Vincent Tam

A distributed multihop time synchronization protocol for wireless sensor networks using Pairwise Broadcast Synchronization

Recently, a time synchronization algorithm called pairwise broadcast synchronization (PBS) is proposed. With PBS, a sensor can be synchronized by overhearing synchronization packet exchange among its neighbouring sensors without sending out any packet itself. In an one-hop sensor network where every node is a neighbour of each other, a single PBS message exchange between two nodes would facilitate all nodes to synchronize. However, in a multi-hop sensor network, PBS message exchanges in several node pairs are needed in order to achieve network-wide synchronization. To reduce the number of message exchanges, these node pairs should be carefully chosen. In this paper, we investigate how to choose these ldquoappropriaterdquo sensors aiming at reducing the number of PBS message exchanges while allowing every node to synchronize. This selection problem is shown to be NP-complete, for which the greedy heuristic is a good polynomial-time approximation algorithm. Nevertheless, a centralized algorithm is not suitable for wireless sensor networks. Therefore, we develop a distributed heuristic algorithm allowing a sensor to determine how to synchronize itself based on its neighbourhood information only. The protocol is tested through extensive simulations. The simulation results reveal that the proposed protocol gives consistent performance under different conditions with its performance comparable to that of the centralized algorithm.

Using Micro-Genetic Algorithms to Improve Localization in Wireless Sensor Networks

Wireless sensor networks are widely adopted in many location-sensitive applications including disaster management, environmental monitoring, military applications where the precise estimation of each node position is inevitably important when the absolute positions of a relatively small portion as anchor nodes of the underlying network were predetermined. Intrinsically, localization is an unconstrained optimization problem based on various distance/path measures. Most of the existing localization methods focus on using different heuristic-based or mathematical techniques to increase the precision in position estimation. However, there were recent studies showing that nature-inspired algorithms like the ant-based or genetic algorithms can effectively solve many complex optimization problems. In this paper, we propose to adapt an evolutionary approach, namely a micro-genetic algorithm, as a post-optimizer into some existing localization methods such as the Ad-hoc Positioning System (APS) to further improve the accuracy of their position estimation. Obviously, our proposed MGA is highly adaptable and easily integrated into other localization methods. Furthermore, the remarkable improvements attained by our proposed MGA on both isotropic and anisotropic topologies of our simulation tests prompt for several interesting directions for further investigation.

Removing Node Overlapping in Graph Layout Using Constrained Optimization

Although graph drawing has been extensively studied, little attention has been paid to the problem of node overlapping. The problem arises because almost all existing graph layout algorithms assume that nodes are points. In practice, however, nodes may be labelled, and these labels may overlap. Here we investigate how such node overlapping can be removed in a subsequent layout adjustment phase. We propose four different approaches for removing node overlapping, all of which are based on constrained optimization techniques. The first is the simplest. It performs the minimal linear scaling which will remove node-overlapping. The second approach relies on formulating the node overlapping problem as a convex quadratic programming problem, which can then be solved by any quadratic solver. The disadvantage is that, since constraints must be linear, the node overlapping constraints cannot be expressed directly, but must be strengthened to obtain a linear constraint strong enough to ensure no node overlapping. The third and fourth approaches are based on local search methods. The third is an adaptation of the EGENET solver originally designed for solving general constraint satisfaction problems, while the fourth approach is a form of Lagrangian multiplier method, a well-known optimization technique used in operations research. Both the third and fourth method are able to handle the node overlapping constraints directly, and thus may potentially find better solutions. Their disadvantage is that no efficient global optimization methods are available for such problems, and hence we must accept a local minimum. We illustrate all of the above methods on a series of layout adjustment problems.

A comparative study of centroid-based, neighborhood-based and statistical approaches for effective document categorization

Associating documents to relevant categories is critical for effective document retrieval. Here, we compare the well-known k-nearest neighborhood (kNN) algorithm, the centroid-based classifier and the highest average similarity over retrieved documents (HASRD) algorithm, for effective document categorization. We use various measures such as the micro and macro F1 values to evaluate their performance on the Reuters-21578 corpus. The empirical results show that kNN performs the best, followed by our adapted HASRD and the centroid-based classifier for common document categories, while the centroid-based classifier and kNN outperform our adapted HASRD for rare document categories. Additionally, our study clearly indicates that each classifier performs optimally only when a suitable term weighting scheme is used All these significant results lead to many exciting directions for future exploration.

Localization in Sensor Networks with Limited Number of Anchors and Clustered Placement

Many localization algorithms have been proposed in recent years. Although different algorithms based on different methodologies, the use of anchors is common to most algorithms. The placement and the density of anchors affect the accuracy of different algorithms to different extent. Location estimates are usually more accurate with a higher density of anchors. When there are only a few anchors, efficient algorithms tend to perform poorly. However, having more anchors will increase the cost of a sensor network. In this paper, we present an algorithm which uses two different localization techniques, multidimensional scaling (MDS) and proximity-distance map (PDM), in a phased approach. MDS has a high complexity but can give good results when there are only very few anchors. PDM, on the other hand, is a distributed algorithm but performs poorly when anchors are scarce. The phased approach has comparable complexity to PDM but less than MDS. With extensive simulations, we demonstrate that the proposed algorithm gives accurate solution with very few anchors or clustered anchors which is intrinsically a difficult challenge to most existing algorithms.

