An iterative calibration method for nonlinear coefficients of marine triaxial accelerometers

Single-axis rotation technique is often used in the marine laser inertial navigation system so as to modulate the constant biases of non-axial gyroscopes and accelerometers to attain better navigation performance. However, two significant accelerometer nonlinear errors need to be attacked to improve the modulation effect. Firstly, the asymmetry scale factor inaccuracy enlarges the errors of frequent zero-cross oscillating specific force measured by non-axial accelerometers. Secondly, the traditional linear model of accelerometers can hardly measure the continued or intermittent acceleration accurately. These two nonlinear errors degrade the high-precision specific force measurement and the calibration of nonlinear coefficients because triaxial accelerometers is urgent for the marine navigation. Based on the digital signal sampling property, the square coefficients and cross-coupling coefficients of accelerometers are considered. Meanwhile, the asymmetry scale factors are considered in the I-F conversion unit. Thus, a nonlinear model of specific force measurement is established compared to the linear model. Based on the three-axis turntable, the triaxial gyroscopes are utilized to measure the specific force observation for triaxial accelerometers. Considering the nonlinear combination, the standard calibration parameters and asymmetry factors are separately estimated by a two-step iterative identification procedure. Besides, an efficient specific force calculation model is approximately derived to reduce the real-time computation cost. Simulation results illustrate the sufficient estimation accuracy of nonlinear coefficients. The experiments demonstrate that the nonlinear model shows much higher accuracy than the linear model in both the gravimetry and sway navigation validations.     

基于改进克隆卡尔曼滤波的视觉惯性里程计实现

视觉-惯性多传感器融合方案可以有效解决GNSS据止环境下的导航问题。为了更有效地处理视觉的相对观测信息,文中研究并实现了一种基于克隆卡尔曼滤波(stochastic cloning Kalman Filter,SCKF)与滑动窗方法相结合的低成本视觉惯性里程计紧耦合算法。开源数据集实验表明,文中算法精度超过了多状态约束卡尔曼滤波(MSCKF)开源代码的效果。     

惯性测量组合零偏的一种现场标定方法

高精度的惯性测量组合(IMU)中,现场标定零偏十分必要.基于输入输出模相等的原则,推导出一个IMU零偏的多位置现场标定方法,该方法是无偏的,且对重力加速度误差不敏感.证明了三个标定位置定理,指出IMU零偏的现场标定至少需要四个位置且IMU测量矢量的顶端不能在同一个平面内,给出了四位置标定精度最高的条件.仿真证明了标定方法的可行性和标定位置定理的正确性.     

加速度计的多位置标定精度分析与比较

在阐述加速度计组件的多位置标定方法原理的基础上,对转台误差进行了分析和建模,推导了误差的传播途径,理论分析了转台误差对多位置标定精度的影响。分析结果表明:转台外框的非正交误差主要对12位置和24位置标定精度产生影响,不水平误差主要对6位置标定精度产生影响,中框的非正交误差及转台坐标系与标定坐标系不重合误差对每种多位置标定方法的精度都存在影响,影响量级为误差量级大小。最后通过仿真验证了结论。     

Matlab在倒摆系统设计中的应用

倒摆系统是一个复杂的机械系统,该文从倒摆的状态空间模型出发,借助Matlab工具,采用LQR（线性二次型最优控制）方法实现倒摆控制系统的设计。     

惯测组合高低温环境简易标定方法

为简化惯性测量组合(IMU)高低温标定过程,基于输入输出物理量模相等的原理,提出了一种高低温环境IMU简易标定方法,并给出了最佳标定编排。高低温环境下对一套激光陀螺捷联惯导系统进行标定试验,结果表明:与传统标定法相比,简易标定法精度满足导航级惯性导航系统的应用要求。试验证实了简易标定法的有效性。     

基于单应矩阵的相对位姿改进算法

从单应矩阵恢复相对位姿在视觉导航、视觉伺服应用中具有重要价值,证明了单应矩阵的新性质,并利用该性质改进了基于单应矩阵分解的相对位姿估计算法。与已有算法相比,该算法的候选解个数减半,并扩大了适用范围。理论分析、合成数据和真实数据测试均表明,改进方法使从候选解中筛选出唯一解的运算时间减少了50%,提高了位姿估计的总体运算效率。     

基于总体最小二乘的捷联三轴磁力仪标定与地磁场测量误差补偿

提出了一种磁力仪标定与地磁场补偿一体化的方法。对捷联三轴磁力仪的测量过程进行误差分析,建立包含磁力仪标定误差与地磁场测量误差的参数化模型。针对测量噪声的分布特性,采用基于总体最小二乘（TLS）算法对模型的参数进行了估计。仿真结果与实验结果均表明,与传统的基于最小二乘（LS）方法相比,基于TLS算法的参数估计精度更高,能够获得更好的补偿效果。     

人工智能中A*算法的局部改进及其实现

本文对人工智能中的经典启发式搜索算法—A     

Research on SINS Alignment Algorithm Based on FIR Filters

An inertial frame based alignment (IFBA) method is presented, especially for the applications on a rocking platform, e.g., marine applications. Defining the initial body frame as the inertial frame, the IFBA method achieves the alignment by virtue of a c