QOATA: QoI-aware task allocation scheme for mobile crowdsensing under limited budget

The pervasiveness of mobile phones and the increasing sensing capabilities of their built-in sensors have made mobile crowdsensing a promising approach for large-scale data collection. In mobile crowdsensing, a specific situation is that the service provider (SP) needs to recruit contributors to fulfill sensing tasks requested by consumers. There are several challenges for contributor selection and task allocation in mobile crowdsensing. Firstly, since some of the contributors may need to travel a certain distance to complete the sensing task, they may need to be compensated for their contributions, in proportion to the distance they need to travel. Secondly, the diversity in sensing devices and contributor behavior means that the sensing data from different contributors will have different Quality of Information (QoI). Moreover, the SP usually only has a limited budget to compensate the contributors. Thus, how to handle the trade-offs between traveling distances of the contributors and QoI of the sensing data in order to maximize the sensing revenue based on limited budget is an important consideration in mobile crowdsensing applications. In practice, the QoI from different contributors are usually unknown to the SP in advance, which makes contributor selection and task allocation even more challenging. In this paper, we propose a QoI-aware task allocation scheme (QOATA) for contributor selection and task allocation in mobile crowdsensing applications. Using an online learning approach to learn the QoI from different contributors, the scheme aims to achieve the highest sensing revenue under a limited budget. The effectiveness of QOATA is evaluated through extensive simulations.