1 Introduction

Wireless Body Area Networks (WBANs) are wireless networks composed of small smart devices, such as sensors and actuators, inside, on or close to the human body [16]. These networks can collect, process, and transmit data over a wireless channel [3], allowing the remote monitoring of the human body conditions without constraining personal life activities [36].

WBANs can enable many applications. For example, a medical patient can be equipped with a WBAN composed of sensors that constantly measure specific physiological variables, such as temperature, blood pressure, heart rate, and electrical cardiac activity, with the advantage that the patient can move freely at home or in a hospital. Beyond medicine, WBAN applications are arising in many other areas, such as sports, military and entertainment [16].

Medium Access Control (MAC) protocols are basic mechanisms for implementing Wireless Networks, such as WBANs. These protocols are important for coordinating network node transmissions, considering specific network requirements [28, 42]. However, developing MAC protocols for WBANs can be a challenge. The implementation of WBANs must meet specific requirements, such as [24, 30]: not affecting body mobility; avoiding that thermal and electromagnetic radiations harm body health; supporting specific-purpose, heterogeneous, and tiny devices as network nodes; optimizing energy use, as WBAN nodes are typically battery-powered; guaranteeing the security of sensitive data; providing timely operations, among others. In addition, WBANs typically must handle network traffic from regular to emergency operations [16]. Dynamic network traffics can impose a significant challenge, especially during body health emergencies, since physiological parameters may be correlated, demanding unexpected concurrent transmissions from different devices [16, 45, 46]. Moreover, MAC approaches for WBANs must also consider physical phenomena that may compromise the quality and reliability of the short-range lower-power wireless transmissions between WBAN nodes, such as radiofrequency interference, channel fading, and body mobility [26, 36]. As a result, several MAC protocols have been proposed to coordinate wireless communication while providing performance and reliability guarantees for WBAN applications.

This paper reviews three decades of strategies proposed for the MAC sublayer of WBANs to ensure quality, efficiency, and reliability during data acquisition, considering dynamic network traffic and body mobility. For that, we conduct a Systematic Literature Review (SLR) on MAC protocols for WBANs and make three major contributions. Firstly, for the publication selection phase of the SLR, we propose a new automatic selection approach that considers quantitative parameters (such as the number of citations) and qualitative indicators (related to the methodological aspects used in reviewed publications). The main novelty in our selection method is the use of cluster analysis techniques to automate the classification of the publications in distinct clusters of methodological quality, relevance for the main research question, and impact in terms of the number of citations. Secondly, we present the result of using the proposed SLR method to overview how existing MAC protocols handle failures and dynamic network traffic and the implications of such protocols for WBAN reliability and efficiency. Lastly, we identify new research challenges and opportunities based on the reviewed and classified literature.

A preliminary work-in-progress version of this paper was previously published [17], which only covered works up to January 2021. This paper is a revised and improved extension, covering publications up to December 2023.

The remainder of this paper is structured as follows. Section 2 presents a general overview of WBANs, including main MAC communication standards. The developed SLR method used in this paper is described in Sect. 3. Then, Sect. 4 describes the conducted SRL for WBAN MAC. Next, Sect. 5 discusses related works. Finally, Sect. 6 concludes this paper with final remarks and opportunities for future works.

2 Wireless body area network (WBAN)

A WBAN is a network composed of small, lightweight and smart devices that are installed inside (implantable), on (wearable) or around a human body and that can communicate wirelessly [36] – Fig. 1 illustrates a typical WBAN. Due to the context where they are used, WBAN devices have to operate using very low transmission power and be energy-efficient [3, 36, 46], imposing specific characteristics to the WBANs.

Fig. 1
figure 1

A typical WBAN

Full size image

Figure 2 illustrates the relationship between WBANs and other wireless networks, such as WPAN (Wireless Personal Area Network) and WLAN (Wireless Local Area Network). While WBANs are restricted to the human body (up to 2 ms), WPANs and WLANs can reach up to 10 and 100 ms away, respectively [11, 31].

Fig. 2
figure 2

Relationship between WBAN and other wireless networks

Full size image

Due to the sensitivity of body sensing data, WBANs demand strict reliable communication with transmission rates ranging from 1Kbps to 10Mbps. WPANs and WLANs may negotiate reliability and transmission rates according to application requirements.

Concerning network structure, WBAN devices must be positioned on the human body according to their manipulation of specific physiological or contextual parameters. Thus, WBAN topologies depend on the positioning of its devices. The distances between the devices and, consequently, the connectivity between them can vary during body movement. WBAN has a limited number of devices in order not to affect body mobility. Moreover, the WBAN devices usually communicate only with a central device, called network coordinator [46]. The three main WBAN topologies are star, mesh and tree (Fig. 3a, b and c, respectively). Star topologies are the most common ones, in which nodes are directly connected to a central device [46]. In the case of WPANs and WLANs, there may be a more significant number of devices, moving more freely and communicating with more than one network coordinator. Thus, some WPANs and WLANs can self-organize more widely and assume a more significant number of network topologies.

