11/11/2005 10:40:00 PM
Kospi Squak
Hi there dear beautiful people. Just got a list of some of the interesting projects being offered to us students here. just though i'd share it with you guys in case any of u wondered what it is that i am supposed to be studying. (Unlikely that i will get any of them cuz of my crap results)
but yeah, they sure do sound interesting. hope it is an interesting read
Weigh in Motion for trucks
---------------------------
Currently a common method of detecting truck weight is to use a force transducer(s) placed across the road. However oscillations of the vehicle load can lead to erroneous readings with accuracy as poor as +- 20%. This project will investigate introducing a known disturbance before the transducer and using a computer vision system to track the resulting movement of the truck body as it passes over several force transducers, in order to better estimate the true load of the truck. This project will involve the use of LabView and one of the extensions will be the incorporation of license plate identification technology to identify vehicles. (Note: prior experience with LabView is not essential).
Design of a moving fluoroscopy system
---------------------------------------
Fluoroscopy is a technique for obtaining live x-ray images in a living patient. This project will involve the costing and design of a system that will move along a fixed track in the biomechanics lab and keep up with a patient as they walk in a straight line. In principle, it is similar to the technology used to take TV images of swimmers from the bottom of the pool or 100m runners during their events. A controller will be used in order to close the loop and ensure the camera moves at the same rate as the patient. The fully built system will replace the need for up to six cameras in a biomechanics laboratory to track movement of limbs. It is anticipated that Labview may be used as the interface, although previous experience is not essential.
Drive cycle implementation through engine computer (with Ford Motor Company)
-----------------------------------------------------------------------------------
During a fuel economy or emissions test vehicles have to follow the drive trace within a certain limit. The tolerance band of the legal drive cycle is relatively small (± 2 kph) but even in this small tolerance band individual deviations of the test driver are inevitable. These deviations strongly influence the accuracy and repeatability of the measurements.
In this project, the vehicle's own engine control modules should control the actuation of electric throttle body and brake during an emission test. For this purpose a calibration tool can be used. These calibration tools are normally used during development tests anyway to record and manipulate data from the control module. The calibration tool needs to be programmed with a script file accordingly and an interface from the calibration tool to the chassis dyno needs to be implemented.
[This project is subject to confirmation]
Set up of a distributed computing environment using Dymola and Corba.
-------------------------------------------------------------------------
Many optimisation techniques require intensive repeated iterations of plant models to develp correct control strategies. This can be extremely time consuming when a single processor is used and the plant models are complex (e.g. an internal combustion engine, or a biomechanical system). This project will attempt to design a framework whereby a plant model in Modelica can be run on multiple workstations that share information over TCP/IP via a CORBA program. This project requires good knowledge of C++, and is primarily for students who are looking for a software based project.
Classification of patient movement
----------------------------------
A hypothesis is that intensive physiotherapy given to stroke patients shortly after their stroke will result in improved limb function in the longer term. Presently however there is no automated way of measuring limb functionality in a stroke patient (the common approach is an observer goes to the patients home and notes every movement they make – which is clearly labour intensive and can violate the integrity of the experiment). In 2005, a project involving the use of a 3-D accelerometer and 3-D gyroscope wireless device was run that sought to measure functional activity in a stroke patient when they have left hospital by analysing the sensor outputs and attempting to classify the data into a type of movement. This project will seek to extend on those results and improve the classification techniques to make the prototype device useful in a clinical setting.
Research projects
------------------
Primarily for students interested in pursuing postgraduate degrees and can be discussed upon request, and tailored to the student(s). Typical areas of research projects I offer are applied control (in automotive and biomedical areas), classification, estimation (intelligent and optimal) and optimisation (typically applied to a particular problem).
