Research on Ship Network Security Situation Awareness and Defense Mechanism Based on Improved Spatiotemporal Attention and Deep Reinforcement Learning

With the development of ship intelligence, network security threats are increasing day by day. This paper proposes a ship network security situation awareness algorithm based on an improved spatiotemporal attention mechanism, and constructs a supporting defense mechanism. The algorithm accurately captures changes in network security situation through dynamic weight allocation and multi-scale feature extraction. In the experimental simulation, OMNeT++ is combined with SUMO to build a ship network simulation environment, and Maritime - CPS - Dataset and other data sets are used for testing. The algorithm in this paper is compared with ARIMA, LSTM, GRU and other algorithms. The results show that in terms of situation awareness accuracy, the algorithm in this paper reaches 95.6%, which is 27.8% higher than ARIMA, 12.3% higher than LSTM, and 10.1% higher than GRU respectively; the average response time of the defense mechanism is shortened to 2.3 seconds, which is 40% faster than the traditional static defense strategy, the attack loss is reduced by 78%, and the resource occupancy rate is reduced by 35%. The experimental data fully verifies that the algorithm and defense mechanism significantly improve the ship network security protection capability, providing reliable technical support for the intelligent and safe operation of ships.