一种新的加权最小二乘支持向量回归机及其对性能退化航空发动机推力估计的应用(英文)

要实现航空发动机的性能退化缓解控制,需要设计一个高精度并具有良好实时性的推力估计器。针对这种需要,本文提出了一种新的加权最小二乘支持向量回归机,并在此基础上设计了性能退化推力估计器。和现有的加权策略相比较,基于此新的加权策略的推力估计器不仅能满足性能退化缓解控制对精度的要求也能满足实时性的要求。最后,仿真实验证明此加权最小二乘支持向量回归机以及基于此回归机设计的推力估计器的有效性和可行性。     

城市道路交通流检测数据精度评价

目前各种检测方法采集到的路段速度数据普遍存在一定的误差,其准确性直接影响各种交通分析模型结果的精度.针对浮动车检测器、微波检测器以及线圈检测器的路段速度数据进行研究,利用GPS(Global Positioning System)设备实时记录试验车位置,分别在快速路和主干路进行了12次实验,对不同采集方式获取的速度信息进行了精度验证和评价.实验结果分析表明,浮动车数据获取的城市快速路和主干路的路段行程速度平均误差分别为13.6%和27.8%,且采集值与实际速度具有相同的分布.快速路微波检测器和主干路线圈检测器的速度数据与真实行程速度数据的误差分别为30%和56%.表明了微波检测器的断面速度值,用来表征所处路段的行程速度时虽然具有一定的误差,但是基本能表征路段的行驶状况.该研究结论为不同精度需求的交通模型选择合适的数据源提供支持,并为利用多源数据的冗余信息提高数据精度和完备性提供了理论依据.      

A Dual-driven Intelligent Combination Control of Heat Pipe Space Cooling System

Effective thermal control systems are essential for reliable operation of spacecraft.A dual-driven intelligent combination control strategy is proposed to improve the temperate control and heat flux tracking effects.Both temperature regulation and heat flux tracking errors are employed to generate the final control action;their contributions are adaptively adjusted by a fuzzy fusing policy of control actions.To evaluate the control effects,describe a four-nodal mathematical model for analyzing the dynamic characteristics of the controlled heat pipe space cooling system(HP-SCS) consisting of an aluminum-ammonia heat pipe and a variable-emittance micro-electromechanical-system(MEMS) radiator.This dynamical model calculates the mass flow-rate and condensing pressure of the heat pipe working fluid directly from the systemic nodal temperatures,therefore,it is more suitable for control engineering applications.The closed-loop transient performances of four different control schemes have been numerically investigated.The results conclude that the proposed intelligent combination control scheme not only improves the thermal control effects but also benefits the safe operation of HP-SCS.     

未来航空发动机控制技术的发展趋势

发动机控制系统在航空发动机系统中占有非常重要的地位,其性能的优劣直接影响到发动机及飞机的性能。本文针对未来航空发动机发展对控制系统的要求,分析了航空发动机控制技术的发展趋势。     

基于迟滞鲁棒性的自动车队队列稳定性控制

基于车辆纵向迟滞动态模型,对自动车队队列稳定性以及控制方法进行分析研究.建立考虑迟滞因素的车辆纵向迟滞动态模型,在此模型的基础上建立基于固定时间间距策略的滑模控制器和比例积分微分(PID,Proportion Integration Differentiation)控制器,然后依据队列稳定性判断准则分别对两种控制器进行队列稳定性分析并分别得到相关的队列稳定性条件.定量比较分析和仿真比较分析结果显示,PID控制器具有更强的迟滞鲁棒性.     

Process-condition-driven Automatic Location of Components

In order to reduce the development cycle of mechanical products and corresponding toolings, a large number of components are utilized in these structures. Accordingly, automatic location method for components is a key issue in the research on intelligent computer aided design (CAD) techniques. Integrated with the investigation into intelligent design techniques for aircraft manufacturing toolings, in-depth researches on this issue are conducted. Formalized automatic location principles, representative methods and implementation algorithms are proposed and established. First, several terms and their formalized representative methods including process condition, location base and location operator are defined. Second, computation modeling of components automatic location and methods of location transformation are proposed. Then, definition of base structure and some key algorithms for automatic location are introduced. These principles and algorithms are developed, applied and verified in research project “Jig Intelligent CAD (JigICAD) System”.     

一种高精度温度控制的复合方法及其应用

介绍了一种用于高精度温度控制的复合方法及其在恒温槽上的成功应用.该方法集Fuzzy逻辑和专家式智能PID(比例、积分、微分)控制于一体,充分运用了专家的人工调节经验和PID控制的定量调节特性,较好地满足了高精度控温中的高精度与快速性要求.采用了简单新颖的PID参数在线修正和温度变化预报方法,易于实现.用于恒温槽时控制精度优于0.01℃.实践证明该方法在高精度控温中具有良好的控制效果.      

自主虚拟人智能驾驶行为模型的研究和实现

描述了自主虚拟人的分层驾驶行为及对虚拟世界的感知.在对自主虚拟人的驾驶行为进行研究的基础上,结合导向行为与专家经验,建立了基于产生式规则的智能行为模型,并给出了规则的具体实现.通过虚拟现实技术对该智能模型进行了验证,证明了其可行性.      

航空发动机的智能神经网络自适应控制研究

针对结构复杂、模型不确定、强非线性的航空发动机对象,提出一种综合模糊推理、神经网络自适应和PID简单控制各自优点的控制方案.在改进模糊PID控制器的基础上,进行了新型智能型神经网络控制器的设计,并提出离线混沌蚁群优化与在线误差反传调整相结合的优化方法.应用具有良好泛化能力的最小二乘支持向量机进行系统辨识,对某型航空发动机进行了设计点处的线性和非线性模型控制仿真.结果表明:控制系统具有满意的动、静态性能和较好的鲁棒性,验证了该方案的可行性和有效性.      

Shadow detection for mobile robots: Features,evaluation, and datasets

ABSTRACT

Shadows have long been a challenging topic for computer vision. This challenge is made even harder when we assume that the camera is moving, as many existing shadow detection techniques require the creation and maintenance of a background model. This article explores the problem of shadow modelling from a moving viewpoint (assumed to be a robotic platform) through comparing shadow-variant and shadow-invariant image features — primarily color, texture and edge-based features. These features are then embedded in a segmentation pipeline that provides predictions on shadow status, using minimal temporal context. We also release a public dataset of shadow-related image sequences, to help other researchers further develop shadow detection methods and to enable benchmarking of techniques.