Location Prediction-based Routing In Delay Tolerant Network

In recent years, with the rapid development of the wireless network and mobile device, many new types of networks emerge, such as deep space network, vehicular network and sensor network in extreme environments. Unlike the traditional Internet that is characterized by short round-trip delay and low bit error rate, the nodes in these kinds of network rely on the opportunistic contact to exchange data, which makes the persistent network connection unavailable and leads to high data delivery latency. These kinds of networks are called delay tolerant networks(DTN).In DTN, data routing usually uses the store-carry-forward strategy. The node in the network sends data package to destination node through a series of contact between two nodes. Predicting when and where these contacts will happen would have huge impact on routing performance. Location predicting based on DTN routing predicts node's location, and calculates the contact probability between nodes pairs in the future according to their location.The nodes in the DTN usually move between a set of interest points in a specific region. We model the DTN region as a grid, and the grid points represent the interest points, the edges represent the paths between the interest points. The movement of nodes is following grid based on mobility model which can be described as a Markov process. Based on the established Markov process model, the nodes' location can be obtained and the contact probability between two nodes can be calculated. In the routing process, the node has a larger contact probability with the destination node will has a higher priority to forward the data package.Performance evaluation shows that the location prediction based routing protocol has a better delivery ratio compared with other routing protocols.