Fig. 3
figure 3

Typical WBAN topologies

Full size image

WBAN devices must transmit low-power radiofrequency signals following a Specific Absorption Rate (SAR) of the human body. It avoids harming the body tissue or changing body health conditions. WPANs and WLANs can vary transmission power to increase network lifecycle and are less concerned with dealing with human tissues than wearable and implanted devices of a WBAN. In medical applications, a WBAN must meet timing requirements.

Regarding power requirements, WBAN devices have battery size constraints, needing to prioritize low power consumption during message exchange, as batteries cannot be easily replaced, especially in implanted devices. In WPANs and WLANs, devices can be more conveniently replaced or recharged. In WBANs, the human body and coexistence with other networks are obstacles during wireless communication. WPANs and WLANs can use Line of Sight (LOS) to facilitate communication with less attenuation and interference [11, 31].

2.1 Device types

WBANs have three main device types [30]: sensors, actuators, and personal devices. Sensor devices are equipped with a sensing unit, a processor, a power unit, memory, and a transceiver. These devices collect data about the human body (e.g., temperature, blood pressure, frequency of cardiac beats, among others) and forward these data to a processing device. Actuators collect data like sensor devices but can also act upon the body. They might, for example, inject some medical substance into a patient following some medical procedure. Personal devices usually have greater processing power, memory, and energy supply than the other nodes in a WBAN. They commonly play the role of the WBAN coordinator (also called body control unit or sink). Sensors and actuators collect and transmit data to the coordinator, which can forward them or make them available to external devices or client applications.

One of the main characteristics of WBANs that differentiates them from other types of wireless sensor networks is the high heterogeneity and specificity of devices. Glucose readers, electrocardiogram monitors, electroencephalogram devices, pulse oximeters, and endoscopy capsules are examples of devices that might exist in WBAN applications [16].

2.2 MAC communication standards

IEEE 802.15.4 [22], IEEE 802.15.6 [19], and SmartBAN [4] are key communication standards for WBANs, outlining physical layers and MAC sublayers tailored to WBAN. These standards underpin numerous MAC protocols detailed in scholarly literature. In the following subsections, we briefly describe the different modes of coordinating access to the medium defined in IEEE 802.15.4, IEEE 802.15.6, and SmartBAN standards. The concepts and information discussed in these subsections guide the method proposed in the SLR presented in Sect. 3.

Other technologies based on IEEE 11073 Personal Health Device (IEEE 11073-10101 or IEEE 11703 PHD) [21], Bluetooth [18], ANT+ [33], Zigbee [8], and others support medical devices in health applications. However, as we briefly discuss below, they focus on something other than the MAC sublayer of WBANs. Thus, they were not directly explained and detailed in this paper but are based on some MAC approaches discussed in the following sections.

Zigbee is an implementation of the IEEE 802.15.4. Technologies such as Bluetooth and Bluetooth Low Energy (BLE) were designed using the IEEE 802.15.1 standard and were not adapted to the requirements of a WBAN. Bluetooth devices consume much more energy than WBAN devices because they must transmit at higher powers to reach longer distances. BLE was designed to meet the power-saving requirement. Still, BLE adopts the same frequency band as telemedicine devices that operate at 2.4 GHz, sharing the bandwidth with other networks and, therefore, being susceptible to interference [10, 46].

ANT is a protocol designed to reduce the power consumption of network devices in a flexible and scalable topology. ANT+ implemented improvements in ANT and was developed to provide interoperability between network devices [33]. However, this technology operates up to approximately 30 ms away and, like Bluetooth, operates at 2.4 GHz, not fitting well with WBAN requirements.

IEEE 11073 PHD and ANT+ do not deal directly with the MAC sublayer. The IEEE 11073 PHD is intended for communication between medical devices and external computer systems. Similar to technologies based on IEEE 11073, ANT+ [33] defines characteristics of communication modes between computer systems..

2.2.1 IEEE 802.15.4

The IEEE 802.15.4 standard was conceived as a general specification for low-power, short-range wireless networks [22]. Amendments and modifications have been incorporated into the original specification to address the distinct needs of different and particular application domains, including medical applications [20].

The specification was developed to allow implementation in devices that might be very constrained regarding computational processing and memory resources. It defines a particular type of device, called network coordinator, that is considered to have more computational power and energy supply than other regular devices in the network. This node is responsible for creating the network and maintaining node synchronization, thus playing a special role in managing the network. The specification supports star and mesh network topologies(see [22] for more details).

At the MAC sublayer, IEEE 802.15.4 specifies the access modes that might use beacons. A beacon is a control message conveying information for the management and synchronization of devices [19].

