Yurong Xu

Detecting Wormhole Attacks in Wireless Sensor Networks

The development of wireless sensor devices in terms of low power and inexpensive data-relaying has been partially achieved because of the rapid progress in integrated circuits and radio transceiver designs and device technology. Due to these achievements, the wireless sensor devices are able to gather information, process them if required and send them to the next sensor device. The captured information might be regarding temperature, pressure or light, and so on. In some applications, these wireless sensor devices must be secured, especially when the captured information is valuable, sensitive or for military usage. Wormhole attacks are a significant type of security attacks which can damage the wireless sensor networks if they go undetected. Unfortunately, these attacks are still possible, even if the communication is secured. The wormhole attack records packets at one point of the network, passes them into another node and this last node injects the packet into the wireless sensor network again. The main objective of this project is to build an actual test bed to simulate the wormhole attack on a wireless sensor network and to implement one of the current solutions which is called packet leashes to protect sensor networks from these kinds of attacks. This project’s test bed consists of a combination of Mica 2 motes and Stargate sensor devices.

Mobile anchor-free localization for wireless sensor networks

In this paper, we consider how to localize individual nodes in a wireless sensor network when some subset of the network nodes can be in motion at any given time. For situations in which it is not practical or cost-efficient to use GPS or anchor nodes, this paper proposes an Anchor-Free Mobile Geographic Distributed Localization (MGDL) algorithm for wireless sensor networks. Taking advantage of the accelerometers that are present in standard motes, MGDL estimates the distance moved by each node. If this distance is beyond a threshold, then this node will trigger a series of mobile localization procedures to recalculate and update its location in the node itself. Such procedures will be stopped when the node stops moving. Data collected using Tmote Invent nodes (Moteiv Inc.) and simulations show that the proposed detection method can efficiently detect the movement, and that the localization is accurate and the communication is efficient in different static and mobile contexts.

A wireless sensor network architecture and its application in an assistive environment

Todays assistive environments place pervasive technologies in the patients living area in order to provide therapeutic or assistive type of human monitoring [13]. One type of pervasive technology commonly used is using wireless sensor motes, which can be deployed in an indoor environment. In this paper, we propose a system architecture for placing a wireless sensor network (WSN) in a home; sensor nodes are placed into different private rooms where people are intended to live. Using photocells sensors in the nodes, our system can detect the movement of people inside of the rooms. Our system also includes a web-based interface which is used to visualize the data flow in our system in real-time. We call such a WSN implementation Assistive Home Monitoring, or AtHoM. This paper studies how we can combine the WSNs architecture with a WWW application to create a system that can visualize the data remotely by reacting to the sensors deployed in the environment. This paper describes an architecture to integrate the Web with Wireless Sensor Networks for an assistive environment and how it was implemented in the context of experiments. We provide a detailed description of the architecture of the AtHoM system we designed by basing our experiments on a real-world apartment, the Heracleia Apartment., that exists inside our laboratory, the Human Centered Computing Lab. We describe two different experiments implemented within the Heracleia apartment. The aim of these experiments is to validate localization algorithms and all related measurements. From the first experiment and the two subsequent applications we implemented in our apartment test bed, we found that the integration of WWW and WSNs can make the sharing of data collected by WSNs easier for people to access and visualize.

Mobile device protection from loss and capture

Mobile devices play a critical role in assistive environments. How to authenticate and secure communications among them has become more important especially against loss and capture of the devices. In this paper, we present an approach to protect signing keys of mobile devices based on mediated RSA introduced by Dan Boneh and others. The important property of our scheme is transparent self-resilience. In other words, in case a device is lost or captured and at the risk of being compromised and impersonated, our scheme can disable the device instantly and the replacement will be transparent to other users. In this way, if an attacker captures a mobile device, he has limited time to use it because it will become soon invalid. If he wants to break our scheme, he must compromise the device and its mediator simultaneously.

Preventing unofficial information propagation

Digital copies are susceptible to theft and vulnerable to leakage, copying, or manipulation. When someone (or some group), who has stolen, leaked, copied, or manipulated digital documents propagates the documents over the Internet and/or distributes those through physical distribution channels many challenges arise which document holders must overcome in order to mitigate the impact to their privacy or business. This paper focuses on the propagation problem of digital credentials, which may contain sensitive information about a credential holder. Existing work such as access control policies and the Platform for Privacy Preferences (P3P) assumes that qualified or certified credential viewers are honest and reliable. The proposed approach in this paper uses short-lived credentials based on reverse forward secure signatures to remove this assumption and mitigate the damage caused by a dishonest or honest but compromised viewer.

