Radar coherent integration algorithm for detection of complexly maneuvering target with extended velocity and acceleration scopes

Abstract

Excess acceleration and Doppler ambiguity produce inaccurate acceleration and radial velocity in long-time coherent integration for complexly maneuvering targets. This study proposes an algorithm based on the product-scaling periodic-modified Lv's distribution (PSPMLVD) and inverse-scaling periodic discrete Fourier transformation (ISPDFT) that solves these problems by eliminating the range migration errors. First, the periodic modified Lv's distribution (PMLVD), which extends the estimation scope of the modified Lv's distribution, is proposed to estimate the excess acceleration and precise jerk. Then, the scaling and product operations are combined with PMLVD to decouple and depress the sidelobes. Subsequently, high-order parameters are compensated for, and SPDFT is employed to extend the period of discrete Fourier transform (DFT). Finally, inverse DFT is applied to achieve coherent integration and estimate the unambiguous velocity and initial range. The algorithm is appropriate for targets following first- to third-order motions as it avoids searching for parameters and approximating models. Unlike several existing algorithms, this algorithm can achieve equilibrium between computational complexity and anti-noise detection performance. Simulation experiments and real data prove its effectiveness.