19.2.1 Unmanned Aircraft Systems

Unmanned Aerial vehicles (UAVs), popularly known as drones, are considered enablers of a completely new way of performing tasks that were previously either unreachable or high cost, so fulfilling gaps in many modern applications (Marshall et al., 2015). As shown in Figure 19.1, these aircraft can be of many different sizes and shapes, and missions can be accomplished by either a single or multiple UAVs. This richness of form allows them to be used in diverse applications such as search and rescue, surveillance missions, and goods delivery.

UAVs are usually considered as part of a bigger unmanned aircraft system (UAS) that includes all the needed elements to accomplish a mission (Fahlstrom and Gleason, 2012). The components of the UAS may vary according to UAV type, size, and mission, but typically include the ground control station (GCS), the communications subsystem, and the safety and recovery mechanisms.

Creating a UAS with multiple cooperating UAVs brings additional, highly demanding connectivity requirements and in order to meet them, a new network approach has emerged, the Flying Ad Hoc Network.

19.2.2 Flying Ad Hoc Networks

Communication is a crucial aspect of, and one of the biggest challenges in, the design of multiple vehicle systems (Bouachir et al., 2014; Chung et al., 2011). In a UAS, there are three main types of communications: (a) internal machine communications (IMC), which encompasses any communication between UAVs' internal modules or devices, such as the automatic pilot or cameras; (b) machine-to-machine communications (M2M), which encompasses communications among UAVs; and (c) machine-to-infrastructure communications (M2I), which encompasses the communication between UAVs and the network infrastructure (which can be, for instance, a ground control station or a satellite (Frew and Brown, 2009), or even a combination of both).

In the simplest scenario, all vehicles are directly connected to a common infrastructure, and this can act as an intermediary for all communications between them. However, this strategy has several problems. First, each vehicle must be equipped with expensive and complex hardware in order to perform the long-distance communication with the control station or satellite. Second, many factors may compromise communication reliability such as changing environmental conditions, the high mobility of vehicles, different terrain topologies, or obstacles. Finally, the typical use of a ground control station (GCS) to provide the communication infrastructure limits the mission target locations to the GCS coverage area, since beyond that vehicles disconnect from the network, and become unreachable.

The implementation of an ad hoc network connecting all vehicles is one of the most feasible alternatives to infrastructure-based communication. An ad hoc network is composed of nodes that also function as routers, forming a temporary network with no fixed topology or centralized administration (Sarkar et al., 2008). This approach increases the mission target area, since communications between vehicles and the GCS can be routed through other vehicles in a series of hops. Also, even if there is no connection to a GCS, the nodes can form an ad hoc network to share information or work in cooperation.

Ad hoc networks are classified according to their implementation, utilization, communication, and mission objectives. If the nodes that compose an ad hoc network are mobile, the network is classified as MANET (Mobile Ad hoc NETwork). For vehicle-specific applications, MANETs are subdivided into UANET (Underwater Ad hoc NETwork) for aquatic vehicles, VANET (Vehicular Ad hoc NETwork) for terrestrial vehicles, or FANET (Flying Ad hoc NETwork) for aerial vehicles (Bekmezci et al., 2013; Sahingoz, 2013), as illustrated in Figure 19.2.

Each type of vehicular network faces different, unique challenges: For instance, a UANET must deal with an underwater transmission medium and VANETs often encounter unexpected road obstacles. However, it has been recognized that FANETs have to address more challenging issues than other ad hoc networks (Bekmezci et al., 2013; Sahingoz, 2013), because of the following specific characteristics:

• Higher Node Mobility. FANET nodes typically have higher mobility than those in other types of MANET. As a result, a FANET's network topology can change more frequently, which increases the overhead caused by connecting and routing operations.
• Multiple Connections. In many applications, the nodes in FANETs collect environmental data and then retransmit it to the control station, similarly to wireless sensor networks (WSN) (Rieke et al., 2011). Therefore, FANETs have to manage multiple communications between UAVs and monitoring stations, as well as providing support to peer-to-peer connections among UAVs.
• Very Low Node Density. Typical distances among nodes in FANETs are usually longer than in MANETs and VANETs (Clapper et al., 2007); thus, the communication range in FANETs must also be greater than in other networks. This imposes more demanding requirements for radio links and other hardware elements.
• Heterogeneity. UAV systems may include heterogeneous sensors, and each of them may require different strategies for data distribution.
• Obstacles. Due to the higher node mobility, obstacles may randomly block links among UAVs, which must be addressed in order to provide different temporary communication paths, avoiding the disconnection of nodes.

19.2.3 Flying Things: Unmanned Aerial Vehicles and More

Common to the many different visions for the Internet of Things is the ubiquitous presence of everyday objects equipped with identifying, sensing, networking, and processing capabilities that communicate with each other to achieve a common goal (Atzori et al., 2010; Whitmore et al., 2015). In the context of emerging network models such as IoT, a new name for UAVs starts to take place. As a Flying Thing, UAVs (and any other elements able to fly, like autonomous or nonautonomous air vehicles) can be integrated into a network of physical interactive objects that are able to communicate with other internet-enabled devices and systems. This integration has been called the Internet of Flying Things and has the potential of addressing some of the issues in unmanned systems, as well as introducing new, powerful applications.

