Shared Wireless Sensor Networks as Enablers for a Context Management System in Smart Cities

Wireless sensor networks (WSNs) are commonly used as a sensing infrastructure for smart city applications. A WSN is easy to use and can cover a wide area at low costs because of its wireless communication capability. The sensor nodes constituting a WSN are usually equipped with one or more sensor devices and can be used for different measurement purposes by reprogramming them. If WSNs could be shared by different smart city applications, they could be even more valuable enablers for smart cities. However, it is not easy to share WSNs. A shared WSN needs to support different kinds of measurement tasks at the same time and be able to accept new tasks at runtime. Even in a traditional closed WSN, its software should be carefully developed to satisfy certain quality requirements despite the severe resource constraints affecting the individual programmable sensor nodes (the sensor nodes of WSNs usually have quite limited resources, e.g. small batteries, low-spec CPU and narrow bandwidth). This issue is much harder to resolve in the case of a shared WSN. To satisfy the quality requirements of different applications, a WSN should be configured carefully according to specifications of the tasks, their quality requirements, and the environment, and should adapt its configuration in response to changes in the environment and the applications. A shared WSN should support various measurements, manage tasks at runtime and adapt to changes in the environment to reduce unnecessary consumption of resources. To develop such a shared WSN, we propose a middleware support for the network. In this chapter, we describe the architecture of our XAC middleware and the issues relevant to the shared WSN from the viewpoints of the task-description language, runtime task management and self-adaptation.