Computer vision often refers to the
visible spectrum, making it easier for
humans to program robots that see
what we see. But AI is expanding the
possibilities, enabling vision beyond the
human-visible wavelengths."
Arnaud Deleule
Director of Application,
STMicroelectronics
However, vision systems are not the only tools used
by machines to understand their surroundings,
nor are they limited to passive functionality.
Active sensors—including ultrasonic, radar, and
light detection and ranging (lidar)—emit energy
and detect how it interacts with the environment.
Vision systems require complex interpretation of
rich visual data, whereas active sensors such as
lidar, radar, and ultrasonic produce more focused
outputs like distance or velocity. Processing these
signals typically requires less computing power than
vision systems, allowing the data to be interpreted
and acted upon more quickly. Speed is important
for obstacle detection and distance management,
which are tasks that depend on quick analysis to
provide rapid actions. By combining different
sensor types, machines create a clearer picture of
their surroundings.
AI and the Use of Complex Data
Data are the biggest drawback when using sensors.
Although ample data are available from multiple
sources, analyzing and then acting on these
data in a timely manner is challenging,
which is where AI comes into play.
C h a p t e r 3 | S e n s i n g a n d Pe r c e p t i o n
20
Autonomy Meets Intelligence: Enabling the Future of Factory Automation Autonomy Meet
