Joengmin Hwang

Revisiting random key pre-distribution schemes for wireless sensor networks

Key management is one of the fundamental building blocks of security services. In a network with resource constrained nodes like sensor networks, traditional key management techniques, such as public key cryptography or key distribution center (e.g., Kerberos), are often not effective. To solve this problem, several key pre-distribution schemes have been proposed for sensor networks based on random graph theory. In these schemes, a set of randomly chosen keys or secret information is pre-distributed to each sensor node and a network is securely formed based on this information. Most of the schemes assumed that the underlying physical network is dense enough, that is, the degree of each node is hig. In this paper, we revisit the random graph theory and use giant component theory by Erdos and Renyi to show that even if the node degree is small, most of the nodes in the network can be connected. Further, we use this fact to analyze the Eschenhauer et. als, Du et. als, and Chan et. als key pre-distribution schemes and evaluate the relation between connectivity, memory size, and security. We show that we can reduce the amount of memory required or improve security by trading-off a very small number of isolated nodes. Our simulation results show that the communication overhead does not increase significantly even after reducing the node degree. In addition, we present an approach by which nodes can dynamically adjust their transmission power to establish secure links with the targeted networked neighbors. Finally, we propose an effcient path-key identification algorithm and compare it with the existing scheme.

Exploring in-situ sensing irregularity in wireless sensor networks

The circular sensing model has been widely used to estimate performance of sensing applications in existing analyses and simulations. While this model provides valuable high-level guidelines, the quantitative results obtained may not reflect the true performance of these applications, due to the sensing irregularity introduced by existence of obstacles in real deployment areas and insufficient hardware calibration. In this paper, we design and implement two sensing area modeling (SAM) techniques useful in the real world. They complement each other in the design space. Physical sensing area modeling (P-SAM) provides accurate physical sensing area for individual nodes using controlled or monitored events, while virtual sensing area modeling (V-SAM) provides continuous sensing similarity between nodes using natural events in an environment. With these two models, we pioneer an investigation of the impact of sensing irregularity on application performance, such as coverage scheduling. We evaluate SAM extensively in real-world settings, using testbeds consisting of 14 XSM motes. To study the performance at scale, we also provide an extensive 1,400-node simulation. Evaluation results reveal several serious issues concerning circular models and demonstrate significant improvements in several applications when SAM is used instead.

Detecting Phantom Nodes in Wireless Sensor Networks

In an adversarial environment, various kinds of security attacks become possible if malicious nodes could claim fake locations that are different from where they are physically located. In this paper, we propose a secure localization mechanism that detects the existence of these nodes, termed as phantom nodes, without relying on any trusted entities, an approach significantly different from the existing ones. The proposed mechanism enjoys a set of nice features. First, it does not have any central point of attack. All nodes play the role of verifier, by generating local map, i.e. a view constructed based on ranging information from its neighbors. Second, this distributed and localized construction results in quite strong results: even when the number of phantom nodes is greater than that of honest nodes, we can Alter out most phantom nodes. Our analysis and simulations under realistic noisy settings demonstrate our scheme is effective in the presence of a large number of phantom nodes.

uSense: A Unified Asymmetric Sensing Coverage Architecture for Wireless Sensor Networks

As a key approach to achieve energy efficiency in sensor networks, sensing coverage has been studied extensively. Researchers have designed many coverage protocols to provide various kinds of service guarantees on the network lifetime, coverage ratio and detection delay. While these protocols are effective, they are not flexible enough to meet multiple design goals simultaneously. In this paper, we propose a unified sensing coverage architecture, called uSense, which features three novel ideas: asymmetric architecture, generic switching and global scheduling. We propose asymmetric architecture based on the conceptual separation of switching from scheduling. Switching is efficiently supported in sensor nodes, while scheduling is done in a separated computational entity, where multiple scheduling algorithms are supported. As an instance, we propose a two-level global coverage algorithm, called uScan. At the first level, coverage is scheduled to activate different portions of an area. We propose an optimal scheduling algorithm to minimize area breach. At the second level, sets of nodes are selected to cover active portions. Importantly, we show the feasibility to obtain optimal set-cover results in linear time if the layout of areas satisfies certain conditions. We evaluate our architecture with a network of 30 MicaZ motes, an extensive simulation with 10,000 nodes, as well as theoretical analysis. The results indicate that uSense is a promising architecture to support flexible and efficient coverage in sensor networks.

