Recruitment selection strategies for crowdsourced sensing

Participatory sensing is a new paradigm that is based on the data collection, such as images, air and noise measurements, through the embedded digital sensors and radios on current smartphones owned by voluntary users. The participants in such a system can either be self-motivated or persuaded through incentives with amounts depending on factors such as quality of the contributed data or sensors, availability, reputation, urgency and priority. In this paper, the problem of selecting a subset of all users to gain the maximum merit is studied in the context of urban sensing optimization. Accordingly, an efficient participant selection, i.e. recruitment, method is proposed that runs online, i.e. without referring to high computational complexity and prohibitive runtime. The performance of our proposal is compared to random selection schemes and observed to outperform the random method in terms of query merits.