Multiple Service Class TDMA Protocol for Healthcare Applications

Sophia Athanasiadou, Georgios Papadimitriou and Petros Nicopolitidis

Dept. of Informatics, Aristotle University of Thessaloniki, Thessaloniki, 54 124, Greece

Keywords:

Classes of Service, Healthcare Applications, Tactile Internet, Time Division Multiple Access.

Abstract:

A novel medium access control (MAC) protocol designed for healthcare environments is proposed. The pro-

tocol is aimed to support different Classes of Service (CoS) with diverse needs, each including different kinds

of medical applications. The highest priority class includes Tactile Internet applications, that require an end-

to-end latency of less than 1 msec, while the rest of the classes are assigned graded priorities.The proposed

protocol provides the highest priority class with the required bandwidth and guaranteed low delay while the

other classes, which also include critical applications, do not starve. Two protocol versions are introduced

with the second version exhibiting very good performance in terms of average delay and throughput.

1 INTRODUCTION

In the recent years many medical applications have

been developed and are increasingly used in health-

care environments, such as hospitals and clinics.

Most of these applications involve data transmission

and thus rely on appropriate networks that are capa-

ble of serving their communication needs. However,

healthcare applications vary and have different Qual-

ity of Service (QoS) necessities. Therefore, to serve

their individual needs in a better way it is important to

classify network trafﬁc into CoS with different QoS

requirements.

On the other hand, TDMA protocol has been

extensively studied ((Tanenbaum, 1996), (Papadim-

itriou and Pomportsis, 2000), (Rubin and Baker,

1990)), as a simple and utterly collision-free method

of coordinating multiple access to the same medium.

TDMA protocol is widely studied today in numerous

ﬁelds, such as vehicular networks ((Tianjiao and Qi,

2017), (Li et al., 2020), (Madi and Al-Qamzi, 2019)),

ad hoc networks ((Zhao et al., 2019), (Liu et al.,

2019b)), wireless sensor networks ((Liu et al., 2019a),

(Chang et al., 2019)), Internet of Things ((Batta

et al., 2019)) and wireless power transfer ((Bayat and

Aissa, 2019)) among others. The support of mul-

tiple CoS in TDMA-based protocols has been stud-

ied in (Capone and Stavrakakis, 1999b), (Wang and

Iversen, 2008), (Tong and Hamdi, 2000) and (Capone

and Stavrakakis, 1999a). In (Capone and Stavrakakis,

1999b) the call admission region for a TDMA system

supporting applications with different QoS require-

ments is determined, where QoS is deﬁned by the

maximum tolerable packet delay and dropping prob-

ability. Also, in (Wang and Iversen, 2008) the Erlang

capacity of a TDMA system which employs adaptive

modulation and coding is evaluated. In (Tong and

Hamdi, 2000) and (Capone and Stavrakakis, 1999a)

the achievable QoS is determined for a TDMA sys-

tem serving competing applications.

Tactile internet has received a lot of attention

lately ((Haddadin et al., 2019), (Kim et al., 2019),

(Gupta et al., 2019)), as the next internet revolution

after Internet of Things. It requires very low latency

and extremely high availability, reliability and secu-

rity. Due to these features it can provide new capabil-

ities in ﬁelds such as remote healthcare, robotics and

education.

In (Soomro and Cavalcanti, 2007) ﬁve CoS are

proposed that include all trafﬁc in a medical facility.

In Table 1 the ﬁve classes appear according to their

priority. Also, their requirements in delay and packet

loss are shown. The most demanding class is Tac-

tile Intenet Medical which includes Tactile Internet

applications, such as tele-diagnosis, tele-surgery and

tele-rehabilitation. These applications require delay

to be lower that 1 msec or else a “cyber sickness” phe-

nomenon, similar to motion sickness, is experienced.

Applications included in MERC class involve trans-

mission of instructions to medical devices such as in-

fusion pumps, which control drug infusion to patients.

