Author:

Abstract:

The last decade has seen a rapid increase in demand for wireless connectivity and therefore wireless communication had become one of the most active research areas in the communication field. The growth in the number of wireless devices in our living environment is affecting our lives by helping us to communicate easier and faster, and to access information from almost anywhere. Most wireless networks to date are designed centrally, as communication is done through central entities such as wireless access points. However, wireless devices can communicate without access points in so-called ad hoc or peer-to-peer mode. These ad hoc networks have many interesting applications including, but not limited to, search and rescue operations, disaster recovery, policing and firefighting, vehicular network and service (e.g. traffic and accident warnings, road and weather condition information exchange), commercial services in airports, stadiums, and shopping malls, educational services in universities, multiplayer games, and outdoor internet sharing. Communication in mobile ad hoc networks faces several challenges: (1) scaling up in terms of the number of devices in one place, as this increases the density of the network and increases collisions, (2) scaling up in terms of the amount of information to be propagated in the network, as this consumes critical resources, such as transmission bandwidth and battery power, (3) devices move around, break the links, change the topology, and invalidate the collected routing information and therefore hindering the routing, (4) dealing with the heterogeneity, as devices have interaction with many other devices at one place, having different wireless technologies and characteristics, and (5) dealing with the irrelevancy of information, as the applications on wireless devices have different topics of interest that can vary on the basis of the geographical area, the role of the applications and so on. In this thesis, we design a communication middleware for mobile ad hoc networks. The reason to choose a middleware is that it provides transparent service to the applications and helps to connect different applications on different devices, using various communication technologies. Our middleware uses probabilistic routing techniques and forms groups of nodes based on their contextual information, when it is feasible. Probabilistic routing techniques prove to be suitable for highly mobile networks, as they do not form any logical structure (such as routing trees) that break with the change in the network topology. Groups of nodes are formed when it can be predicted from the contextual information of the nodes that these groups can stay for a reasonably long period. In order to evaluate the proposed systems, we use network simulations. We use widely adopted mobility patterns for simulating mobile ad hoc networks, and real-life movement data for simulating vehicular networks. For simulating larger networks, in order of hundreds of thousands nodes, we perform abstracted simulations using logic programming tools. The benefits of the abstracted simulations are (1) speeding up the simulation work, (2) giving us a general view of the performance of the proposed systems, and (3) quick gauging of efficient system parameters. The results show that by using probabilistic routing we can tolerate a high mobility rate of nodes. Moreover, our improved gossiping techniques and our dynamic subscription model reduces network traffic overhead and processing power overheard in nodes. Our simulation in vehicular networks show that using context-based grouping reduces network traffic overhead and increases the relevancy of the propagated information. Our simulation results also suggest that by combining the delay-tolerant network communication model with multi-hop communication, we can increase the coverage of information in the sparse part of the network and avoid losing important information. The thesis presents a combination of network communication techniques, together with several improvements, to design a scalable and accurate communication system for mobile ad hoc networks, and employs simulated experiments to evaluate its performance. The research work presented in this thesis can be used as a basis for further design of communication systems for mobile ad hoc networks.