This paper focuses on the weak maneuvering target detection problem in the alpha-stable noise (ASN) environment. A novel coherent integration framework is proposed, where the range migration (RM) is first removed via approximate linear methods to make the noise distribution character remain unchanged, and then the ASN suppression, Doppler frequency migration (DFM) compensation, and coherent integration are simultaneously achieved in the proper transform domain. Based on this framework, a robust detection method is developed. First, the second-order keystone transform (SoKT) is employed to correct the range curve (RC) induced by the target's acceleration. Thereafter, the centroid axis rotation (CAR) is presented to remove the residual range walk (RW) by rotating the slow time axis parallel to the RW line in the fast-time and slow-time domain. Finally, the phase fractional lower-order Lv's distribution (PFLOLVD) is proposed to compensate the DFM and realize the energy accumulation of the target in the ASN environment. Furthermore, the performance of proposed algorithm in aspects of coherent integration time, computational complexity, and multiple targets detection are analyzed in detail. Compared with the existing coherent integration detection methods, the proposed method is both robust for ASN environments and superior in the detection performance.

本文着重于α稳定噪声（ASN）环境中的弱机动目标检测问题。提出了一种新颖的相干积分框架，其中首先通过近似线性方法去除距离迁移（RM），以使噪声分布特征保持不变，然后同时进行ASN抑制，多普勒频率迁移（DFM）补偿和相干积分。在适当的变换域中实现。基于此框架，开发了一种鲁棒的检测方法。首先，采用二阶梯形失真校正（SoKT）来校正由目标加速度引起的距离曲线（RC）。此后，提出了质心轴旋转（CAR）以通过在快时域和慢时域中旋转平行于RW线的慢时间轴来消除残余距离游动（RW）。最后，提出了分数阶低阶LV分布（PFLOLVD），以补偿DFM并实现目标在ASN环境中的能量积累。此外，从相干积分时间，计算复杂度和多目标检测等方面对所提出算法的性能进行了详细分析。与现有的相干集成检测方法相比，该方法在ASN环境下既健壮，又具有优越的检测性能。并详细分析了多目标检测。与现有的相干集成检测方法相比，该方法在ASN环境下既健壮，又具有优越的检测性能。并详细分析了多目标检测。与现有的相干集成检测方法相比，该方法在ASN环境下既健壮，又具有优越的检测性能。