Secure localization with phantom node detection

In an adversarial environment, various kinds of attacks become possible if malicious nodes could claim fake locations that are different from their physical locations. In this paper, we propose a secure localization mechanism that detects existence of these nodes, termed as phantom nodes, without relying on any trusted entities, an approach significantly different from the existing ones. The proposed mechanism enjoys a set of nice features. First, it does not have any central point of attack. All nodes play the role of verifier, by generating local map, i.e. a view constructed based on ranging information from its neighbors. Second, this distributed and localized construction results in strong robustness against adversaries: even when the number of phantom nodes is greater than that of honest nodes, we can filter out most of the phantom nodes. Our analytical results as well as simulations under realistic noisy settings demonstrate that the proposed mechanism is effective in the presence of a large number of phantom nodes.

Realistic Sensing Area Modeling

Despite the well-known fact that sensing patterns in reality are highly irregular, researchers continue to develop protocols with simplifying assumptions about the sensing. For example, a circular 0/1 sensing model is widely used in most existing simulators and analysis. While this model provides high-level guidelines, it could cause wrong estimation of system performance in the real world. In this project, we design and implement a practical Sensing Area Modeling technique, called SAM. By injecting events through regular and hierarchical training, SAM estimates the sensing areas of individual sensor nodes accurately. Especially, this work is the first to investigate the impact of irregular sensing area on application performance, such as coverage scheduling. We evaluate SAM using outdoor experiments with XSM motes, indoor experiment with 40 MicaZ motes as well as an extensive 1000-node simulation. Our evaluation results reveal serious problems caused by circular sensing model, while demonstrating significant performance improvements in major applications when SAM is used.

Exploring In-Situ Sensing Irregularity in Wireless Sensor Networks

The circular sensing model has been widely used to estimate performance of sensing applications in existing analyses and simulations. While this model provides valuable high-level guidelines, the quantitative results obtained may not reflect the true performance of these applications, due to the sensing irregularity introduced by existence of obstacles in real deployment areas and insufficient hardware calibration. In this paper, we design and implement two sensing area modeling (SAM) techniques useful in the real world. They complement each other in the design space. Physical sensing area modeling (P-SAM) provides accurate physical sensing area for individual nodes using controlled or monitored events, while virtual sensing area modeling (V-SAM) provides continuous sensing similarity between nodes using natural events in an environment. With these two models, we pioneer an investigation of the impact of sensing irregularity on application performance, such as coverage scheduling. We evaluate SAM extensively in real-world settings, using testbeds consisting of 14 XSM motes. To study the performance at scale, we also provide an extensive 1,400-node simulation. Evaluation results reveal several serious issues concerning circular models and demonstrate significant improvements in several applications when SAM is used instead.

Energy efficient organization of mobile sensor networks

One of main design issues for a sensor network is conservation of energy available at each sensor node. To increase lifetime of a sensor network, we can organize the sensors into disjoint sets so that only one set is active at a time. The lifetime of the network increases proportionally to the number of disjoint sets. In general, sensors are randomly scattered in the monitored area, thus the number of disjoint sets is significantly smaller than in the ideal case. In this paper, we propose a way to increase the number of disjoint set using mobile sensors. We organize sensors into disjoint sets using the heuristic proposed by Sljjepcevic et al.[8]. Then, we rearrange mobile sensors that are not included in any set. The rearrangement process consists of two phases. In the first phases, we identify the fields that are not covered by any of the remaining sensors. Then we identify the locations from which mobile sensors could cover the fields. In the second phase, our proposed heuristic selects sensors to be rearranged and locations where they will move to. This selection is made by considering the coverage before and after a mobile sensor is moved. Our experiments show that we can effectively increase the number of disjoint set with a small number of mobile sensors rearranged.

A framework for discovering spatio-temporal cohesive networks

A spatio-temporal cohesive network represents a social network in which people often interact closely in both space and time. Spatially and temporally close people tend to share information and show homogeneous behavior.We discuss modeling social networks from spatiotemporal human activity data, and alternative interest measures for estimating the strength of subgroup cohesion in spatial and temporal space. We present an algorithm for mining spatio-temporal cohesive networks.