Content-aware search of multimedia data in ad hoc networks

The infrastructure-free and self-organizing nature of wireless ad hoc networks presents fundamental challenges to the design of content-based multimedia search algorithms that are efficient with respect to search cost and fair across various network setups. In contrast to the wealth of research literature on ad hoc routing protocols, few works have realistically considered the methods of locating multimedia data sources in a highly dynamic ad hoc network. Moreover, multimedia information retrieval strategies proposed in the wired networks are not applicable in the context of wireless ad hoc networks, due to the limitations of bandwidth and energy. In this paper, we describe two probability-based schemes for the efficient multimedia content location in wireless ad hoc networks. The Association-based Content Prediction (ACP) scheme and Bayesian-based Content Prediction (BCP) scheme make use of probabilistic information to lower proactive network traffic while minimizing search cost. The combination of theoretical analysis and simulation results shows that the proposed schemes perform favorably in terms of response time, traffic complexity, and scalability.