Fast estimation method for X-ray pulsar frequency based on waterfall dimensionality reduction and bee colony optimization search

In order to estimate the time delay of X-ray pulsars more quickly and accurately, and to further improve their navigation performance, this paper proposes a fast X-ray pulsar frequency estimation method based on Waterfall Dimensionality reduction and Bee colony Optimization search (WDBO) for high-data-volume characteristics. First, this paper folds the waterfall plot of the observation task and jointly analyzes the candidate frequencies from the two aspects of profile signal strength and segmented phase consistency. Secondly, in order to reduce the computational cost, this paper reconstructs the high-dimensional space sample distance model based on the actual sample distribution characteristics, and performs nonlinear Stochastic Neighbor Embedding (SNE) on the waterfall plot to reduce the dimension. Finally, in order to improve the problem of slow traversal search speed when estimating frequency in high-data-volume observation tasks, a fast pulsar frequency search method based on an improved bee colony optimization algorithm is proposed. The search area is effectively constrained by the pre-search behavior of employed bees, and the exploration behavior of scout bees is optimized by combining the local focusing strategy based on the optimal value to further speed up the iteration speed. At the same time, this paper conducts sensitivity analysis and ablation tests on the proposed method. Verified by NICER satellite observation data, compared with the traditional χ² test and weighted Z² test, the frequency estimation accuracy of our method is improved by 23.24% and 21.48%, and the computational efficiency is improved by 80.01% and 79.68%, respectively.