Available Undergraduate Student Research Opportunity
1. Intelligent Wireless Mobile Sensor Nodes.
Motivation
Sensor networks can be used to provide useful data / information at the right time reliably. One of the challenges in distributed sensor networks is that how a large number of sensors involved can provide useful data /information at the right time reliably, while putting on hold information that is either redundant or not required at the moment. One means of enhancing the reliability of the data is to be able to monitor the health of the sensor node locally and tag the data /information with the confidence value.
Project Description
The goal of the project is to construct a system that permits large number of sensors to share information and make intelligent decisions locally about the set of data that should be transmitted, thus provides access to the data without overwhelming the network bandwidth. For this purpose, students will study different technologies, and design and implement hardware and embedded software for intelligent wireless mobile sensor modules at three different tiers. The system will be able to provide low-cost and highly reliable sensor node with flexibility to be added or removed from the sensor network.
Undergraduate Opportunities
The project will provide a number of compelling opportunities to involve undergraduate students in the research. For example, undergraduate students will learn about different information fusion algorithms and help in the implementing and testing of the new fusion algorithm for various sensor network configurations. Simulation will be carried out using Matlab and Simulink, with hardware implementation. The students will have the opportunity to show off their finished project in nationally design conferences.
2. Filtering Based Texture Analysis for Medical Images.
3. Feature Identification for Medical Images.
Motivation
The accurate and efficient feature identification within images has long been a key research area in image processing community. One way to identify features in images is to segment the image into smallest geographic units and then merge regions with similar property together, and many such image segmentation algorithms have been developed. Clearly, detailed testing of these algorithms and comprehensive evaluations of their strengths and weaknesses are vital to the development of good feature identification solutions. To further enhance the accuracy and efficiency of the feature identification algorithm, support vector machine based feature identification algorithm has also been proposed. The increased accuracy and efficiency would make the feature identification algorithm applicable to real time control and medical systems.
Project Description
Dr. Yuan has developed a feature identification algorithm based on texture information using the evolutionary computation algorithm. This algorithm can find as many prominent objects in the image without image segmentation. It does not require strict initial position, as most optimization algorithms do, nor any prior knowledge of what to find. Furthermore, Dr. Yuan are actively engaged in further testing this novel algorithm: to compare its performance with other image segmentation algorithms, and to apply it to various modalities of medical images to verify its broader usefulness. Anew texture analysis algorithm based on multi-spectrum filtering method. Further research is underway to test this new algorithm and to develop faster implementation strategies.
Undergraduate Opportunities
The above two projects will provide a number of compelling opportunities to involve undergraduate students in the research. For example, undergraduate students will learn about the evolutionary computation algorithms, and will help in the testing of the image segmentation algorithms based on texture information on various images, be it synthetic, natural, or medical. In addition, they will also help in implementing different image segmentation and texture analysis algorithms, testing and comparing the performance of them on various image types.
4. Remote Accessible Electro-Mechanical Experiments Setup.
Students will learn Matlab and Simulink. They can choose projects from building a data acquisition system, to simulate how we can control mechatronic systems using micro processor based embedded system. Possible systems include: coupled tank; 2 DOF (degree of freedom) robot, analog plant simulator, 2 DOF inverted Pendulum, Rotary Flexible Link, Heat flow control, etc.
5. Virtual Electric Circuit Lab.
Students will learn about a very popular DAQ system and the software environment from National Instrumentation, Inc. They can then design, implement a virtual electric circuit lab for any lab they choose using Labview, an object-oriented system development environment.