Fig. 4
figure 4

MAC 802.15.4 superframe structure [22]

Full size image

In beacon mode, the network operates according to a superframe structure (see Fig. 4), which defines how communication is organized in the period between two beacons. A superframe consists of two parts: an active and an inactive portion. The active portion is subdivided into time slots. A Beacon is always sent in the first slot of a superframe. The remaining slots are further structured into two parts (Fig. 4): the Contention Access Period (CAP) and the Contention Free Period (CFP) [22]. During CAP, the devices access the communication channel using Carrier Sense Multiple Access with Collision Avoidance (CSMA-CA) [42]. The CFP starts after the CAP. During CFP, access to the medium is done using Time Division Multiple Access (TDMA) [42]. Nodes have exclusive access to CFP slots by allocating Guaranteed Time Slots (GTS). During the inactive part, nodes pause transmissions and go into sleep mode before the beginning of the next superframe, saving energy [22]. In a non-beacon mode, there are no superframes, and devices access the medium using CSMA-CA [25].

The IEEE 802.15.4 standard specifies additional modes of operation, such as Deterministic and Synchronous Multi-channel Extension (DSME) or Timeslotted Channel Hopping (TSCH). For these modes, additional superframe structures are defined. As these modes are out of the scope of this review, we refer the reader to [22] for more details.

2.2.2 IEEE 802.15.6

The IEEE 802.15.6 standard was explicitly developed for WBANs, incorporating important features to coordinate communication among devices near or inside human bodies [47]. Such features are not present in the IEEE 802.15.4 standard. Similar to IEEE 802.15.4, IEEE 802.15.6 uses the Industrial, Scientific, and Medical (ISM) frequency band but also includes specific bands approved by medical and regulatory authorities [29]. In IEEE 802.15.6, nodes are organized into logical sets, called body area networks (BANs) [19]. Nodes in a BAN are arranged in a star topology around a particular node called hub, whose function is to coordinate the BAN. Networks structured in the so-called extended star BANs are also possible, adding an intermediate node to relay messages from nodes that cannot transmit directly to the hub.

The IEEE 802.15.6 specifies three operation modes: beacon mode with superframes; non-beacon mode with superframes; and non-beacon mode without superframes. Modes with superframes allow the hub to define active and inactive superframes. In an active superframe, the hub sends beacons and activates the access phases. Beacons are not sent in an inactive superframe, and the access phases are not activated [19].

Fig. 5
figure 5

Mode with beacon and superframes in 802.15.6 [19]

Full size image

In the beacon with superframe mode, the following access phases are defined (Fig. 5): two Exclusive Access Phases (EAP), two Random Access Phases (RAP), two Managed Access Phases (MAP), an additional beacon (B2) and a Contention Access Phase (CAP). The hub might set the length of any of these phases to zero, although certain specific structuring rules can be applied. The EAP is used for emergency and high-priority network traffic. RAP and CAP are used for other types of network traffic. During these phases, the access to the medium is based on Slotted Aloha [42] or CSMA-CA. The hub uses the MAP phase to arrange allocated communication links by either using polling or TDMA. In non-beacon with superframes mode, there is only a MAP. In non-beacon without superframes mode, devices access the channel using CSMA-CA. Unscheduled polled or posted links might be defined.

2.2.3 SmartBAN

The SmartBAN is a WBAN communication standard created by the European Telecommunications Standards Institute (ETSI) to improve the PHY-MAC layers performance of healthcare applications [14]. This standard was developed to provide reliable communication, low complexity network configuration, low energy consumption and interoperability with other networks [5, 15].

This standard organizes the network in a star topology, consisting of a hub that acts as a coordinator and nodes that transmit messages with the patient’s physiological data. The hub can temporarily assign nodes as relays to retransmit messages from node to hub or vice-versa [7]. The nodes exchange messages with the hub on two types of logical channels in the ISM 2.4 GHz frequency band: Control Channels (CCH), used for unidirectional communication to transmit beacons from hub to nodes, and Data Channels (DCH), used for bidirectional communication between nodes and hub. DCH also can be used to transmit network control and management messages [6].

In the SmartBAN standard, the term “Inter-Beacon Interval" (IBI) describes the superframe structure, comprising three distinct phases (Fig. 6): the Scheduled Access Period (SAP) for TDMA-based data traffic, the Control and Management (C/M) period for regulating message flow using a Slotted ALOHA policy [4], and inactive period. To accommodate dynamic network conditions, both SAP and C/M can utilize Multi-use Channel Access (MCA) to prioritize the transmission of emergency messages with minimal delay and high priority over a designated channel, employing unused, pre-scheduled time slots through Reuse Channel Access [15].

Fig. 6
figure 6

SmartBAN interbeacon structure [4]

Full size image

3 Method of systematic literature review

The quantitative method we applied in this SLR was based on [9, 12, 27] and comprised the following steps: elaboration of the research questions; and definition of the strategies for publication search and selection. Although this strategy produced good results, we needed to assess the quality and relevance of these publications before extracting their data. Thus, we added new steps to qualitatively evaluate the publications, as described in Sect. 3.3.

3.1 Research questions

This SLR aims to answer the following main research question: “What approaches for the MAC sublayer of WBANs have been proposed to ensure quality, efficiency, and reliability during data acquisition under varied network traffic conditions?".

To achieve an overview of the existing approaches, we defined the following secondary questions:

  • SRQ1: Which approaches have been adopted to improve reliability?

  • SRQ2: Which strategies have been employed to enhance message traffic efficiency (i.e., throughput and latency)?

