IEEE Systems, Man and Cybernetics Magazine - October 2019 - 25

environments [27]. In agriculture, another popular domain
for autonomous vehicles, the operational environment is
the fields of crops where autonomous functionality could
be used to monitor pests and diseases and take actions to
prevent losses in crop yield [28], [29].
Yet another way to visualize OA is through its possible
applications in the command and control (C2) process
[15], [30]. In the armed forces, C2 refers to the passage of
a commander's orders and the interpretation of her/his
intent to take specific action(s). C2 entails a complex
planning process that involves multiple participants, substantial facet knowledge, a robust communications archiInterpretations of OA
tecture, and possibly some hidden or otherwise latent
The interpretation of OA varies among groups of researchprocesses. Currently, the only portion of autonomy that is
ers and stakeholders. To most, it is about enabling the
utilized in such complex processfunctions that are performed withes includes intelligence, surveilout the involvement of a human in
lance, and reconnaissance (ISR)
the loop, which defines the overall
systems, which are commonly
functionality of the AUS, to be
Operationalizing
employed using machine-learning
operational in actual missions
autonomy focuses
tools [31], [32]. This suggests that
[13]-[15]. The behavior of such
on the large-scale
there is a lot that could be accomautonomous functions results from
plished when it comes to OA in C2.
a series of decisions or control
implementation of
A possible scenario is illustrated
loops that perform specific tasks.
autonomous systems
in Figure 2, which shows a largeThe primary challenge regarding
autonomy is the autonomy itself,
scale
system that includes multiand their applications
including the required degrees of
ple manned military assets and
and commercial
independence and which capabiliunmanned autonomous agents
ties or tasks can be characterized
from three different domains.
availability.
and validated for its readiness and,
Additional autonomy technolothus, relied upon in operational
gies, such as automated reasonsettings [16], [17]. As autonomous
ing, artificial intelligence-based
functions evolved, the integration of autonomy algorithms
planning and scheduling, and natural language processinto systems improved their capabilities.
ing can be integrated into the C2 process [13].
Autonomous functions span a range of operational utiliAnother perspective regarding OA is the development of
ties, such as mission planning, goal reprioritization, path
systems that can exhibit human-like decision making, a
planning, task assignment and scheduling, trajectory folrather challenging expectation from a machine/mechanical
lowing, and interloop flight control systems [18]-[22]. A
entity. The autonomous functionalities, such as those used
great example of OA is the Automatic Ground Collision
for ISR, are examples that exhibit a system-wide implemenAvoidance System that was recently deployed in a military
tation of autonomy technologies. An unmanned ISR plataircraft [23], [24]. This function detects the emergency situform with embedded autonomy is privileged in the sense
ation of a manned aircraft that is about to impact the
that it has technical and quantitative information that is
ground and takes control to return the plane to a preavailable in real time in the form of pictures and streaming
defined cruising state. It is one of the most promising
video. This makes it possible for an unmanned ISR asset to
examples of building trust in autonomy when it comes to
initiate intelligent decisions, if enabled, regarding the inforsaving lives [25], [26].
mation more quickly than remotely located humans/perAnother school of thought relative to OA regards the
sonnel. If such an unmanned system were deployed in a
operational environment. In this case, it implies autonovery complex environment that provided little or no informous operation in less constrained environments (beyond
mation, it would become very challenging to achieve the
test cases and laboratories). In such environments, OA is
ISR's rapid, human-like decision making.
manifested in ground assets, air vehicles, and sea vessels
carrying out a part of a mission. From a military perspecConcerns Related to OA
tive, the operational environment is the battlefield or front
There is an ongoing debate related to the concerns
edge of a particular outpost [13]. From a commercial
about OA. For example, there is concern whether an
standpoint, the operational environment is typically closer
autonomous system should be allowed on the battleto centers of civilian life. The environment for autonomous
field (especially a lethal one). A segment of society
vehicles may include Interstate routes and dense urban
believes that the operation of autonomous unmanned
The State-of-the-Art in OA
OA requires an understanding on a broader scale that
addresses technical and nontechnical aspects as well as
its social acceptance and associated policies and regulations. OA is about transitioning autonomy beyond prototypes and integrating them into systems at different scales
in operational industrial, civil, and military settings. This
section covers different interpretations of the concept,
societal concerns, and recent developments toward OA at
large scales.

O c tob e r 2019

IEEE SYSTEMS, MAN, & CYBERNETICS MAGAZINE

IEEE Systems, Man and Cybernetics Magazine - October 2019

Table of Contents for the Digital Edition of IEEE Systems, Man and Cybernetics Magazine - October 2019