Mobile Crowdsensing with Mobile Agents: JAAMAS Extended Abstract

Teemu Leppänen (University of Oulu), Jóse Álvarez Lacasia (University of Tokyo), Yoshito Tobe (Aoyama Gakuin University), Kaoru Sezaki (University of Tokyo), Jukka Riekki (University of Oulu)

Abstract

Mobile crowdsensing applications can be designed as multiagent systems with campaign-specific roles, realized as mobile agents, and role-based interactions. Mobile agents facilitate decentralized and autonomous crowdsesing campaign execution that takes into account dynamic resource availability in the system. Mobile agents address issues in collaborative campaign execution, campaign monitoring, data collection, data analysis and handling participant-related issues. A software framework is presented that seamlessly exposes campaigns and mobile agents as Web resources. Real-world evaluation shows comparable results with existing crowdsensing approaches but with smaller energy consumption. In-network data sharing introduces insignificant overhead.