  • SRQ3: Which methods have been used to increase energy efficiency?

  • SRQ4: Which approaches have been applied to support dynamic network traffic?

  • SRQ5: Which techniques have been adopted to handle emergency network traffic?

3.2 Publication search strategy

Based on the research questions, we defined the terms to guide our literature search. Besides the main terms WBAN, MAC, and synonyms, we also included mobility, interference, collision, path loss, and fading because they are aspects that compromise reliability, energy efficiency, and network traffic. Thus, in our search, we used the following expression:

Expression: (wban OR “wireless body area network" OR wbans OR “wireless body area networks" OR wbsn OR “wireless body sensor network") AND (mobility OR interference OR collision OR “path loss" OR “deep fade" OR fading) AND (mac OR “medium access control")

The search expression was intentionally crafted to target research on “dynamic traffic" and “network reliability", utilizing characteristic terms from these fields such as “collision,” “path loss,” and “deep fade”.

We looked for publications in the scientific repositories listed in Table 1. We selected such repositories by their relevance in the scientific community. We considered all publications in those scientific repositories until December 2023.

Table 1 Selected Repositories
Full size table

We saved the search results in BibTeX files and imported them to the Mendeley bibliographic management tool.Footnote 1 By using this tool, we applied four filters (Filters 1 to 4), selecting 336 from the 14,146 found works (see Table 2).

Filter 1 removed duplicate occurrences. Filter 2 removed publications without the terms (wban OR “Wireless body area network” OR wbans OR “Wireless body area networks" OR wbsn OR “wireless body sensor network") AND (mac OR “medium access control") in the title and/or abstract. Initially, we had not restricted in which parts of a publication the terms used in the search expression could appear, resulting in some publications satisfying the expression but being out of the scope of this SLR. Filter 2 was used to eliminate those publications. Filter 3 removed publications with less than five pages and review papers without new approaches for the MAC sublayer. After Filter 3, we obtained 549 works, joining publications from all scientific bases. Finally, Filter 4 excluded works whose main subject was not the MAC sublayer or presented only a comparative analysis between approaches from other works. This resulted in the 336 publications being thoroughly assessed and scored regarding quality, relevance, and impact (number of citations) – as presented in Sect. 3.3.

Table 2 Publication selection process
Full size table

3.3 Assessment of publications quality and relevance

The assessment of the quality of publications was made by identifying those works that present methodological and structural aspects that meet the following criteria:

  1. 1.

    Do they clearly state the research goals? [9]

  2. 2.

    Do they position the described results in the state-of-art?

  3. 3.

    Do they propose viable solutions (for real-world scenarios)? [9]

  4. 4.

    Are simulations or experiments thoroughly analyzed and explained, and do the results strongly support the ideas presented in the work? [9]

  5. 5.

    Do they qualitatively compare the presented approach to existing related ones?

The relevance assessment consisted of identifying to what extent the works might answer the research questions of this SLR. For that, we adopted the following criteria:

  1. 1.

    Do they propose strategies to improve reliability?

  2. 2.

    Do they present approaches to enhance communication efficiency (throughput and latency)?

  3. 3.

    Do they describe techniques to increase energy efficiency?

  4. 4.

    Do they discuss methods to support dynamic network traffic?

  5. 5.

    Do they provide strategies to handle emergency network traffic?

We adopted the following scale of values for each criterion: 0.0 if it does not meet the criterion; 0.5 if it partially meets the criterion; and 1.0 if it meets the criterion. We analyzed the defined quality and relevance criteria to score the 336 publications selected after applying the four filters. Then, we organized the data about the publications in a sheet and imported them to R,Footnote 2 the software we used for processing and chart generation. Next, we applied the k-means clustering method to group publications with similar quality and relevance scores [38]. In k-means clustering, we specify the number of desired clusters, and each object is assigned to a specific set, using a distance measure based on their scores. We used the Euclidean distance between the quality and relevance values as a measure of dissimilarity to do the clustering.

Fig. 7
figure 7

Clusters based on quality and relevance criteria and the number of citations

Full size image

From the quality scores, we distributed the 336 publications in three clusters (Fig. 7): Very Significant Quality, with 187 publications; Significant Quality, with 84 publications; and Little Significant Quality, with 65 publications. From the relevance scores, we also distributed the 336 publications in three clusters (Fig. 7): Very Significant Relevance, with 139 publications; Significant Relevance, with 106 publications; and Little Significant Relevance, with 91 publications. Among the 336 publications, 92 of them were in both Very Significant Quality and Relevance clusters.

Besides the quality and relevance criteria, we also collected the number of citations of each 336 publications in Google Scholar. Based on the number of citations, we organized the publications into three categories (Fig. 7): in the Little Significant category were 171 publications with up to 9 citations each; in the Significant category, we had 71 publications from 10 to 19 citations each; and in the Very Significant category, we included 94 publications with more than 19 citations each. The number 19 was chosen as a reference, as it is an approximate average of the number of publication citations. Among the 94 publications in the group of Very Significant Number of Citations, 60 of them did not fully meet the Very Significant Quality and Relevance criteria. However, we maintained them in the final list of selected papers solely due to the number of citations they have. The final list of selected papers has 152 publications: 92 that are in the Very Significant groups of all three criteria (Relevance, Quality, and Number of citations); and 60 that were only in the Very Significant Number of Citations group. The list of these 152 selected publications and a summary of the data extracted from them and used to answer the research questions of this SLR are available at http://www.lasid.ufba.br/papers/apendix-rsl-mac-wban.pdf.

