Context-Aware Task Distribution for Mobile Crowdsourcing
Maria Clara Pestana, Vaninha Vieira · 2018
In Mobile Crowdsourcing, tasks are distributed to workers according to sensor data, aiming to solve problems using collective intelligence. However, some tasks are not completed mainly because task subjects are not compatible to user capabilities. This article presents a context-aware approach to improve the task distribution in crowdsourcing systems. In this approach, tasks are selected to each worker according to the contexts previously defined. We developed a conceptual task model which uses sensitive information to select participants accordingly to task's contextual requirements and an architecture that illustrates the distribution of tasks through a mobile application. We also presented Contask, a prototype based on the developed architecture, which was used in a study case for evaluating task distribution. The accuracy of the distribution had a result of 63%, followed by 73% of precision and 63% of revocation. These results may be related to the ubiquity characteristic of context-aware computing since the system automatically adapts to user needs and preferences.