نوع مقاله : مقاله پژوهشی
نویسندگان
1 دانشجوی دکتری ،دانشگاه صنعتی مالک اشتر ، تهران ، ایران
2 استاد ،دانشگاه صنعتی مالک اشتر ، تهران ، ایران
چکیده
کلیدواژهها
عنوان مقاله [English]
نویسندگان [English]
Rotational Motion Compensation (RMC) and determining the Coherent Processing Interval (CPI) in Inverse Synthetic Aperture Radar (ISAR) imaging are among the critical and challenging issues. The CPI duration should be determined in such a way that, first, it prevents image blurring due to the target's rotational motion, and second, it ensures the desired resolution in the cross-range direction. The CPI duration depends on two factors: the number of received pulses and the Pulse Repetition Interval (PRI). In this paper, it is assumed that the PRI is constant, making the problem of determining the optimal CPI duration essentially a matter of determining the optimal number of pulses. To determine the optimal number of pulses, a novel method based on deep reinforcement learning and utilizing a Deep Q-Network (DQN) is proposed. In this approach, the ISAR radar, acting as an agent, interacts with the target as the environment to make the best decision, which involves updating the number of pulses and achieving its optimal value. The ISAR radar evaluates each of its actions based on a reward function that depends on the sparsity of the image. The obtained results confirm the effectiveness of the proposed method as a novel approach for determining the optimal number of pulses and correcting the effects of the target’s rotational motion in ISAR imaging.
کلیدواژهها [English]
)
