Joakim Flathagen

Constrained-based Multiple Sink Placement for Wireless Sensor Networks

A Wireless Sensor Network (WSN) consists of a large number of small, low-cost and low-power wireless sensing nodes. WSNs can gather information about the environment automatically and unattended and are suitable for many applications.The typical characteristic of WSNs is that they are energy and bandwidth constrained. Hence, routing protocols and algorithms for WSN must aim to conserve these two scarce resources. WSNs are also highly application-specific. This mean, firstly, that there is a tight bound between the application layer and the different protocol layers. Secondly, there are some WSN target applications that require certain protocol functionality that is not mandatory for other WSNs. In other words, both the general challenges and the specific application challenges must be addressed.This thesis aims to address routing in WSNs both from a general and an application specific perspective. Among the general energy-and bandwidth related topics the work in this thesis focuses on aggregation and routing-efficiency. Among the application-related topics the work focuses on localization and interoperability.The main contributions are:• A method for letting the routing protocol contribute in node localization.• A method for increasing the energy and bandwidth utilization with passive clustering.• A method for increasing the energy and bandwidth utilization using multiple sinks.• A data-aggregation scheme for WSNs that interoperates with external networks via a standardized interface.• A hybrid routing mechanism that are able to operate in high-interference scenarios.• Lessons learned from a real-world test campaign of a surveillance WSN.

O-CTP: Hybrid opportunistic collection tree protocol for Wireless Sensor Networks

Radio interference or deliberate jamming attacks can cause highly unpredictable communication in Wireless Sensor Networks (WSNs). Most prevalent WSN platforms consist of low-cost hardware with no effective measures against these threats. Most proposed countermeasures require a more advanced hardware design or radical changes to the 802.15.4 MAC protocol. These alternatives can be very difficult or even impossible to apply to existing WSN designs. In this paper we do not attempt to change the hardware or the MAC protocol. Instead we investigate how WSN routing protocols behave when the network is affected by interference. The paper proposes enhancements of CTP, the de-facto tree-based routing protocol for WSN, using opportunistic routing. We compare our approach with a wide range of protocols: CTP, TYMO, MultihopLQI, broadcast and geographic opportunistic routing in a real-life TelosB testbed subjected to different interference levels. The results show that our hybrid protocol, O-CTP, both improves the data delivery rate and reduces the cost when compared to standard routing protocols.

Pervasive service discovery across heterogeneous tactical networks

The Service Oriented Architecture concept is vital in order to interconnect different systems and to achieve a functional Network Centric Warfare (NCW). In order to realize an architecture based on loosely coupled services, a well-designed and efficient means of discovering the available services in the network is crucial. The service discovery mechanism used in any network must take the capabilities and limitations of the network into account. Due to the large variation in network capabilities on different operational levels, one single service discovery mechanism can not be chosen. Thus, there is a need for a toolkit consisting of different service discovery mechanisms so that each network can use the mechanism that is most suited for that particular network. In NCW, information exchange, and thus service discovery, must be available across network boundaries. This means that the different service discovery mechanisms must be able to interact with each other without the need for manual configuration. In this paper we present our novel prototype solution integrating WS-Discovery, a Web Services discovery mechanism, with Mercury, our proposed cross-layer service discovery solution tailored for use in MANETs. By using an interoperability gateway, service discovery is feasible across network boundaries, connecting mobile soldier systems and deployed tactical systems.

Integrating Wireless Sensor Networks in the NATO Network Enabled Capability using Web services

Wireless Sensor Networks (WSNs) are expected to provide greatly enhanced situational awareness for warfighters in the battlefield. Sensors widespread in the battlefield are however, of very limited value unless the sensors are reliable during the entire operation and the information produced is accessed in a timely manner. In this paper we focus on these issues by enabling WSNs as a capability in the NATO Network Enabled Capability (NNEC) using Web services. We demonstrate that Web services is an enabling technology for information-sharing, facilitating presentation of sensed data and alarms to a battlefield management system. In addition, we show the feasibility of using a Web services approach as a query processing tool enabling multi-sensor fusion and data aggregation in the WSN domain. The networking protocols can in this way inherently adjust data-aggregation and -processing criteria according to the requirements posed by external subscriber systems. In this way, energy efficiency, which is paramount in WSNs, is optimized without sacrificing the flexibility of Web services. Our proposed methods are tested using practical experiments with TelosB sensing nodes.

Increasing the Lifetime of Roadside Sensor Networks Using Edge-Betweenness Clustering

Wireless Sensor Networks are proven highly successful in many areas, including military and security monitoring. In this paper, we propose a method to use the edge-betweenness community detection algorithm to determine clusters and to facilitate in-network data aggregation for these applications. To minimize the cost of determining the clusters, the approach is based on exploiting the topology information from the ad hoc routing protocol. Three different schemes are proposed. (1) A distributed clustering scheme using the OLSR routing protocol. (2) A centralized scheme using OLSR. (3) A centralized scheme using an extension to the DYMO-low routing protocol. All schemes support sensor heterogeneity allowing that different data content can use different routing paths. The paper presents simulation results and analysis of the cluster generation for each of the schemes. The results show that our method is a simple and effective method to improve scalability and lifetime of roadside sensor networks.

Localization in wireless sensor networks based on Ad hoc routing and evolutionary computation

We propose evolutionary computation to estimate positions of nodes within a sensor network. The approach uses signal strength measurements between nodes and given positions for a subset of these nodes (anchor nodes). The signal strength measurements and routing requests take place simultaneously. A data collecting unit (sink node) receives distance estimates which are input to the evolutionary algorithm projecting node positions. This evolutionary approach can sort out data outliers and hence produce robust estimates of node positions. The present work contributes to decrease the cost and complexity of applying sensor networks. The approach also provides redundancy for the node positioning where alternative methods fail. The present simulations show examples of network generation and routing combined with estimation of node positions.