Cooperative Topology Sensing of Wireless Networks With Distributed Sensors

Abstract

As a key technology of network inference, topology sensing plays an important role in understanding, explaining and predicting network behavior. Majority of existing studies focus on topology inference with complete and perfect observations, which is mainly suitable for wired networks rather than wireless networks with kinds of channel randomness and uncertainty. This article investigates the issue of topology sensing of wireless networks with distributed sensors to handle the situation of limited and unreliable information, which include a time series of the signal presence instants and the corresponding transmitter identification. Firstly, a cooperative topology sensing framework is developed to exploit multi-sensor spatial diversity for coping with wireless randomness. Then, based on the observation that in wireless networks, events at one node are likely to induce a response at some other node in the form of for example an acknowledgement message or a forwarding action, we formulate the communication events among multiple wireless transceivers as a multi-dimensional Hawkes process. Further, we propose and analyze four cooperative topology sensing schemes with distributed sensors from the perspective of hard fusion and soft fusion, respectively. Moreover, simulation results show that the signal power and the event number have a great influence on sensing accuracy and cooperative topology sensing is an effective measure to deal with the issues of wireless channel randomness.