4 Results of the systematic literature review

Figure 8 shows the percentages by year of the citations and the assessed and selected publications, represented by a red line and blue and orange bars, respectively. The works that best meet the criteria of this SLR were published since 2006 (Fig. 8). This is due to the emergence of new smart devices for WBANs [1, 40].

Fig. 8
figure 8

Publications per year until December 2023

Full size image

Until 2018, many selected publications in this SLR were included using citation numbers as a criterion. This was important because older publications are subject to more readings and analyses and, consequently, more citations. Since 2018, there has been a decrease in the citation percentage, which does not necessarily imply a reduction in the quality or research interest of publications. As this SLR also adopts qualitative criteria, it is possible to include qualified publications with few citations. From 2020, we observed a drop in the selected publications number, which may be a reflection of the emergence of the COVID-19 pandemic.

The selected publications were analyzed, and several challenges related to the MAC sublayer of WBANs were identified. Table 3 lists these challenges with the related research questions and shows the percentages and quantities of publications that address the identified challenges. As some publications might address more than one challenge, the percentages in Table 3 add up to more than \(100\%\). Among the challenges, the following stand out: energy consumption; handling of dynamic and emergency traffic flows; latency; and collision.

Table 3 Relationship between the challenges addressed in works and the research questions with the percentage and quantity of publications selected
Full size table

Interference, addressed in \(11.8\%\) of the works, occurs when devices from different networks compete to use the same wireless communication channel or due to ambient noise that generates loss or corruption of transmitted data.

The attenuation of the transmission signal (fading) and the disconnection between the network devices, which are in the works, respectively \(7.2\%\) and \(6.6\%\) of the publications, directly impact WBAN reliability and efficiency. They are consequences of body mobility, obstacles in the environment, and weather conditions.

Overload due to control data was only addressed in \(3.9\%\) of the works, despite increasing the possibility of collisions and delays and decreasing network use efficiency.

The temperature of devices in WBAN is of particular concern as high temperatures might damage human body tissues. Additionally, heat dissipation is responsible for part of the energy consumed by the devices, mainly during transmissions. Buffer overflow, caused by storage limitations in devices, is one of the causes of data loss. Although essential, issues related to the temperature of devices and buffer overflow are only addressed in less than \(1\%\) and \(3.9\%\) of the works, respectively.

Table 4 Approaches adopted for each research question
Full size table

To deal with these WBAN challenges, MAC protocols can combine two or more of the following approaches: traffic prioritization, use of specific synchronization mechanisms, use of more than one type of transceiver, and the definition of particular media access policies. Table 4 associates the research questions to the approaches adopted and the percentages and quantities of publications that address these questions.

4.1 Traffic prioritization

The coexistence of different types of traffic and devices in a WBAN increases the complexity of medium access control. In particular, the medium access control strategy must adequately support dynamic and emergency traffic. An inappropriate strategy might result in higher communication latency or not satisfying time constraints. The primary strategy adopted to address these challenges is traffic classification and prioritization, obtaining low latency in transmissions, higher reliability and energy efficiency, thus relating to the SRQ2, SRQ3, SRQ4 and SRQ5 research questions.

Table 5 Traffic prioritization
Full size table

From the selected works, \(52\%\) define until two traffic categories (see Table 5): do not specify any priority category; or specify two categories (regular, and critical or emergency) like in [35]. More than two categories, as proposed in [43], were found in \(48\%\) of the selected works. As WBAN devices have limited computer resources, most works have adopted up to two traffic categories, obtaining better performance trade-offs accommodating dynamic and emergency traffics.

4.2 Synchronization mechanisms

Interference, collision, and data overload impact the reliability (SRQ1), efficiency (SRQ2) and power consumption (SRQ3) of WBAN. Communication synchronization mechanisms have been adopted to deal with these challenges.

Table 6 Synchronization mechanisms
Full size table

As discussed in Sect. 2.2, the communication standards for WBAN specify synchronization modes with and without beacons. When used, beacons can be transmitted periodically or on-demand. In the former case, a beacon is transmitted at the beginning of each superframe. Some approaches use additional beacons and short notification messages (e.g., ACK). These messages are sent during the superframe to adapt resource usage to changing traffic and reduce power consumption. Modes with periodic beacons were investigated in \(84.9\%\) of the works (Table 6). In on-demand mode, used in \(4.6\%\) of the publications, a coordinator node transmits a beacon when it notices priority changes in the network or when a change in the device clock is detected.