19.3.1 Fog and Cloud Computing

There are two main trends shaping our networking today: cloud computing and the proliferation of mobile computing. Around 90% of global Internet users now rely on cloud-based service, and that is mainly due to the fact that smartphones worldwide shipment overtook that of PCs in 2011 (Luan et al., 2015). An emerging wave of Internet deployments, such as the notable IoT, requires not only mobility support and geodistribution, but also location awareness and low latency.

Cloud computing is a key concept for the provision of IoFT. A cloud is an on-demand computing model composed of autonomous, networked IT (hardware and/or software) resources (Hassan et al., 2012). As most of the communication managed by IoFT is transmitted via IoT infrastructure, quality of service (QoS) becomes an important aspect, which meets the same requirements of cloud computing. Service providers offer cloud services with predefined QoS terms based on the Internet as a set of managed scalable, easy-to-use, and inexpensive services to gather clients on a subscription basis (Hassan et al., 2012). Therefore, it is essential for the IoFT infrastructure to meet QoS requirements on the provision of services from and to the flying thing, since the criticality of its operations might be higher than ordinary fixed IoT elements. Moreover, some IoFT applications need real-time response, which is done by delegating computational tasks to the cloud, due to limited capacity of mobile devices (Luan et al., 2015). Such fact amplifies the necessity of assertive services provision.

On the other hand, in case of remote areas with limited or no internet connectivity, it would not be always possible to arbitrarily access cloud services with acceptable QoS response times and real-time requirements. However, some applications could still benefit from IoFT by temporarily working with local copies of data and services, imitating the structure of a cloud computing, and then allowing the provision of services that do not rely on frequent updates. This is the case where fog computing takes place as a key element for IoFT networks. Fog computing consists of a highly virtualized platform to compute, store, and provide networking services between end devices and traditional cloud computing data centers, which are typically, but not exclusively located at the edge of the network (Bonomi et al., 2012). On the IoFT perspective, for instance, a smart farm with limited access to the Internet could benefit from a fog infrastructure processing most of the data and services needed for the usual tasks, uploading and downloading only relevant and/or essential information to the Cloud.

As discussed in Chapter 4, cloud/fog computing characteristics have important intersections with IoT scenarios. The main requirements of the IoFT are met by fog computing due to following characteristics: (a) low latency and location awareness allowing the IoFT network to operate at least within a limited range and time, providing services and performing tasks and missions; (b) mobility, a main necessity of IoFT applications; (c) very large number of nodes, which meets the requirements of both IoT and IoFT networks; (d) predominant role of wireless access; (f) strong presence of streaming and real-time applications; and (g) heterogeneity (Bonomi et al., 2012).

A flying thing might play two different roles in an IoFT scenario supported by fog computing, as seen in Figure 19.3. The first is as processing units in a fog layer, serving as providers for the Internet of Flying Things. To give two examples, in remote or disaster areas where connectivity is hard to be achieved, a flexible mobile structure that could provide or extend fog services would be an important enabler. Equally, UAV's could have a role as fog providers in highly connected environments, where some applications might work well being offline for short or even long periods of time, reducing data usage, and creating batches of information to be uploaded to the cloud all-in-once. In a different role, UAVs could function as edge nodes (end users) benefiting from a fog infrastructure.

19.3.2 Characteristics of the Internet of Flying Things

The IoFT is indeed flexible. This characteristic is important for the provision of almost every feature in such model. It helps increase overall cooperation and collaboration, is ready for real-time operations, is usually up-to-date due to the highly connected environment and easy access to the Internet, and is assisted by a powerful remote cloud and/or local fog structure. In terms of cost, since the IoFT merges the benefits of two well-known paradigms, namely, IoT and UAVs, which may vary from cheap to expensive commercial off-the-shelf products, there will be an affordable and adaptive solution for most needs.

Cooperation and collaboration are desired features for most of modern computing systems. Many modern applications distribute tasks and share information in real time, providing better results quickly. In particular, an IoT-ready environment is usually designed to be equipped with more than one way of acquiring data, interacting, and automating specific tasks (Ungurean et al., 2014; Andreev et al., 2015). Although IoT is a scalable model, its expansion can mean high costs for relatively small returns. If an environment does not expand easily, it might have its flexibility compromised, resulting in limited cooperation and collaboration. IoFT addresses this tension in different ways—such as by setting up UAVs in strategic areas serving as gateways, fog, or cloud data link providers, and also by being capable of replacing sensors and actuators in more active and inexpensive ways (for instance, if a traffic light fails, a flying thing might be used to temporarily replace its task). Moreover, real-time operations are also a priority of the model, since it can be reconfigured to meet requirements and provide the best connection to servers and services available locally or through the Internet.

The IoFT integration with IoT infrastructures achieved by strategically positioning UAVs helps the model to meet some key features. Linking to internet-based information processing can collate services from all around the world in real time, providing valuable up-to-date accurate information. This in turn can facilitate interactive decision-making in response to dynamic situations. Those decisions can be processed in powerful datacenters available as cloud providers. The net result is to allow more reliable and more adaptive missions, maximizing their potential benefits, or broadening their applicability.