Monitoring of patients’ physiological functions, such

as blood pressure and heart rate, is included in RTCH

applications. Trafﬁc class RTNH involves real time

162

Athanasiadou, S., Papadimitriou, G. and Nicopolitidis, P.

Multiple Service Class TDMA Protocol for Healthcare Applications.

DOI: 10.5220/0009895301620166

In Proceedings of the 17th International Joint Conference on e-Business and Telecommunications (ICETE 2020) - DCNET, OPTICS, SIGMAP and WINSYS, pages 162-166

ISBN: 978-989-758-445-9

medical image transfer, teleconference for medical is-

sues and VoIP. Finally, OMIT applications have the

lowest QoS requirements, since they are ofﬁce appli-

cations such as email, web browsing and ﬁle transfer

of patient records.

In this paper a distributed random TDMA proto-

col for healthcare environments is introduced, called

Medical-Class TDMA (MC-TDMA), that is specif-

ically designed to support the ﬁve CoS depicted in

Table 1. MC-TDMA protocol provides TIM class

with guaranteed low delay, while also meeting the

other classes’ needs. This is achieved by dividing the

TDMA frame in two phases. During the ﬁrst phase

only packets of TI class are transmitted, while dur-

ing the second phase packets of the rest of the CoS

are sent. Each of the non-TI classes is assigned a

different probability of transmission according to the

hierarchy appearing in Table 1. Two versions of this

protocol have been developed, the simple MC-TDMA

and the Highest available class MC-TDMA (HMC-

TDMA), with the second version improving through-

put signiﬁcantly by minimizing idle slots. The oper-

ation of the two protocols was simulated and delay

of TI class was found to be very low and far within

bounds in both versions. Non-TI classes’ delay is

also within restrictions, however, HMC-TDMA pro-

tocol induces less delay than MC-TDMA. Addition-

ally, the two versions’ throughput results differ re-

markably, with HMC-TDMA protocol performing a

lot better.

The paper is organized as follows. In Section 2 the

two versions of MC-TDMA protocol are described.

In Section 3 the two proposed protocols as well as

RTDMA are compared in terms of average delay and

throughput via simulation results. Finally, Section 4

concludes the paper.

2 THE MC-TDMA PROTOCOL

The system model is a wireless network with one base

station and N stations, which could be the end users of

a healthcare facility such as hospital staff and medical

devices. The nodes are able to transmit and receive

packets of any service category from the base station.

The network control is distributed and time is divided

in timeslots.

The protocol operation is based on Random Time

Division Multiple Access protocol (RTDMA, (Ganz

and Koren, 1991)). According to that protocol, times-

lots are grouped into frames. During every timeslot a

station is randomly selected to transmit a packet.

Table 1: Healthcare environment CoS and their QoS re-

quirements.

Medical

Application

- CoS

Delay

Packet

Loss Rate

Tactile Internet

Medical (TIM)

- Class A

< 1 msec

< 10

−5

Medical

Equipment Remote

Control (MERC)

- Class B

< 3 sec ~0

Real Time Critical

Health Care

(RTCH)

- Class C

< 300 msec

~ 10

−6

Real Time non

Critical Health

Care (RTNH)

- Class D

<10 msec

< 10

−4

Ofﬁce Medical

IT (OMIT)

- Class E

< 1 sec

< 10

−2

2.1 MC-TDMA

Next, the simple version of the MC-TDMA protocol

is described. According to this protocol, each frame

is divided in two phases, each consisting of N times-

lots. Phase 1, which includes the ﬁrst N timeslots of

a frame, is dedicated to transmission of category A

trafﬁc (Table 1). During timeslot i the next category

A packet in queue of station i is transmitted. If station

i does not have a category A packet in its queue then

no transmission takes place and this slot remains idle.

During Phase 2, which consists of the next N

timeslots of a frame, categories B, C, D and E are

served. During each timeslot two random selections

are made. The ﬁrst selection concerns the category

to be served and the second concerns the station that

receives permission to transmit. The service cate-

gory selection is random but different categories are

assigned different probabilities of selection. These

probabilities are chosen according to each category’s

needs. Speciﬁcally, probability of selection must be

higher for higher priority categories. Next, the station

that receives the right to transmit is randomly cho-

sen. All stations have equal probability of selection.

