Energy efficient device discovery for reliable communication in 5G-based IoT and BSNs using unmanned aerial vehicles

Abstract

Connectivity among real-world entities is one of the primary requirements of the upcoming Fifth Generation Public Private Partnership (5G-PPP). Both Internet of Things (IoT) and Body Sensor Networks (BSNs) are major applications of 5G networks. However, over-consumption of energy for device discovery, which includes registration, removal, querying, routing etc, quickly depletes the resources of a node, which may further influence the whole network. There are a number of approaches which provide energy efficient mechanisms for the selection of devices in a network operating with different types of nodes; however, these approaches are unable to maintain a high transmission capacity along with energy conservation and fault-tolerance. In this paper, an energy efficient approach for device discovery in 5G-based IoT and BSNs using multiple Unmanned Aerial Vehicles (UAVs) is presented. A functional architecture is proposed, which utilizes XML charts to perform device discovery on the basis of networks state cost and available energy. The significant gains achieved in energy consumption, end to end delays and packet loss show that our solution is capable of providing energy efficient device discovery with 78.4% reduction in the overall energy consumption compared to existing solutions. The advantage of UAVs in energy efficient networking is illustrated using numerical analysis which suggests 75% enhancement in the energy-asymptote of the existing networks.

Introduction

With the ever increasing demand of the users to seek all the information on the go, Internet of Things (IoT) and Body Sensor Networks (BSNs) have evolved as two of the promising areas of research. IoT aims at connecting all the devices on the network to provide a common platform for information sharing. It is predicted that the number of devices for IoT and BSNs will be more than 50 billion by 2020 (Higginbotham, 2011). This is a huge number and will certainly require efficient network approach for handling so many devices.

BSNs comprise multiple sensors which can be installed as on-body or out-body devices which can support the efficient management of the organs to maintain a healthy life. A simple chip implanted on a body can send regular updates to your smart devices or even to a nearby hospital in the case of emergency by connecting body sensors to the network as a part of IoT. Thus, BSNs can be considered as an integral part of the IoT. BSNs have already been studied and implemented as a part of stand-alone applications using specially designed standard (IEEE 802.15.6) for transmission (Kwak et al., 2010). However, with the advent of IoT, it is necessary to include BSNs as body sensors are the crucial devices which are to be connected as a part of IoT.

With the advancement of the Fifth Generation Public Private Partnership (5G-PPP), these devices are considered as an important part of the application layer (Akyildiz et al., 2016). The improvement in wireless networks not only enhances the data rate for users but also enhances the chance of connecting more and more devices. From household appliances to vehicles to personal gadgets, everything will be connected via the internet as shown in Fig. 1. The circles in the figure represent the zones of different home gateways comprising various devices connected to the internet via the common access point. IoT has emerged as a promising area, which will decrease the normal life complexities by providing fast and rapid services (Atzori et al., 2010). Also, the connectivity between most of the devices allows efficient control and management. The idea of smart life can be put into practicality with the installation of devices that can form the part of IoT.

Although computerized devices and intelligent systems have been there for a long time, these are not really connected to the real time, thus limiting the amount of knowledge and control over such devices. But, IoT brings a change in the way such devices will be connected and used (Giang, March et al., 2015). Especially, information processing and availability of control over each device are the key requirements of IoT. Many organizations have already started working on defining the common architectures for the devices that will be a part of IoT. The main focus has been on the development of a common firmware which will facilitate a user to bring a device on the network with ease as well as at low cost (Sharma et al., 2017).

The main difference between BSNs and Wireless Sensor Networks (WSNs) is the design of topology as the sensors are located on a body which serves as a common Region of Interest (ROI) (Lai et al., 2013). An illustration of BSNs is shown in Fig. 2. The design of nodes, placement, fault-diagnosis, data support, reduction in power consumption are the key issues with the BSNs. However, with the continuous demand for data acquisition, designing the low-power rating sensors is the primary focus of researchers. Emphasis on selecting energy-efficient strategies for data dissemination can prevent re-calibrations of sensors.

With the upcoming 5G networks, it will be possible to support a large number of devices in IoT and BSNs simultaneously at higher data rates. IoT and BSNs require separate channel or bandwidth to communicate, which is easily attainable through 5G networks. As stated by Mobile and wireless communications Enablers for Twenty-twenty (2020) Information Society (METIS) (Tullberg et al., 2014), 5G networks are formed by the inclusion of several devices placed in a layered formation such as macrocells, microcells, small cells, femtocells, and picocells. The division of an entire network into smaller segments allows easy control and coverage over the demand areas. With a large number of access points, more devices and users can be supported with continuous connectivity. Such division of networks can prove handy in handling applications related to IoT and BSNs.

Devices and sensors in IoT and BSNs require continuous, robust and fault-tolerant connectivity for their efficient utilization (Rathore et al., 2016, Galov and Korzun, 2014). The services for managing devices and sensors are at the application layer that interacts with either a femtocell or picocell for network connectivity. The architecture of IoT uses a separate gateway (GW) for connectivity to the core network, whereas the BSN uses a GW and an authentication server (AS) which is connected via femtocells or picocells to the main network. The separate deployment of AS puts a heavy load on a network, which can be managed by using the 5G networks because of its vast capacity and coverage. With minor amendments to initial architecture of IoT and BSNs, the devices and sensors in these networks can be connected to the core network. A representative illustration of 5G-based IoT and BSNs is shown in Fig. 3.