Table 19.1 summarizes these IoFT features in comparison with UAVs and IoT paradigm. IoT and UAS segments are limited by their inherent infrastructure characteristics. Although they can be expanded, the setup cost, for instance, is a con that must be considered, especially if such infrastructure might end up being underused. In such case, the use of flexible flying things for sensing and actuating is an advantage.

19.3.3 General Modern Applications of the Internet of Flying Things

UAVs and IoT are popular subjects that have gained attention due to their flexibility and low-cost achievements. The benefits of the Internet of Flying Things go beyond the traditional applications being seen on media press lately. The following paragraphs will present some examples of how applications could be taken to a new level by exploiting the flexibility of IoFT.

19.3.3.2 Applications in Smart Cities

• Surveillance. Flying things can be used to add additional capacity to existing systems for monitoring crowds or responding to emergency signals. For instance, in a smart city, people can be connected to the city's services via wearable technology; in a dangerous situation, a person may notify the infrastructure in several different ways, and trigger an appropriate action such as assigning a surveillance flying thing to monitor the area (see A in Figure 19.4).
• Traffic Monitoring. Although smart cities have devices and sensors to monitor the traffic, these devices are costly and might be unreliable, so coverage is not always complete. Flying things can be used in two ways when planning a smart city infrastructure: to provide flexible additional coverage that might be planned (for example, to monitor the occasional, but predictable high-traffic density around a sports arena) or unplanned (for example, following a traffic incident in an area that is not, or poorly covered by fixed sensors). Thus, flying things can be used to monitor these areas and help the car traffic management (see B in Figure 19.4).
• Commercial Package Delivery. Flying things provide a flexible resource for delivery services. They could be used both to augment vehicles for meeting existing demand types and also to generate new revenue streams for high-speed delivery in hard-to-access or congested areas (see C in Figure 19.4).

19.3.3.3 Applications in smart farms

• Surveillance. Considering that farms are private areas, they may be vulnerable to invasions. The use of flying things to monitor the farms' border might help the surveillance by reporting potential problems in real time using the IoT infrastructure of the smart farm. Moreover, the IoFT serves as a backend infrastructure to IoT-based applications (see A in Figure 19.5).
• Service Integration. Different tasks are required in smart farms, and this is expensive if each is implemented separately. A flexible flying thing can monitor crops, animals, spread products, load supplies, and other tasks. In particular, it can also be connected to other flying things through the Internet even with no line of sight. This feature can improve services in a smart farm by overcoming limitations of traditional communication media (see B in Figure 19.5).
• Identification of Fire or Other Issues. A crop can be damaged by several factors, such as wild animals, fire, and frost. Either through their own sensors or by collected data from fixed sensors, flying things may detect problems in crops (see C in Figure 19.5).

19.3.4 Novel Applications of the Internet of Flying Things

FANETs are enablers of inter-UAV connectivity that simplify the cooperation and collaboration among flying things. However, they are not necessarily connected to the Internet since they operate over swarms of UAVs, which cooperate sharing resources and data in a local offline network. Taking the flying things to a new paradigm, such as IoFT, provides “wings” to the concept of FANETs and makes them more powerful. Imagine flying things geographically distant with no possible line of sight (LoS), but still sharing data, network resources, and services just like if they were all in the same local network. That becomes possible with the introduction of IoFT-oriented networks.

The Wide FANETs (WFANETs) merge the concept of FANETs with IoT, and take wide area networks (WAN) as an inspiration to benefit from all paradigms. A visual representation is seen in Figure 19.6, which illustrates a real scenario with WFANETs.

The main characteristic of ad hoc networks is the existence of at least one path to reach every single node in the network. It is usually possible if nodes are within a limited range and if there are strategic LoS links, guaranteeing that the main links will not be interrupted. The new concept of WFANETs benefits from the connection to an IoT infrastructure already available due to smart environments nearby (e.g., smart cities, smart roads, smart homes, smart farms, and even smart cars). Once connected to an infrastructure that can provide broadband Internet access to the local FANET, a tunnel can be created for information exchange and cooperation with remote FANETs. The act of connecting flying things from different geographical locations through a transparent cloud-based tunneling, providing the same features as a local FANET, creates a WFANET.

WFANETs allow for different approaches and applications. A good example is big events, for example, football matches, Olympic games, and concerts, which require high security in order to manage large crowds, monitor suspicious activities, manage ongoing events, and so on.

Another example of application improved by WFANETs is the pavement scanning for distress, helping the provision of better quality roads. A flying thing scanning roads may be helpful for spreading warnings to intelligent transportation systems, reducing car accident rates, improving security and safety services, and also providing information about possible trip delays and/or personal agenda-based notifications. Additional information such as weather conditions, for example, could also be integrated into the scenario, improving services provision.

By logical extension, one can envisage that the sensors carried on UAVs could also be considered as flying things in their own right, that happen to be colocated with a set of others on board. If these services become advertised through an automatic broker, many other agencies might choose to pay to access them. For example, a food retailer might choose to pick video streams from any UAV-mounted cameras that happened to be in the area if they could be used to predict and respond to incoming customer flows (see Figure 19.7).

