No abstract

No abstract

Due to the rapid development of wireless ad hoc networks, routing protocols in wireless ad hoc networks must have better adaptability to reduce system complexity, save routing overhead and effectively use limited network resources. Currently, the traditional path addressing method using a route-based routing mechanism will cause a large amount of network load and greater delay. To this end, we propose a new method based on convolutional neural network (CNN) to identify, evaluate and guide network features. Using the service characteristics and link status characteristics of the network, comprehensively using the network characteristic matrix as the signal of the network, and adding the measurement of the network performance, the network characteristics can be more completely and accurately identified. The best choice for network performance. Through simulation experiments, it is verified that the method can correctly identify the current network characteristics, make correct predictions about the current network conditions, and make appropriate path selections as needed, thereby effectively reducing network congestion. And can effectively solve the network congestion problem.