Artificial Intelligence and Tactile Healthcare for Mitigating the Impact of COVID-19
By Ali Nauman, Rashid Ali (Member, IEEE), Yousaf Bin Zikria (Senior Member, IEEE), and Sung Won Kim
The COVID-19 epidemic, and the recent related waves of variant outbreaks, have had a significant impact on every field of life. One of the major impacts of COVID is the increased stress on the already exhausted healthcare system. Tactile healthcare is revolutionizing healthcare systems. The 5G and Beyond-5G (B5G) technologies are expected to enable Tactile Healthcare applications, which are time-sensitive and critical. The 5G and B5G communication networks were constructed to support high data rates on an energy-efficient platform and to provide ultra-reliability and low latency. The use of Reinforcement Learning (RL) algorithms, a type of Machine Learning (ML), can enhance the capability of 5G and B5G networks by optimizing latency and reliability in terms of data delivery. The idea of RL algorithms is to make a system capable of mimicking the human brain and enhancing its abilities. The New Radio (NR) in B5G offers flexible Medium Access Control (MAC) frame structures and scalable numerology. The efficient MAC scheduling approaches are of prime importance for wireless networks. The efficient MAC scheduling protocols and RL algorithms ultimately solve the problems of reliability and latency for Tactile Healthcare applications.
This article provides an overview of how ML can improve the scheduling protocols for MAC layers, which can increase the performance of tactile healthcare applications in B5G networks, and in turn, contribute to resolving the challenges of tele-surgery in order to mitigate the impact of COVID-19 on healthcare systems. Moreover, this article provides open research issues due to NR’s flexible frame structure and scalable numerology in future directions.
Importance and Issues Related to Tactile Healthcare
COVID-19 had a huge impact on daily life in every aspect. The already exhausted healthcare system was the first and most highly impacted industry affected by COVID. According to Statista, healthcare systems in the United Kingdom (UK) were worsened by 56% for non-infected persons due to prioritization of COVID-19 . Normal surgical practices are experiencing major challenges due to limited resources, patients with diminished mental stability, and risk of potential viral infection. Tactile Internet (TI) is revolutionizing healthcare systems. The word TI was first used in 2014 by G. Fettweis . The ultra-high reliability, ultra-low latency, edge intelligence, data integrity, etc., are some of the key features of the B5G to enable TI-based applications . These features, provided by the B5G, form the basis for Tactile Healthcare to strengthen the process. In the current era of B5G, Tactile Healthcare has shown immense potential to measure up with the shortage of surgeons and provide healthcare services in remote localities. Tele-surgery requires ultra-reliability and extremely low latency to provide an efficient service to the end-user. However, the existing systems that are using tele-surgery possess high communication and latency overheads. Indeed, these overheads put a big question mark on the future applications of such a beneficial service. For instance, a report from the University of Illinois states a record of 1,391 injuries and at least 144 deaths, within the US since 2004, due to robotic surgeries/procedures. The main reasons behind these failures were power supply, network latency, and system errors . These issues are well-addressed by the Third-Generation Partnership Project (3GPP) for B5G based-communication, so it is foreseen that Tactile Healthcare can benefit from 5G to effectively reduce the number of robotic failures. Figure 1 shows in detail how B5G networks are involved in tactile tele-surgery.
Preliminary Background on B5G Supporting Ultra-Reliable and Low Latency Communication (URLLC)
The B5G is evolving rapidly to cater to the high data rate demands of the ever-growing number of subscribers. The 3GPP has narrowed down the applications of 5G into the following three main categories :
- Enhanced mobile broadband (eMBB)
- Massive machine type communication (mMTC)
- Ultra-reliable and low latency communication (URLLC)
The latency and reliability requirements for these applications vary dramatically to satisfy the needs of end-users. Tactile Healthcare falls into the URLLC category, and the B5G network is expected to support URLLC applications with a 1000x increase in data rates, 99.999% reliability, and a latency of 1 ms (over the radio), hence satisfying the requirements of Tactile Healthcare . One of the solutions proposed by 3GPP is to shorten the Transmission Time Interval (TTI) from 14 OFDM symbols to fewer –that is seven, four, or two symbols . To achieve this, 3GPP standardizes the concept of mini-slots, which allows flexibility in selecting the OFDM symbols. Moreover, the main difference between Long Term Evolution (LTE) and NR is scalable numerology. NR offers scalable SubCarrier Spacing (SCS) with 2n x 15 kHz, where (n = 0, 1, 2,…, N). Increasing the SCS decreases the TTI. Figure 2 depicts the flexible frame structure and scalable numerology in detail. Moreover, it depicts how flexibility and SCS numerology affect the latency and, in return, the reliability. Queueing delay in MAC scheduling occurs due to the statistical multiplexing of data flow for multiple users. Furthermore, the 3GPP standardized the mechanism of puncturing in NR for the URLLC traffic to meet stringent latency and reliability requirements, which categorically halts the ongoing eMBB traffic to transmit URLLC traffic in mini-slots without notifying eMBB user equipment (UE) . The puncturing mechanism degrades the Quality-of-Service (QoS) for an eMBB user. Sometimes, the degrading in QoS experienced by eMBB UE is more than the gain achieved by transmitting URLLC traffic. Increasing the number of devices increases the queueing delay, which degrades reliability by dropping data packets from the queue. Efficiently choosing the size of a mini-slot for URLLC applications with an appropriate SCS can minimize the MAC scheduling delay and maintain the eMBB QoS .