BP-neural network based-characterization of electrographic magnetohydrodynamic signals in MR

Electrocardiographs (ECG) signal collected during magnetic resonance (MR) imaging is affected by signal artifact because magnetic fields produce competing signals, from moving conductors in the large vessels. That is called the magnetohydrodynamic effect, which makes it difficult to recognize ST-T changes from ECG signal collected in a magnetic field (MRI). Resolving that problem is important both for accurate triggering (elimination of false triggers from tall peaked T waves) and for monitoring (identifying if or when patient develops ischemia or myocardial injury). This work describes an algorithm based on neural network that is designed to cancel this artifact for ECG signal acquired during MR imaging.

Requirements for implementation of localization into real-world assistive environments

Accurate and efficient localization methods in sensor networks are critical to enabling a robust assistive environment where tracking human actions and interactions are needed to predict human behavior and prevent accidents. In this paper we describe an anchor-free localization approach where the sensor motes themselves determine their location without any given starting point or additional hardware. Instead, the location is discovered by allowing sensors to branch out through their connections to each other to establish maps that define their surroundings. We describe a Geographical Distributed Localization (GDL) algorithm which consists of a set of motes that compute local maps based on their hop counts from a special mote called bootstrap. In this paper, we provide a set of requirements for real world conditions, since GDL was developed and tested using the NS2 simulation system using synthetic data. It is now desired to test GDL in a real world assistive environment and generate a set of requirements that are useful in this and other settings. To do this, we chose Tmote Invent wireless sensors and designed ways to transfer the system from simulation to laboratory. Later, we used SunSPOT motes to continue the system. In this paper we report on specific features and requirements discovered that need to be taken into consideration to account for physical limitations of the sensors, when trying to move the system from one environment to another. Also, we provide new directions of research when mapping sensor localization to real-world environments, based on the given resources and the components available.

A Feature Extraction Metod for Multimedia data Analysis in Robot Wireless Sensor Networks

Multimedia sensor networking technologies have developed very rapidly in the last ten years. Robot wireless sensor networks are an increasingly capable platform and show great potential in environments from future battlefields to disaster discovery. Robot wireless sensor networks can enhance multimedia surveillance and provide the foundation for strategies based on multi-modal sensor integration. We illustrate this concept with an example feature extraction approach for image or video data: lines are extracted as image features from images transformed using the Hough transform technique. This approach enables features to be combined and constrained by their structure, allowing for example region correspondence mapping between images from different sensors or the same sensor in different time phases.

A BP-Neural Network Improvement to Hop-Counting for Localization in Wireless Sensor Networks

GDL (Geographic Distributed Localization) is a distributed localization technique useful for letting nodes in a network determines their relative locations without special hardware or configuration. As compared with previous proposed measurement techniques, which require additional hardware (such as special antennas), this technique is based on hop counting and only requires the connectivity of network. In this paper, we propose to incorporate neural networks into the standard hop counting-based approach to improve the accuracy of localization over the existing method. We use one neural network at each node to compute its hop-coordinates, after a training procedure. Using network simulations, we show that incorporating our technique improves the accuracy of localization results generated by GDL by 15%. The application of this method is not limited to the GDL algorithm; it can also be used in other hop-counting based localization algorithms, such as MDS-MAP. It is also designed to be suitable for implementation in current embedded systems, such as standard motes like Moteiv’s Invent.

An adaptive approach for image subtraction

Image subtraction is widely used in angiography as a means of highlighting differences induced by contrast agents. New knowledge of previously unsuspected causes of disease, in particular, secondhand smoke exposure, spurs interest in pushing the limits of early accurate diagnosis. Simple image subtraction induces artifacts causing problems for ensuing measurements and 3D reconstruction. Image registration techniques have been used to partially solve this problem. However, a complete registration is slow, and misregistration often occurs in images where bones are surrounded by vessels with similar image characteristics. We propose an approach based on the idea of global match followed by local refinements. In the global match, an image pair is aligned using a similarity measure so as to reduce overall difference. In the local refinements, localized displacements and deformations of tissue are handled by a combination of techniques: image registration, region growing, erosion, and dilation. This approach is fast compared to registration based image subtraction and it can find vessels abutting a bone. It is designed to be especially suitable for large cross-section image stacks. With additional vessel connectivity analysis between adjacent slices, the algorithm provides a good foundation for 3D vessel reconstruction.