IEEE - Aerospace and Electronic Systems - July 2021 - 66

Artificial Intelligence and Data Fusion at the Edge
Table 1.
Comparison of Data Fusion and AI Architectures Symbols: @-Yes; $-Being Used; n-Been Adapted; tu-Could
Be Adapted; *-Not Used
Model and/or
Architecture
DFD
DIG
WSN
ELF
GAN
JDL/DFIG
Data Fusion
(Static)
@
$
$
$
Real-Time
(Sensing)
$
@
@
Human
Centered
ut
n
*
IoT/
CPS
*
$
n
AI
(Big
Data)
n
ut
ut
$ ** *
@
$
@
CPCG @@ @
operating picture (UDOP) in the Joint Directors ofLaboratories
(JDL)/Data Fusion Information Group (DFIG) model
[17]. We note that the UDOP concept extends the common
operating picture to help enable the human operator to supervise
the processing, exploitation, and dissemination ofinformation
for situation awareness. The UDOP enables
rendering and visualization of data analytics services customized
to the human operator's needs for efficient command
decision making for a given mission. While the DFIG
model supports multiple sensors and distributed users, it still
relies on a common location for command and control. Other
communities have been looking at fusion architectures that
emphasize fusion-focused data analysis, such as the datafeature-decision
(DFD) [12] model, wireless sensor networks
[14], and the more recent efforts from the AI community
with the early and late fusion (ELF) methods [15] and
generative adversarial networks (GANs). Since 2000, the
DFD and the JDL/DFIG have remained similar, while the
WSN community has continually utilized the distributed
approaches that extend to IoT and CPS. While each of these
methods are important for the fusion community, there is a
desire for a data fusion architecture that is able to perform
distributed command and control amongst a variety of
human teams and distributed sensors. Hence, in this section,
we elaborate cyber-physical command guided (CPCG)
architecture that is able to perform distributed command and
control amongst various human teams and distributed sensors
and IoT devices. We further briefly discuss DFIGmodel
as it help provide an understanding offusion levels.
DATA FUSION INFORMATION GROUP (DFIG) MODEL
Figure 2 depicts a JDL/DFIG model. The JDL/DFIG
model defines multiple levels of fusion where each level
exploits AI developments to support assessment (level 0,
66
Figure 2.
DFIG model for information fusion (IF).
IEEE A&E SYSTEMS MAGAZINE
JULY 2021
ut ** @
@
@
@
Connected
(Distributed)
ut
@
@
$
*
@
Reference
Dasarathy[12]
Hall et al.[13]
Chair and
Varshney[14]
Snoek et al.[15]
Goodfellow etal.[16]
ut $$ Blasch et al.[17]
Cruise et al.[18]
" yes " indicates that (yes) research supports the directions/usage while " used " implies that it has been implemented in real systems.
1, 2 and 3) and refinement (level 4, 5, and 6). System management
(level 6) incorporates contextual constraints
based on mission, objectives, and goals. The DFIG model
describes different levels of fusion as follows.
Level 0-Data Assessment. Provides estimation and
prediction of signal/object observable states based on
pixel/signal level data association.
Level 1-Object Assessment. Provides estimation and
prediction of entities/objects based on data association
and (both continuous and discrete) state estimation.
Level 2-Situation Assessment. Provides estimation
and prediction of relations between entities/objects.
Level 3-Impact Assessment. Provides estimation and
prediction of impact on planned or estimated actions by
the participants.
Level 4-Process Refinement. An element of resource
management and provides adaptive data acquisition and
processing to furnish estimation of sensing objectives and
prediction of impact on planned or estimated actions by
the participants.
IEEE - Aerospace and Electronic Systems - July 2021

Table of Contents for the Digital Edition of IEEE - Aerospace and Electronic Systems - July 2021
