Norman Dziengel

A system for distributed event detection in wireless sensor networks

Event detection is a major issue for applications of wireless sensor networks. In order to detect an event, a sensor network has to identify which application-specific incident has occurred based on the raw data gathered by individual sensor nodes. In this context, an event may be anything from a malfunction of monitored machinery to an intrusion into a restricted area. The goal is to provide high-accuracy event detection at minimal energy cost in order to maximize network lifetime. In this paper, we present a system for collaborative event detection directly on the sensor nodes. The system does not require a base station for centralized coordination or processing, and is fully trainable to recognize different classes of application-specific events. Communication overhead is reduced to a minimum by processing raw data directly on the sensor nodes and only reporting which events have been detected. The detection accuracy is evaluated using a 100-node sensor network deployed as a wireless alarm system on the fence of a real-world construction site.

Cooperative event detection in wireless sensor networks

Event detection in wireless sensor networks is a sophisticated method for processing sampled data directly on the sensor nodes, thereby reducing the need for multihop communication with the base station of the network. In contrast to application-agnostic compression or aggregation techniques, event detection pushes application-level knowledge into the network. In-network event detection - especially the distributed form involving multiple sensor nodes - has thus an exceptional potential to increase energy efficiency, thus prolonging the lifetime of the network. In this article, we summarize recently proposed system architectures and algorithms employed for event detection in wireless sensor networks. On the example of the AVS-Extrem platform, we illustrate how energy-efficient event detection can be implemented through a combination of custom hardware design and distributed event detection algorithms. We then continue to present a brief evaluation of the detection accuracy and the energy consumption that is achievable by current systems.

Quality estimation based data fusion in wireless sensor networks

The main purpose of wireless sensor networks (WSNs) is to obtain information about their environment. However, WSNs often produce imprecise and incorrect sensor data, e.g. because of sensor failure or unreliable radio communication. We propose a system for WSN applications that allows to assess the quality of sensor data and further allows to fuse data based on their estimated quality. Our system comprises local and distributed heuristics to estimate the quality of sensor data, with a focus on data accuracy and data consistency. In the fusion step, the most plausible value of the measured quantity is inferred from multiple sensor readings by use of the Dempster-Shafer theory of evidence. Both quality assessment and data fusion are carried out within the network and thus do not rely on a powerful sink node. We demonstrate the effectiveness of our system by means of a wireless game controller for the game Pong, built from multiple sensor nodes. The controller can detect and reject incorrect sensor readings and thus improve the players control over the in-game paddle.

Deployment and evaluation of a fully applicable distributed event detection system in Wireless Sensor Networks

Versatility and real world applicability are key features of Wireless Sensor Networks (WSNs). In order to achieve these benefits we have to face the challenges of high practical relevance during application. We deploy and evaluate the concrete example of a fence monitoring task to reveal how our distributed event detection system is able to perform under real application conditions.The challenge is to bring both opposing aspects-high event detection accuracy and long service life-into one applicable ubiquitous system. If sensor nodes compose parts of events cooperatively, a comprehensive event assessment with low energy demands is possible. We propose a classifier based distributed event detection system consisting of two frameworks. The evaluation framework delivers a classification model and enables the theoretical evaluation of a given training set. The distributed event detection framework subsequently applies the classification model to assess, filter and classify events within the network. We evaluate our system by training and detecting events with our WSN composed of 49 nodes which are integrated in construction site fence elements.We compare four different data application scenarios with varying data processing concepts and varying network sizes to analyze the resulting communication load as well as the system lifetime. We compare the results of our evaluation framework with the results of our application to show that the evaluation framework reflects the real world deployment results in a credible way. We show the full applicability of our approach by comparing the resulting increased event detection accuracy against our previous work. Compared to other information fusion scenarios, our distributed event detection system reduces energy consumption beyond a communication distance of two hops which yields a prolonged lifetime of the network while additionally achieving an improved event detection accuracy.

Energy-aware distributed fence surveillance for wireless sensor networks

Fences are used all over the world to protect areas against unauthorized access. While most fences meet these requirements by building a physical and psychological barrier against intruders, this is not sufficient for areas of particular interest like restricted areas of an airport or construction sites with expensive goods. To detect intruders we integrate wireless sensor nodes into a fence. We fulfill the requirements for real-world energy-awareness by using active sensors with configurable logic, power saving modes like WOR and power down modes. The sensor nodes are capable of differentiating between several events, with the appliance of a distributed classification algorithm. We present an energy-aware platform for a distributed fence surveillance system in a wireless sensor network. We report on our experience in creating a platform that is concerning energy and performance demands and specifically tailored for the purpose of fence surveillance including fully integrated housing.

In-network training and distributed event detection in wireless sensor networks

In order to avoid transmitting raw data to a base station, sensor nodes are trained to cooperatively recognize deployment-specific events based on the data sampled by their sensors. As both training and event detection are performed without the need for central coordination or processing, only information about the detected event needs to be reported.

Towards Motion Characterization and Assessment Within a Wireless Body Area Network

The combination of small wireless sensor nodes and inertial sensors such as accelerometers and gyroscopes provides a cheap to produce ubiquitous technology module for human motion analysis. We introduce a system architecture for in-network motion characterization and assessment with a wireless body area network based on motion fragments. We present a segmentation algorithm based on biomechanics to identify motion fragments with a strong relation to an intuitive description of a motion. The system architecture comprises a training phase to provide reference data for segmentation, characterization and assessment of a specific motion and a feedback phase wherein the system provides the assessment related to the conduction of the motion. For fine-grained applicability, the proposed system offers the possibility of providing a motion assessment on three different evaluation layers during the motion assessment process. We evaluate the system in a first practical approach based on a dumbbell exercise.

A developer and a reference board for distributed motion evaluation in Wireless Sensor Networks

Wireless Sensor Networks continuously extend their practicability in real world applications and are passing their break-even to hit the consumer market. An increasing number of miniaturized wireless sensor nodes is acquirable, mostly intended to stream environmental data as well as tracking health with single node devices. A lack of distributed processing of sensor data within a cooperative network, a short lifetime or an extremely specialized design limit the use of wireless sensor nodes to specific applications. In this paper we target these drawbacks and introduce two consecutive hardware platforms - one platform for flexible development and a second miniaturized hardware platform for ambient integration in the environment. The goal is to enable applications for state-of-the-art Wireless Body Area Networks and Wireless Sensor Networks that fit the contradicting demands of a small form factor and durable network for distributed and cooperative processing. Furthermore, we focus with these platforms on optimization of energy consumption in order to extend the lifetime without maintenance. We support distributed movement evaluation with a multi-modal sensor technology for fine-grained motion tracking of a human body, for observing the structural health of buildings, for observing fences and borders or for any other versatile system.