Artificial Perceptual Sensors - Artificial Human Sensors: Science and Applications - page 143

human acting as a sensor. The frequency and vibration instruments are superior
in performance and may provide much better result in industrial measurement
applications. This aspect will of course affect the human participation in measure-
ment situations in industrial plants, at least in countries where the labour cost is an
important part of the production costs. Striving towards fully automated produc-
tion plants has been on the focus for some years now, and we can already observe a
state of semi-labour-automation situation. For example, in a general car manufac-
turing plant for mounting cars there was approximately half of labour compared
with working robots, indicating also that the main task for the staff was to keep the
robots working properly. The general trend of the high-tech industry has focussed
more on the maintenance and control of the automated work than direct human in-
volvement in the production. We may expect that the human still involve to confi-
dently use her ability (in some cases still exhibiting outstanding abilities) to act and
interact with the production. For some years now, the trend has been to actively
search for human absence environment and to find corresponding automated or
autonomous solutions. For example, in the rock drilling industry in extracting
gas leakage sound from a noisy environment, Kotani (2001), and by developing a
matching pursuit approach to predict small drill bit breakage, Fu (1999). Another
illustrative example will demonstrate that the human skill of operator perception
abilities may change to automated, and technical cost-effective mining production
solution.
In the segment of rock drilling vehicles as shown in Fig. 6.9, the experience
shows a trend towards unmanned autonomous equipment. Similar trends can
be seen in the advanced technology production industry, where the strain of the
operator in a hard environment may be stressful. A main issue is however that
the drilling process still exhibits manual operations in the process of drilling in
rocks. For example, the retraction of the drilling steel has after the automatically
performed drilling process, to be manually handled to get it safely out of the rock.
This phase is still often maintained by the operator auditory perceptual system,
when hearing to the characteristic sound phenomenon appearing, indicating when
the splices between the different drill steel parts are opened-up enough to allow
retraction.
The sound detection ability by the human ear in noisy environments is
strained and will in the long term have a destructive effect on an operator's ability.
Therefore, artificial systems that substitute the human sensing factor in a produc-
tion site, especially when the environment is not healthy, will always be welcomed.
The paper by Bergstrand (2004) introduces a Fast Fourier Transform (FFT) solu-
tion that analyses the detected audio information to find specific behaviours. The
analysing method makes it feasible to locate a specific sound related to the opening
of the drill steel, following the retraction.
By artificially recording and analysing sound, there has been comparison
between an implemented “electronic ear” and an experienced operator's auditory
system, i.e., human ears, with sufficient results, and a commercialised system is
now available.

The following standard process is used in a continuous loop of