Research

Provisioning sensors-as-a-service in sensor-cloud

The Sensor-cloud architecture has depicted improvement over Wireless Sensor Networks (WSNs) for realizing Internet of Things (IoT) applications. Sensor-Cloud enables improved processing and management of sensed information, while supporting multiple users and heterogeneous applications in a single platform for a large-scale deployment of IoT. This framework provides a virtualized platform for offering sensors-as-a-service (Se-aaS), which facilitates dynamic management of sensing and computing resources by information sharing among the stakeholder entities such as Sensor-Cloud Service Provider (SCSP), sensor owners (SOs), and end-users. Using this framework, the SOs can achieve energy-efficiency of the deployed nodes, while the SCSP can provide services to various end-users, considering the user-specific requirements. The abstraction provided by the framework ensures secure information sharing among the SOs, and helps in better collaboration and coordination among the deployed nodes. In my PhD thesis, different schemes were designed for delivering Sensor-Cloud-based IoT platform to enable energy-efficient dynamic provisioning of the deployed sensors, while maintaining user- and service-specific Quality-of-Service (QoS) parameters. Specifically, the participation of the different stakeholders in the decision making process was considered while designing these schemes. The economic aspect of the framework was also studied, while devising the revenue distribution technique.

Architecture and localization in underwater sensor networks

Underwater Sensor Networks (UWSNs) pose challenges that differ from those of the terrestrial wireless sensor networks (WSNs) in many respects — passive node mobility is one of them. Due to the effect of passive node mobility, the network topology changes rapidly with time. This spatial variability of the network topology affects the connectivity between the sensor nodes, and the hop-to-hop data delivery based schemes face temporary losses of connectivity. As a result of this, the basic network functionalities, such as reporting of sensed data of the UWSN, are affected. Accordingly, an energy-efficient UWSN architecture, which is capable of providing communication guarantee between the source sensor nodes and the surface sinks was proposed.

In UWSNs, it is important to tag the sensed data with location information, in order to have better insight of the sensed information. Consequently, a sensor node needs to know its time-varying location. The existing localization schemes exhibit performance challenges such as high energy consumption, high localization error, reduced localization coverage, and high beacon message overhead. Motivated by this, a static anchor-based and a mobile anchor-based localization scheme exhibiting low energy consumption, high localization coverage and less location estimation error was proposed. The traditional approaches for iterative localization do not work specifically for sparse node deployment scenarios. In such scenarios, the sensor nodes lack the required number of anchor nodes for localization. Also, the mobile-anchor based schemes are deficient in achieving network-wide localization coverage in such scenarios. Accordingly, two localization schemes – a static anchor-based and a mobile anchor-based – was designed specifically considering the sparse UWSN scenario.