雷达在工作过程中所应对的干扰场景复杂且多变,所具有的反干扰措施难以穷举。人工设计的反干扰流程与抑制策略在面对这些对抗场景时,由于受限于专家的经验知识,其反干扰性能难以保证。对此,文中从雷达抗干扰的应用需求出发,通过引入强化学习方法,提出一种基于强化学习模型的智能抗干扰方法。分别利用Q 学习与Sarsa 两种典型的强化学习算法对反干扰模型中的值函数进行了计算并迭代,使得反干扰策略具备了自主更新与优化功能。仿真结果表明,强化学习算法在训练过程中能够收敛并实现反干扰策略的优化。相比于传统的反干扰设计手段,雷达反干扰的智能化程度得到了有效提升。

Radar has to deal with complex and changeable jamming scenarios in the process of work, and it is difficult to exhaust anti-jamming approaches. In the face of these confrontation scenarios, the artificial anti- jamming flow and suppression strategy cannot be guaranteed due to the limited experience and knowledge of experts. Based on the application requirements of radar antijamming, this paper introduces reinforcement learning method and proposes an intelligent anti-jamming method based on reinforcement learning model. Two typical reinforcement learning algorithms, Q-learning and Sarsa, are respectively used to calculate and iterate the value function in the anti-interference model, so that the anti-interference strategy has the function of self-updating and optimization. Simulation results show that the reinforcement learning algorithm can converge and optimize the anti-jamming strategy. Compared with the traditional anti-jamming design method, the intelligence of radar anti-jamming is improved effectively.

TN972