ID 原文 译文
5824 实验结果表明,所提方法在高斯信道环境下的分类精度比现有方法要高,在Low-SNR的不同信道环境下有较高的识别率,且使模型在时间、相位和频率偏移量方面具有更好的鲁棒性。 The experimental results show that the proposed method in gaussian channel environment of higher classification accuracy than the existing methods, in the Low SNR has higher recognition rate under different channel environment, and make the model in terms of time, phase and frequency offset has better robustness.
5825 针对具有参数跳变的非线性系统,联合聚类算法和神经网络提出新的多模型自适应控制方法。 In view of the non-linear system with parameter jump, combined clustering algorithm and neural network is a new multiple model adaptive control method is proposed.
5826 首先对系统的输入输出数据进行模糊聚类,然后基于递推最小二乘法建立多个固定模型。 First of all, the system input and output data are fuzzy clustering, and then based on the recursive least squares method to establish multiple fixed models.
5827 为提高系统的暂态性能,同时建立两个自适应模型,并在此基础上设计鲁棒自适应控制器。 In order to improve the transient performance of the system, at the same time, two adaptive model is established, based on the robust adaptive controller is designed.
5828 此外,为了补偿系统的非线性部分,建立非线性预测模型,并设计非线性神经网络自适应控制器。 In addition, for the sake of the nonlinear part of the compensation system, establish a nonlinear prediction model, and nonlinear neural network adaptive controller is designed.
5829 所提方法可使控制切换系统具有稳定性保证。 The proposed method can make the control switch system has stability guarantee.
5830 最后,通过性能指标对控制器进行平滑切换。 Finally, a performance index to smooth switching controller.
5831 仿真结果表明,所提方法能够保证系统具有良好的控制性能。 The simulation results show that the proposed method can guarantee the system has good control performance.
5832 针对无人水面艇(unmanned surface vessel,USV)集群在路径规划中的协同避碰问题,提出了基于滚动优化策略结合粒子群优化算法的USV集群协同避碰方法。 For unmanned surface vehicle (unmanned surface vessel, USV) cluster in collaborative collision avoidance problem of path planning, based on the rolling optimization strategy is proposed combined with particle swarm optimization algorithm of USV cluster collision avoidance method together.
5833 首先,通过已有雷达、光电等传感器参数指标建立综合视域模型; First of all, through the existing radar and photoelectric sensor parameters to establish a comprehensive vision model;