| 引用本文: | 周长霖，王春阳，宫 健，谭 铭，包 磊，刘明杰. 基于DQN和功率分配的FDA⁃MIMO雷达抗扫频干扰[J]. 雷达科学与技术, 2024, 22(2): 155-160.[点击复制] |
| ZHOU Changlin, WANG Chunyang, GONG Jian, TAN Ming, BAO Lei, LIU Mingjie. Anti⁃Sweep Interference of FDA⁃MIMO Radar Based on DQN and Power Allocation[J]. Radar Science and Technology, 2024, 22(2): 155-160.[点击复制] |
|
| 摘要: |
| 频率分集阵列（Frequency Diversity Array, FDA）雷达由于其阵列元件的频率增量产生了许多新的特性，包括其可以通过发射功率分配进行灵活的发射波形频谱控制。在以扫频干扰为电磁干扰环境的假设下，首先，通过引入强化学习的框架，建立了频率分集阵列?多输入多输出（Frequency Diversity Array?Multiple Input Multiple Output, FDA?MIMO）雷达与电磁干扰环境交互模型，使得FDA?MIMO雷达能够在与电磁环境交互过程中，感知干扰抑制干扰。其次，本文提出了一种基于深度Q网络（Deep Q?Network, DQN）和FDA?MIMO雷达发射功率分配的扫频干扰抑制方法，使得雷达系统能够在充分利用频谱资源的情况下最大化SINR。最后，仿真结果证实，在强化学习框架下，FDA?MIMO雷达能够通过对发射功率分配进行优化，完成干扰抑制，提升雷达性能。 |
| 关键词: 频率分集阵列 扫频干扰 强化学习 深度Q网络 功率分配 |
| DOI：DOI:10.3969/j.issn.1672-2337.2024.02.005 |
| 分类号:TN974 |
| 基金项目:国家自然科学基金（No.62201580）；陕西省自然科学基金（No.2021JM‐222） |
|
| Anti⁃Sweep Interference of FDA⁃MIMO Radar Based on DQN and Power Allocation |
|
ZHOU Changlin, WANG Chunyang, GONG Jian, TAN Ming, BAO Lei, LIU Mingjie
|
|
1. Air and Missile Defense College, Air Force Engineering University, Xi’an 710051, China;2. College of Information and Communication, National University of Defense Technology, Wuhan 430010, China;3. Test Center, National University of Defense Technology, Xi’an 710106, China
|
| Abstract: |
| The frequency diversity array （FDA） radar has many new characteristics due to the frequency increment of its array elements, including the flexible spectrum control of the transmitted waveform through the transmission power allocation. Aiming at the problem of suppression of radar performance by sweep interference, firstly, the interaction model between frequency diversity array?MIMO （FDA?MIMO） radar and electromagnetic interference environment is established by introducing a reinforcement learning framework, so that FDA?MIMO radar can sense interference and suppress interference in the interaction process with electromagnetic environment. Secondly, a sweep interference suppression method based on deep Q?network （DQN） and FDA?MIMO radar transmission power allocation is proposed, so that the radar system can maximize SINR while making full use of spectrum resources. The simulation results confirm that under the framework of reinforcement learning, FDA?MIMO radar can achieve interference suppression and improve radar performance by power allocation. |
| Key words: frequency diversity array （FDA） sweep interference reinforcement learning deep Q⁃network （DQN） power allocation |