The selected station transmits the next packet of the

selected category in its queue. If a packet of this cat-

egory does not exist in the selected station’s queue,

then no transmission takes place and this slot remains

idle.

Multiple Service Class TDMA Protocol for Healthcare Applications

163

100

120

140

160

180

0,001 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1

Average delay (slots)

Network load

Average delay

RTDMA

MC-TDMA

HMC-TDMA

Figure 1: Average delay of protocols RTDMA, MC-TDMA

and HMC-TDMA in slots versus network load.

0,001

0,1

0,2

0,3

0,4

0,5

0,6

0,7

0,8

0,9

Average delay (slots)

Network load

Class A - Average delay

MC - TDMA

HMC-TDMA

Figure 2: Average delay of CoS A with protocols MC-

TDMA and HMC-TDMA in slots versus network load.

2.2 Highest Available Class MC-TDMA

This version of MC-TDMA protocol differs from the

simple MC-TDMA in the following points.

• During Phase 1, if station i does not have a cat-

egory A packet in its queue, then it transmits a

packet of the next higher category available. The

category hierarchy is depicted in Table 1.

• During Phase 2, if a packet of the selected cate-

gory does not exist in the selected station’s queue,

then a packet of the next higher available category

in the queue is sent. The priority of the categories

is again as depicted in Table 1, with the difference

that, at this phase, packets of category A have the

lowest priority.

Therefore, in both phases the only case in which a slot

remains idle is when the queue of the selected station

is empty.

Average delay (slots)

Class B - Average delay

0,001

0,1

0,2

0,3

0,4

0,5

0,6

0,7

0,8

0,9

Average delay (slots)

TDMA

HMC

TDMA

0,001 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1

Network load

Figure 3: Average delay of CoS B with protocols MC-

TDMA and HMC-TDMA in slots versus network load.

100

0,001 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1

Average delay (slots)

Network load

Class C - Average delay

MC-TDMA

HMC-TDMA

Figure 4: Average delay of CoS C with protocols MC-

TDMA and HMC-TDMA in slots versus network load.

3 SIMULATION RESULTS

In the following, Highest available class MC-TDMA

protocol is compared to simple MC-TDMA proto-

col and to RTDMA in terms of average delay and

throughput. Simulations have also been made for

each service category separetely, for N = 10 stations

and network load 0.1 to 1. All stations have equal

probability of a packet arriving at them. Also, each

arriving packet has equal probability to belong to any

of the ﬁve CoS. During transmission Phase 2, classes

B, C, D, and E have probabilities 0.4, 0.3, 0.2 and

0.1 respectively of being selected for transmission of

a packet.

In Figure 1 it is seen that the average delay of the

system is much higher in simple MC-TDMA proto-

col than in RTDMA and HMC-TDMA, with the lat-

ter having slightly better results than RTDMA. The

maximum average delay of HMC-TDMA protocol is

21 slots for network load 1. Assuming a network

throughput of 500 Mbps and packet size of 2000 bits,

the slot size would be 4μsec and HMC-TDMA proto-

WINSYS 2020 - 17th International Conference on Wireless Networks and Mobile Systems

164

100

120

140

160

180

0,001 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1

Average delay (slots)

Network load

Class D - Average delay

MC-TDMA

HMC-TDMA

Figure 5: Average delay of CoS D with protocols MC-

TDMA and HMC-TDMA in slots versus network load.

100

200

300

400

500

600

0,001 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1

Average delay (slots)

Network load

Class E - Average delay

MC-TDMA

HMC

TDMA

Figure 6: Average delay of CoS E with protocols MC-

TDMA and HMC-TDMA in slots versus network load.

col’s average delay would be at most 84μsec. In Fig-

ures 2- 6 the average delay of classes A to E are plot-

ted to network load. Delay of class A is very low (<

13 slots) in both protocols. Again, assuming the same

slot size, class A delay would be 52μsec at maximum

network load, which is far lower than the bound of 1

msec.

