university of bridgeport

University of Bridgeport

School of EngineeringSchool of Engineering

Prof. Tarek Sobh

Vice President for Graduate Studies and Research Dean, School of Engineering

University of Bridgeport School of EngineeringSchool of Engineering

Capabilities and research facilities / groups

Multimedia Information Group Multimedia Information Group (MIG) Laboratory (MIG) Laboratory

COMPUTER SCIENCE & ENGINEERING DEPARTMENTCOMPUTER SCIENCE & ENGINEERING DEPARTMENTUniversity of Bridgeport, Bridgeport, CT

MIG@UB

• Multimedia Information Group @ UB– Department of Computer Science and Engineering at UB, CT

• Location: Charles A. Dana Hall #234

• People– Prof. Jeongkyu Lee– 2 PhDs and 2 MS students

• Collaborators– Prof. JungHwan Oh at UNT– Prof. Shou-jiang Tang at UTSW – Prof. Dongwon Lee at PSU

http://www.bridgeport.edu/~jelee/mig

UNIVERSITY OF BRIDGEPORTUNIVERSITY OF BRIDGEPORT

Multimedia Information GroupMultimedia Information Group

Our Vision

Modeling and Design Data Analysis and Mining Index and Query Process System and Application

Low-levelapproach

High-level approach

Database Research

Multimediaunderstanding

raw video miningvideo segmentation

video surveillance miningvideo parsingEarly Research

medical video analysisSTRG-IndexBLASTed Image Matchinggraph-based model Video copy detection

SmartView (WCE)

Current ResearchTaeKwonDo videos

graph-based SBD

GVDBMSApplications- medical DB- spatio-temporal DB- UAV

STRG-QLknowledge discovery from Videos

Future Research

MM Ontology




Active Research


• Developing algorithms and S/W for event detections from WCE (Wireless Capsule Endoscopy) videos• Funded from IntroMedic, Co. Ltd., Korea• Collaborator: Dr. Tang at UTSW, Dr. Oh at UNT

TaeKwonDo project

Vlinkage: Video Linkage project Multimedia Ontology

SmartView project• Developing algorithm and prototype system for automatic judge of TaeKwonDo poomsae• Sponsor: International College at UB, WTF• Collaborator: Prof. Kim of Martial Art dept. at UB

• Developing a noble method for video and image matching• Applicatoin1: Video Copy Detection for YouTube• Applicatoin2: BIM (BLASTed Image matching)• Collaborator: Dr. Dongwon Lee at Penn State University

CNN

NBC

MLB

Movie

Search Policy

Find copied videos from Very Large Video Database

Scalability+

Search Content

based Video Search

• Developing a new algorithm for automatic generation of Multimedia Ontology• Target applications: Medical videos, Surveillance camera, and Military videos• Looking for sponsors

MIG



Proposal Title: Active Stability System using Visual Information for Gun Launched Hybrid Projectiles

Overview of Active Balancing System Operational Capability

• Operate over a wide range of general environmental conditions. • Capture visual information using CCD mounted on top of UAV.• Generate signal using visual features, such as color signal.• Compute the stability score using FFT and Energy function.• Control motors to balance UAV.

Proposed Technical Approach

Task1: Installation of Visual Equipments• Mount CCD camera on the tip of UAV• Install SoC (System on a Chip)

Task2: Development of Visual Analysis Module• Develop and implement feature signals generation. • Develop and implement FFT and Energy function.• Develop and implement stability scoring function.

Task3: UAV Control Module• Control UAV for stability.• Test the active stability system.

Cost and Schedule

Total Cost: $ ??????Deliverables• Technical report describing compiled results of active stability system for Gun Launched Hybrid Projectiles.• Prototype system integrating developed system.Corporate Information•.Department of Computer Science and Engineering University of Bridgeport, Bridgeport, CT 06604• Phone: (203) 576-4397, Fax: (203) 576-4765• Email: [email protected] (Dr. Jeongkyu Lee)

LaunchWing

Unfolded Stabilizing

CCD camera

Videos Visual Analysis

Control UAV

M1-2 M3-4

Project Period

M5-6Tasks

Task 1

Task 2

Task 3

Robotics, Intelligent Sensing and Control Lab (RISC)

University of BridgeportSchool of Engineering

Outline of Outgoing Project• Online Automation and Control: An Experiment in Distance Engineering

Education

• E-Learning: Case Studies in Web-Controlled Devices and Remote Manipulation

• Prototyping Environment for Robot Manipulators

• Manipulator Workspace Generation and Visualization in the Presence of Obstacles

• Kinematic Synthesis of Robotic Manipulators from Task Descriptions

• New concept in optimizing the manipulability index of serial Manipulators using SVD method

Outline of Outgoing Project• Recovering 3-D Uncertainties from Sensory Measurements for

Robotics Applications

• Industrial Inspection and Reverse Engineering

• Sensing Under Uncertainty for Mobile Robots

• Robot Design and Dynamic Control Simulation Software Solutions From Task Points Description.

• RISCBot II

Online Distance Laboratories

• Using Automation and Telerobotic (controlling devices from a distance) systems

• Real-time laboratory experience via the internet

1. Tele-operation of Mitsubishi Movemaster2. RISCBOT – A Web Enabled Autonomous Navigational

Robot3. Tele-operation of the FESTO Process Controller

Capabilities and Research Facilities

• Sensing under uncertainty.• Sensor-based distributed control schemes.• Control and planning for autonomous mobile systems.• Modeling and recovering uncertainty in 3-D structure and motion. • Dynamics and kinematics generation and analysis for multi-DOF robots.• Active observation and control of a moving agent under uncertainty.• Automation for genetics application.• Manipulator workspace generation in the presence of obstacles.• Turbulent flow analysis using sensors within a DES framework

Other Projects

Workspace Dimensionsand Coordinates of the Task-Points

Velocity and AccelerationRequirements

Obstacles, Working Medium, and Trajectory Biases

Restrictions on ManipulatorConfiguration

Daniel Y. Toundykov

Computing optimal geometry for robotic manipulators is one of the most intricate problems in contemporary kinematics. Especially when the specification provides little or no instructions on what the mechanism should look like, but rather what this mechanism will be required to do.Basic source data may include workspace dimensions, and coordinates of the task points. More elaborate models should take into consideration restrictions on velocity and acceleration at each point. To complete the picture we may add description of the obstacles, peculiarities of the working medium (for instance, if the machine will be submerged into liquid), and possible trajectory biases, in case some paths are more preferable than their alternatives.

