CAJ | 학술논문

针对系统动力学模型不准确可能导致滤波精度下降，以及系统状态协方差阵可能出现的负定性问题，提出一种新的高斯过程回归平方根分解无迹粒子滤波（Gaussian process regression square-root decomposition unscented particle filter，GPSR-UPF）算法。在该算法中，采用高斯过程回归求取 UPF 的重要性密度函数。当系统模型不准确时，通过高斯过程回归学习训练数据，进而获取系统的回归模型及系统噪声协方差，同时引入平方根变换抑制系统状态协方差阵的负定性。将提出的 GPSR-UPF 算法应用到捷联惯导／全球定位系统（strapdown inertial navigation system／global positioning system，SINS／GPS）组合导航系统中进行仿真验证。结果表明，所提出滤波算法的性能优于基本的无迹粒子滤波算法，能提高组合导航系统的解算精度。
침대계통동역학모형불준학가능도치려파정도하강，이급계통상태협방차진가능출현적부정성문제，제출일충신적고사과정회귀평방근분해무적입자려파（Gaussian process regression square-root decomposition unscented particle filter，GPSR-UPF）산법。재해산법중，채용고사과정회귀구취 UPF 적중요성밀도함수。당계통모형불준학시，통과고사과정회귀학습훈련수거，진이획취계통적회귀모형급계통조성협방차，동시인입평방근변환억제계통상태협방차진적부정성。장제출적 GPSR-UPF 산법응용도첩련관도／전구정위계통（strapdown inertial navigation system／global positioning system，SINS／GPS）조합도항계통중진행방진험증。결과표명，소제출려파산법적성능우우기본적무적입자려파산법，능제고조합도항계통적해산정도。
In view of the uncertainty of the system dynamic model may reduce the filtering effect and the system state covariance matrix is negative definiteness,a new unscented particle filter(UPF)based on Gaussian process regression and square-root decomposition(GPSR)is proposed.The importance density function of UPF is gotten by Gaussian process regression.When the system model and observation model are inaccurate,Gaussi-an process regression is used to learn the training data,the regression models and noise covariance of the dynam-ic system are gotten;square-root decomposition is used to restrain the negative definiteness of the system state covariance matrix.The proposed algorithm is applied to the integrated navigation system of strapdown inertial navigation system/global positioning system (SINS/GPS).The simulation results show that the proposed al-gorithm is better than UPF,and also effectively improves the positioning precision of the navigation system.