Collection Tree for Wireless Coverage Problem in Mobile Crowdsensing

Abstract

Mobile devices play an important role in crowdsensing with the help of storage infrastructure. In the event of emergencies, a wireless sensor network can he constructed for people to sense the environment. When devices’ mobility is out of control, collection trees are considered to route all the data to the sink. We propose a new wireless coverage problem in mobile crowd-sensing when the core network is corrupted, and our aim is constructing collection trees to minimize the average delay of information. We propose two kinds of algorithms to construct the collection trees. We first propose Greedy Collection Tree (GCT) and Dynamic Greedy Collection Tree (DGCT) with two different objective functions, transferring probability and entropy increment. Then we propose Improved Greedy Collection Tree (IGCT) based on the assumption of low population fluidity. The effectiveness of our algorithms is testified in aspects of transmission average delay, delay’s standard deviation and entropy evolution.