School of EngineeringUniversity of Bridgeport

Manipulability Bands of Puma 560 in 2D workspace

RISCbot II

CT Post

Wireless & Mobile Communications Wireless & Mobile Communications (WMC) Laboratory (WMC) Laboratory

COMPUTER SCIENCE & ENGINEERING DEPARTMENTCOMPUTER SCIENCE & ENGINEERING DEPARTMENTUniversity of Bridgeport, Bridgeport, CT

WMC Current Research Projects WMC Current Research Projects

1. Wireless Multiuser Communications for Cellular and Mobile Networks

• BER and SNR Analysis of DS-CDMA Cellular Networks

• Multiple Access interference (MAI) Cancellation for Wireless Multiuser Receivers

• Analysis of Processing Gain for Wireless Multiuser DS-CDMA Systems

• Computational Complexity and Algorithm Optimization for 3G Cellular Networks

WMC Research ProjectsWMC Research Projects

2. Wireless Mesh Networks• The Use of Orthogonal Frequency Code Division

(OFCD) in Broadband Wireless Mesh Networks

• Efficient Routing Algorithms for Wireless Mesh-Hypercube (M-H) Networks

WMC Research Projects (Cont..)WMC Research Projects (Cont..)

3. Mobile Ad Hoc Networks (MANET)• The Best and Worst Case Capacity Analysis of MANET

• Efficient DSR Based Routing Scheme for MANET

• Minimizing the Malicious Behavior of Mobile Nodes for Maximizing the MANET Data Throughput

4. Wireless Sensor Networks1. Resource Optimization in Wireless Sensor Networks Via

Self-Adaptive Methodology

2. Minimizing the Energy Consumption of Wireless Sensor Nodes Using Active Node Optimization Method

Related Recent ResultsCDMA Receiver: Multiuser Receiver

• A class of CDMA receivers known as multiuser receivers

• It exploit the available information about the spreading sequences and mobile channel impulse responses of all the CDMA users

• The goal is to improve the performance of the wireless CDMA users

Classification of CDMA detectors

WMC research group focuses on this part

Related Recent Results Antenna Design for Cellular Networks

• Advantages– Co-channel interference

reduction– Collect multipath components– Delay spread reduction– reduce handoff rate– stand alone technique

• Disadvantages – Linear increase in Interference– Cancel only L-1 interference– Difficult to achieve convergence

in low SNR

Correlator ormatched

filter

Training

-+

Reference

AdaptiveAlgorithim

C1

CL

C2

Elem ent1

E lem entL

E lem ent2

Array Output

One per path

Figure: Adaptive Antenna ArrayWMC research group focuses on the utilization of adaptive antenna array with CDMA systems for achieving optimal performance

Related Recent Results Multiple Access interference Cancellation

for Wireless Multiuser Receivers• Applications: Military RF

Consideration– High-power, fixed-frequency

transmitters make easy targets.• Easy to jam• Easy to destroy

90.9WETA

Missile seeker head locks-on RF transmitters

AGM-88High-speed Anti-Radiation (HARM)missile

University of Bridgeport Faculty of EngineeringFaculty of Engineering

Other relevant faculty expertise

Ausif Mahmood

• Video analysis, capture and processing.

• Signal processing.

• Video and data compression technologies.

• Electronics circuit design.

Ausif Mahmood

• Image processing projects such as Object recognition, Face Recognition using Eigen Faces and 2-D FFT.

• Developing various hardware prototypes based on microcontrollers such as controlling image capture.

• Hardware and software development, including having developed some commercial projects involving wireless communication, GPS tracking, and RS-232 communication.

Navarun Gupta

• Programming a real-time digital signal processing (DSP) chip to do specific tasks like filtering, image processing and mathematical calculations.

• Implementing such an autonomous image tracking / detection system.

Neal Lewis and Elif Kongar

They can assist with various forms of simulation including:

• Fluid dynamics of airborne objects.• Sensitivity analysis using Monte Carlo techniques • Simulation of multi-step operations using Arena

software. • Simulation and Six Sigma techniques which are

recommended in the design stage of the small size product to maximize product reliability.

Xingguo Xiong

• Low power VLSI circuit design and VLSI testing. For the flying UAS bullet project, low power circuit design reduces the power consumption of the control circuits so that the battery life can be extended.

• MEMS (Microelectromechanical Systems) and nanotechnology. By using MEMS technology, the size, weight and cost of the UAS bullet can be greatly decreased. For example, he can develop the MEMS micromotor and other various MEMS sensors/actuators for the UAS bullet.

• Performing shock and vibration simulation/testing for the UAS bullet.

Saikat Ray

• Target demo: A UAS hanging by a thread with (micro)-motors installed for bending the wings. The wings are controlled from a laptop/PC through wireless link.

• Building the wireless link (system) with relatively small form factor.

Lawrence Hmurcik

• Electronics circuit design • Microelectronics • Signal processing • Controls • DSP • Circuit Design / Analysis • Simulation • Micro Electronics/ MEMS

university of bridgeport

Documents