Dual-steering mode based on direct yaw moment control for multi-wheel hub motor driven vehicles:Theoretical design and experimental assessment

One of the main challenges for multi-wheel hub motor driven vehicles is the coordination of individual drivetrains to improve mobility and stability in the steering process.This paper proposes a dual-steering mode based on direct yaw moment control for enhancing vehicle steering ability in complex environ-ments.The control system is designed as a hierarchical structure,with a yaw moment decision layer and a driving force distribution layer.In the higher-level layer,the objective optimization function is con-structed to obtain the slip steering ratio,which represents the degree of vehicle slip steering in the dual-steering mode.A yaw moment controller using active disturbance rejection control theory is designed for continuous yaw rate control.When the actual yaw rate of the vehicle deviates from the reference yaw rate obtained by the vehicle reference model and the slip steering ratio,the yaw moment controller is actuated to determine the yaw moment demand for vehicle steering.In the lower-level layer,there is a torque distribution controller based on distribution rules,which meets the requirement of yaw moment demand without affecting the total longitudinal driving force of the vehicle.For verifying the validity and feasibility of the dual-steering mode,simulations were conducted on the hardware-in-loop real-time simulation platform.Additionally,corresponding real vehicle tests were carried out on an eight-wheel prototype vehicle.Test results were generally consistent with the simulation results,thereby demon-strating that the proposed dual-steering mode reduces steering radius and enhances the steering per-formance of the vehicle.     

基于波束域MUSIC方法的高分辨方位估计

现代声纳系统普遍采用水听器基阵和一定的信号处理来提高对目标的检测和定位能力,而基阵的波束形成则在其中起着核心作用。文中研究了窄带波束域高分辨方位估计技术,分析了波束域MUSIC方位估计的构造过程和具体实现方法。仿真计算表明,基于波束域MUSIC的方位估计算法是一种分辨空间小角域内多个目标的有效方法。     

无人机虚拟维修训练系统及其HLA设计

以某型无人机虚拟维修训练系统的设计为实例,采用Creator对联邦实体建模,结合动态视景仿真软件Vega Prim e进行环境与目标的驱动,基于高层体系结构HLA对系统的分布式网络互联进行仿真设计,无人机虚拟维修训练系统有效地改善了目前无人机结合实装进行操作训练的局限性。     

现代文本分类技术研究

分析了当前的文本分类技术,总结了文本分类需要解决的问题;在详细比较了各种不同的分类算法以及文本模型以后,为进一步开展优化以及提高文本分类技术的研究工作提供了一些参考和建议。     

基于特征结构配置的车辆主动悬架控制器设计

基于特征结构配置参数化方法,提出了车辆主动悬架控制器设计方法,其目的是设计一组状态反馈控制器,使得闭环结构系统具有希望极点和特征向量。该方法提供的自由参数,可用来满足系统的鲁棒性能等指标。该方法直接基于车辆悬架系统的参数矩阵,故便于工程应用。对车辆悬架系统进行仿真分析,结果表明该设计方法简单有效。     

PMESI：一种优化进程私有数据访问的缓存一致性协议

并行应用程序中绝大部分的访存是对私有数据的访问,在cache一致性协议上不会产生冲突。传统一致性协议没有根据程序私有数据的访问模式进行针对性设计,存在着很大的优化空间。针对以上的问题,提出了一种支持私有状态的cache一致性协议PMESI,通过动态关闭和激活内存空间的cache一致性目录,优化私有内存空间的访问延迟和功耗。通过时钟精确模拟器的测试,PMESI协议优化了程序中54%的访存,并行程序的执行时间平均缩短了9%。     

高超声速飞行器倾侧转弯耦合控制策略

针对高超声速飞行器倾侧转弯过程中俯仰、偏航和滚动通道间的强烈耦合,提出一种耦合控制策略。针对高超声速飞行器快时变、非线性和强不确定性的控制问题,基于解析形式的非线性最优预测控制方法,采用分层设计思想设计了飞行器姿态控制系统,可较好满足高超声速飞行器的快速性要求;在分析了倾侧转弯飞行控制过程的主要影响因素及其影响规律的基础上,提出一种"降低攻角—快速滚转—拉起攻角"的耦合控制策略。对该控制策略对于高超声速飞行器的适用性进行了仿真分析,结果表明:所提耦合控制策略有效降低了偏航通道的控制需求,降低了倾侧转弯控制过程的失控风险,提高了控制系统的可靠性。     

Transformations of node‐balanced routing problems

This article describes a polynomial transformation for a class of unit‐demand vehicle routing problems, named node‐balanced routing problems (BRP), where the number of nodes on each route is restricted to be in an interval such that the workload across the routes is balanced. The transformation is general in that it can be applied to single or multiple depot, homogeneous or heterogeneous fleet BRPs, and any combination thereof. At the heart of the procedure lies transforming the BRP into a generalized traveling salesman problem (TSP), which can then be transformed into a TSP. The transformed graph exhibits special properties which can be exploited to significantly reduce the number of arcs, and used to construct a formulation for the resulting TSP that amounts to no more than that of a constrained assignment problem. Computational results on a number of instances are presented. © 2015 Wiley Periodicals, Inc. Naval Research Logistics 62: 370–387, 2015     

不同时间信息条件下多目轨迹交会测量点目标运动的方法

针对多相机对点目标的运动轨迹测量,提出不同时间信息条件下的多目轨迹交会法。对点目标运动轨迹进行时间多项式参数化描述,再将多个相机观察目标的系列视线与目标的参数化运动轨迹进行交会,通过最小化物方残差,确定出目标的运动轨迹。相对于传统的多目交会测量模式,本方法不仅能够有效地提高测量精度,而且能够在多相机之间观测不同步,或者时间未对准,甚至无时间信息的情况下仍得到目标的运动轨迹参数。相对于单相机测量模式时要求相机必须运动,多目情况下不要求相机自身运动。仿真实验和真实实验验证了该方法的有效性和高精度。     

临近空间飞艇内部自然对流的流场特征仿真研究

针对临近空间飞艇,对其内部氦气的自然对流特征进行了数值模拟。首先采用计算流体力学方法,并利用自编的用户自定义函数将外部的温度边界条件导入壁面网格。进而,在不同的稳态条件下,通过对内部氦气压力、温度、速度的分布等流场特征参数的研究,分析了临近空间飞艇内部气体自然对流的运动特性及其影响规律,并对自然对流非稳态变化过程进行了初步的探索。仿真结果表明,在临近空间环境下,飞艇内部氦气的自然对流,对于内部氦气自身的热交换具有一定程度的促进作用,而对蒙皮受力和结构安全性影响很小。