A framework for designing AED drone networks

(1)

A framework for designing AED drone networks

Master Thesis

MSc Technology & Operations Management

University of Groningen, Faculty of Economics and Business

by

J. Ruijfrok

2379120

j.ruijfrok@student.rug.nl

Supervisor / university

Dr. ir. D.J. van der Zee

Co-assessor / university

Dr. N. Szirbik

Supervisor / field of study

Ir. J. Hatenboer

UMCG Ambulancezorg

(2)

2 Abstract

Purpose - Sudden cardiac arrest is one of the leading causes of death in Europe. For Out-of-Hospital

Cardiac Arrest (OHCA) patients, early defibrillation is critical to optimize a patient’s probability of survival. Late response times of emergency service (EMS) providers however, has lead to the need of alternative methods to facilitate timely defibrillation of these patients, particularly in rural areas. Current alternative methods such as publicly accessible Automatic External Defibrillators (AED) are not able to be highly responsive in these OHCA situations in rural areas. Drones can therefore be a new alternative approach as studies have shown that drones are capable of reaching these locations within a few minutes. AED drones need to be supported by a comprehensive network consisting of operational processes, network resources and a control structure. Existing studies offer little support in designing AED drone networks. This study therefore aims to develop and evaluate a framework for designing a responsive AED drone network for the use in rural regions facilitating treatment of Out-of-Hospital Cardiac Arrests. This framework will be a stepwise structure that support EMS providers to design AED drone networks.

Method – A design science method is adopted to develop this framework. First the literature is

reviewed to identify the basic elements of the OHCA treatment, drones and AED drone network design. Then the literature, interviews with domain experts and observations of relevant documents are used to characterize and describe AED drone networks. Building on this, a framework to design AED drone networks will be developed and a limited evaluation in a pilot study will be performed.

Findings – The characterization of AED drone network shows the network elements and its sub

elements for operational processes, network resources and the control structure. Key decisions have been provided for all sub elements. The framework shows that these decisions should be approached by aiming for optimal benefits while ensuring tolerable costs. Besides, the key decisions are externally driven by the impact or need of region-specific characteristics. In doing so, optimal AED drone networks can be designed. In essence, the framework input is the choice of region with its corresponding region-specific characteristics. The framework itself consists of four phases. It starts with defining the impact of the region-specific characteristics on AED drone network design. Then, the physical infrastructure will be determined in terms of operational processes, network resources and the control structure. Based on these decisions, the network resource dimensions will be determined. Finally, initial solution designs can be put to the test in order to assess their performance and potentially reveal the need for adjustments. The framework output is the final AED drone network design that entails all made decisions on operational processes, network resources, the control structure and the network resource dimensions. The evaluation results of the framework showed that many critical elements of the AED drone network design are currently still not available. Especially certain technology and regulations do not enable a viable design and implementation process of AED drone networks. A full evaluation has, however, not been performed given time constraints, existing regulations and unavailability of technology; the framework therefore needs adaptions on regular basis.

Conclusion – The provided structure to AED drone networks results in many starting points for EMS

providers that tend to design such a network. The framework helps to identify the key network elements and alternative decisions. In doing so, the framework offers support for making them. Therefore, the framework helps to design AED drone networks effectively.

(3)

3 Preface

This master’s thesis is the final step in completing the program Technology and Operations Management. Furthermore, it marks the end of my student career. The goal of this thesis is to develop a framework that helps EMS providers to design AED drone networks.

Performing research in an emerging world of drone applications has been a great and challenging experience to me. This research allowed me to travel the whole country for interviewing a host of experts in several distinctive fields. The research would not have been possible without their help. I really admired the enthusiasm about AED drones and the willingness to help with everyone I met. In particular, I would like to thank Mr. Hatenboer of UMCG Ambulancezorg for his guidance, time and his valuable insights. He really involved me and introduced me to many people who are somehow related to drones. I travelled for instance with him to The Hague to attend a meeting with the regulatory authority of The Netherlands that is responsible for regulations on drones. Furthermore, I would like to thank Mr. Heidekamp of ANWB Medical Air Assistance for helping me with his expertise and view on my research. Besides, I am thankful to all other people that helped me during the research.

