The following algorithms have been proposed and tested in the thesis : 1 frequency selective fading : combine the isomorphism between the input space and the output space and propose a new approach to blind equalization of the channel . compared with conventional methods , the new approach offers lower computational complexity , better performance , and more robust against the over - determination of the system order ; 2 time selective fading : a new approach to the equalization of time selective channel based on the zero - forced equalizer is proposed which is more simple in its structure of algorithm ; 3 time - varying channel : using the instantaneous mean value changes of the output signal to extract the information of channel variations and model it using ar model , kalman filter is then employed to track channel variations , it bears faster ability in tracking the variation of tv channels ; based on the isomorphism between the inputs and the outputs and some of the approaches using in mimo system , a new algorithm of equalization of simo time - varying channel is proposed , which also share the merits of being robust against the over - determination of the system order ; model the time - varying channel using the multi - resolution decomposition wavelets , and then a blind identification method based " on the model is proposed ; at last , a new model for equalization and identification of mimo system is proposed 主要工作在以下几个方面： 1 、针对频率选择性衰落信道：结合输入输出空间同构关系提出一种新的频率选择性信道均衡方法，与传统方法相比，该方法计算量更小，收敛速度更快，性能更优，且对系统阶次的过确定表现稳健，具有实际均衡应用价值； 2 、针对时间选择性衰落信道：提出一种基于迫零均衡的时间选择性信道均衡方法，算法结构简单； 3 、针对时变色散信道：利用瞬态均值曲线提取信道时变信息，对之ar建模，利用卡尔曼滤波器跟踪时变信道抽头变化，可以快速跟踪信道变化；基于输入输出空间之间的同构关系以及多输入多输出系统的处理方法，提出了新的单输入多输出色散时变信道均衡与识别算法，同样具有对信道阶次过确定保持稳健的优点；结合小波多分辨率分析提出一种基于小波模型的信道盲识别算法；研究时变的多输入多输出系统的盲均衡与盲反卷积问题，给出一种时变系统处理模型。
The second part brings forward a new ar - model - based arm detection method based on features of arm and characteristics of velocity and acceleration of the radar echoes . by establishing two - order ar models , this new detection method computes the poles of models representing different targets , thus estimates the acceleration of the targets so as to determine the nature of the targets . furthermore , this method can not only be applied when the arm is being tracked by radar , but also be put into use in radar ’ s seeking and scanning times 第二部分则是根据反辐射导弹( arm )的特点及其雷达回波信号中速度、加速度等特征,提出了一种新的基于ar模型的arm检测算法,该方法通过建立二阶ar模型、估计代表不同目标的模型极点,并由此估算目标加速度来判断目标性质,该方法不仅可以在雷达已跟踪上arm后采用,而且适用于雷达搜索扫描时使用,仿真结果表明该方法具有分辨率高,对雷达脉冲重复频率( pulserepetitionfrequency , prf )及积累脉冲数要求不高的特点,在低的prf及少的积累脉冲下,利用该方法仍可有效地识别检测arm 。
This paper discuss a modeling and predicting means for nonlinear systems proceeding from nonlinear systems modeling and predicting theory , whch is based on drnn model . this means overcomes the fact that ar model is used only in linear systems , at the same time it connects itself with approximation theory symbolic statistics and conjugate gradient algorithm , and formulate a system of large watercrafts motion modeling and predicting which is based on drnn model , and simulate it 本论文从非线性系统建模与预报的理论及应用观点出发，系统地阐述了一类适用于非线性系统的建模预报方法? ?基于drnn模型的建模预报方法，克服了ar模型仅局限于线性的情况，同时结合逼近论、数理统计等知识，运用共轭梯度算法，提出并建立了基于对角回归神经网络的大型舰船运动建模预报系统，并进行了仿真。
