05-09-2012, 09:47 AM
An Automated Cleaning System for Hospitals
An Automated Cleaning.pdf (Size: 908.04 KB / Downloads: 192)
Abstract
Insufficient hygienic practices in Irish hospitals coupled with one of the highest
number of reported cases of MRSA in Europe have highlighted the need for solutions to
aid in the task of cleaning.
This automated cleaning system consisted of two robots: a core robot
developed separately with navigational and task scheduling capabilities integrated. The
cleaning task was carried out by making use of a commercially available Roomba
vacuum cleaner which had been adapted to operate in conjunction with the core robot.
A uni-directional communications was established; commands were sent from the core
robot to the Roomba.
A visual analysis software, by the name of RoboRealm, was integrated into the
system as the primary component. The initial role of the software was to allow the
vacuum robot to orientate itself in order to enable transport from location to
destination by means of visually tracking an object of interest. The object was to be
located on the rear of the core robot.
Subsequently the visual recognition aspect took on a greater role and
encompassed a system by which commands were issued by the main robot and visually
interpreted by the Roomba. This enabled the cleaning system to issue uni-directional
commands and therefore carry out regular cleaning of any room, spot cleaning on a
small spillage, following from one location to a destination or pause at any point during
transport for emergency reasons.
Introduction & Background
The IWARD project; Intelligent Robot Swarm for Attendance, Recognition,
Cleaning & Delivery [1] is comprised of a number of mobile robotic platforms
constructed with the ability to autonomously navigate throughout a hospital facility in
which they were intended to operate. As is suggested in the name, the robots were
designed for multiple tasks. This was achieved by developing modular robots where
necessary components were interchangeable in a quick and easy manner.
The IWARD base robot consisted of an aluminum chassis that held all essential
drive components. In order to navigate, the robots utilized a semi-circular array of
sonar sensors on the front, which was located above a laser scanner capable of
measuring numerous distances along a single horizontal axis. An impact sensor was
attached along the leading edge of the robot in the event of a malfunction of the
wireless sensors occurring or in the rare instance where an obstacle was, for any
reason, not detected in advance.
Problems in Irish hospitals
In early 2008 results were published [3 & 4] of a hygiene audit that had been
carried out in all publically run hospitals throughout Ireland. This was the 2nd such audit
that was conducted by HIQA (Health Information and Quality Authority) in as many
years, and the 4th survey of this kind that took place since 2005. That first survey, run
by the Health Service Executive (HSE), declared that 91% of hospitals examined were
found to have inadequate hygiene standards. 2007 saw the first independent study
carried out by the HIQA where no public hospital in the country was found to have a
rating of ‘very good’ or better, once the results were published. Out of the 51 hospitals
included in the survey, nine were rated as ‘poor’ while thirty-five hospitals achieved a
rating of ‘fair’. A mere seven hospitals in 2007 achieved the status of ‘good’ while none
received any higher rating.
Imaging analysis for distance measurement
“Design of Distance Monitoring Algorithm for Robotic Applications” was the
name of a project carried out in the University of Iowa in spring of 2009 [7]. The aim of
the research was, as the name suggests, to develop an algorithm for distance
measurement making use of a generic web-camera as well as a visual analysis software
known as RoboRealm. According to the background theory the initial concept for the
system was to release a set of three robots into an unknown environment where they
would subsequently roam throughout the area and “produce a map of the
surroundings, based on the information obtained from various sensors in the system”
As stated by the authors, Chen & Schelin, the intention of this publication was
to develop a visual system that could be incorporated into the existing hardware,
previously made available to the authors, without the addition of many new hardware
components. The basic operating principle was to use the camera in conjunction with a
laser diode and then to process the visual information by means of a software program
to locate the laser dot. Based on the location of that laser dot, information regarding
linear distance to a surface in front of the camera could be calculated.