飞控系统传感器数据融合技术研究

【作者】 王军；

【导师】 吴成富；

【作者基本信息】 西北工业大学 ， 导航、制导与控制， 2007， 硕士

【摘要】 本文以提高无人机在各种状态下的稳定性和机动性、改善飞行品质为目标，研究了无人机控制增稳系统所需要的两个重要的反馈信号—迎角和侧滑角的三种获取方法。三种方法分别是非线性观测器方法，几何关系法，以及多传感器数据融合估计方法。 主要研究工作包括： 1．建立了无人机的非线性模型。给出了无人机系统中的几个仿真模块的数学模型，并给出了飞行控制律。 2．利用非线性观测器方法来估计无人机状态参数。对该方法的原理进行了分析研究。利用Simulink中的NCD Blockset模块进行观测器增益矩阵优化设计。运用Matlab／Simulink进行了离线设计非线性观测器，在线实时估计无人机状态参数的仿真验证。 3．利用几何关系法估计无人机的状态参数。先对INS／GPS组合导航系统的原理进行了介绍，然后研究了将组合导航系统用于几何关系法估计无人机状态参数的原理，最后进行了实时估计无人机状态参数的仿真验证。 4．提出了利用多传感器数据融合的方法估计无人机的状态参数。先介绍了数据融合的理论及方法，接着研究了利用多传感器数据融合方法来估计迎角和侧滑角的原理和方法，最后通过仿真进行了验证。 仿真结果表明，三种估计方法都可以达到要求的精度。几何关系法和多传感器数据融合估计方法都可以直接应用于实时估计无人机的迎角和侧滑角。本文提出的利用多传感器数据融合方法估计迎角和侧滑角具有一定的创新性，采用该方法可以充分发挥已有的传感器的功用，避免研制价值昂贵的高精度复杂传感器，具有较大的工程应用价值。更多还原

【Abstract】 In order to enhance stabilization, improve flight quality and maneuverability of the UAV, three methods for gaining two important feedback signals—angle-of-attack and sideslip angle that are needed in UAV control and stability augmentation system are studied in this dissertation. They are nonlinear observer, geometry relation, and multisensor data fusion estimation. Major research is listed as follows:1. The UAV’s nonlinear model is built. Several mathematic models of simulation modules of UAV and the flight control laws are given.2. UAV state parameters are estimated by nonlinear observer, and the principle of this method is analyzed. Gain matrix of observer is optimized using NCD Blockset of Simulink. Nonlinear observer is designed offline and the real-time estimation of UAV state parameters is verified by the simulation using Matlab/Simulink.3. UAV state parameters are estimated using the method of geometry relation. Principle of INS/GPS system is introduced firstly, then principle of applying INS/GPS system to geometry relation to estimate UAV state parameters is studied, finally simulation results verify the real-time estimation of UAV state parameters.4. The method of utilizing multisensor data fusion to estimate UAV state parameters is proposed. Principle and method of data fusion are introduced firstly, then principle of utilizing the method of multisensor data fusion to estimate angle-of-attack and sideslip angle is studied, and verified by simulation finally. The simulation results indicate that three methods all can obtain requiredprecision. The methods of geometry relation and multisensor data fusion can both directly be applied to estimate real-time angle-of-attack and sideslip angle of UAV. The method of utilizing multisensor data fusion to estimate UAV state parameters proposed in this dissertation is innovative in certain extent, and this method enables the existing sensors to function effectively and avoid developing complicated sensors with high precision. So this method takes on the better value of project application.更多还原