Passive-based Neighbor Detection In Wireless Opportunistic Networks

On the basis of the traditional WLAN, kinds of peer-to-peer, self-organizing and multihop wireless ad hoc networks without fixed infrastructure arise. However, reasons such as nodes’ fast mobility and sparse distribution, signal attenuation caused by obstacles and energy restriction result in frequent link interruptions, which is unable to meet the end-to-end network connectivity requirement in traditional ad hoc networks. Thus, wireless opportunistic networks with intermittent connectivity was advanced in which nodes forward the data per hop to communicate using opportunistic contacts brought by mobility.Nodes must continually detect neighbors in its communication range to forward data in opportunistic networks. Accurate information of neighbor contacts is of importance for improving the delivery ratio of opportunistic forwarding and forwarding efficiency. Neighbor detection can be active or passive. The active detection requires nodes broadcasting probing messages constantly according to their own needs, which can only detect neighbors in communication range with large overhead; while the passive approach keeps sensing the carrier to estimate the number of neighbors and their directions. The existing passive detection schemes are based on the correct demodulation about complete message packet, not relating to the signal analysis, leading to weak ability in detecting neighbors.Comparing to the "clear" neighbors detected by the complete message demodulation using active approaches, this paper innovatively proposes to utilize the signals that can not be demodulated to estimate more "fuzzy" neighbors. Our approach makes use of carrier sensing to detect more potential neighbors to the best extent. Without extra resources consumption and introducing interference in passive detection, this paper uses signal analysis to evaluate the number and the directions of "fuzzy" neighbors. Our passive detection approach consists of two steps. Firstly, from the perspective of statistics, nodes use the minimal risk criterion to judge the existence and the number of "fuzzy" neighbors according to the monitored carrier. Secondly, the beam numbers of antennas are adopted to represent the neighbors’ positions using directional antennas. The feasibility of the proposed method in this paper is verified by numerical analysis and simulations.