Research Topics
Some of the important issues in wireless sensor networks and other networks that I have been involved on over the years include (please see my publication list for related publications):
3D Coverage in Wireless Sensor Networks: An important problem addressed in literature is the sensor coverage problem. This problem is centered on a fundamental question: “How well do the sensors observe the physical space?” Sensor networks pose a number of challenging conceptual and optimization problems such as location, deployment, and tracking. One of the fundamental problems in sensor networks is determining how many sensors are really needed to cover a specific area during an interval of time.
3D Connectivity in Wireless Sensor Networks: Connectivity implies that the location of any active node is within the communication range of one or more active nodes such that all the active nodes form a connected communication backbone, while coverage requires all locations in the coverage region are within the sensing range of at least one active node. Obviously, the relationship between coverage and connectivity depends on the ratio of sensing radius to communication radius. The problem of wireless sensor networks considers a large number (in the order of thousands) of identical nodes which possess limitations in available energy, computational power, memory, and communication range. In potential sensing applications, the sensor nodes may be randomly deployed in a hazardous or dangerous environment where the nodes are physically inaccessible after deployment. In this way, the design of the network needs to consider energy conserving schemes to account for a limited energy supply, low memory/computation and resilient networking schemes to account for the hazardous environment.
Tracking using Wireless Sensor Networks: One of the most important areas where the advantages of sensor networks can be exploited is for tracking mobile targets. Scenarios where such network may be deployed can be both military (tracking enemy vehicles, detecting illegal (border crossings) and civilian (tracking the movement of wild animals in wildlife preserves). Typically, for accuracy, two or more sensors are simultaneously required for tracking a single target, leading to coordination issues. Additionally, given the requirements to minimize the power consumption due to communication or other factors, we would like to select the bare essential number of sensors dedicated for the task while all other sensors should preferably be in the hibernation or off state. In order to simultaneously satisfy the requirements like power saving and improving overall efficiency, we need large scale coordination and other management operations. These tasks become even more challenging when one considers the random mobility of the targets and the resulting need to coordinate the assignment of the sensors best suited for tracking the target as a function of time.
Using wireless Sensor Network for querying the physical world: The present web has a certain shortcoming that although it is successful at describing conceptual space, it does a poor job of describing physical space. Let us say, for example that we are interested in answers to queries such as:
"Find the closest Fast food Restaurant with the minimum waiting time",
"Let me know when my college opens. When there are more than 25 students in my class"
"Let me know when the enemy crosses the border and how many are they?".
"Estimate how long would it take me to go from where i am to Downtown taking into consideration all the closed roads and traffic".
The current web architecture does not support such location dependent queries, depending on dynamically generated data from various kinds of sensor networks. Obtaining such rapid, accurate and useful information in response to location-dependent questions as these would provide a whole new level of convenience and empowerment to individuals who ask them. Our perceived architecture forms an innovative information infrastructure capable of providing any service that obtains, maintains, manages and delivers spatially-constrained data. It proposes that all information resides at the physical location in which it is produced, and assumes that all the information sources have a network presence. A common scenario in our vision is that of information sources and the users who query them to be mobile, thus making the response to queries dynamic in nature. This architecture also assumes that information sources can have unbounded heterogeneity; they may be as simple as embedded sensor devices that store & communicate only a single bit of information, or they may be as complex as full-scale databases.
Data Mining in Smart Sensor Networks: Sensors streaming their data online are turning the Internet into a global sensor network. Software platforms that integrate and mine these data streams may create a world in which sensors become pixels and we browse reality as easily as we browse Web pages today. The evolution of low cost, networked sensors, often directly Internet-enabled, is bringing sensors out of their traditional closed-loop realms into the rest of our reality. Consider cell phones: There are 1.4 billion active cell phones in use today with more than half a billion units sold last year. As cameras become a standard cell phone feature, we're becoming the most connected and instrumented people in history. As sensor and communications technology continues to develop, we can envision a very different Internet than the one we use today. Rather than sending messages and browsing Web pages, we may experience new interactions such as experience sharing and browsing reality. Data mining, defined broadly as extracting useful information and insights from data, may be the untold half of the sensor networks story. Given the potentially huge amount of data streamed by live sensors, algorithms to fuse, interpret, augment, and present information will become an increasingly important part of networked sensor applications. I will be involved in this exciting new research area.
Localization and Topology Control in Sensor Networks: The locations of sensor nodes are important for the meaning of collected data and for routing. The localization problem is to determine the node positions based only on the information obtained through the nodes’ interactions, such as connectivity, edge lengths and angles. Topology control is the process of controlling the topology of a wireless network by adjusting the coverage ranges of its wireless nodes. Using rigorous combinatorial and probabilistic analysis, my goal is to study various topological properties (e.g., connectivity, routing path, degree, local minimum for geographical routing, etc.), and present several variations of the topology control method. The findings should show balance between the various aspects of network performance.
Generalized geographical routing: Geographical routing is a very important routing method, especially for ad hoc networks, mainly due to its high scalability. It is widely used for wireless networks, where it guides routing by using nodes’ coordinates. It is also being used for overlay networks in the Internet, where overlay nodes are embedded into certain metric spaces to obtain their virtual coordinates. (A typical example is the Chord peer-to-peer network.) Currently, the performance of geographical routing is limited by the hardness of the embedding process, the metric distortion caused by embedding, and its sensitivity to particular network models (such as the UDG model for wireless networks). I am interested in studying new routing methods that generalize geographical routing, which use new backbone structures to guarantee message delivery and new shortcut links to guarantee efficiency.
Next generation ubiquitous and pervasive healthcare systems: Although still in is its infancy, the sensor network technology holds great potential for a significant impact on next generation ubiquitous and pervasive health care. Today, hospitalized patients are tethered to instrumentation, even though some attention has been given to highly customized, high-cost wireless devices. Instances of un-tethered, wireless devices have been few in number, but advances in bioengineering, biochemistry and biotechnology hold the promise of an ever-expanding pool of knowledge in this emerging discipline. Moreover, with the continued advanced in microchip technology, more and more functionalities are being placed on smaller and smaller chips, paving the way for wireless devices to be implanted within the body and operate at the molecular level. This technology will simplify testing, monitoring and treatment, while also improving patient quality of life by minimizing time spent in the hospital, and enabling automatic, un-tethered and continuous treatment of chronic conditions. The realization of the next generation ubiquitous and pervasive healthcare systems is a challenging task, as these systems area likely to involve a complex structure consisting of various devices, ranging from resource-constrained sensors and actuators to complex multimedia devices, supporting time critical applications. This is further compounded by cultural and socio-economical factors that must be addressed for next generation healthcare systems to be widely diffused and used. My future work will involve research projects in the field of distributed control and sensing of networked wearable and implantable medical devices.
