University of Engineering and Technology Taxila, Pakistan

RESEARCH PROJECTS

Project Title:
Vehicle Future Position Overlay Over Wind Screen: Head-up-Display and Gaze Estimation Framework.

Project Director:
Dr. Hafiz Adnan Habib

Junior Project Director:
Afshan Jamil

Start Date: 1st May 2010
End Date: 30th April 2011

Abstract:
For safe driving, driver should maintain his/her control on the road. Maintaining lateral and longitudinal control on the road has key role in the safe driving. This is particularly important for novice drivers. The high crash involvement of young novice drivers is recognized as one of the most intractable problems in road safety. This research and development work focuses on designing a display system for novice drivers by displaying future position of vehicle, assuming the movement of vehicle in straight line. System captures image of driver in real-time, estimates head pose of driver, and provides timely information to the driver about future destination points of wheels on the road. The information about future points is projected on windscreen of car to help driver follow straight path on the road. Projection angle varies depending on the position of driver, distance of driver from windscreen etc. Future destination points projected on the windscreen are actually reflections from the road received by the eye. The reflection passes through windscreen to reach the human eye. Proposed system highlights those positions on the windscreen through which driver receive road reflections. End product will be software installable in car PC along with hardware (begal board for projecting points on the windscreen of car and IR camera focusing driver). The proposed framework is a novel extension in driving assistance systems. The proposed system can be installed in automotives and will be a major contribution in automotive industry.

For more information please contact
[email protected]
Project Title:
Video Analytics Framework for Quantitative Employee Performance Evaluation.

Project Director:
Dr. Hafiz Adnan Habib

Start Date: 1st May 2010
End Date: 30th April 2012

Abstract:
Performance of each employee has key role in the success of any employer: enterprise, company, industry, organization. Employer adapts several tools for performance evaluation of employees such as amount of work, time spent at workplace, quality of work etc. Information technology offers support to employers at various levels such as automated attendance and workplace time spent calculations by biometrics and RFID, etc. This research and development work proposes intelligent algorithmic framework for activity analysis of employees. Real time video is acquired at the workplace of employees and analyzed by algorithms that generate report about the performance of employees in a quantifiable way. Employees will be identified in the video sequences; their activities will be recognized and classified in desired and undesired activities and statistical report will be generated. Proposed system will be trainable so that it can be installed and utilized at any workplace. Developed system will have stationary as well as portable architecture. It will be able to monitor complex activities of employees. End product will be software installable at some server with capability to read video sequence from multiple cameras: wired/wireless and analog/Ethernet interfaces. An embedded system based camera version (smart camera) will also be developed possessing the same capability to perform video analytics.The proposed framework is a novel extension in video analytics arena. The proposed framework has immense applications and can be installed in small office to large corporate offices, small manufacturing industry to multi thousand employees industries, construction and mining sites, point of sales and shops, education etc.

For more information please contact

[email protected]

PhD Research Title:
VFGIA: Vision based Feet Gesture Interface for Automotives

Research Scholar:
Muhammad Haroon Yousaf

Research Supervisor:
Dr. Hafiz Adnan Habib

Research Laboratory:
Video and Image Processing Lab

Abstract:
This report proposes a new perceptual interface for primary controls of automotives. Vision based foot gesture interface for automotive control is a project of human machine interaction. Main objective is to design a location adaptive system to interface the primary controls of automotives. System will replace the traditional system of the primary controls of the automotives i.e. brake, clutch and accelerator with a vision based location adaptive system. Gesture recognition and human machine interaction techniques are used to develop the system. Gestures generated by the Human feet will be captured on regular intervals and processed to identify the current action by the driver i.e. acceleration mode, brake mode etc. Respective commands against each of the action will be passed to the vehicle in form of electric signals. These electric signals will take place the human lower limb movement to press the primary controls.

Start Time: November, 2007
End Time: ..November, 2010
PhD Research Title:
Security Provider Layer for Ahoc Networks Networks

Research Scholar:
SYED SOHAIL AHMED

Research Supervisor:
DR. ADEEL AKRAM

Discipline: COMP ENGG.

With the advent of mobile devices in our daily life, the need for more reliable high performance network has increased many folds. Traditional Infrastructure networks provide limited coverage for limited users. On the other hand Ad hoc networks promise greater coverage and rapid deployment. Both types of networks have their limitations.Researchers have proposed the use of Ad hoc with infrastructure networks to reap the benefits of both types. This kind of networks is termed as heterogeneous or hybrid networks.

The aim of this research is to design a security infrastructure for hybrid networks that will provide a security layer for network aware applications and operating systems of mobile devices owned by the users. The objective is to provide sufficient security to evade intruders into the hybrid networks while keeping the security mechanism as transparent and user friendly as possible.

Date of Regn: 20-11-2007