【作者】 潘献飞；

【导师】 黄新生；

【作者基本信息】 国防科学技术大学 ， 控制科学与工程， 2008， 博士

【摘要】 随着国内激光陀螺水平的不断提高,机抖激光陀螺捷联惯导系统在国内惯性技术应用领域正日益受到重视,发挥越来越重要的作用。机抖陀螺特有的机械抖动特性,在消除陀螺锁区的同时,也使捷联系统具有一系列新的更为复杂的动态误差特性,系统的优化设计与误差补偿研究具有尤其重要的理论和现实意义。论文以机抖激光陀螺捷联惯导系统为研究对象,提出了捷联系统的算法设计必须与系统的信号特性和应用环境相适应的思想,基于机抖激光陀螺在不同条件下的信号频域特性,开展了捷联惯导系统动态误差与补偿算法研究。论文主要完成了以下研究工作：1.研究了激光陀螺的数字控制特性,设计了新的数字控制方法,为陀螺信号频域特性研究奠定了基础。根据陀螺的抖动偏频机理,研究了陀螺静态和动态锁区的误差特性,探讨了抖动参数与陀螺精度的关系,得到了陀螺抖动参数的选取原则；设计了一种新的不同于模拟抖动的非线性抖幅控制方法、随机注入方法和一种新的适于数字实现的直流稳频控制算法,建立了机抖激光陀螺数字控制系统。新算法大大简化了陀螺控制系统的硬件实现,提高了系统可靠性和可扩展性,改善了控制精度。2.研究了机抖激光捷联系统的动态误差特性,建立了系统动态误差的优化处理原则。首先比较研究了整周期同步和数字滤波两种抖动解调方法,指出数字滤波法为提高捷联系统精度提供了更多更灵活的手段和可能性,已逐步取代前者得到了广泛应用；之后分类研究了捷联系统的动态特性,根据运动来源和运动性质的不同,对系统可能存在的圆锥和划摇运动进行了分类,建立了以减振器滤除干扰运动、以滤波器滤除干扰噪声的系统动态误差优化处理原则；最后把上述结论应用于机抖激光捷联系统,结合实际信号的频谱特点,研究了静态环境中陀螺抖动偏频和抖动耦合引入的各种动态误差及处理方法,振动环境中高频谐波振动引入的动态误差及处理方法,指出主振动的高频谐波是振动环境下影响系统精度的主要因素。3.基于陀螺数字滤波的信号处理方式,提出应把陀螺信号滤波后的幅频和相频畸变作为影响捷联系统姿态解算精度的重要误差因素进行研究,指出通常的滤波器通带指标一般不能满足姿态解算的信号稳定性要求,可能严重影响捷联系统姿态解算精度。设计了一种新的能与信号频域特性相匹配的圆锥优化算法,推导了经典圆锥运动下的圆锥优化算法公式,研究了算法的误差特性,证明了该算法在任意运动形式下的普适性,并扩展了其在消除伪圆锥误差和补偿陀螺自身频率特性时的应用。优化算法不增加任何算法实现难度和计算量,优化效果与滤波器性能相关。仿真和实验研究表明：优化算法能有效补偿信号滤波引入的圆锥误差,随信号滤波条件的不同能提高数倍到数个量级的姿态解算精度。4.圆锥算法受陀螺信号滤波影响,划摇算法同时受陀螺和加速度计两者的信号滤波影响。提出把陀螺和加速度计两者信号的滤波畸变作为影响捷联系统导航解算精度的重要误差因素进行研究,设计了一种新的能与信号频域特性相匹配的划摇优化算法,推导了经典划摇运动下的划摇优化算法公式,揭示了在相同的信号滤波条件下的划摇优化算法和圆锥优化算法依然满足算法的“对偶”关系,并证明了优化算法在任意运动形式下的普适性。仿真和实验表明,划摇优化算法具有与圆锥优化算法相似的误差特性,能有效补偿信号滤波引入的划摇误差,显著改善捷联系统导航精度。5.针对角振动环境,设计了一种新的基于固定频率运动优化的圆锥算法。不同于标准算法误差随圆锥频率单调变化的特性,新算法在设定频点处具有误差极小点,因而能有效改善系统在特定频带上的姿态解算精度。当系统运动环境已知且运动频带较窄时,优化算法具有良好的应用效果。最后,把基于信号频域特性的圆锥和划摇优化算法组成一套完整的优化导航算法,通过Matlab捷联系统仿真平台和某型机抖激光捷联系统在转台晃动、环形车载和远距离车载等动态环境中的实验,研究了信号频域特性对捷联系统导航精度的影响,验证了优化算法在不同运动环境和不同信号滤波条件下的补偿性能。更多还原

