Will you accept my job? A new approach towards predicting human participation in mobile crowdsensing

Abstract

The exponential growth of wireless devices have paved the way for several new and innovative mobile driven paradigms. One area that has started to receive attention in literature is mobile crowdsensing. Though promising, one specific aspect of mobile crowdsensing that has been mostly ignored in literature is human participation. In this paper, we take on this issue and present a novel probabilistic approach to predict whether we can expect a response from the crowd or not. The proposed candidate selection algorithm takes its inspiration from statistics and handles the typical uncertainties in human behavior. Validation of the proposed framework is done in two ways: 1) we implement a prototype and deploy it over an enterprise service bus; 2) we perform numerical testing on real datasets. With this experimental testbed, we show the efficacy of the framework in actual deployment scenarios.