IEEE Circuits and Systems Magazine - Q3 2020 - 17

V. Conclusions
The integration of ML within IoT devices will provide them
with a flexibility and processing capabilities never seen before. The TinyML paradigm proposes to adapt advanced
ML techniques to the constraints of MCUs in order to convert them in real smart objects. In this work, a comprehensive review and discussion of this novel and vibrating ecosystem have been presented. The related challenges and
opportunities have been identified as well as the potential
services that will be enabled by these intelligent small elements. The TinyML frameworks available in the market
that currently permit the integration of complex ML mechanisms within MCUs have been also reviewed and their applicability has been discussed. Finally, a realistic use case
has been explored. Concretely, a multi-RAT smart object architecture has been proposed and the ML integration process has been explained. A number of TinyML frameworks
have been employed for solving the problem of selecting
the most adequate communication interface considering
both the status of the device and the characteristics of the
data to be sent. We have demonstrated that several classification algorithms can be integrated within an Arduino Uno
board for tackling this issue, namely, SVM, MLP, decision
trees, and RF. From the attained outcomes, we can conclude
that decisions trees and RF present the best performance
in terms of accuracy, memory footprint, and classification
speed. Other algorithms also showed good results in terms
of classification accuracy but their implementation should
be carefully addressed given the memory restrictions of
the considered MCU. As future work, we plan to consider
other constrained devices in order to benchmark additional
TinyML frameworks. Besides, other relevant aspects such
as decision safety and model correctness in sensitive fields
of application, e.g., eHealth, will be considered.
Acknowledgment
This work has been supported by the European Commission, under the projects IoTCrawler (Grant No. 779852),
CyberSec4Europe (Grant No. 830929) and Fed4IoT (Grant
No. 814918), and by the Spanish Ministry of Science, Innovation and Universities, under the project PERSEIDES
(Grant No. TIN2017-86885-R with ERDF funds).
Ramon Sanchez-Iborra received the
B.Sc. degree in telecommunication engineering and the M.Sc. and Ph.D. degrees
in information and communication technologies from the Technical University
of Cartagena, in 2007, 2013, and 2016, respectively. He is currently an Assistant Professor and a
Researcher with the Information and Communications
Engineering Department, University of Murcia. His main
research interests are evaluation of QoE in multimedia
THIRD QUARTER 2020 		

services, management of wireless mobile networks, green
networking techniques, and IoT/M2M architectures.
Antonio Skarmeta received the B.S. degree (Hons.) from the University of Murcia, Spain, the M.S. degree from the University of Granada, and the Ph.D. degree
from the University of Murcia, all in computer science. He has been a Full Professor with the University of Murcia, since 2009. He has been
part of many EU FP projects and even coordinated some of
them. He has published more than 200 international articles. His main interests include the integration of security
services, identity, the IoT, 5G, and smart cities.
References
[1] L. Columbus, "10 Charts that will change your perspective of big
data's growth," Tech. Rep., 2018. [Online]. Available: https://software
strategiesblog.com/2018/06/02/10 - charts-that-will- change -your
-perspective-of-big-datas-growth/
[2] Y. Sun, M. Peng, Y. Zhou, Y. Huang, and S. Mao, "Application of machine learning in wireless networks: Key techniques and open issues,"
IEEE Commun. Surveys Tut., vol. 21, no. 4, pp. 3072-3108, 2019. doi:
10.1109/COMST.2019.2924243. [Online]. Available: https://ieeexplore
.ieee.org/document/8743390/
[3] T. Baltrusaitis, C. Ahuja, and L.-P. Morency, "Multimodal machine learning: A survey and taxonomy," IEEE Trans. Pattern Anal.
Mach. Intell., vol. 41, no. 2, pp. 423 - 443, Feb. 2019. doi: 10.1109/
TPAMI.2018.2798607. [Online]. Available: https://ieeexplore.ieee.org/
document/8269806/
[4] Q. Chen et al., "A survey on an emerging area: Deep learning for
smart city data," IEEE Trans. Emerg. Topics Comput. Intell., vol. 3, no. 5,
pp. 392-410, Oct. 2019. doi: 10.1109/TETCI.2019.2907718. [Online]. Available: https://ieeexplore.ieee.org/document/8704334/
[5] G. Li, R. Gomez, K. Nakamura, and B. He, "Human-centered reinforcement learning: A survey," IEEE Trans. Human-Mach. Syst., vol. 49,
no. 4, pp. 337-349, Aug. 2019. doi: 10.1109/THMS.2019.2912447. [Online].
Available: https://ieeexplore.ieee.org/document/8708686/
[6] T. Luo et al., "DaDianNao: A neural network supercomputer," IEEE Trans.
Comput., vol. 66, no. 1, pp. 73-88, Jan. 2017. doi: 10.1109/TC.2016.2574353.
[Online]. Available: http://ieeexplore.ieee.org/document/7480791/
[7] B. Garbinato, R. Guerraoui, J. Hulaas, M. Monod, and J. H. Spring, "Pervasive computing with frugal objects," in Proc. IEEE 21st Int. Conf. Advanced
Information Networking and Applications Workshops (AINAW'07), 2007, pp.
13-18. [Online]. Available: http://ieeexplore.ieee.org/document/4224076/
[8] C. Bormann, M. Ersue, and A. Keranen, Terminology for ConstrainedNode Networks, IETF RFC 7228, 2014.
[9] M. S. Hadj Sassi, F. G. Jedidi, and L. C. Fourati, "A new architecture
for cognitive internet of things and big data," Proc. Comput. Sci., vol. 159,
pp. 534-543, 2019. doi: 10.1016/j.procs.2019.09.208. [Online]. Available:
https://linkinghub.elsevier.com/retrieve/pii/S1877050919313924
[10] R. Roman, J. Lopez, and M. Mambo, "Mobile edge computing, Fog
et al.: A survey and analysis of security threats and challenges," Future Gener. Comput. Syst., vol. 78, pp. 680-698, Jan. 2018. doi: 10.1016/j.
future.2016.11.009. [Online]. Available: https://linkinghub.elsevier.com/
retrieve/pii/S0167739X16305635
[11] B. Martinez, M. Monton, I. Vilajosana, and J. D. Prades, "The power of
models: modeling power consumption for IoT devices," IEEE Sensors J.,
vol. 15, no. 10, pp. 5777-5789, Oct. 2015. doi: 10.1109/JSEN.2015.2445094.
[Online]. Available: http://ieeexplore.ieee.org/document/7122861/
[12] TinyML. 2020. [Online]. Available: https://www.tinyml.org/
[13] C. MacGillivray and M. Torchia, "Internet of Things: Spending
trends and outlook," Tech. Rep., 2019. [Online]. Available: https://www
.idc.com/getdoc.jsp?containerId=US45161419
[14] H. Doyu, "TinyML as a Service and the challenges of machine learning at the edge," Ericsson, Tech. Rep., 2019. [Online]. Available: https://
www.ericsson.com/en/blog/2019/12/tinyml-as-a-service
IEEE CIRCUITS AND SYSTEMS MAGAZINE	

17
IEEE Circuits and Systems Magazine - Q3 2020

Table of Contents for the Digital Edition of IEEE Circuits and Systems Magazine - Q3 2020
