Wireless embedded systems

Wireless embedded systems combine sensors and actuators with processing, storage, and communication capabilities. Networks of small, lowpower wireless devices, so called nodes, offer the potential to collect observations of the physical world at unprecedented fidelity and scale. Time and location are of fundamental importance in the context of wireless sensor networks. Accurate time and location information are crucial for many tasks, e. g., sensor data fusion, spatiotemporal coordination, lowpower operation, and wireless medium access. In the first part of this volume, we study the problem of time synchronization. As clock sources in wireless embedded systems often exhibit severe drift, and message exchange is subject to a delay, sophisticated algorithms are mandatory to keep clocks in synchronization. Existing time synchronization algorithms are designed to provide networkwide synchronization, i. e., between arbitrary nodes in the network. However, large synchronization errors may become visible with increasing distance from a reference node, but also closeby nodes in a network may be synchronized poorly. To address these issues, we propose two clock synchronization protocols: the Gradient Time Synchronization Protocol to achieve local synchronization, and the PulseSync protocol for global synchronization. We evaluate the performance of both protocols by simulations and experiments in different wireless sensor network testbeds. The second part of this volume is dedicated to the question how we can provide accurate location information in the context of wireless embedded systems. We present the SpiderBat platform, which provides node localization for sensor nodes using ultrasound pulses. By employing multiple transmitters and receivers, we can estimate the distance and angle between neighboring nodes. Using both distance and angle information allows to localize nodes with an accuracy of a few centimeters, even in sparse networks where other approaches fail. In the third part, we discuss design considerations for the architecture of future wireless embedded systems, as they have evolved from pure research tools to ubiquitous smart things. To this end, we present a resourceoriented approach to facilitate the connection between heterogeneous devices and services, and a novel wearable sensor platform that utilizes mobile phones for personalized sensing applications.