Furthermore, the integration of flying things into smart environments allows improvements to the UAV's operation. For instance, GPS (global positioning system) location is a critical information that is essential for a flying thing. With the existence of an infrastructure to which the flying thing is connected, high accurate location information may be available, helping the UAV operate more precisely. Other sensors such as weather stations mounted on buildings might provide real-time data about flying conditions ahead on the flight path. Some security issues can also be solved with IoFT itself, such as GPS spoofing.

19.4.1 General Issues

The new Internet of Flying Things model combines the best of IoT and UAVs into one solution that inherits characteristics, but it also introduces new challenges that must be overcome. In short, these fall into three categories.

The first set of issues relates to public safety, and to ethical concerns around the collection and distribution of data. Other obvious concerns would be around snooping—either by physically capturing images or by moving into the proximity of local secure wireless networks. It seems likely that many countries will enact legislation requiring certification and authorization to fly UAVs and benefit from the possibilities of IoFT.

The second set of issues arises from a desire to avoid the past mistakes of the computing industry and lower the thresholds to engagement for key stakeholders, so that the entire population can reap the benefits of this technology. These issues are related to standardization, and must involve both the industry and governments. From a technical point of view, standardization is necessary to make all things actually able to “talk” to one another, instead of creating a “Babel Tower Effect” in which devices become split into disjoint subsets (for instance, all devices from the same manufacturer) that can only talk to others from the same subset. Hardware and software design must take into account limitations of memory, storage, processing capabilities, and power source. This is likely to lead to different solutions for flying objects being developed by different manufacturers, using various hardware architectures, platforms, and communication protocols. Thus, there is a pressing need for the emerging standards for communication and data transfer between devices to be implementable in highly resource-efficient algorithms. From a social and economic point of view, standardization will favor the entrance of small and medium companies in the market, stimulating entrepreneurship and competition, benefiting the final customer, and spreading the use of the technology.

The third set of issues that arises is mainly related to security and Big Data. The heterogeneity will also extend to the data being collected, since it is expected that IoT—and consequently IoFT—will bring a large amount of new, most likely nonstructured data. How and where (locally or remotely) this data will be stored and recovered following real time and security requirements will involve the use of Big Data techniques and technologies.

These considerations apply to a single IoFT device, but in many scenarios more than one device may be deployed. As the number of connected devices grows, addressing and managing all of them without prejudicing quality of service become a critical issue. This is compounded by the fact that new devices, offering or requesting new services can join the ad hoc network at any time. Hence, applications must be designed from the ground up to enable extensible services and operations. Moreover, IoFT and WFANETs have a common issue to deal with, which is the possible service unavailability regarding an Internet connection. This raises questions such as what measures should be taken if the LTE/3G/4G connection is lost without prejudicing the mission and data privacy?

19.4.2 Security Issues at Different Internet of Flying Things Conceptual Layers

In general, security challenges are big concerns of computing systems. There are several security issues that could exist in IoFT-based applications, which are mostly inherited from the underlying networks and technologies (e.g., UAS, IoT). The main security issues that could be imposed by the different network layers are listed as follows:

