I. Introduction
Inverse synthetic aperture radar (ISAR) has the capability of providing two-dimensional (2D) high-resolution images of non-cooperative moving targets, which can be utilized in many military and civilian applications including target recognition, aircraft traffic control, and air/space surveillance [1]–[5]. The stepped-frequency waveform (SFW) radar system utilizes a continuous series of short monotone pulses transmitted at a high pulse repetition frequency (PRF), with only a single frequency component in each pulse. A sequence of pulses, stepped by a fixed frequency step from pulse to pulse is called a burst. Thereafter, a set of bursts is collected during a coherent processing interval (CPI). High range resolution is determined by the bandwidth of the transmitted signal, while a change in the relative aspect angle (e.g., the number of bursts) between the target and the radar results in high cross-range resolution [6]–[9]. The motion of a target can be divided into two types: rotational motion and translational motion (TM). The former is essential to achieve high-resolution imaging along the cross-range direction, owing to the Doppler gradient that makes the cross-range resolution possible. The latter causes severe blurring in the ISAR images due to unwanted phase error terms that yield distortions along the range and cross-range directions [10]. Thus, the TM must be accurately compensated in order to obtain a well-focused ISAR image [11]–[14].