基于数字锁相环的晶振频率同步模块设计

为满足现代通信技术、雷达技术、电子测量以及光电应用领域对高稳定度高准确度时钟的要求,设计了一种基于数字锁相环的晶振同步系统.系统以基于FPGA数字延迟线的高分辨率鉴频鉴相器以及在MicroBlaze核中实现的卡尔曼数字环路滤波器为核心,通过16 bit DAC微调本地晶振振荡频率,使其同步于GPS秒脉冲,从而获得了高准...     

基于KALMAN滤波和深度学习的机器人飞拍方法研究

针对传统机器人飞拍系统存在视-控分离,导致系统实时性和效率不高的问题,提出视-控一体的设计方案,并对机器人飞拍的相关算法展开研究。为提高系统的实时性,操作系统采用基于Windows的硬实时系统,机器人控制模块和视觉模块在同一操作系统里通过共享内存的方式实现数据交互;基于改进型卡尔曼滤波算法,提出一种无传感器的精准时间触发算法,减小相机触发时机器人在同一位置的重复度误差;结合深度学习算法,提高图像处理的处理速度,保证在机器人下一节拍前完成纠偏值的计算,最终实现精准的机器人飞拍定位。实验表明:较之传统视控分离方案,该机器人飞拍纠偏系统具有简捷易用、精度和效率高等优势。     

基于H∞的捷联系统初始对准及参数γ选取方法研究

针对KALMAN滤波在系统模型不确定或噪声统计特性不准确的情况下滤波精度不高甚至发散的问题,引入了H∞滤波技术,介绍了H∞滤波基本理论和递推公式,与KALMAN滤波在本质上做了比较。建立了12维状态量的初始对准模型,在此基础上做了KALMAN滤波和H∞滤波在白噪声条件和有色噪声条件下的对比仿真,为了说明H∞滤波中参数γ的选取对对准速度和精度的影响,还分别选取3个有代表性的参数γ做了对比仿真,仿真结果表明：在白噪声条件下,两种滤波的精度相当,KALMAN滤波相对于H∞滤波要稍好些;在有色噪声条件下,KALMAN滤波显得有些发散,而H∞滤波明显对有色噪声的抗干扰性强于KALMAN滤波。同时通过对参数γ的适当选取,对系统的平均性能（精度）和鲁棒性能进行一定的平衡。     

A Novel High-Speed and High-Accuracy Manipulator of Planar Five-Link Structure: Modeling and Calibration

This paper describes a novel manipulator of planar five-link structure, which is designed to be characterized as high speed and high accuracy. The manipulator's mathematic models are built for purposes of analysis and control. Due to the lack of high-accuracy angular measurement equipment during manufacturing, the exact initial joint positions of the manipulator are unknown. An extended Kalman filter-based calibration algorithm is proposed to estimate those initial angles and a laser interferometer is adopted as the measurement equipment. To satisfy the high-speed requirement, a new Cartesian trajectory which combines the trapezoidal and S-shape trajectories is proposed. The testing results show that the requirements with respect to velocity, acceleration and accuracy are successfully achieved. Finally, an application example is given to demonstrate its potential uses.