Multipoint communications for wireless ad hoc networks

Wireless ad hoc network refers to a class of wireless networks in which devices are connected in a fully distributed manner. Devices in such networks tend to have short transmission ranges, and communicate with other devices through wireless links over multiple hops. Because devices are not directly connected, some sort of routing is needed for data packets to be transmitted from one device to another.;Multipoint communications refers to a communication mode in which the participating parties involved are more than just a sender and a receiver. It could mean the communication among a special group of devices in the network (i.e. multicasting) or among all devices in the network (i.e. broadcasting). This research work concentrates on the design and analysis of multipoint routing algorithms and protocols, including multicast and efficient broadcast routing, for wireless ad hoc networks.;We started our work with evaluating and enhancing the famous Multicast Ad hoc On-Demand Distance Vector (MAODV) [1] routing protocol. Our study showed that MAODV has problems on multicast tree maintenance and robustness. We enhanced the protocol by introducing new mechanisms and algorithms to the protocol to make it workable.;In the second part of our work, we designed a dynamic RTT-based algorithm which can dramatically reduce the route discovery latency of reactive ad hoc routing protocols. We evaluated our algorithm by applying it to MAODV and achieved great improvement on route discovery latency.;In the third part of our work, we developed a very efficient multicast routing protocol for wireless ad hoc networks with resource constrains. This algorithm fully utilizes the "logical tree + local link state" information so that a multicast algorithm with very low overhead can be achieved.;In the last part of this work, we study the "Broadcast Strom" problem in wireless ad hoc networks and developed a "smart" broadcast scheme that is simple, highly efficient, robust and scalable. This fully distributed algorithm introduces no control overheads (in the terms of control traffic), and can adapt to node density and network movement at no extra costs.