Markov Chain-based Mobility Prediction and Relay-node Selection for QoS Provisioned Routing in Opportunistic Wireless Network

Abstract

In the fields of networking and communication, wireless sensor networks are a commonly used technology. Although Mobility in Wireless Networks or Mobile Ad hoc Networks (MANETs) were developed to solve most critical problems in a disaster scenario. However, data transmission in a disaster area is a difficult operation due to irregular connections between sensing devices and the sink. Opportunistic network-based MANETs are utilized to overcome this problem. There are many routing issues such as communication overhead ratio, packet loss, and packet delivery ratio at mobile nodes. There are several traditional approaches used in opportunistic networks for node localization and routing. The drawback of these traditional routing methods is their poor reliability. A Markov chain-based opportunistic routing protocol is proposed in this paper. A Markov chain model is used to predict the node mobility for the random mobility model, and a neighbor table-based relay node selection procedure is proposed. A Markov prediction model uses the history of a node’s movement without requiring additional information. By analyzing simulation metrics including delay, overhead rate, packet drop, and packet delivery rate, the proposed routing protocol is evaluated and verified based on the Packet Delivery Ratio (PDR) in an opportunistic network environment (ONE) simulator. In addition, existing routing protocols such as Spray and Wait (SNW), Epidemic, Prophet, and MaxPro differ from the proposed routing protocol in terms of buffer size, node count, and message generation rate. Corresponding result analysis demonstrates the superior performance of the proposed routing protocol over existing protocols.