Research On Intelligent Optimization Of MAC Layer Anti-Jamming Method In MANET

With the rapid development of wireless communication technology and microprocessor technology,mobile ad hoc network that is infrastructure-free and easy to deploy has been widely used in various fields such as smart transportation,military operations,and disaster relief,playing an increasingly important role.However,due to the fact that nodes in the network communicate via wireless channels and are limited by the communication capabilities of single node,mobile ad hoc network is vulnerable to malicious jamming attacks from external sources,making it difficult to ensure reliable network services.As the controlling layer for accessing channels,the Media Access Control(MAC)layer can achieve efficient anti-jamming effects by avoiding jammed channels.However,traditional MAC layer anti-jamming methods have gradually lost their effectiveness due to increasingly complex and dynamic jamming environments.To improve the reliability of mobile ad hoc network under jamming,this paper focuses on the intelligent optimization research of MAC layer anti-jamming methods.Firstly,this paper focuses on the challenging problem of predicting channel state in mobile ad hoc network with limited node perception and rapidly changing jamming conditions.Combining deep reinforcement learning technology with strong perception and decision-making abilities,an anti-jamming method based on the Dueling Double Deep Recurrent Q-Network(D3RQN)is proposed.By using Long Short-Term Memory(LSTM)networks to aggregate historical channel perception data,the overall channel state can be inferred more accurately.At the same time,the dueling network structure and double Q-networks are used to overcome the overestimation and unstable convergence defects of the Deep Q-Network(DQN),improving algorithm training efficiency and enabling faster and more stable development of optimal anti-jamming strategies.Simulation analysis results show that this method has better convergence performance under various jamming patterns.In complex jamming patterns,the channel selection result of this method is 20% higher than that of DQN algorithm,achieving better anti-jamming effects.This method also exhibits good adaptability and robustness.Subsequently,this article focuses on the application of large-scale node mobile ad hoc network,considering the impact of network clustering structure on intelligent anti-jamming methods,and proposes a Weighted based Clustering Algorithm for Anti-Jamming(WCA-AJ).The WCAAJ algorithm is based on the requirements of anti-jamming tasks,using node jamming perception,ideal node degree difference,energy,and relative mobility as clustering weight measurement indicators.At the same time,by applying the Analytic Hierarchy Process and Criteria Importance Though Intercrieria Correlation(CRITIC),subjective analysis of decision-makers is combined with the characteristics of objective data to set more reliable weight factors,thereby selecting more efficient nodes as cluster heads within the network.Through the algorithm process and example analysis,the clustering process and working principle based on the algorithm are elaborated in detail.Simulation analysis results show that the average channel selection accuracy of the cluster head selected by WCA-AJ algorithm is 5% higher than that of other clustering algorithms,which effectively enhances the anti-jamming performance of the network,has better stability of the cluster structure,and extend the network’s lifespan.