平面刀具加工面齿轮的理论分析

采用平面刀具加工面齿轮可提高刀具的通用性和降低设计制造成本.首先在对格林森方法仿真的基础上指出仅用高阶滚比是不能获得理想齿面的.然后提出了附加运动，通过这个附加运动与高阶滚比刀具在运动中可以实现虚拟渐开线的精确模拟.其三提出了确定高阶滚比、附加运动多项式系数的方法.其四建立了平面刀具的产形面和平面刀具加工的面齿轮的数学模型.最后进行了滚比与运动规律、轮齿接触分析等数值仿真，结果表明:算例中的最大齿面误差为-1.05mm，接触路径倾斜，接触椭圆长度为11.2mm，传动误差约为0″，被加工面齿轮与标准小轮的啮合表现出良好的“准共轭”特性.因此这种可粗切，亦可精磨的加工方法具备良好的可行性与实用性.      

基于数值微分计算的SERF原子自旋惯性测量动态仿真

基于量子调控技术的SERF原子自旋惯性测量仪表与传统惯性测量仪表相比,具有精度高、无活动部件、对加速度不敏感等优点,被公认是下一代高精度惯性测量的发展方向。分析了SERF原子自旋惯性装置Bloch微分方程组模型下磁场、激光、角速度输入对碱金属电子极化率和惰性气体核子极化率的影响。针对实验条件下原子系综Bloch微分方程组难以解析求解的问题,基于可变阶次数值微分计算,建立了相应的Simulink模型对实验条件下的原子系综进行数值求解,并分析了极化率对光场、磁场、角速度输入的瞬态响应与稳态响应。仿真结果与理论计算结果相吻合,验证了Bloch方程数值求解的可行性与准确性。该方法可得到惯性测量系统的动态响应,可用于模拟复杂输入下的SERF原子自旋惯性测量系统的输出响应。     

分数阶Cohen-Grossberg神经网络的Mittag-Leffler稳定性

在整数阶 Cohen-Grossberg神经网络与分数阶理论及分数阶神经网络的基础上,提出了分数阶 Cohen-Grossberg神经网络。为了研究该类型神经网络,引入 Mittag-Leffler函数并利用 Mittag-Leffler函数及分数阶导数的相关性质,进而通过构造 Lyapunov函数的方法,研究了分数阶 Cohen-Grossberg神经网络的 Mittag-Leffler稳定性,并最终给出了相应的充分性条件。最后,通过实例仿真验证了结论的正确性。     

Fractional order adaptive robust formation control of multiple quad-rotor UAVs with parametric uncertainties and wind disturbances

In recent times, multiple Unmanned Aerial Vehicles (UAVs) are being widely utilized in several areas of applications such as agriculture, surveillance, disaster management, search and rescue operations. Degree of robustness of applied control schemes determines how accurate a swarm of UAVs accomplish group tasks. Formation and trajectory tracking controllers are required for the swarm of multiple UAVs. Factors like external environmental effects, parametric uncertainties and wind gusts make the controller design process as a challenging task. This article proposes fractional order formation and trajectory tacking controllers for multiple quad-rotors using Super Twisting Sliding Mode Control (STSMC) technique. To compensate the effects of the disturbances due to parametric uncertainties and wind gusts, Lyapunov function based adaptive controllers are formulated. Moreover, Lyapunov theorem is used to guarantee the stability of the proposed controllers. Three types of controllers, namely fixed gain STSMC and fractional order Adaptive Super Twisting Sliding Mode Control (ASTSMC) methods are tested for the swarm of UAVs by performing the numerical simulations in MATLAB/Simulink environment. From the presented results, it is verified that in presence of wind disturbances and parametric uncertainties, the proposed fractional order ASTSMC technique showed improved robustness as compared to the fixed gain STSMC and integer order ASTSMC.     

表贴式永磁直线同步电机控制系统参数自动整定方法

针对表贴式永磁直线同步电机矢量控制系统参数设置需反复测量和调试的问题,提出一种自动整定的方法。测量电流控制器输出波形并计算相应的沃尔什一阶系数,反复迭代反电动势常数和定子电感,直至沃尔什一阶系数收敛到零,自动辨识表贴式永磁直线同步电机的电气参数；随后让电机自由运行辨识出机械参数；最后计算出速度和电流控制器的PI参数。样机测试结果验证了提出方法的有效性。     

Performance analysis of the vectorized high order expansions method in the accurate landing problem of reusable boosters

This paper addresses the design of the terminal phase maneuver of the booster landing in the presence of initial deviations and uncertainties. The primary purpose of this study is to evaluate the effectiveness of the vectorized high order expansions method to extract commands and near-optimal trajectory, while the accurate satisfaction of the terminal conditions is of great importance. After reviewing high order expansions methods, the Vectorized High Order Expansions method is briefly discussed. Then, the solution for the booster landing problem is extracted up to the third order. To study the solution quality in the presence of uncertainties, mass, density, and gravity are assumed to be constant in extracting the 3rd order solution. The assumptions however are eliminated in the simulations, and the effectiveness of high order expansions solution is studied subjectively and numerically in the presence of initial deviations and uncertainties. After implementing simulations and considering different assumptions, it will be shown that the 3rd order solution holds a desirable quality in accurately satisfying final conditions in presence of uncertainties and initial deviations and provides far better results than the 1st order solution. Moreover, in this paper, a novel method is proposed for online updating, and the performance and effect of this modification are studied using simulations. Reviewing the results of the 3rd order solution combined with the online update strategy, it becomes clear that the implementation of the online update significantly reduces landing point errors and improves the performance of the high order method. Furthermore, the study compares the performance of this strategy with the well-known state-dependent Riccati equation method to better evaluate the effectiveness of the proposed approach.     

基于二维紧凑型四阶格子模型的激波管流动模拟

为了提高格子的稳定性，使用Hermite展开方法，构建了新的二维四阶紧凑型格子模型，即D2Q37A。比较了D2Q37A和与Philippi给出的紧凑型格子模型(D2Q37B)的稳定性。在相同的碰撞频率下，与D2Q37B相比，D2Q37A可以模拟初始密度比更高的一维激波管流动。这表明D2Q37A与现有格子模型相比，具有更好的稳定性。详细给出了适用于高阶格子模型的边界条件实现方式。此边界条件实现方式保留了体现LBM（lattice Boltmann method）粒子特性的迁移 碰撞机制。用以上给出的格子模型和边界条件处理方式模拟激波管流动，得到的模拟结果和解析解吻合得很好。这表明所给出的边界处理方式是可行的。此边界格式同样可以用于其他类型的流动和边界。      

基于短时分数阶傅里叶变换的 含噪信号特征提取

通过对比噪声背景下短时傅里叶变换和短时分数阶傅里叶变换的时频分辨能力,进一步明确了短时分数阶傅里叶变换对频率时变信号的独特优势。研究了短时分数阶傅里叶变换的窗口滑动方式对瞬时频率估计的影响,可以发现:首先,在调频率较大或是变化剧烈时,短时分数阶傅里叶变换具有更好的鲁棒性;其次,在基于短时分数阶傅里叶变换的瞬时频率估计,实际应用中应主要采用以第一采样时刻作为估计值的逐点滑动方式。