Some synchronization mechanisms adopt a specific frame, named preamble, commonly used to allow a node to prepare itself to receive a message and synchronize its clock with the sender’s clock [42]. In conjunction with contention-based MAC policies, this mechanism is used in \(5.9\%\) of the publications. Operation modes without beacons are used in \(4.6\%\) of the works, mainly in contention-based MAC approaches. In this mode, node synchronization is done using short messages (e.g., ACK) with clock values.

The selected publications indicate that modes with periodic beacons are more suitable when used with contention-free or hybrid MAC policies, as continuous synchronization between devices is required to reserve transmission slots. Modes without beacons, with the transmission of beacons on demand, or based on preambles are primarily used with contention-based MAC policies.

4.3 Types of transceivers

Each WBAN device has at least one transceiver. The use of additional or particular types of transceivers impacts how the MAC sublayer handles collisions (SRQ1 and SRQ2), power consumption (SRQ3), and dynamic (SRQ4) and emergency (SRQ5) traffic. Regarding transceivers, we can classify the approaches found in the selected works into two categories: Special approaches, which include some special transceiver arrangements or types – like directional antennas, cognitive radios, and wake-up radios; and Ordinary approaches, which use ordinary transceiver arrangements.

Table 7 Use of special transceiver approaches
Full size table

Most approaches described in the selected publications, \(87.5\%\) of them (i.e. 133 works), are based on ordinary transceiver approach and the most works adopt IEEE 802.15.4 or IEEE 802.15.6 compliant transceivers (Table 7). Among the selected publications, only one work (\(0.7\%\)) is based on the SmartBAN-compliant transceiver.

In 14 works, \(9.2\%\) of the selected publications, a wake-up radio is used. A wake-up radio is a secondary transceiver that consumes very little energy and solely aims to detect the beginning of a transmission. When that happens, the primary transceiver used to receive and transmit data is switched on, thus saving energy and improving response time, especially for emergency traffic. Wake-up radios are not used in implanted devices and in situations with high physical space limitations, as they represent additional components in the WBAN nodes.

In 4 works (\(2.6\%\)), there are descriptions of approaches with directional antennas. As these antennas concentrate the energy of transmissions in specific directions, their use improves spatial reuse and allows simultaneous transmissions without collisions. However, their benefit might be limited, as the gain in performance might be compromised when body movements modify the direction of the antennas.

Only one work, \(0.7\%\) of the selected publications, adopted Cognitive radio, an adaptive transceiver that adjusts its parameters to improve the use of the radio frequency spectrum and avoid interferences. Cognitive radios can choose the best radiofrequency spectrum for transmitting and receiving messages. The main advantage of this approach is the possibility of improving message throughput once it can find a radiofrequency spectrum not used by other devices. However, Cognitive radios are still little explored in WBAN, as they demand greater complexity and energy consumption in monitoring and searching for unused radiofrequency spectrums [37, 48].

4.4 Medium access policies

The reliability and efficiency of protocols depend highly on the policies adopted to control the wireless medium access by devices. These policies can be broadly categorized into (Table 8): contention-based, contention-free, and hybrid policies. Additionally, MAC approaches can be categorized according to whether they are based or not on IEEE and ETSI standards. In the latter case, MAC policies are generally based on TDMA, FDMA, Polling, CSMA/CA, or some combination of them.

Table 8 Medium access policies
Full size table

Contention-based policies are studied in 29 works (\(19.1\%\)) of the selected works. These policies are based on CSMA-CA and better accommodate dynamic (SRQ4) and emergency (SRQ5) traffic. However, they are susceptible to collisions and retransmissions, being more suitable in applications without strong energy restrictions and strict reliability requirements. The works describing contention-based approaches can be further divided into four groups: ten works, \(6.6\%\) of the selected publications, propose improvements to 802.15.4 CAP; eleven works (\(7.2\%\)) suggest modifications to EAP, RAP, and CAP of the IEEE 802.15.6 standard; one work (\(0.7\%\)) is based on SmartBAN; and seven works (\(4.6\%\)) do not use the IEEE or the ETSI specifications.

In MAC protocols that use contention-free policies, devices are scheduled following a pre-established order for data transmission. In this case, there is no competition for the channel, and each device is assigned to specific time slots or frequency bands (i.e., multiple channels) for transmission. Contention-free policies are generally energy efficient, avoiding collisions and reducing retransmissions (SRQ1 and SRQ3). The most used techniques are TDMA, FDMA, and Polling. Contention-free protocols are described in 43 works, \(28.3\%\) of the selected publications. From those, 20 works (\(13.2\%\)) are based on IEEE 802.15.4, 08 works (\(5.3\%\)) based on IEEE 802.15.6, and 15 works (\(9.9\%\)) do not use the IEEE or the ETSI specifications.

Most protocols, 80 works or \(52.6\%\) of the selected publications, are based on hybrid policies, combining the flexibility for accommodating variable traffic, as provided by contention-based policies, with the energy efficiency of contention-free policies. However, hybrid approaches can lead to higher control data overhead. Among the hybrid protocols described in the selected publications, 802.15.4 and 802.15.6 standards are the basis for 36 works (\(23.7\%\)) and 31 works (\(20.4\%\)), respectively. Whereas only 13 works (\(8.6\%\)) are not based on these IEEE standards.