• Physical Layer. Both jamming and tampering attacks are known issues for this layer. Jamming is a well-known attack that causes interferences to the radio frequencies that network's nodes are using. It can interrupt the network if a single frequency is used throughout the network: at worst, interrupting communications with flying things; and at best causing excessive energy consumption. On the other hand, if an attacker can physically tamper nodes, a tampering attack takes place which damages, replaces, and electronically “interrogates” the nodes to acquire information (Sen, 2012). GPS spoofing attacks, which happen in this layer, consist in the use of a signal that is stronger than and mimics the attributes of a genuine GPS signal to take over a GPS receiver, have become more frequent. Such attacks can cause the aircraft to completely lose control, which is a very critical issue.
• Data Link Layer. Collision, exhaustion, and unfairness are the most likely attacks at this level. A collision happens when two nodes simultaneously attempt to transmit on the same frequency, resulting in either partial or complete packet disruption, which will cause an erroneous data transmission through a communication channel. In an IoFT, it is a big issue because the possibility of having an intermittent network condition is likely to cause chaos in a FANET. Therefore, each flying thing can be affected and could crash with each other due to the loss of information on the fly. If repetitive collisions take place, the system may suffer from resource depletion, that is, an exhaustion attack. Finally, rather than blocking access to a service outright, an attacker can degrade it to gain an advantage such as causing other nodes in a real-time MAC protocol to miss their transmission deadline, characterizing an unfairness attack (Sen, 2012).
• Network Layer. There are several attacks found on this layer. In a selective forwarding attack, malicious nodes attempt to stop the packets in the network by refusing to forward or drop messages passing through them (Khan et al., 2012), which could compromise a FANET that relies on strategically placed flying things to reach all the destinations within the network. If an attacker makes the compromised flying thing look more attractive to surrounding ones, which is considerably an easy task since IoFT applications usually take place in open environments, the selective forwarding attack becomes very simple. Then, through the affected flying thing, a data transfer situation may be started leading to a sinkhole attack (Dener, 2014). Another attack on network layer is the Sybil attack, in which a flying thing exhibits multiple identities to other flying things in the network (Dener, 2014). On the other hand, a wormhole is an out-of-band connection between two nodes using wired or wireless links, which can facilitate forwarding packets faster than via normal paths (Wallgren et al., 2013). According to Wallgren et al. (2013), a HELLO flood attack refers to overriding “HELLO” messages by broadcasting a stronger signal, allowing an attacker to introduce himself as a neighbor to many nodes, possibly the entire network, which is very likely to be seen in scenarios with flying things. In operation, the communications systems must balance the desire to maximize throughput (the amount of data transferred per unit time—which suggests large packets) and minimize latency (maximum delay accepted for the data transmitted and received—which suggests small packets). The final issue when multiple devices are deployed is that nodes in FANETs can be fixed or not and the distance among nodes can vary while in motion. To assure the route among nodes, FANETs rely on broadcast techniques, which becomes a problem that can be interpreted as an attack to the communication system: the growing amount of data traffic between flying things and IoFT-infrastructure elements; the increasing number of manned aircraft sharing the same airspace; and the progressive forthcoming of flying things to dispute nonsegregate airspace. To solve the problem of broadcast packets to all nodes and avoid packet redundancy and its associated problems (i.e., Broadcast Storm Problem - BSP), techniques to mitigate BSP must be taken into account (de Melo Pires et al., 2016).
• Transport Layer. Flooding attack causes immense traffic of useless messages on the network. It may result in congestion, and eventually lead to nodes exhaustion (Sastry et al., 2013). The desynchronization attack is made when the adversary repetitively pushes messages, which convey sequence numbers to one or both of the endpoints (Dener, 2014). The GPS spoofing may help attackers to hijack flying things, which is another issue that is strongly related to situations where an attacker secretly relays and possibly alters the communication between two parties, also known as man-in-the-middle attacks. This kind of attack allows an attacker to land the flying thing in an unauthorized place, and taking advantage of its legitimate network access. However, such problem can be neutralized by the high connectivity in IoFT environments: In most cases, a solution lies in techniques that check the accuracy of GPS signals by comparing to the ones provided by access points and other fixed known infrastructures.

From another point of view, some researchers have organized IoT in a new security architecture that also applies to IoFT (Jing et al., 2014). Such architecture is mainly divided in three layers: perception, transportation, and application. This organization with three layers only implements most of the features of Open Systems Interconnection model (OSI model), despite the fact it is not as well separated. Figure 19.8 shows a comparison between the IoT model proposed by Jing et al. (2014) and OSI model. Each layer of IoT model faces specific challenges, which will be discussed further.

• Perception Layer. This layer is mainly about information collection, object perception, and object control. In the perception layer within the Internet of Flying Things, tasks related to security of RFID (radio-frequency identification), WSN, RSN (RFID Sensor Network), GPS technology, and so on will be performed (Wu et al., 2010). The heterogeneity of flying things and ordinary IoT/IoFT-ready devices is one of the main problems that may arise in this layer, which can lead to compatibility problems. Another issue is the limitation of power, computing ability, and storage capacity, especially in flying things, that make them more vulnerable to attacks, allowing for physically stealing information and also functioning modification (Jing et al., 2014).
• Transportation Layer. Also referred to as network layer, the transportation layer's main function is transmitting information obtained from the perception layer (Wu et al., 2010). This layer encompasses the Wi-Fi, establishing and maintaining MANET/FANET and 3G/4G/5G networks, leading to a heterogeneity problem for the exchange of information among different networks, which is even more challenging when it comes to the IoFT and its inherent integration among different networks (IoT, FANETs). Moreover, it also leads to new vulnerabilities on the relevant segments for the implementation of IoFT networks. For instance, the main issues related to Wi-Fi are phishing access attacks, malicious AP, DDoS/DoS attacks, and so on. On MANET/FANET side, the security problems that may be faced are data security, network routing, and DDos/DoS issues. Finally, related to the 3G/4G/5G networks, data security and unlawful attacks are the main concerns (Jing et al., 2014).
• Application Layer. This layer supports all sorts of business services and realizes intelligent computation and resources allocation in screening, selecting, producing, and processing data. The security issues it faces cannot be solved in other layers of the IoT model, such as privacy protection issue, which can become a real demand in certain special contexts. Thus, in operation, the application support layer must be able to recognize untrusted data (e.g., spam data and malicious data), and filter them in real time. The application layer can be organized in different ways according to different services, and usually includes middleware, M2M, cloud computing platform, and service support platform (Jing et al., 2014). The privacy concerns in flying things have been recently discussed as a big threat as storage systems get more sophisticated. There is a tendency to store as much as possible into the flying thing memory, so guaranteeing that the needed information is always available (Hartmann and Steup 2013). However, such approach ends up being a very critical security threat (Jing et al., 2014). As the flying thing is inserted in a highly connected environment, it can be physically stolen or its control can be taken. Thus, the attacker may still use it as a gateway to obtain information from the network, since the flying thing is authorized and will be able to access private and confidential information. This is a consequence of the high connectivity of things and the increased contact surface that generates more possible threats to be explored by malicious entities.

