engleski

Energy-efficient mobile crowd sensing in the Internet of Things domain

Mobile crowdsensing is a new paradigm that empowers ordinary citizens to collectively sense the environment by using their mobile devices and share data of common interest. Since MCS applications can cause significant energy consumption and quickly drain user’s battery, it is vital to achieve energy-efficient sensing and data transmission from mobile devices to the cloud so that valuable data is collected when it is indeed required by an MCS service. In this thesis we first provide a comprehensive overview of existing MCS applications, identify their unique features and define a functional MCS architecture which is mapped to existing IoT reference architectures. Then, we propose a mathematical model for MCS that describes the basic crowdsensing interactions. The proposed model is used to define the algorithms for energy-efficient data acquisition in both centralized and decentralized MCS systems. In particular, we present a centralized MCS ecosystem which utilizes a quality-driven sensor management function to continuously select the k-best workers for a predefined sensing task in the area of interest. The algorithm is designed to obviate redundant worker activity and consequently reduce overall system energy consumption. Our results show that by using the proposed algorithm the overall energy consumption for an MCS task can be significantly reduced if compared to a baseline approach which acquires all generated sensor data. Next, we propose a hierarchical MCS ecosystem which assumes the usage of edge computing resources to decentralize MCS services and improve their performance. The proposed architecture allows users to autonomously decide when to contribute data to the edge MCS service by using a Bloom filter structure. Finally, we extensively evaluate the proposed decentralized algorithm by using a real data set containing crowd sensed data. Our analysis shows that a Bloom filter structure is indeed applicable for MCS and its usage can greatly reduce energy consumption in decentralized MCS environment.

mobile crowdsensing ; MCS formal model ; energy-efficient data transmission ; QoS Manager ; Mobile Edge Computing ; Bloom filter ; decentralized algorithm

nije evidentirano