ID 原文 译文
1883 实验结果表明，当面向实际环境农作物病害识别时，本文方法在识别精度和鲁棒性上均优于其他方法。 Extensive experiment results demonstrate that the proposed HORPSF approach significantly out-performs other competing methods in terms of recognition accuracy and robustness, especially demonstrating superior per-formance when dealing with the real-world examples of crop-disease recognition.
1884 为平衡粒子群算法勘探与开发能力，本文提出混合均值中心反向学习粒子群优化算法。 In order to balance the exploration and exploitation of particle swarm optimization, this paper proposes ahybrid mean center opposition-based learning particle swarm optimization.
1885 算法将所有粒子和部分优质粒子分别构造的均值中心进行贪心选择，得出的混合均值中心将对粒子所在区域进行精细搜索。 The algorithm performs greedy selection on the mean center of all particles and some high-quality particles respectively, and the obtained hybrid mean center will search the region in detail where the particles are located.
1886 同时对混合均值中心进行反向学习，使粒子能探索更多新区域。 At the same time, the hybrid mean center is using opposition-based learning, so that the particles can explore more new regions.
1887 将本文算法与最新改进的粒子群算法、人工蜂群算法和差分算法在多种测试函数集上进行比较，实验结果验证了混合均值中心反向学习策略的有效性，算法的综合优化性能更强。 The proposed algorithm are compared with the latest improved particles warm optimization, artificial bee colony algorithm and difference algorithm in various test function sets, and the results veri-fy the effectiveness of the hybrid mean center opposition-based learning and the overall optimization performance of the al-gorithm is stronger.
1888 本文基于现代控制理论观点下的量测模型的能观测性分析，提出绝对时钟状态量测模型的能观测性问题。 Based on the observability analysis of the measurement model under the modern control theory, this paper proposes the observability problem of the absolute clock state measurement model.
1889 启发于量测状态向量空间与状态向量运算法则，转化为本质上向量空间的同构映射原理，建立了以基本量测单元(Basic Measurement Unit，BMU)构建最小均方误差(Minimum Mean Square Error，MMSE)等价变换下的能观测性解耦量测模型。 It is inspired by measured state vector space and state vector algorithm, and transformed into the isomorphic mapping principle of vector space essentially. It establi-shes the minimum mean square error (MMSE)observable decoupling measurement model under equivalent transformation with basic measurement unit (BMU).
1890 该方法揭示了在双向信息交换下对称量测性能的本质，在量测模型满足能观测性的必要条件下实现时钟状态追踪的 Kalman filtering 算法。 This method reveals the essence of symmetric measurement performance under two-way message exchange, and realizes Kalman filtering algorithm for clock state tracking under the necessary conditions that the measurement model satisfies the observability.
1891 本文算法不依赖于优化的初始点设置，初始点选择具有鲁棒性，并且对于网络连接性的变化具有稳健性。 The proposed algorithm does not depend on the optimized initial point set-ting, the initial point selection is robust, and it is also robust to changes in network connectivity.
1892 仿真结果表明，能观测性量测模型能实现规模化扩展，设计的算法具有局部和全局一致的 MMSE 量测性能，接近于贝叶斯 CRLB(Cramer-Rao Lower Bound)量测性能边界。 The simulation results show that the observable measurement model can achieve scale expansion, and the proposed algorithm has local and global uniformof MMSE measurement performance, which is close to the boundary of the Bayesian Cramer-Rao lower bound (CRLB)measurement performance.