Deep Learning based Channel Assignment with Load Balancing in MANET for Improved Performance

In Mobile Ad Hoc Networks (MANETs), the effective channel assignment is important for optimizing the performance of the network. The proposed system uses deep learning with CNN, RNN and LSTM for channel assignment with load balancing. The spatial parameters from the input data are obtained using CNN. This helps in analyzing the relationship between network nodes with communication patterns. The important features that help in channel assignment decisions are obtained through feeding the signal strength information to the CNN network. The sequential data in the network are handled using RNN networks. They help in analyzing the temporal dependencies in the channel assignment process. This helps in obtaining informed decisions based on communication links. LSTM networks are used for recollecting information regarding the previous channel assignments. This helps the system to learn from past circumstances and make better decisions with load balancing. The optimal channel assignments are based on the contemporary state of the network. This helps in reducing congestion with improved resource utilization. The proposed system are designed to adapt for real time variations in the network. This is done through considering various factors such as signal strength, interference and active nodes. Continuous improvement of decisions are obtained through feedback for improved performance of the network.