IMPROVING EVOLUTIONARY ALGORITHMS FOR EFFICIENT CONSTRAINT SATISFACTION

Hard or large-scale constraint satisfaction and optimization problems, occur widely in artificial intelligence and operations research. These problems are often difficult to solve with global search methods, but many of them can be efficiently solved by local search methods. Evolutionary algorithms are local search methods which have considerable success in tackling difficult, or ill-defined optimization problems. In contrast they have not been so successful in tackling constraint satisfaction problems. Other local search methods, in particular GENET and EGENET are designed specifically for constraint satisfaction problems, and have demonstrated remarkable success in solving hard examples of these problems. In this paper we examine how we can transfer the mechanisms that were so successful in (E)GENET to evolutionary algorithms, in order to tackle constraint satisfaction algorithms efficiently. An empirical comparison of our evolutionary algorithm improved by mechanisms from EGENET and shows how it can markedly improve on the efficiency of EGENET in solving certain hard instances of constraint satisfaction problems.

Integrating the Kinect camera, gesture recognition and mobile devices for interactive discussion

The Microsoft Kinect camera is a revolutionary and useful depth camera giving new user experience of interactive gaming on the Xbox platform through gesture or motion detection. Besides the infrared-based depth camera, an array of built-in microphones for voice command is installed along the horizontal bar of the Kinect camera. As a result, there are increasing interests to apply the Kinect camera for various real-life applications including the control of squirt guns for outdoor swimming pools. In additional to the Kinect camera, mobile devices such as the smartphones readily integrated with motion sensors have been used for different real-time control tasks like the remote control of robots. In this project, we propose to integrate the Microsoft Kinect camera together with the smartphones as intelligent control for interactive discussion or presentation for the future e-learning system. To demonstrate the feasibility of our proposal, a prototype of our proposed gesture recognition and command specification software is built using the C# language on the MS .NET platform, and will be evaluated with a careful plan. Furthermore, there are many interesting directions for further investigation of our proposal.

A fast and flexible framework of scripting for Web application development: a preliminary experience report

The World-Wide Web represents an opportunistic Internet-based platform on which many new applications are developing rapidly. To facilitate fast Web applications, a scripting approach is one possible solution for software developers. However, then are two major drawbacks for many existing scripting tools. First, most of these scripting tools are proprietary, and only available on chosen platforms. Second, most scripting tools use the Hyper-Text Markup Language (HTML) as the basis for extension. To provide extra functionality such as database connection, various extended codes or commands are mixed with HTML tags to produce expected results. However, this dirty approach will make the design and development process more complicated. We propose a systematic and flexible scripting framework for fast Web application development. The two major components, layout script and information-processing script, naturally reflect the static and dynamic properties of a Web application. In addition, the information-processing script can flexibly be split into different component scripts to suit different applications. We built a prototype of a script parser and an Integrated Development Environment (IDE) tool to quickly develop an intelligent Web-based Personal Computer Configuration Advisor. Clearly, our proposal opens up many possible directions for investigation including the uses of the IDE tool to develop other applications, and the integration with other components to handle more complicated Web applications.

Enhancing Learning Paths with Concept Clustering and Rule-Based Optimization

Finding a good learning path with respect to existing reference paths of closely related concepts is very challenging yet important for effective course teaching and especially adaptive e-learning systems. There are various approaches including ontology analysis to extract the key concepts which could then be correlated to one another using an implicit or explicit knowledge structure for relevant courses. With the available correlation information, an effective optimizer can ultimately return a good learning path according to its predefined objective function. In this paper, we propose to obtain more thorough correlation information through concept clustering, which will then be passed to our rule-based genetic algorithm to search for better learning path(s). To demonstrate the feasibility of our proposal, a prototype of our ontology analyser enhanced with concept clustering and rule-based optimizer was implemented. Its performance was thoroughly studied and compared favorably against the benchmarking shortest-path optimizer on actual courses. More importantly, our proposal can be easily integrated into existing e-learning systems, and has significant impacts for adaptive or personalized e-learning systems through enhanced ontology analysis.

A Comparison of MCQ Assessment Delivery Methods for Student Engagement and Interaction Used as an in-Class Formative Assessment

Student interaction and formative assessment are two agendas which hold equal importance in guiding the student approach to learning in engineering and in many disciplines. With MCQs being a familiar assessment tool for most teachers, it would be ideal if there were a combined approach which involved using MCQ to assess formatively and at the same time, to encourage student engagement and in-class interaction. In this paper, we investigate a number of MCQ delivery methods which may be able to step up to both concerns.