As we can observe, no hybrid or contention-free policies are based on the SmartBAN standard among the selected works. Moreover, only one work (\(0.7\%\)) used this ETSI specification, which also is very low representative.

Contention-based policies improve reliability (SRQ1) and accommodate dynamic (SRQ4) and emergency (SRQ5) traffics well, mainly with adaptions in CSMA/CA, as proposed in [34, 43]. However, data traffic (SRQ2) and energy efficiency (SRQ3) are better addressed with contention-free policies. Finally, hybrid approaches handle dynamic and emergency traffics well and are energy efficient (SRQ3), presenting better performance trade-offs in a more unpredictable and dynamic environment.

4.5 Discussion

We cataloged the results in 4 approaches to answer the main research question: traffic prioritization, synchronization mechanisms, use of more than one type of transceiver, and definition of medium access policies. In addition, we analyzed and organized the data, considering the most used solutions in the selected publications. Thus, we could observe some trends in these approaches, as seen in Fig. 9.

Fig. 9
figure 9

Publications percentage according to the adopted strategies

Full size image

Firstly, most selected works adopt ordinary transceiver approaches (\(87.5\%\)) and are based on IEEE 802.15.4 (\(37,5\%\)) or IEEE 802.15.6 (\(30,3\%\)) compatible transceivers. As the IEEE 802.15.4 specification was published years before IEEE 802.15.6, there are still more works based on IEEE 802.15.4 than on IEEE 802.15.6. These two specifications are references for WBANs.

SmartBAN, on the other hand, has been adopted in very few works (\(0.7\%\)). This standard emerged in 2013 following the publication of 802.15.4 and 802.15.6 standards. Both SmartBAN and 802.15.6 were specially designed for WBANs, but 802.15.6 is an evolution of 802.15.4, widely adopted in the industry with a history of maturity and continuous development. SmartBAN, in contrast, lacks such a consolidated history.

Furthermore, several transceivers already employ the 802.15.4, simplifying its incorporation into experiments. Regarding the MAC sublayer, SmartBAN exhibits a similar structure to 802.15.4, with subtle differences in the superframe organization. Since IEEE standards are widely adopted in most communication devices, especially transceivers, transitioning to a new standard is not trivial and requires investment. Consequently, given the comprehensive similarity between these standards, research tends to adopt standards with already established devices and simulation tools, such as IEEE 802.15.4 and 802.15.6.

Secondly, there is a variance in traffic prioritization when combined with synchronization mechanisms and medium access policies (Fig. 9). We observed that most works (\(30.3\%\)) have chosen to use more than two traffic prioritization categories together with a hybrid medium access policies and periodic beacons. This combination is made because, in a hybrid policy, a portion of the superframe is allocated for contention-based media access. Thus, the devices can compete with each other according to the established priorities or use the contention-free part of the superframe to reduce the effects of frame collisions that may occur in the previous part. Despite being an adequate strategy for accommodating varied traffic, the competition for communication media during contention increases the energy consumption caused by frame retransmissions. However, such a combination has presented better performance trade-off in most scenarios, but the main challenge verified is how to properly adjust the size of contention-free and contention-based periods.

In addition, the other strategies are combined with up to two traffic prioritization categories: regular and emergency. These are more used with contention-based and contention-free media access policies. In the selected publications, \(13.2\%\) chose to use periodic beacons with two traffic categories and a contention-free access policy. In this approach, the superframe slots are distributed by the WBAN coordinator, observing the priority of each device and synchronized with each beacon. However, as the WBAN communication medium is not free from interference, the delay in message delivery can be increased due to the device waits for new slots in the next beacon to retransmit its data. To reduce the effects of this interference, some publications use wake-up radio. However, as previously explained, incorporating an additional transceiver is not trivial, mainly due to the small size of WBAN devices.

5 Related work

Initially, many published works used Wireless Body Sensor Networks (WBSN) to name a kind of WBAN [23]. However, WBSNs have incorporated actuators and implanted devices, becoming more generically called WBAN. There is a set of literature reviews addressing different aspects of WBANs, ranging from characterization and basic concepts to the discussion of specific implementation issues [1, 2, 10, 13, 16, 32, 36, 39,40,41, 44, 46].

A review of the characteristics, problems, and challenges of WBAN-based medical applications is presented in [40]. The work present the main types of devices used in these applications, such as wearable and implantable sensors, and discuss trends in the literature from 2008 to 2018. WBAN applications in telemedicine are described in [2] with focus on patient screening. General aspects of WBANs are also presented in [36], where the authors discuss the significant technological advances in the area until 2013.

General functional as well as non-functional aspects of WBANs are discussed in [10] and in [39]. In particular, [10] focuses on general communication requirements, quality of service and safety issues for e-health applications. In [39], the authors discuss fault tolerance aspects and issues related to network coexistence and interference.

Specific communication network issues are more directly addressed in [13, 32, 41, 44, 46]. The authors in [13] and [46] present a review of protocols for the physical, MAC, network and application layers, but [13] only covers the literature from 2003 and 2004, while [46] includes only works until 2010.