For classes B, C, D and E it is seen in Figures

3 to 6 that HMC-TDMA protocol leads to a signiﬁ-

cantly lower delay than MC-TDMA. The maximum

delay with noted for these CoS with HMC protocol

is 47 slots for class E under network load 1, which is

approximately 0.2 msec for slot size 4μsec. However

both protocols are within bounds of Table1.

In Figure 7 it is seen that throughput of HMC-

TDMA protocol is slightly higher than that of RT-

DMA and reaches 0.85 packets/slot under network

load 1. The throughput of MC-TDMA under the same

conditions is much lower at 0.22 packets/slot.

In ﬁgure 8 throughput of each CoS is appeared for

HMC-TDMA and simple MC-TDMA protocols. All

CoS have almost identical throughput performances,

while performance for HMC-TDMA protocol is a lot

0,0

0,1

0,2

0,3

0,4

0,5

0,6

0,7

0,8

0,9

0,001 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1

Throughput (packets/slot)

Network load

Throughput

RTDMA

MC-TDMA

HMC

TDMA

Figure 7: Throughput of protocols RTDMA, MC-TDMA

and HMC-TDMA in packets/slot versus network load.

0,02

0,04

0,06

0,08

0,1

0,12

0,14

0,16

0,18

0,001

0,1

0,2

0,3

0,4

0,5

0,6

0,7

0,8

0,9

Throughput (packets/slot)

Network load

All classes- Throughput

…

Figure 8: Throughput of all CoS with protocols MC-TDMA

and HMC-TDMA in packets/slot versus network load.

higher that that of simple MC protocol. Here, the sys-

tem throughput appearing in Figure 7 is basically di-

vided equally in ﬁve parts. This derives from the fact

that, in all simulations, every packet arriving at a sta-

tion has equal probability of belonging to any of the

ﬁve classes.

4 CONCLUSIONS

In this paper MC-TDMA protocol, based on random

TDMA, that is designed for healthcare applications is

proposed. This distributed protocol manages to meet

the TI class’s high requirements while not neglect-

ing the other CoS. This is achieved by dividing the

TDMA frame in two phases, with the ﬁrst phase serv-

ing only TI class and the second phase serving the rest

of the classes. Two versions of this protocol are in-

troduced, with HMC-TDMA performing signiﬁcantly

better than MC-TDMA in both delay and throughput

aspects. Protocol HMC-TDMA accomplishes that by

exploiting idle slots with transmission of the highest

Multiple Service Class TDMA Protocol for Healthcare Applications

165

available class packet. Therefore, the proposed pro-

tocol manages to fully serve all medical CoS require-

ments with HMC-TDMA version offering extremely

low average delay and very high throughput. Future

work will be aimed towards developing a learning al-

gorithm that further reduces idle slots. This will be

achieved by increasing probability of selecting station

i for transmission if station i is already transmitting,

thus exploiting the fact that trafﬁc from a station usu-

ally comes into bursts.

ACKNOWLEDGMENT

This research has been co-ﬁnanced by the European

Union and Greek national funds through the op-

eration program Competitiveness, Entrepreneurship

and Innovation under the call RESEARCH-CREATE-

INNOVATE (project code: TIEDK-02489).

REFERENCES

Batta, M. S., Aliouat, Z., and Harous, S. (2019). A dis-

tributed weight-based tdma scheduling algorithm for

latency improvement in iot. In IEEE 10th Annual

Ubiquitous Computing, Electronics & Mobile Com-

munication Conference (UEMCON).

Bayat, A. and Aissa, S. (2019). Wireless power transfer

scheduling: Comparative study of tdma and sdma un-

der harvesters nonlinearity. In IEEE Global Commu-

nications Conference (GLOBECOM).

Capone, J. M. and Stavrakakis, I. (1999a). Achievable

qos in an interference/resource limited shared wireless

channel. IEEE Journal on Selected Areas in Commu-

nications, 17:2041–2051.

