MEMS加速度计混合误差标定补偿方案

针对微机电系统(MEMS)加速度计在实际使用过程中存在非正交零偏误差和温度漂移误差的问题,提出了一种混合误差标定补偿算法。算法通过分析加速度计温度与误差的关系,在不同温度区间下建立加速度计输出的误差模型,在每个温度区间采用十二位置校准法对加速度计的非正交零偏误差进行标定补偿,得到精确的零偏和刻度因子,同时采用最小二乘法拟合零偏和刻度因子与温度的一维关系函数,最终实现不同温度区间下的动态误差补偿。实验结果表明,本算法可使加速度计输出的精度提高1个数量级,补偿效果明显。     

A whispering gallery mode strain sensor based on microtube resonator

A novel whispering gallery mode (WGM) strain sensor based on microtube has been proposed, where perceiving strain variations are reported via the dynamical regulation of a whispering gallery mode. The WGMs in the microtube resonator were evanescently excited by a micro-nano fiber fabricated by the fusion taper technique. The structural changes of microtubes under axial strain were simulated with finite element software, and the effect of microtube wall thickness on strain sensitivity was systematically studied through experiments. The experimental results show that the strain sensitivity of thin-walled microtube is found to be 1.18 pm/με and the Q-factor in the order of 4.4×104. Due to its simple fabrication and easy manipulation as well as good sensing performance, the microtube strain sensor has potential applications in high-sensitivity optical sensing.     

基于因子图的自主导航多源异构信息融合算法

为了解决惯性导航系统长航时累计误差和实时性问题,提出了一种基于因子图的多源信息融合算法。论文基于和积算法,采用了因子图模型进行基于最大后验概率的状态估计,计算变量的均值和方差完成数据融合,利用超宽带技术和视觉传感器对发生累积误差的惯性导航系统进行校准,完成误差修正。对信息融合算法进行了仿真,结果为:该算法各轴向的误差分别为0.36%和0.31%,解算时间为50 ms。该算法误差仅为惯性导航累计误差的1/3,实时性相较于其它算法更高。     

A temperature-insensitive polarization filter and a neotype sensor based on a hybrid-circular-hole microstructured optical fiber

A temperature-insensitive polarization filter and a neotype sensor based on a hybrid-circular-hole microstructured optical fiber (MOF) are proposed. Numerical investigations demonstrate that the x polarized component of silica core mode can couple to the cladding mode in the researched wavelength, while the y polarized component would not. Furthermore, the resonant region can be controlled by changing the diameters or coordinates of the air holes, and the MOF has good performance on stability of temperature. Moreover, the hybrid-circular-hole structure is propitious to selectively integrate different functional materials. Two different materials are integrated into this proposed MOF, the application of the Sagnac interferometer in temperature sensing is studied, and two groups of dips would be observed in the transmission spectra, which have different temperature sensitivities. Therefore, the proposed MOF can be used as a flexible temperature-insensitive polarization filter or potentially applied to a two-parameter sensor.     

Design and characteristic research on side-polished fiber fluid sensing system

A novel fluid sensing system based on side-polished optical fiber (SPOF) is proposed, which realizes the fluid replaceability and effective refractive index (RI) sensing characteristics. Numerical investigations demonstrate that the photonic bandgap effect can be obtained if the RI of liquid is higher than that of substrate material in the wavelength range studied. The relationship between bandgap edge wavelength and RI is studied theoretically. The SPOF with a depth of 57 μm is used in the experiment to realize the construction of the fluid channel. After filling three different liquids, the result shows that the wavelength of the bandgap edge has a red shift with RI increased, which is nearly linear in the RI range of 1.56—1.6 with a sensitivity about 5 543.64 nm/RIU. The proposed sensing system can be flexibly applied to the field of fluid characteristic sensing such as biochemical solution characteristic detections.     

基于加速度时域-斜率特征的多向行走识别算法

为了解决行人多向行走模式识别中实时性差和识别精度不高的问题,提出了一种基于加速度时域-斜率特征的多向行走识别算法,通过计算和分析z轴加速度波峰或波谷时刻行人运动方向加速度斜率大小,准确识别行人前进、后退、左移、右移多个方向的行走模式。结果表明,本算法能实时识别行人每一步行走模式,平均识别精度在96%以上。     

基于MEMS传感器组合的行人室内高度定位算法

针对气压计易受环境温度影响的问题,该文提出了一种基于微机电系统(MEMS)加速度计和气压计融合的行人室内高度定位算法。该算法分别利用x轴加速度计的峰值特征和z轴加速度计的四分位距特征进行上、下楼判定;然后利用滤除掉气压突跳点的气压计数据进行高度解算;最后行人平走时,将解算高度修正到半层楼高的整数倍,减少高度误差。实验结果表明,该文提出的行人室内高度定位算法不易受环境温度的影响,解算高度与实际楼层高度的误差小于2 m,可将人员定位到正确楼层。     

一种基于AIKF的姿态测量算法

惯性测量单元中传感器具有较强的非线性和噪声的不确定性,导致使用常规卡尔曼滤波时误差大,容易出现发散,针对此问题,该文提出了一种改进的自适应增量卡尔曼滤波（AIKF）算法。该算法使用互补滤波将加速度计、磁力计和陀螺仪的数据进行融合,利用滤波后的数据增量作为卡尔曼滤波器的观测量,同时对系统噪声进行自适应在线估计,以获得精准的姿态输出。实验结果表明,该算法能够实现姿态的精准测量,摇摆台试验中俯仰角、横滚角误差小于0.05°,航向角误差小于0.15°,具有较好的噪声抑制能力。     

A novel high birefringence equal diameter circular-hole photonic crystal fiber

A novel equal diameter circular-hole photonic crystal fiber (PCF) with high birefringence is proposed and numerically analyzed by employing the finite-element method. The proposed PCF’s birefringence is 10-3, which can reach 2 orders higher than that of traditional high birefringence fiber, and this equal diameter circular-hole structure reduces the difficulty of the actual drawing process. The effect of different parameters on the birefringence of this PCF is investigated, and the application of the Sagnac interferometer based on fiber filling technology in temperature sensing is studied. The result shows that the high birefringence PCF can be used in both optical communication and optical sensing fields.     

基于MEMS惯性传感器时域特征的人体行为识别

提出了一种基于微机电系统(MEMS)惯性传感器组合系统的高精度实时人体行为识别算法。算法选取一个2 s的滑动时间窗作为特征提取窗口,提取惯性传感器组合系统输出的时域特征作为特征参量,采用基于平衡决策树的支持向量机对人体不同行为模式进行分类识别。在实验室自主研发的可穿戴智能终端平台上进行测试,结果表明,在识别时间缩短到2 s/次的条件下,对5种行走类行为模式和5种非行走类行为模式的识别率均可达88%。与现有算法相比,该算法的实时性和精度得到明显提高,且拓展了模式识别的种类。