Q be avoided. 
t0 b « deter, 
"«nbefe 
nanu al at the 
lve WOf k. One 
Rüttle time 
This is an im. 
' n tbis section 
worn 
can 
- 1 to use infor 
ms technology 
be inspection 
: lining (not 
t^t the staff 
le to this, no 
sen created to 
ie data arrmi. 
Crane control Add. Sensors 
Sensors: 
Audio 
Infra-rot 
Mobile Assistant 
Server: 
Information Warehouse 
N '** 
,3D-Dlstance 
camera 
Vibration 
• Simple applications 
- data reduction 
- simple analysis 
- additional information 
• Long-term analysis 
- data mining 
- complex/demanding analysis 
- quantitive —qualitative data 
hie within 
ap- 
Figure 3. Use of WINSPECT-key technologies for the inspection of cranes 
le station can 
uple: 
essary for the 
ity is the pre- 
:omputer and 
wireless net 
work premises 
; general pur- 
Applications 
react flexible, 
ving the well 
inologies. In 
their control 
ditions of the 
bit-rates with 
ovide a work- 
ve use within 
e this goal we 
icult environ- 
;1. Within the 
and server it 
jje amount of 
t h a transfer- 
¡xainplc» it 
;a transferred 
for the actua 1 
hniques» such 
sis[Boronow* 
Another objective is to analyse the data collected 
from the inspection. This includes large amount and 
different types of data (e. g., vibration data, infrared 
pictures of temperature, audio-sensor-data with high 
resolution). Aided with knowledge discovery methods 
we are seeking new knowledge which we integrate in the 
cycle of maintenance. Figure 4 provides the overall con 
cept of the project. The specific problems of inspection, 
planning of inspections and analysis of inspections are 
the main objectives of our project. The concepts include 
interfaces so that information of the reparation process 
will be considered at the planning level and therefore, 
can be transferred to the service team. 
A main topic in this project is, the option to feed 
back information collected and analysed from inspec 
tion and repair to the planning level automatically. Aside 
from knowledge discovery methods, we use numerous 
tools for the implementation of this feedback, e. g. plan 
ning, assessment and data mining tools. 
3.3 Used technologies and methods 
Wearable Computing Tasks within the maintenance 
process, installation, problem solving or inspection are 
quite complex. Technicians need efficient and effective 
support in order to fulfill their job. A new level of tech 
nology which came up over the last couple of years in 
the area of wearable computing is contributing posi 
tively to this problem. Wearable computers are a new 
hardware technology, where the user works on a high 
performance PC, but has his hands, legs, and eyes free. 
Therefore, the user can fulfill his obligations as before 
and has the option to access local, with the help of wire 
less LAN even global, information resources (server). In 
literature the term ’Augmented Reality’ (AR) can be 
found, a research topic within the virtual reality (VR) 
community [Vallino, 1998]. Vallino defines AR as a com 
bination of a real scene which the user is viewing at, and 
a virtual scene which will be generated with the help of 
a computer. One extends the real scene with the virtual 
one, e. g. with additional information such as construc 
tion plans of a technical device. 
The miniaturisation of these fully equipped wear 
able computer, the new orientation of displays, and the 
additional option of giving input via voice provides the 
working world new potentials within the integration of 
information technology. This technology first was dis 
played at the Comdex in December 1996 [Gates, 1996]. 
On the 2nd international conference on wearable com 
puting in May, 1998 the vertical and horizontal markets 
for wearable computer were shown [Xybernaut, 1997], 
[Visser, 1998]. Big machine industry was identified and 
placed first, for the vertical market and maintenance 
processes as well as collection of data are topics of the 
horizontal market. This new technology has found its 
place to date only in a few big industrial companies 
(e.g., DaimlerChrysler, IBM, Boeing). The use of this 
technology demands thought about the potential role 
of computers within business processes. 
Knowledge Discovery & Data Mining In recent 
years new areas of research have been developed; knowl 
edge discovery in large (KDD), data bases and data 
mining. In general, methods in this area are investi 
gation to acquire to find implicit knowledge in large 
data bases and then make them explicit. Fayyad et al. 
[1996] define KDD as the ’’discovery of knowledge as 
a non-trivial process of the identification of valid, new, 
potentially useful, and comprehensive pattern in large 
data bases”. Fayyad et al. consider data mining as a 
7B-3-5