Can Reinforcement Learning offer Low latency and Ultra reliability in Wireless Networks?
RL is a type of ML, which follows the framework of the Markov Decision Process (MDP). Unlike other ML model-driven approaches, RL learns by directly interacting with the environment. Moreover, RL works with a lot less complexity and computation . RL algorithms can learn the presence of heterogeneous services and their requirements in the network. Furthermore, RL techniques adapt to the changes in the network in real-time applications without the need to train data sets. To optimize MAC scheduling, RL reduces latency and increases reliability further for URLLC Tactile Healthcare applications. The use of RL in choosing the size of a mini-slot, to optimize the latency and reliability, reduces the cost of heavy computation and complexity compared to other heuristic approaches or exhaustive search mechanisms. Incorporating RL enables the network to work in a self-organizing and self-sustaining manner. On the contrary, the use of RL can significantly improve the Tactile Healthcare applications and can help the research community make the healthcare industry more useful and reliable. Figure 3 shows RL agent-environment interaction in detail and demonstrates how RL can be used to select the optimal size of mini-slots and SCS numerology for URLLC applications.
Future Direction and Conclusion
Two additional open research challenges exist due to the flexible frame structure of NR—the Inter-Numerology Interference (INI) and position of Signal Synchronization Block (SSB), which directly depends on the SCS numerology. Choosing higher numerology shortens the cyclic prefix, which causes INI . The SSB constitutes cell and sector identification, critical for initial cell search and channel access . Both these issues could lead to degradation of reliability, which will augment latency as well. RL-enable INI-aware resource allocation and mapping of SSB to OFDM symbols provide viable solutions.
The concept of Internet-of-Everything (IoE) in B5G and 6th generation (6G) of cellular communication involves the massive number of connected devices, which forms an ultra-dense network environment. In order to meet the stringent latency and reliability requirement of such a massive number of devices, NR offers a grant-free random-access process. However, in grant-free random access, the channel quality of the user equipment is unknown, which degrades the QoS. Integration of AI and device-to-device communication (D2D) acts as a piece of the B5G jigsaw puzzle and offers to fill the QoS gap for massive channel access problems [13, 14].
The use of AI in every aspect of life raises ethical questions. However, when we discuss the improvement of applications, which are directly involved in improving the healthcare system and reducing the number of death counts, we recommend the use of AI. AI can reduce the complexity of future generation communication, which requires the coexistence of heterogeneous applications with different service requirements within a network. This article presents how AI can improve the B5G networks to augment the efficiency of tactile tele-surgery in order to mitigate the impact of COVID-19 on non-infected people who are suffering from life-threatening diseases and require timely treatment.
- Frederic Michas. “Extent of COVID-19 pandemic affecting the care of patients in the UK as of April 2020,” in https://www.statista.com/statistics/1114759/how-covid-19-pandemic-affects-care-to-other-patients-in-the-uk/, Accessed [2021-05-26].
- Gerhard Fettweis. “The Tactile Internet: Applications and Challenges”. IEEE Vehicular Technology Magazine, 2014;9(1): 64-70. DOI: 10.1109/MVT.2013.2295069.
- Gupta, R. et al. “Tactile-Internet-Based Telesurgery System for Healthcare 4.0: An Architecture, Research Challenges, and Future Directions”. IEEE Network, 2019;33(6): 22-29. DOI: 10.1109/MNET.001.1900063.
- Zhang, Q. et al. “Towards 5G Enabled Tactile Robotic Telesurgery”. 2018. DOI: arXiv:1803.03586v1
- Ali, R. et al. “URLLC for 5G and Beyond: Requirements, Enabling Incumbent Technologies and Network Intelligence”. IEEE Access. 2021;9: 67064-67095. DOI: 10.1109/ACCESS.2021.3073806.
- Yu, H. et al. “Tactile Internet: Technologies, test platforms, trials, and applications”. Future Generation Computer Systems. 2020;106: 685-688. DOI: org/10.1016/j.future.2020.01.057.
- Li, C. et al. “5G-Based Systems Design for Tactile Internet”. Proceedings of the IEEE. 2019;107(2): 307-324. DOI: 10.1109/JPROC.2018.2864984.
- Korrai, P. et al. “A RAN Resource Slicing Mechanism for Multiplexing of eMBB and URLLC Services in OFDMA Based 5G Wireless Networks”. IEEE Access. 2020;8: 45674-45688.
- Sachs, J. et al. “Adaptive 5G Low-Latency Communication for Tactile Internet Services”. Proceedings of the IEEE. 2019;107(2): 325-349. DOI: 10.1109/JPROC.2018.2864587.