Capone, J. M. and Stavrakakis, I. (1999b). Delivering qos

requirements to trafﬁc with diverse delay tolerances

in a tdma environment. IEEE/ACM Transactions on

Networking, 7:75–87.

Chang, C., Chang, C., Chen, S., Tu, S., and Ho, K. (2019).

Optimisation-based time slot assignment and synchro-

nisation for tdma mac in industrial wireless sensor net-

work. IET Communications, 13:2932–2940.

Ganz, A. and Koren, Z. (1991). Wdm passive star-protocols

and performance analysis. In IEEE INFCOM ’91.

The conference on Computer Communications. Tenth

Annual Joint Comference of the IEEE Computer and

Communications Societies Proceedings.

Gupta, R., Tanwar, S., Tyagi, S., and Kumar, N. (2019).

Tactile-internet-based telesurgery system for health-

care 4.0: An architecture, research challenges, and fu-

ture directions. IEEE Network, 33:22–29.

Haddadin, S., Johannsmeier, L., and Ledezma, F. D. (2019).

Tactile robots as a central embodiment of the tactile

internet. Proceedings of the IEEE, 107:471–487.

Kim, K. S., Kim, D. K., Chae, C. B., Choi, S., Ko, Y. C.,

Kim, J., Lim, Y. G., Yang, M., Kim, S., Lim, B., Lee,

K., and Ryu, K. L. (2019). Ultrareliable and low-

latency communication techniques for tactile internet

services. Proceedings of the IEEE, 107:376–393.

Li, S., Liu, Y., Wang, J., and Sun, Z. (2020). Scmac: A slot-

ted contention based media access control protocol for

cooperative safety in vanets. IEEE Internet of Things

Journal, page 1.

Liu, L., Cao, Y., Ding, L., Yang, F., Qian, L., and Zhi, C.

(2019a). A priority-enhanced slot allocation mac pro-

tocol for industrial wireless sensor networks. In 25th

Asia-Paciﬁc Conference on Communications (APCC).

Liu, L., Peng, L., Xu, R., Zhao, W., and Shi, H. (2019b). A

requirements-oriented mac protocol in directional ad

hoc networks using collision-free dynamic time slot

assignment algorithm. In IEEE 5th International Con-

ference on Computer and Communications (ICCC).

Madi, S. and Al-Qamzi, H. (2019). A tdma-based mac

protocol for hybrid vehicular sensor networks using

practical mobility models with real maps. In Interna-

tional Conference on Advanced Electrical Engineer-

ing (ICAEE).

Papadimitriou, G. I. and Pomportsis, A. S. (2000).

Learning-automata-based tdma protocols for broad-

cast communication systems with bursty trafﬁc. IEEE

Communications Letters, 4:107–109.

Rubin, I. and Baker, J. E. (1990). Media access control for

high-speed local area and metropolitan area commu-

nications networks. In IEEE.

Soomro, A. and Cavalcanti, D. (2007). Opportunities and

challenges in using wpan and wlan technologies in

medical environments. IEEE Communications Maga-

zine, 45:114–122.

Tanenbaum, A. S. (1996). Computer Networks. Englewood

Cliffs, NJ: Prentice-Hall.

Tianjiao, Z. and Qi, Z. (2017). Game-based tdma mac pro-

tocol for vehicular network. Journal of Communica-

tions and Networks, 19:209–217.

Tong, K. M. and Hamdi, M. (2000). Achievable qos for

multiple delay classes in cellular tdma environments.

In IEEE Wireless Communications and Networking

Conference.

Wang, H. and Iversen, V. B. (2008). Erlang capacity of

multi-class tdma systems with adaptive modulation

and coding. In IEEE International Conference on

Communications.

Zhao, M., Liu, Q., Sun, Y., Wang, Q., and Li, P. (2019).

A cluster-dynamic tdma slot assignment protocol for

large-scale mobile ad hoc networks. In IEEE 5th In-

ternational Conference on Computer and Communi-

cations (ICCC).

WINSYS 2020 - 17th International Conference on Wireless Networks and Mobile Systems

166