IEEE Robotics & Automation Magazine - December 2022 - 46

We propose two
comparatively nonpopular
future directions that
contribute to deploying
robots within our society.
perception modules. The scope of robotic scene perception is
low since fundamental commonsense skills are missing.
Commonsense skills can be generated with scene perception
frameworks that combine several specific methods
to achieve a holistic understanding. Second, although existing
re search on robot perception
achieves good
performance on specific
tasks, the flexibility for
adjusting the perception
within the robot usage is
missing. It must be possible
to extend initially
deployed perception skills
and to adjust the perception
at any time to
the scene. Therefore, we
address in particular our
second and third identified gaps to future research that
underlines the importance of perceptual learning being fundamental
in everyday life. It is a central technique toward and
beyond human-like perceptual performance.
Conclusion
This article transfers human scene perception to mobile
robots for comparison since the performance of human
perception is superior in many tasks. Current research in
robotics presents specific perception skills evaluated in a
predefined and constricted manner, whereby they already
achieve promising results for everyday applications. However,
for lifelong unsupervised and autonomous usage,
multifunctional mobile robots in particular need to perceive
the scene holistically to reliably handle unknown
and changing situations as well as uncertainties and
dynamics. Human perception has been extensively studied
since the 1970s, offering mature neuroscience studies
and theories that define human perception as the process
of recognizing, representing, and interpreting scene information.
This article uses this threefold division for a
transfer from human to robot scene perception to identify
similarities and differences in the process. The transfer
revealed that the robotic approaches partly mirror
human-like perception.
However, much research investigates specific methods,
such as object classification, that outperform human perception.
This prerequisite toward and beyond human-like perceptual
performance is promising. A new research area
integrates multiple state-of-the-art methods in frameworks
that aim toward a holistic scene understanding. However,
these frameworks lack trivial commonsense perception
skills and therefore cannot substantiate a holistic scene
understanding. Moreover, only nonpopular research contributes
to the perceptual learning of robots, which is needed
to learn, adjust, and customize their perception.
Therefore, these two major gaps need to be addressed by
future research.
46 * IEEE ROBOTICS & AUTOMATION MAGAZINE * DECEMBER 2022
Acknowledgment
This research has received funding from the German Ministry
of Economic Affairs, Labor and Tourism Baden-Württemberg
by the AI Innovation Center Learning Systems and
Cognitive Robotics under Grant Agreement 017-180036
(841022).
References
[1] C. Müller, B. Graf, and K. Pfeiffer, " World robotics 2021 - Service robots, "
International Federation of Robotics, Frankfurt, Germany, Tech. Rep., 2021.
[Online]. Available: https://ifr.org/ifr-press-releases/news/service-robots-hit
-double-digit-growth-worldwide
[2] A. Gasparetto and L. Scalera, " A brief history of industrial robotics in the
20th century, " Adv. Historical Stud., vol. 8, no. 1, pp. 24-35, 2019, doi: 10.4236/
ahs.2019.81002.
[3] U. Neisser, " Without perception, there is no knowledge: Implications for
artificial intelligence, " in Natural and Artificial Minds, R. G. Burton, Ed. Albany,
NY, USA: SUNY Press, 1993, pp. 174-164.
[4] M. Naseer, S. Khan, and F. Porikli, " Indoor scene understanding in 2.5/3d
for autonomous agents: A survey, " IEEE Access, vol. 7, pp. 1859-1887, 2019,
doi: 10.1109/ACCESS.2018.2886133.
[5] G. L. Malcolm, I. I. A. Groen, and C. I. Baker, " Making sense of real-world
scenes, " Trends Cogn. Sci., vol. 20, no. 11, pp. 843-856, 2016, doi: 10.1016/j.
tics.2016.09.003.
[6] S. Garg et al., " Semantics for robotic mapping, perception and interaction:
A survey, " Found. Trends® Robot., vol. 8, nos. 1-2, pp. 1-224, 2020, doi:
10.1561/2300000059.
[7] H. Levesque and G. Lakemeyer, " Cognitive robotics, " in Handbook of
Knowledge Representation (Foundations of Artificial Intelligence), vol. 3, F. v.
Harmelen, V. Lifschitz, and B. Porter, Eds. New York, NY, USA: Elsevier,
2008, pp. 869-886.
[8] M. Potter, " Meaning in visual search, " Science, vol. 187, no. 4180, pp. 965-
966, 1975, doi: 10.1126/science.1145183.
[9] I. Biederman, " Perceiving real-world scenes, " Science, vol. 177, no. 4043,
pp. 77-80, 1972, doi: 10.1126/science.177.4043.77.
[10] R. A. Epstein and C. I. Baker, " Scene perception in the human brain, "
Annu. Rev. Vis. Sci., vol. 5, no. 1, pp. 373-397, 2019, doi: 10.1146/annurevvision-091718-014809.
[11]
J. Beck, B. Hope, and A. Rosenfeld, Human and Machine Vision, 1st ed.
New York, NY, USA: Springer-Verlag, 1983.
[12] A. Oliva and A. Torralba, " Modeling the shape of the scene: A holistic
representation of the spatial envelope, " Int. J. Comput. Vis., vol. 42, no. 3, pp.
145-175, 2001, doi: 10.1023/A:1011139631724.
[13] C. M. Funke, J. Borowski, K. Stosio, W. Brendel, T. S. A. Wallis, and M.
Bethge, " Five points to check when comparing visual perception in
humans and machines, " J. Vis., vol. 21, no. 3, p. 16, 2021, doi: 10.1167/
jov.21.3.16.
[14] L. San Roque et al., " Vision verbs dominate in conversation across cultures,
but the ranking of non-visual verbs varies, " Cogn. Linguistics, vol. 26,
no. 1, pp. 31-60, 2015, doi: 10.1515/cog-2014-0089.
[15] M. Ward, G. Grinstein, and D. Keim, Interactive Data Visualization:
Foundations, Techniques, and Applications. Natick, MA, USA: A. K.
Peters, 2015.
[16] L. Seminara, P. Gastaldo, S. Watt, K. Valyear, F. Zuher, and F. Mastrogiovanni,
" Active haptic perception in robots: A review, " Frontiers Neurorobot.,
vol. 13, p. 53, Jul. 2019, doi: 10.3389/fnbot.2019.00053.
https://www.ifr.org/ifr-press-releases/news/service-robots-hit-double-digit-growth-worldwide https://www.ifr.org/ifr-press-releases/news/service-robots-hit-double-digit-growth-worldwide
IEEE Robotics & Automation Magazine - December 2022

Table of Contents for the Digital Edition of IEEE Robotics & Automation Magazine - December 2022
