Synthetic aperture radar (SAR), provides high resolution, day-and-night and weather-independent images for several applications including surveillance, targeting, navigation, moving target indication, change detection, environmental monitoring, and 3-D mapping. The high range resolution is derived by transmitting a high bandwidth signal such as wideband linear frequency modulation (LFM), and appropriate compression of the received echoes. If stepped-LFM is applied instead of wideband LFM, in addition to high range resolution, the instantaneous bandwidth and sampling rate are reduced. In ideal imaging, it is assumed that the radar platform moves at a constant velocity along a nominal trajectory. However, atmospheric turbulence, short manoeuvres, and platform perturbations cause trajectory deviations and lead to phase error. This error is called motion error, which can severely impair the image quality. In order to compensate the motion error, platform position, utilizing navigation sensors installed on radar platform, such as GPS and INS, is measured and most of the error is compensated. To compensate the residual motion error, autofocus techniques are adopted. In autofocus techniques, required parameters for focusing the image are derived from the raw data. One of the most popular autofocus techniques is PGA which has been designed for spotlight mode SAR. This technique, with some modifications in raw data or algorithm, can be applied in stripmap mode as well. In this thesis, the spotlight and stripmap autofocus techniques are reviewed. Then, a UAV-based SAR with stepped-LFM signaling in stripmap imaging mode is considered and three important autofocus methods, WPGA, PWE-PGA, and LML-WPGA are adopted for this system. Moreover, the performances of these autofocus methods are evaluated through computer simulation. Based on the simulation results, in most cases LML-WPGA has the best performance in motion error compensation. This method, assuming a three dimensional motion error, estimates the range-invariant error by WPGA and the range-dependent error by a local maximum-likelihood WPGA algorithm. Keywords: Synthetic Aperture Radar (SAR), Stepped-LFM, Residual motion error, Autofocus
