CAJ | 학술논문

针对 SAR 图像匹配及定位需要耗用不等的计算时间而造成的量测不等间隔输出和量测信息滞后问题，提出一种新的 SAR 时延补偿算法。该算法在标准卡尔曼滤波（KF）基础上，当 SAR 有量测信息生成时，根据多模型方法进行量测预测，利用预测值修正 SINS 状态；而 SAR 无量测信息输出时，通过插值方法生成量测信息来改善系统滤波精度。仿真结果表明，采用基于多模型量测预测的 KF 算法可以将位置误差由45 m 减小到10 m 以内，航向角稳态误差值小于5.8"；而在此基础上叠加插值预测算法可以将位置误差进一步控制在6 m 以内，航向角稳态误差小于4.7"，证明了本文提出的算法能够有效补偿 SAR 的随机时延并提高组合导航系统的解算精度。
침대 SAR 도상필배급정위수요모용불등적계산시간이조성적량측불등간격수출화량측신식체후문제，제출일충신적 SAR 시연보상산법。해산법재표준잡이만려파（KF）기출상，당 SAR 유량측신식생성시，근거다모형방법진행량측예측，이용예측치수정 SINS 상태；이 SAR 무량측신식수출시，통과삽치방법생성량측신식래개선계통려파정도。방진결과표명，채용기우다모형량측예측적 KF 산법가이장위치오차유45 m 감소도10 m 이내，항향각은태오차치소우5.8"；이재차기출상첩가삽치예측산법가이장위치오차진일보공제재6 m 이내，항향각은태오차소우4.7"，증명료본문제출적산법능구유효보상 SAR 적수궤시연병제고조합도항계통적해산정도。
The image matching and positioning of SAR need non-constant computation time, which result in unequal interval and delay characters of measurement output. In view of measurement prediction thoughts, a new algorithm is proposed to compensate the delay of SAR, which is based on the basic Kalman filtering. As the measurement information of SAR generated, the measurement prediction is made by a multiple model method, then the predicted value is used to correct the state of SINS; and when there is no measurement output, the measurement prediction is generated by the interpolation method and used to improve the system filtering accuracy. The simulation results are as follows: the position error can be reduced to 10 m from 45 m and the heading angle error is less than 5.8" when using the improved KF algorithm which is based on multi-model measurement prediction method. On the premise of the above algorithm, and by adding on the interpolation prediction algorithm, the position error is less than 6 m and the heading angle accuracy is superior to 4.7", which prove that the proposed algorithm can effectively compensate the random delay of SAR and improve the calculating precision of the integrated navigation system.