CAJ | 학술논문

对于异步多传感器观测数据，基于先同步，再去相关的思想，提出了一种改进的左同步提升异步观测融合算法。采用左同步法避免了右同步过程中的系统状态矩阵求逆和可能出现的非因果问题；对同步后的系统基于Cholesky分解进行噪声去相关处理，理论上分析了去相关处理前后的算法计算量；用信息滤波器进行预测估计，简化了滤波增益的计算过程。仿真结果表明：改进算法能够在不减小跟踪精度的基础上减小计算量，增强了算法的实时性。
대우이보다전감기관측수거，기우선동보，재거상관적사상，제출료일충개진적좌동보제승이보관측융합산법。채용좌동보법피면료우동보과정중적계통상태구진구역화가능출현적비인과문제；대동보후적계통기우Cholesky분해진행조성거상관처리，이론상분석료거상관처리전후적산법계산량；용신식려파기진행예측고계，간화료려파증익적계산과정。방진결과표명：개진산법능구재불감소근종정도적기출상감소계산량，증강료산법적실시성。
As for asynchronous multisensor observation datas,a modified left synchronously lifting fusion algorithm based on the ideal of synchronism firstly and decorrelation next is proposed. Left synchronism is used to avoid the state matrix inversion and the possibly emerged non-causality problem in the process of right synchronism;decorrelation operation based on Cholesky factorization is conducted to the system after synchronism,the computation cost of the algorithm is analyzed theoretically;information filter is used for prediction and estimation, simplify the computation process of filtering gain. Simulation result shows that,the modified algorithm can reduce computation cost without changing tracking precision and enhance the realtime of the algorithm.