CAJ | 학술논문

针对GPS/INS与JTIDS组合的相对导航算法设计，采用无轨迹卡尔曼滤波方法（UKF）实现了非线性导航参数估计。为了解决相对导航中UKF算法在定点DSP中计算量大、不能够满足实时性要求的问题，研究了基于定点DSP芯片的UKF算法的优化方法。将程序中的变量进行定标，采用Q表示法将浮点数转换成定点数。借助TMS320C64x＋IQmath函数库，将程序中的C语言浮点算法转换成定点算法。实验结果表明，该方法能保证相对导航定位精度，提高在定点DSP芯片中的计算效率，满足导航处理中的实时性要求。
침대GPS/INS여JTIDS조합적상대도항산법설계，채용무궤적잡이만려파방법（UKF）실현료비선성도항삼수고계。위료해결상대도항중UKF산법재정점DSP중계산량대、불능구만족실시성요구적문제，연구료기우정점DSP심편적UKF산법적우화방법。장정서중적변량진행정표，채용Q표시법장부점수전환성정점수。차조TMS320C64x＋IQmath함수고，장정서중적C어언부점산법전환성정점산법。실험결과표명，해방법능보증상대도항정위정도，제고재정점DSP심편중적계산효솔，만족도항처리중적실시성요구。
Unscented Kalman filter was used to estimate non-linear navigation parameters in the design of relative navigation sys-tem integrated by GPS/INS and JTIDS.To solve the problem that large computation of UKF algorithm in the Fixed-point DSP could not meet the real-time processing,the UKF algorithm optimization was proposed.Firstly,the variables were scaled in the C program,and Q notation was used to convert floating point to fixed point.Then,the floating point C program was converted to fixed point algorithm using TMS320C64x+IQmath library.Experimental results indicated that the optimized UKF algorithm could guarantee the relative navigation' s position accuracy,and the computational efficiency was improved to satisfy the real-time processing.