The purpose of the thesis is to design and develop a network of automated, distributed,
living cell-based sensors, called CytoSensors. Their main role is to detect a
variety of biological and chemical toxins. The system is designed to help researchers
to carry out multitude of experiments, in order to build a practical knowledge base in
toxin detection. The network is developed in accordance with industry standards, to be
used and deployed for prevention in inhospitable environments such as battlefields, toxic
urban locations or polluted agricultural regions.
The sensor is composed of a processing unit (processor and memory), an archiving
unit (permanent data storage), a communication unit, input devices attached to a data
acquisition unit, and control devices. The CytoSensor is specifically designed to acquire
and analyze visual information about the living cells: hence cameras are used as input
devices and frame grabbers are used as the digitizers. The control devices are additional
external devices developed to help control and automate the process of data acquisition:
they comprise light intensity control USB boards to provide the correct amount of light
to view the cells, touch panels for user-instrument interaction, and bar code readers to
identify vials and experiments. The software, on the other hand, is a complex mosaic
of different elements, each of which has a specific task to accomplish. These building
blocks include the real-time acquisition, archiving, networking, processing, modelling,
sensor output presentation and user interfaces. Our goal is to develop, integrate and
optimize all these components to produce a viable and working device. The prototypes
evolved from an offline, portable sensor equipped with a single high-resolution CCD
camera and high-quality optics, to distributed online sensors with multiplexed CCD
cameras and affordable optics.
The acquisition board digitizes in real time the images from one to twelve multiplexed
high resolution cameras. Several operational requirements must be met. First,
a fault-tolerant and stable control over the input devices and control devices must be
provided. Secondly, acquisition timing errors should be minimized as a trade-off between
performance and the use of a low-cost, general-purpose, industry-standard operating system
such as Microsoft Windows NT. Finally, in order to reduce development time and
increase code reusability, a common abstraction layer is designed to provide for flexible
use with various types of digitizers and cameras.
As part of a distributed detection network, each sensor is able to exchange data
with other "trusted" sensors and users, and to allow remote control of certain tasks. The
sensor may be seen as a node capable of transmitting and receiving acquired or processed
data to a distant device (another sensor, a workstation or a PDA) for visualization, inspection
and decision-making by a front-end user. Each node on the network provides a
set of complementary services including data acquisition, data processing, communication
and system. The mandatory system service monitors the local system performance
and manages data archiving. The communication service connects the various services on
the network by enabling message-passing, file transfer and caching. The sensor network
integrates a lightweight, interoperable and flexible RPC (Remote Procedure Call) protocol
to achieve real-time control and monitoring of these distributed resources. A reliable
embedded database system is used to store metadata bound to acquired and processed
images. This database is also used to maintain information on neighbor nodes, and to
check access credentials of available local services. Finally, by adding store-and-forward
messaging capabilities, the application can be extended to work in wireless and mobile
networks. / Graduation date: 2003
| Identifer | oai:union.ndltd.org:ORGSU/oai:ir.library.oregonstate.edu:1957/32478 |
| Date | 28 March 2003 |
| Creators | Boichon, Bertrand |
| Contributors | Kolodziej, Wojtek J. |
| Source Sets | Oregon State University |
| Language | en_US |
| Detected Language | English |
| Type | Thesis/Dissertation |
Page generated in 0.0027 seconds