19.4.3 Safety Issues of the Internet of Flying Things

The growing popularization of UAVs has increased the research in this field and is fostering the use of such technology in many applications. There are roadmaps published periodically by military and civil organizations—for example, United States Army (US Army), American Department of Defense (DoD), European RPAS Steering Group (ERSG), and Federal Aviation Administration (FAA)—that outline the expected advances for UAVs (US Army, 2010; Yearbook, 2011; UK Civil Aviation Authority, 2012; DoD, 2013). However, there are not enough studies on safety for the specific integration of UAVs and IoT, which is one of the most important topics to be discussed and an open opportunity for researches in safety-critical systems.

There are five challenges of UAS integration, as stated by Dr. Wilson Felder, the Director of the William J. Hughes Technical Center of the FAA, reported by Stark et al. (2013): procedural, technical, aircraft safety, crew credentials, and public acceptance. Sense and Avoid Systems remain as one of the largest obstacles for the safe integration of UAS into airspace. Any person or computer-based system that meets the three mandatory activities to operate an aircraft (flight, navigation, and communication) should assume the command of an aircraft, be it manned or unmanned (Baraldi Sesso et al., 2016). Regulations regarding collision detection and resolution efforts must be met by any UAS designed for nonsegregated airspace.

In short, FAA requirements demand that UAS meet safety levels equivalent to those of manned aircraft. It includes the frequency of collision of a UAS being operated in a FAA-controlled airspace, which is currently 1 × 10⁻⁷ events per hour of operation for manned aircraft (Asmat et al., 2006). As for IoFT applications, there will be merged approaches that may include both manned and unmanned aircraft at the same time, leading to the necessity of using the nonsegregated airspace. This fact increases the need for meeting safety requirements and the first step is to point them out and discuss possible solutions.

Despite the security applied to each layer, there is also the need to find joint approaches that ensure safety individually and collectively. Reducing the breaches throughout the layers will consequently reduce the overall chances of attacks to the network. Such characteristics regarding security and safety make it more difficult to ensure the system safety, composing one of the main threats in the area, which is finding approaches that deal with both concepts at once.

The concept of safety has a long tradition for vehicles (ISO, 2011). This is a mature area, and several standards exist for creating safe systems such as RTCA/DO-178C for UAV software (RTCA Inc., 2011) and RTCA/DO-254 for UAV hardware (RTCA Inc., 2000). Safety deals with minimizing the frequency of accidents or failures in a system, mainly when related with loss of life, high-value assets, and it is related with incautious actions or events.

Providing a safe wireless communication means ensuring that the information transmitted is received without any transmission error and loss of the information. Due to noise, interference, and fading effects, wireless network cannot have zero transmission error, since there is no system with zero risk. For wireless network, transmission error and loss of information cannot be avoided, but they can be overcome by reducing or by detecting them. In order to guarantee communications with safety, Pendli (2014) listed a number of requirements that must be satisfied. Communication links should be reliable and immune against noise, jamming signal, interference, and fading effects in order to provide a link without errors and losses. Since flying things are safety-critical systems, communication channels must be continuously available and provide timely delivery information without failure. Ensuring real-time performance means that the technology used must take into consideration the delay during information transmission and retransmission and be able to cope with burst errors. Device mobility, and the changing external environment requires communication links to be robust even under adverse conditions against channel fading, low SNR (signal-to-noise ratio) conditions, and channel losses.

For flying vehicles in general, there are regulation and legislation requirements, which demand that any factors affecting safety must be taken into account, and incorporated into a risk model meeting certain minimum standards before certification is granted. For IoFT devices, the possible failure rates of solutions to all the summarized communications issues must be incorporated. Indeed, in some applications where the flying thing operates above cities or farms, which are critical fields, the need for certification is even higher. Although there is some way to go, it is hoped that the robustness of the IoT infrastructure may help achieving the low failure rates needed.

In short, developers and users of IoFT systems must take safety as one of the main concerns. However, the combination of high mobility and wireless communications highly increases the exposure of these systems to malicious threats and to faults deriving from uncertain connectivity or communication timeliness. Nonfunctional requirements like security have thus become harder to fulfil, creating new challenges to such safety-critical embedded systems (Bloomfield and Lala, 2013). In fact, further research on the development and assurance of both safety and security must be performed, addressing the needs of multidisciplinary approaches like integrated control systems, communication, security mechanisms, artificial intelligence, neural networks, safety assets, and other technological concerns. A key challenge is that the architectural solutions to ensure safety may open further weaknesses from the point of view of security. On the other side, security weaknesses may lead, if exploited by attackers, to safety violations and the implementation of a given security mechanism may impact safety. In the few cases where security is taken into account, the only problem that is addressed is the open network communication, such as wireless systems. Security at large is not handled (or handled in a very general way) without a full support to identify and mitigate security threats. Therefore, the relationship between safety and security seems to be still an open issue in the community.

