冗余捷联惯组信息容错管理算法

对于单机多表斜置冗余捷联惯组的故障诊断、隔离和重构，提出无迹卡尔曼滤波结合三交叉容错融合方法。考虑各误差项，对不同的冗余配置建立统一数学模型。通过三交叉重构模型的设计可快速隔离一度故障并输出正确结果，对被隔离的故障表各误差系数采用无迹卡尔曼滤波方法标定，可辨识因误差系数改变而引起的软故障。以单机五陀螺的两种误差系数软故障为例仿真验证了冗余捷联惯组信息容错管理算法的正确性。
A new algorithm based on intersection fault-tolerant fusion and unscented Kalman filter was proposed for fault diagnosis, isolation and reconstruction of skew redundant strapdown inertial navigation system. Considering the error coefficients of gyros, a unified model was built for redundancy configuration. A single point of failure can be isolated rapidly with the intersection fusion model, and an unscented Kalman filter was used for the error coefficients of isolated gyros, which can determine soft failure caused by error coefficients′ change. The simulation with configuration of 5 gyros shows the validity of algorithm under 2 soft failure caused by error coefficients′ change.