- Chien, W. et al. “Intelligent Architecture for Mobile HetNet in B5G”. IEEE Network. 2019;33(3): 34-41. DOI: 10.1109/MNET.2019.1800364.
- Yazar, A. et al. “Reliability enhancement in multi-numerology-based 5G new radio using INI-aware scheduling”. EURASIP Journal on Wireless Communications and Networking. 2019;110(2019). DOI: doi.org/10.1186/s13638-019-1435-z
- Chen, F. et al. “Design and implementation of initial cell search in 5G NR systems”. China Communications. 2020;17(5): 38-49. DOI: 10.23919/JCC.2020.05.005
- Nauman, A. et al. “Reinforcement learning-enabled Intelligent Device-to-Device (I-D2D) communication in Narrowband Internet of Things (NB-IoT)”. Computer Communications. 2021;176: 13-22. DOI:org/10.1016/j.comcom.2021.05.007
- Zhang, S. et al. “Envisioning Device-to-Device Communications in 6G”. IEEE Network. 2020;34(3): 86-91. DOI: 10.1109/MNET.001.1900652
Dr. Ali Nauman is a doctoral candidate in the Wireless Information Networking Laboratory (WINLab) at the Department of Information and Communication Engineering (ICE), Yeungnam University, Republic of Korea. He received his B.E. degree in electrical (telecommunication) engineering from COMSATS University Islamabad, Pakistan, in 2013, and his M.S. degree in wireless communications from the Institute of Space Technology Islamabad, Pakistan, in 2016. The main domain of his research is in the field of artificial intelligence-enabled wireless networks for healthcare, multimedia, and industry 5.0. The research interest also includes radio resource management and allocation for 5G and Beyond-5G (B5G) networks, routing protocols, Internet-of-Everything (IoE), URLLC, Tactile Internet (TI), and artificial intelligence.
Rashid Ali (Member IEEE) received his Ph.D. degree from Dept. of Information and communication Engineering, Yeungnam University, Republic of Korea, in February 2019. He is working as an Assistant Professor at the School of Intelligent Mechatronics Engineering, Sejong University, Republic of Korea. His research interests include next-generation wireless local area networks (IEEE 802.11ax/ah), unlicensed wireless networks in 5G, Internet of Things, performance evaluation of wireless networks, Named-Data/Information-Centric Networking, Deep Learning techniques for wireless networks, and URLLC requirements for Tactile Internet.
Yousaf Bin Zikria (Senior Member, IEEE) is currently working as an Assistant Professor in the Department of Information and Communication Engineering, Yeungnam University, South Korea. He authored more than 80 articles, conferences, book chapters, and patents. He published papers at the top venue, including IEEE Communications, Surveys, and Technologies, IEEE Wireless Communications Magazine, Elsevier Future Generation Computer Systems, Elsevier Sustainable Cities and Society, etc. He has managed numerous FT/SI in SCI/E indexed journals. His research interests include IoT, 5G, Machine Learning, wireless communications and networks, WSNs, routing protocols, CRAHN, CRASN, transport protocols, VANETS, embedded systems and, network and information security. He also held the prestigious CISA, JNCIS-SEC, JNCIS-ER, JNCIA-ER, JNCIA-EX, and Advance Routing Switching and WAN Technologies certifications.
Sung Won Kim received his B.S. and M.S. degrees from the Department of Control and Instrumentation Engineering, Seoul National University, Korea, in 1990 and 1992, respectively, and his Ph.D. degree from the School of Electrical Engineering and Computer Sciences, Seoul National University, Korea, in August 2002. In March 2005, he joined the Department of Information and Communication Engineering, Yeungnam University, Gyeongsangbuk-do, Korea, where he is currently a Professor. His research interests include resource management, wireless networks, mobile computing, performance evaluation, and machine learning.
Dr. Jerry Chun-Wei Lin received his Ph.D. from the Department of Computer Science and Information Engineering, National Cheng Kung University, Tainan, Taiwan in 2010. He is currently a full Professor with the Department of Computer Science, Electrical Engineering and Mathematical Sciences, Western Norway University of Applied Sciences, Bergen, Norway. He has published more than 400+ research articles in refereed journals (IEEE TKDE, IEEE TCYB, IEEE TII, IEEE TITS, IEEE TIAS, IEEE TETCI, IEEE SysJ, IEEE SensJ, IEEE IOTJ, ACM TKDD, ACM TDS, ACM TMIS, ACM TOIT, ACM TIST) and international conferences (IEEE ICDE, IEEE ICDM, PKDD, PAKDD), 11 edited books, as well as 33 patents (held and filed, 3 US patents). His research interests include data mining, soft computing, artificial intelligence and machine learning, and privacy-preserving and security technologies. He is the Editor-in-Chief of the International Journal of Data Science and Pattern Recognition, the Guest Editor/Associate Editor for several IEEE/ACM journals such as IEEE TFS, IEEE TII, ACM TMIS, ACM TOIT, and IEEE Access. He has recognized as the most cited Chinese Researcher respectively in 2018 and 2019 by Scopus/Elsevier. He is the Fellow of IET (FIET), senior member for both IEEE and ACM.