A Systematic Analysis of Current Architectures and Apps in the Field of Mobile Crowdsensing

The broad usage of mobile devices, along with the growing number of sensors embedded in them, has considerably facilitated the implementation of crowdsensing systems. Therefore, Mobile Crowdsensing has received noteworthy attention throughout the years, which has lead to an increasing number of publications with different approaches and solutions. These solutions require further examination and classification, since there are not many scientific papers which have analyzed a large number of systems. This thesis performs a systematic analysis of current architectures and apps in the field of Mobile Crowdsensing, providing a comprehensive overview. The reviewed literature was retrieved from Google Scholar, PubMed, PsycINFO and Web of Science in accordance with the PRISMA guidelines. This resulted in 158 unique records on Mobile Crowdsensing systems, as well as 292 distinct records on Ecological Momentary Assessment apps for further analysis. There are 93 different Mobile Crowdsensing systems, whereas 69 distinct Ecological Momentary Assessment apps were determined. A detailed overview of the years in which the literature was published is presented, indicating a growing research trend in recent years. Subsequently, descriptive data about the systems, such as participatory and opportunistic sensing, website, GitHub, Ecological Momentary Intervention as well as current availability in the Google Play Store and Apple App Store is provided. Additionally, operating systems, used sensors for data acquisition, databases, computation including scalability and context-awareness are also presented. Furthermore, assessment types, e.g., signal-contingent, interval-contingent and event-contingent, followed by notification methods, e.g., push notification, auditory alerts, vibrations, SMS and flash of the camera are described. In addition, reminders, response limit and adjustable prompting are presented as well. Moreover, different incentive mechanisms are inspected, followed by categorization of use cases. Lastly, certain systems are described in more detail.