sparse double propagation channel, the channel estimation method and the constraint optimization estimation method of the cross ambiguity function, the sequential compression reconstruction method of the dynamic delay Doppler sparse double extension channel are discussed, and the foundation of the channel estimation is laid down for the later section. The two part studies the best transient observer in the communication problem, namely the Turbo communication system, and analyzes its effectiveness by comparing with the capacity of the Shannon channel. The Turbo communication system under the double expansion channel is obtained. The third part studies the great improvement of the communication effectiveness and reliability of the space time array communication system, and puts forward the reliable work. An air time frequency generalized sidelobe offset receiver under the double spread channel is used to obtain the final topic of this paper: the space time Turbo array communication system under the underwater acoustic dual expansion channel. The simulation and experimental data analysis of the sea test verify the reliability and effectiveness of the system.
【学位授予单位】：浙江大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TN929.3

【参考文献】