With the advent of 5G networks, the focus has been on the development of standard architectures, which can improve the network range and connectivity. The rising demands of devices can be handled by deploying more number of access points, which itself require new deployment sites and network planning (Sharma et al., 2016a, Sharma et al., 2016b, Sharma et al., 2015). An alternative to this has been suggested by many researchers, which emphasis on the use of on-demand nodes such as Unmanned Aerial Vehicles (UAVs) for serving as the Access Point (AP) to devices as shown in Fig. 4. Since a lot of researches have been concentrated on the efficient utilization of UAVs in the next generation wireless networks, it becomes important to understand their impact on the services in the IoT, especially in 5G environments.

UAVs can play a key role in the management and control over the 5G-IoT and BSNs by serving as the active gateway to the network. UAVs themselves operate on batteries, which further adds to the issue of energy management in the 5G-based IoT and BSNs. However, efficient resolution of this issue can lead to the formation of a network, which can be better than the traditional architectures suggested for IoT and BSNs (Park et al., 2016, Mozaffari et al., 2016). UAVs can be used to determine the efficient server center which can enhance the services to users and can lower the latency in the case of congestion or sudden increase in the number of end devices in 5G-based IoT and BSNs.

The devices in IoT are operable in continuous mode, thus the overall amount of energy consumption is too high, which may result in power failures for the nodes operating on the battery leading to a large number faults or network shutdown. Device discovery refers to the registration, removal, querying and route formations between the network nodes. With an efficient selection of the next serving hop, the data across the network can be rapidly disseminated. However, with a large number of devices operating together to share the information, the amount of data to be processed on its reception is also very high. The massive reception of data requires efficient processing, which comes at the cost of more consumption of energy resources.

The energy consumed in processing can be optimized to some extent only, but the load and route can be further optimized to select a path which can prevent overload of computations contributing to the formation of a network with high lifetime. Since the network devices in 5G networks are dependent on the battery power, it becomes more important to select devices with an energy-efficient strategy for the formation of a robust and fault-tolerant network. Thus, three major aspects which are crucial for the formation of a reliable 5G-based IoT and BSNs are fault-tolerant connectivity, energy-efficient device discovery and efficient offloading. The network offloading can help reducing the over-consumption of energy over the same device. Further, it is an easy and efficient way to achieve load balancing in the network which can enhance the lifetime of the network. The improvement in these factors allows the formation of a network with better lifetime and vast coverage.

The existing approaches for IoT and BSNs primarily focus on the selection of energy efficient route for sharing data so as to enhance the lifetime of the network, but at the cost of transmission capacity and efficiency. The existing solutions are not able to withstand the tradeoff between the energy and transport efficiency of the network. Thus, considering the energy aspects of the devices in IoT and BSNs, the approach should not only provide energy efficient selection of devices and route, but should be capable of handling extra load within the energy limits of the network.

The major limitation of the existing solutions given by Kandhalu et al., 2010, Jiang et al., 2016, Weng and Lai, 2013 and Zhou et al. (2015) for efficient device discovery is the no consideration of parameters and features of 5G networks. Sharma et al., 2016c, Yu et al., 2016 and Yoo et al. (2016) considered aerial vehicles to coordinate routing between the sensor networks, but did not focus on the transmission capacity, reliability, and fault-tolerance issues. Further, these approaches did not emphasize much on the offloading which is an important metric in handling energy efficient networks. Thus, the problem deals with the efficient discovery of devices in 5G-based IoT and BSNs with better lifetime as shown in Fig. 5. The lifetime of a network also includes the efficient selection of devices to form a robust and fault-tolerant path which can withstand failures caused by energy breakdowns.

The work presented in this paper uses UAVs to support 5G-based IoT and BSNs. An architecture is developed on the basis of energy and traffic models which allow the formation of a fault-tolerant network that is capable of providing efficient offloading. The proposed architecture offers energy efficient solution for device discovery and load balancing in the network. The key highlights of the proposed work are:

• An energy-efficient architecture for device discovery in 5G-based IoT and BSNs using multiple UAVs.

• Efficient data offloading using on-demand nodes such as UAVs.

• Enhancement of the network lifetime and reduction in per-device energy consumption

• Selection of the energy efficient route between the end devices.

The solution proposed in this paper is capable of reducing energy consumption by 78.4% with 21.3% lower delays and 57.9% lower packet loss. Further, use of UAVs lowers the energy asymptote by 75% and provides more flexibility in implementing energy efficient device discovery.

The rest of the paper is structured as follows: Section 2 gives insights to the related work. Section 3 presents the detailed system, offloading, energy, and fault-tolerant models followed by the energy-efficient architecture for device discovery and theoretical analyses of the proposed approach in Section 4. Section 5 gives simulation results. Finally, Section 6 concludes the paper.