IEEE Technology and Society Magazine - March 2023 - 46

Feature
Background: Autonomous systems
In the 19th century, the first remote control system
with limited autonomy capability was proposed
[14]. In the 1930s, primary types of ASs were used for
travel, including auto tiller to protect sailboats from
wind and autopilot to preserve altitude and speed
using a gyroscope [15]. Von Neumann et al. [16]
introduced a new field termed cybernetics in 1940
while developing automated range finders for antiaircraft
guns. In the 1950s, General Motors inspected
and tested new cars driven by special auto constructions
[15]. An autonomous rover system, which was
able to walk through the corridors of a building and
charge its battery by identifying a power outlet on
the wall and plugging it into it, was introduced in
1964 [17]. Digital control electronics was developed
in the 1970s. AI technology created automated perception
and cognition. With the integration of these
technologies, significant advances have been made
in ASs to perform complicated operations without
human intervention [17]. ASs are widely used to
decrease human labor as well as enhance productivity
and safety [18], [19].
Levels of autonomy
Human-robot interaction (HRI) defines the types
of robot categories, with each category requiring a
different level of autonomy [20]. Developers should
evaluate levels of autonomy suitable for their robot
according to the framework guidelines and investigate
the effects of autonomy on HRI [20].
According to SAE International's new standard
J3016, there are six levels of driving automation
for an autonomous vehicle (AV): level 0: no automation
(providing warnings and momentary assistance,
such as automatic emergency braking, blind
spot warning, and lane departure warning-requires
supervising steer, brake, or accelerate systems constantly);
level 1: driver assistance (providing steering
OR brake/acceleration support to the driver,
such as lane centering, OR adaptive cruise control-requires
supervising steer, brake, or accelerate
systems constantly); level 2: partial automation
(providing steering AND brake/acceleration support
to the driver, such as lane centering AND adaptive
cruise control simultaneously-requires supervising
steer, brake, or accelerate systems constantly);
level 3: conditional automation (the vehicle is
driven under limited conditions and is operated
when all conditions are met, such as traffic jam
46
chauffeur-requires a driver); level 4: high automation
(the vehicle is driven under limited conditions
without a driver and is operated when all requirements
are met, such as a local driver-less taxi-does
not require a driver); and level 5: full automation
(driving the vehicle everywhere under all conditions-does
not require a driver). At levels 0, 1, and
2, a human driver monitors the driving environment,
while at levels 3-5, an automated driving system performs
surveillance [21].
Aeronyde defines drones with seven levels of
autonomy. Level 0: no automation (the pilot controls
every movement and the platforms manually);
level 1: pilot assistance (overall vehicle operation
and safety remain under the pilot's control. A drone,
however, can temporarily perform at least one vital
function. The applications of this type of vehicle
are inspection and maintenance, localization and
detection, photography and filming, protection and
security, and monitoring); level 2: partial automation
(safety is still the pilot's responsibility, from monitoring
airspace and flight conditions to reacting in an
emergency. Nevertheless, the drone is capable of
taking over control of heading, altitude, and speed
in certain situations. The applications of this type of
vehicle are mapping, surveying, spaying and seeding,
and measuring); level 3: conditional automation
(at this level, the drone is capable of performing all
functions " under certain conditions. " Nevertheless,
the human pilot must always remain alert and ready
to take over. The applications of this type of vehicle
are mapping, surveying, and delivery); level 4: high
automation (there is no need for a human pilot to
always control a drone. The drone can fly on its own
full-time if the right conditions are met. The applications
of this type of vehicle are photography and
filming); level 5: adaptive autonomy (the drone's
function allocation needs to be adapted to changes
in environmental conditions. This is referred to as
adaptive autonomy); level 6: augmented autonomy
(the augmented autonomy drones perform better
than both manually and autonomously controlled
drones) [22].
The Air Force Research Laboratory has outlined
levels of autonomy for different systems. In
this regard, Los Alamos National Laboratory has
described six autonomy levels for mobility, acquisition,
and protection (MAP) survival space. Level
" - " : motion only occurs under the application of an
external force (the energy of this type of system is
IEEE Technology and Society Magazine
IEEE Technology and Society Magazine - March 2023

Table of Contents for the Digital Edition of IEEE Technology and Society Magazine - March 2023
