ID 原文 译文
55878 针对循环前缀(cyclic prefix,CP)编码单载波(single carrier,SC)空间复用多输入多输出(multipleinput multiple-output,MIMO)系统,基于变分推理,本文提出一种新的低复杂度压扩变换Turbo频域均衡(Turbo frequency domain equalization,TFDE)算法。 For cyclic prefix (cyclic prefix, CP) encoding single carrier (single carrier, SC) spatial multiplexing multiple input multiple output (multipleinput multiple - output, MIMO) system, based on the variational inference, this paper proposes a new low complexity Turbo pressure expansion transformation frequency domain equalization (Turbo frequency domain equalization and TFDE) algorithm.
55879 相对于传统的TFDE算法,文中算法有两点明显不同,一是采用解码器压扩后验信息,计算软输入软输出(soft-in soft-out,SISO)频域均衡器(frequency domain equalizer,FDE)的数据先验均值和方差矩阵,二是采用SISO FDE外部数据均值、方差以及解码器压扩外信息,计算解码器先验信息。 Compared with traditional TFDE algorithm, the algorithm has two obvious different, one is the decoder pressure expansion posterior information, calculate soft input soft output (soft - in soft - out, SISO) frequency domain equalizer (frequency domain equalizer, FDE) prior mean and variance of data matrix, the second is using SISO FDE external data mean value, variance and decoder pressure expansion outside information, calculate the a priori knowledge about the decoder.
55880 分析和仿真结果表明,这种新的TFDE算法在误码率(bit error rate,BER)性能、迭代收敛速度和总体计算复杂度方面均优于传统的MIMO TFDE算法。 Analysis and simulation results show that the new algorithm of TFDE the BER (bit error rate and BER) performance, overall iterative convergence speed and computational complexity is superior to the traditional MIMO TFDE algorithm.
55881 针对非线性系统中不可观测故障参数估计问题,提出基于多重渐消因子强跟踪平方根容积卡尔曼滤波(multiple fading factors strong tracking square-root cubature Kalman filter,MSTSCKF)的状态和参数联合滤波算法。 Aiming at an observation fault parameter estimation problem in nonlinear systems, based on multiple fading factor strong tracking Kalman filter square root volume (multiple fading factors strong tracking the square - root cubature Kalman filter, MSTSCKF) states and parameters of the combined filter algorithm.
55882 MSTSCKF基于强跟踪滤波器理论框架,通过引入多重渐消因子实时调整增益矩阵,克服平方根容积卡尔曼滤波(square-root cubature Kalman filter,SCKF)在故障参数变化函数未知或者突变时滤波精度下降甚至发散的缺点,并兼具SCKF在非线性拟合精度和数值稳定性等方面的优点。仿真结果表明,相比SCKF和强跟踪无迹卡尔曼滤波(unscented Kalman filter,UKF),本文提出的方法具有更高的估计精度。 MSTSCKF theoretical framework based on strong tracking filter, real-time adjustment of gain matrix by introducing multiple fading factor, Kalman filtering overcome square root volume (the square - root cubature Kalman filter, SCKF) at the time of fault parameters change function or unknown mutations filtering precision falling even divergent faults, and both SCKF in nonlinear fitting accuracy and the advantages of numerical stability, etc. Compared the simulation results show that SCKF and strong tracking no trace Kalman filtering (unscented Kalman filter, UKF), the proposed method has higher estimation accuracy.
55883 针对传统盲分离算法对宽带信号不适用的问题,提出了一种基于阵列接收模型的宽带盲源分离算法。 In view of the traditional blind source separation algorithm of broadband signals shall not apply to the problem, puts forward a model based on array receiving broadband blind source separation algorithm.
55884 该算法以子带分解的方法实现了瞬时复值盲分离方法在宽带情形下的扩展。 The algorithm to subband decomposition method to realize the complex blind source separation method in the situation of broadband instantaneous extension.
55885 针对扩展过程中可能出现的子带间次序模糊及子带内幅度模糊的问题,利用阵列接收情况下分离矩阵与混合矩阵的特点,提出了一种基于波达方向(direction of arrival,DOA)估计的次序调整及幅度去模糊方法。 For extension may occur in the process of subband order between fuzzy and subband within the range of problems, using the array receiving cases separation matrix and the characteristics of the mixing matrix, this paper proposes a based on DOA (direction of concatenated, DOA) to estimate the order of the adjustment and amplitude fuzzy method.
55886 仿真结果表明,该算法不仅能有效地分离宽带信号,而且可准确地恢复出信号幅度。 The simulation results show that the proposed algorithm can not only effectively separate broadband signals, and can accurately recover the signal amplitude.
55887 针对认知移动自组网中认知用户(cognitive user,CU)移动性和主用户(primary user,PU)优先通信导致网络不稳定,基于改进的移动模型,运用概率原理预测链路平均保持时间,并给出邻域拓扑保持时间的计算公式。 Cognitive mobile ad-hoc network of cognitive users against (cognitive user, CU) mobility and primary user (primary user, PU) preferred communication network is not stable, based on the improved mobile model, using the principle of probability prediction link average stay time, and neighborhood topology is given to keep the calculation formula of the time.