Real-time and accurate pulsar time-of-arrival estimation using GA-optimized EMD-CS

For the estimation methods of the pulsar TOA (time-of-arrival) based on CS (compressed sensing), it is very important to construct the measurement matrix, which affects accuracy and computation load directly. Subjected to computational load, the matrix size is also an important performance indicator. To reduce the matrix size, we propose a pulsar TOA estimation method using GA (Genetic Algorithm)-optimized EMD (empirical mode decomposition)-CS. Firstly, the standard pulsar profile is decomposed by the EMD to get IMFs (Intrinsic Mode Function). After the decomposition and the phase shift, these IMFs are performed to obtain the set of IMFs with multiple phases. And then, we take the error of the pulsar TOA estimation as the optimization goal, and use the GA to select IMFs from this set. The IMFs selected by the GA forms the optimized measurement matrix. Finally, the CS with the GA-optimized measurement matrix is used to obtain the pulsar TOA estimation. Theoretical analysis and simulation results show that the GA-optimized EMD measurement matrix has unlimited dimensions and small size. The pulsar TOA estimation using GA-optimized EMD-CS has a high accuracy and small computational load.