Peter van der Stok

Adaptive Streaming of MPEG-based Audio/Video Content over Wireless Networks

This paper describes a method for robust streaming of combined MPEG audio/video content (encoded either with MPEG-2 or MPEG-4/H.264) over in-home wireless networks. We make use of currently used content distribution format (MPEG Transport Stream) and network protocols (either RTP or TCP/HTTP). The transmitted bit-rate is constantly adapted to the available network bandwidth, such that audio and video artifacts caused by packet loss are avoided. Bit-rate adaptation is achieved by using a packet scheduling technique called I-Frame Delay (IFD), which performs priority-based frame dropping upon insufficient bandwidth. We show an implementation using RTP and an implementation using TCP. Measurements on a real-life demonstrator set-up demonstrate the effectiveness of our approach.

Accurate Timeliness Simulations for Real-Time Wireless Sensor Networks

The use of wireless sensor networks is rapidly growing in various types of applications that benefit from spatially distributed data collection. Some of these applications, such as industrial automation, fire detection or health monitoring, have strong timeliness constraints. Since field deployments are difficult to monitor and debug, the development of real-time communication protocols for wireless sensor networks necessitates accurate simulation models. This paper presents open source Omnet++ simulation models based on the IEEE 802.15.4 standard. Four evaluation scenarios are used to compare simulation timeliness and packet error rate results with experimental measurements. The small scale of the scenarios allows to isolate the effect of each system component. The comparison validates the models for timeliness estimates in sensor networks and pinpoints the variability of software implementations in embedded systems as a major cause of differences between simulated and measured results.

QoS-based resource management for ambient intelligence

Future homes will probably be equipped with in-home networks, combining a backbone of wired networks and a large number of devices connected through wireless links, to provide a responsive and supportive environment, known as Ambient Intelligence. Many of the applications provided to the user lean heavily on media processing and streaming data. Therefore it is expected that consumer terminals will play an important role in providing new experiences to the users. Most of the devices have to be very cost and power effective, while digital media processing is able to consume all the resources a device can offer, and more. Typically, the number of applications and the resource needs of the applications change over time. To adapt to these variations, applications have the ability to trade resource usage for quality of service (QoS). QoS based resource management enables these tradeoffs in resource-constrained systems. In this paper we present our QoS approach, and we explore an integrated approach that addresses terminal and network resources, and takes power issues into account.

Efficient time synchronization for wireless sensor networks in an industrial setting

This paper outlines an efficient variation of the Flooding Time Synchronization Protocol (FTSP) on real hardware with an IEEE 802.15.4 MAC layer. The paper compares the performance of two clock drift estimation techniques: Least Squares Linear Regression and the Kalman filter. In assessing the viability of time synchronization in an industrial setting, our reliance on real hardware gives confidence for the deployment in future products. The concrete result is a high time synchronization accuracy, with an achievable average error of 0.4 μs, and a repetition interval of a few seconds, while using a star network topology.

Distributed Task Synchronization in Wireless Sensor Networks

A WSN is envisaged to consist of a large number of nodes, each combining sensing, processing and wireless communication capabilities. The nodes have to cooperate to achieve a global function required by the user. Towards achieving this goal, multiple tasks are performed on all nodes or a subset of nodes in the network. The subject of this paper is the synchronized execution of periodic tasks at different nodes, for example: sensing tasks or a periodic on/off switching of radio transceivers. The method is based on synchronizing timer interrupt occurrences at different nodes to a single reference, leading to tight synchronization of tasks over nodes. Measurements show that task synchronization is achieved on our hardware platform with an average time difference of 5 μ s between the starts of temporally related tasks on different nodes.

WASP – Wirelessly Accessible Sensor Populations: A Project Overview

This article presents an overview of the integrated project WASP that deals with the research and development of Wirelessly Accessible Sensor Populations. WASP is a European project of the Embedded Systems Strategic Objective in Framework Programme 6 of the European Union. WASP is covering the complete technology chain from sensor node hardware to the implementation of applications. The article illustrates the integrated approach of WASP by subsuming the variety of research and development issues that are addressed to reach the ambitious project goal: the provision of a complete system view for building large populations of collaborating objects.

Performance evaluation of network coding and packet skipping in IEEE 802.15.4-based real-time wireless sensor networks

In a number of application domains, the volatility of the monitored environment where Wireless Sensor Networks (WSNs) operate engenders timing constraints on the generation, processing, and communication of sensory data. In this paper, we primarily focus on the use of an information theoretic approach, namely network coding, to improve the on-time delivery of messages in IEEE 802.15.4-compliant networks. We further study the real-time gain that packet skipping can provide on its own and in combination with network coding. Subsequently, we investigate the potential benefits of introducing deadline-awareness into the coding mechanism through the use of an Earliest Deadline First (EDF) - based scheduling policy. Simulation results show that network coding and packet skipping are by themselves effective techniques to increase on-time goodput, with additional gain obtained when the two techniques are used concurrently. Our results further provide insight into the performance of deadline-aware scheduling in a network coding environment.

A Framework for Time-Controlled and Portable WSN Applications

Body sensor network applications require a large amount of data to be communicated over radio frequency. The radio transceiver is typically the largest source of power dissipation; improvements on energy consumption can thus be achieved by enabling on-node processing to reduce the number of packets to be transmitted. On-node processing is facilitated by a timely control over process execution to sequence operations on data; yet, the latter must be enabled while keeping high-level software abstracted from both underlying software and hardware intricacies to accommodate portability to the wide range of hardware and software platforms. We investigated the challenges of implementing software services for on-node processing and devised constructs and system abstractions that integrate applications, drivers, time synchronization and MAC functionality into a system software which presents limited dependency between components and enables timely control of processes. We support our claims with a performance evaluation of the software tools implemented within the FreeRTOS micro-kernel.

Adaptive streaming of combined audio/video content over wireless networks

This paper describes a method for robust streaming of combined MPEG-2 audio/video content over in-home wireless networks. We make use of currently used content distribution formats and network protocols. The transmitted bit-rate is constantly adapted to the available network bandwidth, such that audio and video artifacts caused by packet loss are avoided. Bit-rate adaptation is achieved by using a packet scheduling technique called I-Frame Delay (IFD), which performs priority-based frame dropping upon insufficient bandwidth. We show an implementation using RTP and an implementation using TCP. Measurements on a real-life demonstrator set-up demonstrate the effectiveness of our approach.

Packet skipping and network coding for delay-sensitive network communication

We provide an analytical study of the impact of packet skipping and opportunistic network coding on the timely communication of messages through a single network element. In a first step, we consider a single-server queueing system with Poisson arrivals, exponential service times, and a single buffer position. Packets arriving at a network node have a fixed deadline before which they should reach the destination. To preserve server capacity, we introduce a thresholding policy, based on remaining time until deadline expiration, to decide whether to serve a packet or skip its service. The obtained goodput improvement of the system is derived, as well as the operating conditions under which thresholding can enhance the performance. Subsequently, we focus our analysis on a system that supports network coding instead of thresholding. We characterize the impact of network coding at a router node on the delivery of packets associated with deadlines. We model the router node as a queueing system in which packets arrive from two independent Poisson flows and undergo opportunistic coding operations. We obtain an exact expression for the goodput of the system and study the achievable gain. Finally, we provide an analytical model that considers both network coding and packet skipping, capturing their joint performance. A comparative analysis between the aforementioned approaches is provided.