19.5.1 Case Study 1: WFANETs for Surveillance Tasks in Smart Farms

Farms of hundreds of hectares with varied topology can be formed and subjected to different climatic conditions along their area. They can perform many different tasks in agriculture like raising livestock, becoming specialized units (e.g., vegetable or fruit farm, dairy, pig and poultry farm or even used for the production of natural fibers, biofuel, and other commodities). All of them also have a large infrastructure that may include plantations, pasture, feedlots, orchards, greenhouses, silos, barns and other buildings, and the farmhouse. In modern times, such term has been extended to include industrial operations as wind farms and fish farms, both of which can operate on land or sea. For more information about the various applications of IoT in agriculture, please refer to Chapter 18.

The concept of smart farms has been used to denote a ubiquitous IoT-ready farm environment for increasing general productivity and optimization of everyday tasks. Internet connectivity can be achieved in such environments mostly on strategic administrative buildings, but also in some dedicated fields with specialized technologies for sensing and monitoring. This case study analyzes a scenario from the IoFT perspective through the application of Wide-FANETs.

The concept of Wide-FANETs is an ad hoc network that benefits from the connection to an IoT infrastructure already available due to smart environments nearby. Such connection can provide broadband Internet access to local FANETs and, thereafter, a tunnel can be created for information exchange and cooperation among remote FANETs. More than one flying ad hoc network from different geographical locations connected through a transparent cloud-based tunneling consists of a WFANET.

19.5.1.2 Proposed Solutions

Following topics will discuss the hypothetical application of FANETs and WFANETs to three farm scenarios: border control, cattle monitoring, and crop monitoring. Figure 19.9 illustrates the scenarios discussed further.

The surveillance of farm borders to identify potential threats, such as intruders (humans or not), and trigger appropriate procedures is one of the most important tasks for safety purposes. In a typical scenario, the task would be performed by human patrol and/or surveillance cameras. This solution, besides costly, is not likely to be effective, since the risk of human failure is always present.

A WFANET would provide border control at a lower cost. The UAVs would gain from the available IoT infrastructure to provide surveillance images or data in real time to a control central (e.g., for security guards, a specialized security company or a police station). The use of UAVs would reduce the necessity of surveillance cameras installed over the property border and its required infrastructure (such as batteries or solar cells for power), again reducing the task cost¹. The advantages would be even bigger in bigger farms, since the use of WFANETs would allow the exchange of relevant information among more than one local FANET, in a cooperative live border monitoring. Additionally, data from official governmental web services may be used to identify fugitives.

Another situation is related to cattle monitoring. In big farms, it is normal to specify strategic land areas intended for cattle raising. Such areas must be constantly monitored in order to avoid escapes or unauthorized human intervention. Once again, if the solution relies on human resources, the coverage of monitoring might be subject to failures or be too slow to take appropriate actions with little damage to the cattle or other farm fields.

An IoT infrastructure would provide means of cattle tracking by installing cheap sensors in each single animal. Moreover, by surveying the area with UAVs, data acquisition would flow fast from the cattle raising area to monitoring centers, providing means of taking timely actions. As an alternative, automatic actions could also take place as soon as an unwanted behavior is identified, for example, closing backup gates in case of cattle escape. For that, FANETs (in small areas) or WFANETs (in bigger areas) would identify and trigger the appropriate action.

The surveillance of crop fields can be motivated by several factors, such as early identification of pests and plagues, soil checking, and weather conditions monitoring. Once again, appropriate local actions could take place as soon as an atypical situation is identified. In such case, more accurate information could be acquired by starting specialized missions with FANETs and/or WFANETs based on the additional information required to precisely recognize a countermeasure^2,3.

UAVs can collect data from wireless sensor networks placed on the ground at strategic frequency, being able to transmit such data in real time to monitoring centers. Moreover, such data collection frequency could be based on weather conditions, originating from both local sensor and web services.

19.5.2 Case Study 2: Internet Access and IoT Services Provision in Remote and Peripheral Areas with IoFT as Fog Enabler

Rural areas and city peripheries might be the trickiest areas for provision of Internet/IoT-based services, especially due to the lack of appropriate infrastructure. In most of the cases, it is not worth installing a full infrastructure that will be rarely used in remote areas. Moreover, as the demand is usually low, there will be no interest by government and business companies to update such infrastructure in order to match the newest technologies requirements.

On the other hand, there are situations when the support of Internet/IoT-based services would positively improve tasks related to search and rescue, life quality levels, special events, people/objects tracking, and so on. This case study will investigate the benefits of using the Internet of Flying Things to amplify the coverage of common smart city services. By doing so, there is a potential opportunity of providing Internet connectivity and IoT-based services for the temporary improvement of special tasks in an inexpensive manner.

19.5.2.2 Proposed Solutions

Here, three main applications of this case study will be highlighted. The first investigates how an IoFT network would be important for smart cities peripheries. The second addresses the environmental monitoring in rural areas helping the environmental police to identify illegal actions and take appropriate countermeasures. Last, natural phenomena and emergency operations support are discussed considering that in such situations connectivity becomes an issue due to the loss of infrastructure nearby. Figure 19.10 illustrates the scenarios that will be discussed further.