【Abstract】 With the improvement of ring laser gyro (RLG) in China, the dither RLG strapdown inertial navigation system (SINS) is playing a more and more important role in domestic inertial area. The gyros’unique mechanic dither property which is used to alleviate the gyro lock-in brings a lot of new and complicated dynamic error characteristics to the SINS, so system optimization and error compensation have important theoretic and practical meaning. Focusing on dither RLG-based SINS, the dissertation proposes that SINS algorithms must be designed to match the system signals and application environments, and researches on the dynamic error and compensation algorithms of dither RLG SINS based on the frequency domain characteristics of the RLG signals in various conditions. The main contents of this dissertation are:1. New digital control methods of RLG are designed through studying the gyro digital control characteristics, which sets a sound foundation for the ensuing research on the gyro signal frequency domain characteristics. According to the dither biasing mechanism, error characteristics of gyro static and dynamic lock-in are investigated. The relation between the dither parameters and the gyro precision is argued to obtain the design principles of dither parameters. Different from the analog methods, the new control methods of nonlinear dither magnitude and random dither are designed, and DC frequency-stabilization algorithm is implemented digitally. We then set up the whole digital control system of dither RLG. The new algorithms simplify the hardware and increase the reliability, expandability and precision of the control system.2. The dynamic errors of the dither RLG SINS are investigated and its optimization process principles are constituted. Firstly, it is underlined through peer comparison that the digital filter method, which gives more flexible ways and possibilities to increase the SINS accuracy, has been used widely instead of the whole period-synchronization method. Then, the SINS dynamic characteristics are investigated. The coning and sculling motions which might exist in SINS are assorted according to the source and property of motions. Optimization principles of system dynamic errors that the shock absorbers alleviate the disturbed motions and the filters eliminate the noise signals are set up. Finally, these conclusions are applied to the dither RLG SINS. The SINS dynamic errors introduced by the gyro dither bias, dither coupling in the static environment and the high frequency resonance in the vibration environment are studied, along with their process methods. It is pointed out that the high frequency harmonic vibrations are the major dynamic error source of SINS in the vibration environment.3. Based on the gyro signals process way of digital filter, it is proposed that the magnitude-frequency and phase-frequency distortion of gyro filtered signals should be studied as the important error factors of attitude computation for the SINS. The passband standard of the normal filters can not generally meet the requirement of signals stabilization for the attitude calculation, which may seriously decay the attitude precision. New optimal coning algorithms are thus designed to match the signals’ frequency domain characteristics. Algorithm formulas are derived for the classical coning motions and its error characteristics are studied. It is proved that the optimal algorithms can also be used in general environments. Additionally, the optimization algorithms are extended to eliminate the pseudo coning errors and compensate the gyro frequency characteristics itself. The new algorithms do not complicate the algorithm realization and computation burden. The optimization effect depends on the filter performance. Simulations and experiments show that the optimized coning algorithms can compensate significantly coning errors introduced by signal filter and increase the attitude algorithm accuracy from times to orders depending on various filtering conditions.4. Sculling algorithms are influenced by both gyro and accelerometer signals, in contrast to coning algorithms that is affected only by gyro signals. Regarding the distortion of gyro and accelerometer filtered signals as the important error factors of navigation computation, new optimization sculling algorithms are designed to match the signals’frequency domain characteristics. Algorithm formulas are derived in the classical sculling motions. It is revealed that the optimal coning and sculling algorithms still keep the duality in common filtering conditions and can be also used in the general environments. Simulations and experiments show that the optimal sculling algorithms have similar error characteristics with the optimal coning algorithms and can improve remarkably SINS navigation accuracy by compensating the filter-dependence sculling errors.5. For angle vibration environments, a new coning algorithm is designed which is optimized to the fixed frequency motion. Different from the monotonic error characteristics of the standard algorithm with the coning frequency, the optimal coning algorithm has a minimization point at the specified frequency and can improve the attitude accuracy remarkably for those motions with the specified frequency band. The optimal algorithm has good effect for the motions with the known narrow frequency band.Finally, the optimal coning and sculling algorithms are integrated to form the whole navigation algorithms. SINS Matlab models and experiments (turn-table swinging, land runs in circles and in a long way) are carried out to investigate the effect of signals frequency domain characteristics on the SINS navigation accuracy. The performance of the optimization algorithms are verified favorably for various environments and signal filters.更多还原