We study an efficient broadcast scheme in mobile ad hoc networks (MANETs). The objective is to determine a small set offorward nodes to ensure full coverage. We first study several methods that guarantee coverage when thelocal viewof each node on itsneighborhood information is updated in a timely manner. Then we consider a general case where nodes move even during thebroadcast process, making it impractical to maintain up-to-date and consistent local views. A formal framework is used to modelinaccurate local views in MANETs, where full coverage is guaranteed if three sufficient conditions, connectivity, link availability, andconsistency, are met. Three solutions are proposed to satisfy those conditions. First, we give a minimal transmission range thatmaintains the connectivity of the virtual network constructed from local views. Then, we use two transmission ranges, one forneighborhood information collection and the other for actual data transmission, to form abuffer zonethat guarantees the availability oflogical links in the physical network. Finally, we propose a mechanism calledaggregated local viewto ensure consistent local views.By these, we extend Wu and Dai’s coverage condition for broadcasting in a network with mobile nodes. The effectiveness of theproposed scheme is confirmed via both performance analysis and simulation study. Efficient broad casting Coverage in Mobile Ad Hoc Networks
BROADCASTINGa packet to the entire network is a basicoperation and has extensive applications in mobile adhoc networks (MANETs). For example, broadcasting isused in the route discovery process in several routingprotocols, when advising an error message to erase invalidroutes from the routing table, or as an efficient mechanismfor reliable multicast in a fast-moving MANET. In MANETswith the promiscuous receiving mode, the traditional blindflooding incurs significant redundancy, collision, andcontention, which is known as the broadcast storm problem[23]. Efficient broadcasting in a MANET focuses onselecting a small forward node set while ensuring broadcastcoverage.
Broadcast protocols can be classified intodeterministicandprobabilisticapproaches. The probabilistic approach[11], [23] usually offers a simple solution in which eachnode, upon receiving a broadcast packet, forwards thebroadcast message with probabilityp. However, theprobabilistic approach cannot guarantee full coverage. Thedeterministic approach guarantees full coverage and can befurther classified based on the type of neighborhoodinformation used:location-information-basedandneighbor-set-based. In location-information-based broadcast protocols,location information of neighbors is available, whereas inneighbor-set-based broadcast protocols, only neighbor setinformation is available. Location information facilitatesefficient broadcasting in terms of generating a smallforward node set; however, it comes with a cost—location information requires additional hardware, such as GPS.Other types of information can also be used which fall inbetween the above two models: directional information,where messages arrive from a certain angle-of-arrival(AOA), and distance information based on the signalstrength received. All these models assume some sort ofspecial hardware. In addition, location/direction/distanceinformation may not be accurate. In this paper, we limit ourconsideration to deterministic broadcast protocols that useneighbor set information only, which corresponds to theweakest assumption on neighborhood information used.