For the dynamic process of ship rolling movement , this paper analyses its dynamic date with time series analysis method and brings up this system ' s the most excellent autoregressive model ( ar model ) according to least aic criterion ( akaile , information criterion ) . it reveals the regular pattern of ship rolling movement and forecasts the future value of roll angle and pitch angle , then transforms it to adjusting value of object and adjusting it according to appropriate control rules 对于船舶摇荡运动这一动态过程，采用时间序列分析的方法，建立系统的自回归模型( ar模型) ，并根据最小aic信息量判定准则保证建立的系统模型为最优化模型。利用参数模型的方式对船舶横摇、纵摇运动的动态数据进行分析处理，揭示船舶摇荡运动的规律，预测船舶横摇角、纵摇角的未来值。
The bic method generalized from ar model was adopted to determine the number of input neurons in grnn prediction model . the grnn was applied to single - step and multi - step ahead prediction of the vibration time series of a rotating machine , and its performance was compared with that of 3 - layers perceptrons network with error back propagation training algorithm ( bpnn ) . it is indicated that the grnn is more appropriate for prediction of time series than the bpnn , and the performance of grnn is qualified even with sparse sample data 研究了基于广义回归神经网络( grnn )的大型旋转机械振动状态预测，提出了应用bic准则确定grnn预测模型输入神经元数目的方法，将grnn用于大型机组振动峰?峰值时间序列的预测，与采用误差反向传播学习算法的三层前馈感知器网络( bpnn )的预测结果对比表明， grnn的预测性能优于bpnn ，而且，即使样本数据稀少，也能获得满意的预测结果。
This article discussed ar modeling method , capon method , music method and mn method four kinds of high resolution algorithms principle and has separately made the computer simulation in the ideal situation to the above four algorithms . through the comparison , the beam width of high resolution algorithm is much narrower than conventional beamforming method . then unifies the actual utilization , discussed the influence of time delay unbalance , time domain sampling , plane wave supposition and array element with scope and phase not identically to the beamforming and made the computer simulation 本文首先讨论了ar模型法、 capon法、 music法和mn法四种高分辨率算法的原理,在理想的情况下分别对上述四种算法作了计算机仿真,通过比较,高分辨率算法的束宽比常规波束形成法窄得多;然后结合实际运用时的情况,重点探讨了延时失配、时域抽样、平面波假设和基阵的幅相不一致对波束形成的影响并做了计算机仿真。
According to the needs of gps / sins integrated navigation algorithm , the error models of gps and sins are studied respectively . the autoregressive ( ar ) models and autoregressive moving average ( arma ) models of gps positioning error are established based on the analysis of the properties of static gps positioning error data . and the neural network method to determine the ar model parameters is given 根据gps / sins组合导航算法的需要，分别对gps和捷联系统的误差模型进行了研究，在对gps静态定位误差数据特性分析的基础上，建立了gps定位误差的自回归( ar )模型和自回归滑动平均和( arma )模型，并用神经网络方法确定了ar模型参数。
Relative to general ar model , the setar model was more suitable to describe the price movement during the manipulation . to help the practice of the guarding system , the author try to set up a discrimination model for insider trading and market manipulation , based on the samples of daily - average return and turn - over ratio during the time the equity changed . the criterion is to try to correctly find out the offense and the normal event 除了日均换手率之外的日均内容摘要收益率、日均收益率平方项、日均收益率与日均换手率的乘积项能提高对内幕交易、市场操纵的发生以及非内幕交易、非市场操纵事件的正确判别率，这些变量对内幕交易、市场操纵的判别影响不一样，并且，内幕交易和市场操纵判别模型中的最优阀值也不一样，最优阀值的选取可以体现监管层监管水平的成熟程度。
The robust system can prevent abnormal factors from entering the flood system , so as to ensure the stability of the system and the accuracy of flood forecast . firstly , the theory of robust estimation is introduced into parameter estimation of the auto - regressive model ( ar model ) . also some estimation methods commonly used , including the huber estimation and igg estimation , are introduced and compared with the least square method ( lsm ) 洪水预报系统的抗差性研究，就是把抗差理论引入洪水预报中，利用抗差系统具有的抗差能力，使许多严重的不正常因素误差影响，根本就不能进入系统，这样，减少了系统的污染机会，降低了不正常因素的影响，可大大提高系统的稳定性和洪水预报的精度。
