NS-Based Hybrid Mobility Model for Urban Environments

Abstract

Mobile Adhoc Networks (MANET) has been an emerging eld for research in wireless communications during the last decade. It is the study of spontaneous and self-organizing network of mobile nodes without any infrastructure. Other specialized elds in this area include Vehicular Adhoc Networks (VANET), Flying Adhoc Networks (FANET), Mobile Wireless Sensor Networks (MWSN) and few others. Network Simulators are required for performance analysis, improvement and development of routing or communication protocols, algorithms and models designed for complex networks. e development of a mobility model that provides an accurate and realistic node mobility description is the key to the simulation of mobile networks. Existing generic mobility models lack realism and su er from a number of limitations. Conducting evaluations of MANET routing protocols using variants of random models doesn't provide meaningful results due to the lack of detail and being over simpli ed. For example, traces generated from RandomWaypoint Model (RWP) and Random Direction Model (RDM) lack the resemblance to human movement pa erns. ere are several challenges posed to the simulation of MANET. Most signi cantly, for the meaningful evaluation of protocols, there should be a mobility model that records the key characteristics of its node mobility and is a representative of real world scenarios unlike simple mathematical models. Furthermore, the simulations of mobile networks usually require an external tra c simulator for scenario and movement pa ern generation. e traces generated require conver3sion into a network compatible format before being fed to the network simulator. is becomes a hectic task in network simulators like Network Simulator 3 (NS3). NS3 , however provides very fewmobility models such as Random Direction, Random Waypoint, Random Walk and Gauss Markov. Also, there's no exibility given to the users to make mobility model scenarios of their own choice. In this thesis, I proposed a realistic hybrid mobility model design from a combination of well known mobility models for urban environments. e model's design is generic, but implementation is done in NS3. e key characteristics of the model are the accuracy and realism of movement pa erns of heterogeneous tra c modes such as bike and car to be generated for further simulation. e simulation results however appear as though the testing was conducted in real urban scenario itself.