CAJ | 학술논문

为了提高时变稀疏信号的重构性能，提出了融合压缩感知和卡尔曼滤波的快速收敛重构算法。通过伪测量（ PM）技术以及扩展卡尔曼滤波形式的线性化将近似l0范数约束有效地融合到卡尔曼滤波架构中，求解常规的卡尔曼滤波方程，重构出稀疏信号。此外，针对PM阶段的扩展卡尔曼滤波形式，提出了快速收敛算法，有效地保证了算法的收敛和精度。仿真结果表明，相比原有基于卡尔曼滤波的恢复算法，所提算法的重构精度更高、收敛更快。
위료제고시변희소신호적중구성능，제출료융합압축감지화잡이만려파적쾌속수렴중구산법。통과위측량（ PM）기술이급확전잡이만려파형식적선성화장근사l0범수약속유효지융합도잡이만려파가구중，구해상규적잡이만려파방정，중구출희소신호。차외，침대PM계단적확전잡이만려파형식，제출료쾌속수렴산법，유효지보증료산법적수렴화정도。방진결과표명，상비원유기우잡이만려파적회복산법，소제산법적중구정도경고、수렴경쾌。
In order to improve the reconstructed performance of time-varying sparse signal, a fast convergent reconstruction algorithm with the fusion of compressive sensing and Kalman filter was proposed. The approximate l0 norm constraint was effectively integrated into Kalman filter framework by Pseudo Measurement ( PM ) technology and the linearization in the form of extended Kalman filter, for solving the conventional Kalman filter equations to reconstruct sparse signal. Furthermore, a fast convergent algorithm was proposed against the form of extended Kalman filter in the stage of PM, to ensure the convergence and high-precision of the algorithm. The simulation demonstrates that the proposed algorithm has higher reconstruction accuracy and faster convergence, compared to the original recovery algorithm based on Kalman filter.