Multi Level Bridging Emergency Multimedia Communication Network Congestion Control Algorithm

In multi-level bridging emergency multimedia communication networks, the high dynamism, burstiness, and heterogeneity of network traffic pose severe challenges to congestion control. Traditional methods are difficult to adapt to complex traffic characteristics, resulting in low network resource utilization, high packet loss rate, and poor transmission performance. Therefore, this article proposes a congestion dynamic control strategy based on LSTM. Firstly, Poisson distribution, peak to mean ratio (PMR), and utility function are used to analyze traffic characteristics and reveal network performance bottlenecks. Secondly, a comprehensive feature vector is constructed through multidimensional extraction of statistical features, time-domain features, frequency-domain features, and entropy features, and key features are screened using LASSO regression. Finally, based on the LSTM model to predict future traffic, dynamically adjust bandwidth allocation, and combine token bucket algorithm to smooth out sudden traffic. The experimental results show that the bandwidth utilization of the proposed method remains stable in complex environments, with packet loss rates reduced to below 0.3 %. The transmission rates for video, voice, and data traffic are 95.2 Mbps, 98.6 Mbps, and 97.8 Mbps, respectively. These figures demonstrate that the method significantly enhances network resource utilization efficiency and transmission performance. Moreover, it offers an effective solution for congestion control in multi-level bridging emergency multimedia communication networks.