In [32] and [44], the authors present reviews of the main MAC protocols used in WBANs, including approaches to handle emergency traffic (mainly in [32]). However, these works only include publications until 2016. MAC protocols and energy efficiency are addressed in [41], but the authors do not address emergency traffic. Energy efficiency in the network layer is addressed in [1].

Finally, in [16], the authors discuss many aspects of WBANs, such as traffic characterization, particular types of transceivers, like wake-up radio, and IEEE 802.15.4 and 802.15.6 standards regarding energy consumption. Moreover, aspects related to the implementation of applications based on IEEE 802.15.6 are also discussed. However, the work focuses on software-defined networks and energy harvesting.

Although the existing related works present and discuss many issues related to different aspects of WBANs, they do not systematically address the specific challenges related to the MAC sublayer. We could only identify three works in the literature that are systematic reviews [2, 13, 40]. These reviews, however, use the same quantitative criteria (number of citations on Google Scholar) to select the publications. All other related works are narrative reviews.

Our SLR systematically maps the existing approaches in the literature to MAC in WBANs, concentrating on the quality and efficiency guarantees under different traffic conditions and body mobility. Additionally, this SLR is based on both qualitative and quantitative criteria. While the former criterion reflects the quality and relevance of the publications, the latter reflects the number of citations, reflecting the impact on the scientific community. A preliminary, work-in-progress version of this paper was previously published [17], covering only works until January 2021. In these last three years (from Jan 2021 to Dec 2023), 4, 750 new works related to WBAN were found, which after applying the filters, 50 new works were analyzed and 31 new works were selected with better impact, quality, and relevance indicators.

6 Final remarks

This paper systematically reviews existing approaches for MAC in WBANs, focusing on efficient data acquisition under varied network traffic, interference, and body mobility. This work also introduced a new method to select publications during systematic literature reviews. Our proposed method considers not only quantitative parameters, like the number of citations but also qualitative indicators, like the methodological aspects used in the selected publications. The publications selected by the SLR described approaches to traffic prioritization, synchronization mechanisms, the use of different types of transceivers, and medium access policies.

Traffic prioritization is a fundamental mechanism for dealing with dynamic (SRQ4) and emergency (SRQ5) traffic. The greater the number of priority categories, the more difficult it is to allocate resources to handle changes in traffic. As a result, existing approaches generally tend to use only two prioritization categories, which improves efficiency in dealing with dynamic traffic with a limited complexity for resource allocation in the communication channels.

The synchronization mechanism is required to avoid communication interference and collisions, particularly when combined with contention-free media policies, improving reliability (SRQ1) and optimizing network traffic (SRQ2). However, sending too many control messages for synchronization, particularly during dynamic traffic, increases network load and can jeopardize energy efficiency (SRQ3).

Increasing the number of transceivers, such as with the use of wake-up radios, is another way to improve device synchronization. While this approach can improve energy efficiency (SRQ3), particularly in emergency traffic, on the other hand, it can also increase device complexity and may not be feasible when there are physical space constraints or in applications with implanted devices. Cognitive radios can be a good solution in interference and dynamic traffic scenarios, but they have still been explored little because of the device’s complexity and costs, demanding more research.

Contention-based medium access policies handle dynamic (SRQ4) and emergency traffic (SRQ5) well because devices can access the channel anytime. However, even with the improved CSMA-CA proposals, contention may increase collisions, resulting in more retransmissions, increased energy consumption (SRQ3), and decreased network reliability (SRQ1). Contention-free medium access strategies may result in more predictable message delivery delays with improved energy efficiency (SRQ3) and reliability (SRQ1). On the other hand, handling dynamic traffic may require frequent rearranging of transmission schedules, resulting in additional transmission delays and difficulty accommodating dynamic traffic. Some strategies reserve superframe slots for emergency traffic, wasting such slots when emergencies do not occur frequently.

In hybrid medium access policies, the advantages of contention-based policies for handling dynamic (SRQ4) and emergency (SRQ5) traffic are combined with the reliability advantages (SRQ1) of contention-free policies. However, developing a proper mechanism that dynamically switches between these policies to optimize performance, reliability, and energy consumption remains challenging.

Although some publications propose combining one or more of these approaches to increase the reliability of message transmission, the inherent challenges of each one make this integration difficult, leaving gaps open for future research.

Most revised works address the issues of energy consumption, latency, and collisions. Other issues, such as channel fading, overload, disconnections, and thermal effects, have received little attention, showing the necessity of more research and development so that WBANs can effectively be used in healthcare applications with specific patient mobility patterns.

In addition, there is a tendency to combine approaches. Some approaches adopted hybrid medium access policies with several categories of traffic priorities. Others chose at least two categories with contention-free medium access policies.

Finally, the number of selected approaches by our RSL is too large (152 publications), so a more in-depth comparison between each selected work would significantly extend our paper and go beyond its proposed scope. A narrative review with a more in-depth comparison between each selected work can be done as a future work.