An Artificial Perceptual Model Approach - Artificial Human Sensors: Science and Applications - page 122

added value to a user, e.g., an operator. The time factor is of most importance for
the interfacing procedure in order to achieve control of the essential data received
and to deliver the right data on time. The timing is an essential internal function
in a system that is frequently described in the manufacturing industry exhibit-
ing similar structure of parts/data to be at the right location in an expected time
slot. The concept of lean production manages the benefits of a just in time deliv-
ery concept when optimising products for each other. The production concept has
similarities, which can be used, in designing complex data handling processes, for
more information about lean production, see e.g., Hirano (2006). In sensor systems
architectures and certainly in fusion procedures of sensor data, the need for syn-
chronisation of time dependent information, can usually be improved. The lack
of an effective management organisation of data, often varying in priority and in
time, may decrease the system performance, and in some disadvantageous cases
even cause a contradictory information flow.
The time aspect is one of the most important features and the synchronised
acquisition of sensor data and ought to be of primary interests in most applications.
This is a concern when dealing with different types of sensors which exhibit a dif-
ference in conversion time. In this case, the slowest sensor will decide the time slot
for the measurement process. Many situations may be optimised by effective tim-
ing and of course the right decision of when to take the “snap shot” of the measur-
ing spot. The cause of possible discrepancies may depend on several factors. The
sensor data may be directed with a time lag, the fusing procedure is time-related
or the decisions are made by different “fresh” information. Time procedures in a
sensor fusion system can be identified in two basic time schedules, Biel (2000).
The time schedule, as shown in Fig. 5.13, refers to a procedure that considers
a general translation of the data into a common time scale. The important feature
is that the time scale can relate occurring data to each other in a predetermined
schedule. The structure is organised in a direct perception procedure, that transfer
the data into a common time scale where the appearing sensor data is fused to
specific information. This specific perceptual sensor data flow compromised in
a sequence was introduced by fusion through vision and auditory experiments,
Bothe (1999).
The time schedule, as shown in Fig. 5.14, illustrates the process that sequen-
tially affects the following data as a dependent of the received sensor data is related
to the sequence of sensor data appearance. The sequential information procedure
is crucial to the appearance of the following information and may exhibit an ac-
tive perception procedure. In this example, illustrated in Fig. 5.14, the first fusion
procedure is performed before information added from other sensor systems con-
cludes the overall perceptual information. Thereby, the fusion process may also
introduce the weighting effects between different sensor systems that may pro-
vide various priorities to the output data from the systems. In Fig. 5.14 above,
we consider a sequential perception procedure that as a fist sub-process fuse vi-
sion and chewing sensing data followed by a similar activity when fusing smell
and taste data. The final sub-process will merge the fused results from the two