By strategically moving drones to the edges of smart cities, a powerful connection to the Internet will be available for restricted areas helping with the provision of connectivity to city peripheries. From this connection, a complete range of IoFT services will be available for end users nearby for a specific period, allowing the execution of relevant tasks⁴.

For instance, the smart city's electric power company might automate the reading process of residential energy consumption. Such task is usually performed by a person taking note about the consumption in each residence/building, which takes more time and is susceptible to misreading. Applying appropriate identifiers to each residence (such as RFID tags), which in turn will be recognizable by the IoFT infrastructure, would provide means of reading the energy consumption of a big area in several minutes. That is possible due to the existence of a FANET/WFANET flying over the area, providing such class of services and being able to provide real-time information to both the customer (e.g., bad debt warnings) and the electric power company (e.g., reading issues in specific residences that would require a technical visit).

Moreover, the existence of a single flying thing or a FANET/WFANET over a neighborhood that is not yet equipped with the smart city infrastructure can be used as a traffic support to drivers in that regions. Smart cars would benefit from IoFT warnings about the traffic. Regular cars could also get important updates about the traffic via smartphones connected to the IoFT infrastructure. Actions like these would contribute for a safer and more accurate use of streets, not to mention the emergency situations that could take place, and fully benefit from IoFT infrastructures.

Rural areas usually do not demand Internet connectivity at all times. In some cases, data mules can be used, allowing the transfer of data between remote locations to effectively create a data communication link. Similarly, an IoFT network can be created to transport data to a desired location, but with an intermittent link provided by a mobile ad hoc network connecting the city and the remote area.

The concept of fog computing is clear in this situation, but the IoFT network emerges as an enabler of the model. Section 19.3.1 discussed the application of a UAV as a fog computing infrastructure using the IoFT elements to provide an updated, flexible solution. Such application fits well for rural areas that need better services delivery for more advanced tasks performing.

This model also meets the requirements of environmental monitoring by governmental official agencies, which is a trend due to the global issues being reported lately. The real-time monitoring by flying things can be remotely analyzed by specialists in central offices that will be able to use updated images to identify suspicious activities. If an environmental illegal practice is taking place, the environmental police can promptly move to the region being supported by the IoFT infrastructure at all times, allowing efficient red-handed operation.

A situation that is also frequent in remote areas is related to natural phenomena and emergency operations support. Natural disasters and catastrophic situations can be predictable but usually not controllable. Their destruction might lead to the unavailability of essential services from local infrastructure, such as cellular networks, causing chaos situations.

The application of IoFT to provide means of connectivity from the heart of the city throughout a FANET that reaches the disaster area can be an efficient temporary action to support rescue agents to do their job as qualified as possible⁵. For instance, if a disaster has led to deaths, one might consider the real-time recognition of bodies by relatives. However, it is not always safe or appropriate to transport people to the disaster area. Also, it is not possible in some situations to remove all the bodies right away, resulting in panic and people flooding agencies with calls demanding information.

IoFT in emergency situations will provide connectivity for basic services provision, governmental central offices, and news agencies to be updated with recent discoveries, and the complete support for ambulances being moved to the target area.

19.5.3 Case Study 3: Targeted Services Delivery on Big Events with IoFT

Big events take place throughout the year basically everywhere. They can be of many types such as local exhibitions, big theatre presentations, and hundreds of simultaneous concerts and presentations in several different venues in a music festival. These big events can be crowded by people coming from multiple places with different accessibility demands.

General safety is an issue that can be supported by an IoFT infrastructure, especially when an event takes place in a remote location (e.g., gigantic music festivals). It is similar to the situation described in Case Study 1, when surveillance tasks may be demanded. However, this case study will focus on targeted services delivery for big events provided by IoFT networks.

19.5.3.2 Proposed Solutions

The applicability of IoFT in the cited situations might be a simple and efficient solution. The following subsections will discuss two cases specifically. First, the application of finding people will be discussed. Second, real-time information about multiple venues in music festivals will be addressed. Figure 19.11 presents the scenario of this case study.

It is common to suddenly find yourself separated from your friends or family in crowded places. If a cellular network is not available for any reason or your phone ran out of battery, then you might have a problem. Plus, the situation gets even worse if a kid is lost by any reason and the parents are desperately looking for him/her.

The identification of people using fixed cameras coupled to the infrastructure of venues is efficient most of the times. However, in crowded places the angle of the camera might be a problem. In case of lost children, it is even more problematic, since the height would negatively influence the camera coverage. For cases like this a reasonable solution would be a flying thing connected to a FANET/WFANET for accurate, fast search on site.

The real-time streaming can be monitored by trained people and also relatives. Such effort would significantly reduce the chances of any hurt to an unattended child and also provide effective means to supporting families.

Another relevant topic in big events, especially the ones with simultaneous attractions in multiple venues, is the possibility of getting access to information in real time. One can wonder whether a concert meant to be played at a specific time has started or is delayed. Moreover, why not have access to the audio or a video streaming to check how is the performance going or how crowded it is?

An IoFT network plus fixed cameras and sensors could provide a set of information about multiple venues in real time to a central server that would redistribute such information all over the network. Personal smartphones or special stations could get access to such information, improving the delivered services during the days of the event.