Optimal Design Of Jamming-Detection Shared Signal Based On Deep Reinforcement Learning

With the continuous development of modern electronic warfare technology,radar and jammer,as "spears" and "shields" in modern warfare,both oppose and protect each other.By designing jamming-detection shared signals,jammer and radar can be built into integrated equipment,which can greatly reduce the equipment size and power consumption of the combat platform,and enhance the survivability of the own combat platform.To solve the above problems,this paper proposes two kinds of jamming-detection shared signal optimization design methods based on improved deep reinforcement learning algorithm.The main research contents of this paper are as follows:1.From the perspective of integrated system,a chaotic coding based non-coherent multi-carrier-frequency multi-phase coding interference detection shared signal is designed,which has good noise-like performance and good detection ability.In order to improve the overestimation problem of traditional deep reinforcement learning algorithm,a composite reward-dueling deep Q-learning network based on regularization was proposed.This algorithm solved the overestimation problem of the original algorithm,and introduced the state value function into the original reward value to form the compound reward value,which enhanced the internal relationship between the state and the action.The algorithm is used to optimize the initial phase coding value of multi-carrier frequency multi-phase coded signal,improve the signal interference and detection performance,and the effectiveness of the algorithm is verified by simulation analysis.At the same time,the traditional deep reinforcement learning algorithm is taken as a comparison algorithm,and the simulation shows that the proposed algorithm has better optimization effect under the same conditions.2.A coherent interference detection shared signal based on the improved deep reinforcement learning algorithm is designed based on the non-uniform interrupted sampling periodic repeater jamming signals.Based on the dueling double deep Q network,the state value function is introduced,which is called composite reward-dueling double deep Q network,and the optimal sampling and forwarding coding sequence is obtained by using it,and then the optimal shared signal of coherent interference detection is obtained.The simulation results show that the pseudo target distribution generated by the coherent shared signal optimized by the proposed algorithm is more random,which has the dual effect of jamming suppression and jamming spoofing,and has the detection performance at the same time.Compared with the traditional deep reinforcement learning algorithm,the algorithm proposed in this paper has better optimization effect and the optimal solution is more stable when the amount of coding state increases.