I would also like to thank my supervisor Mr. Van der Zee for his on-point, constructive and insightful feedback throughout the whole process. The structural sessions with provision of feedback by him and two other students, Roy and Justine, have provided me with significant help in finishing this thesis. Last but not least, I would like to thank my girlfriend, family and friends for always supporting and motivating me to finish the thesis, especially in times when I experienced a setback.

(4)

4

5

5. Framework design ... 32 5.1 Framework approach ... 32 5.2 Framework input ... 33 5.3 Framework ... 35 5.4 Framework output ... 52 6. Evaluation of framework ... 53 6.1 Evaluation approach ... 53 6.2 Framework input ... 53 6.3 Framework ... 55 6.4 Framework output ... 63 6.5 Summary of findings ... 63 7. Discussion ... 64 7.1 Main findings ... 64 7.2 Limitations ... 64 8. Conclusion ... 65 8.1 Main conclusion ... 65 8.2 Future research ... 65 References ... 66

(6)

6 List of Abbreviations

AED Automatic External Defibrillator

ALS Advanced Life Support

ANWB Algemene Nederlandsche Wielrijdersbond ATM Air Traffic Management

BVLOS Beyond Visual Line of Sight CPR Cardiopulmonary Resuscitation EASA European Aviation Safety Agency EMS Emergency Medical Service EVLOS Extended Visual Line of Sight GPS Global Positioning System

ICAO International Civil Aviation Organisation IFR Instrumental Flight Rules

ILT Inspectie Leefomgeving en Transport I&W Ministerie van Infrastructuur en Waterstaat INS Inertial Navigation System

JARUS Joint Authorities for Rulemaking on Unmanned Systems NASA National Aeronautics and Space Administration

OHCA Out-of-Hospital Cardiac Arrest SESAR Single European Sky ATM Research SORA Specific Operations Risk Assessment UAV Unmanned Aerial Vehicle

UMCG Universitair Medisch Centrum Groningen UTM Unmanned Traffic System

VFR Visual Flight Rules

VLL Very Low Level

(7)

7 1. Introduction

Sudden cardiac arrest is one of the leading causes of death in Europe (Perkings et al., 2015). Patient outcomes depend on the so-called chain of survival (Perkings et al., 2015). According to the chain of survival, early recognition, early cardiopulmonary resuscitation (CPR), early defibrillation and early advanced life support are the primary care services that should be provided by an emergency medical service (EMS) provider in order to optimize a patient’s probability of survival (Perkings et al., 2015). For out-of-hospital cardiac arrest (OHCA) patients, early defibrillation, in particular, can increase survival rates dramatically (Nordberg et al., 2015). Defibrillation is the administration of an electrical shock in order to restore a normal rhythm of the heart and is usually performed using an automated external defibrillator (AED) (Fredman et al., 2016).

The potential for saving lives is one of the main reasons why European countries have implemented dense EMS networks with high operational costs (Van de Voorde et al., 2017). However, even with the support of other first responders such as volunteers, fire fighters and police, EMS providers generally do not arrive within the critical first 5 to 10 minutes after an OHCA occurs (Van de Voorde et al., 2017). Given the importance of early CPR and defibrillation, EMS providers have attempted to improve response times by implementing AED and volunteer networks (Fredman et al., 2016). Despite these efforts, however, the responsiveness of the individuals who use AEDs remains unacceptable, as access to publicly accessible AEDs is subject to the opening hours of buildings; in addition, distance may be a factor, particularly in rural regions (Van de Voorde et al., 2017). An increase in AED responsiveness would therefore improve survival rates (Larsen et al., 1993). Therefore, alternative approaches to AED facilitation are required in order to increase the responsiveness of AED services. A novel approach to increasing AED responsiveness in the OHCA treatment could involve the use of unmanned aerial vehicles (UAVs), better known as drones. This could be achieved by using a drone to deliver an AED to an OHCA location. Studies have shown that drones are capable of arriving at OHCA locations within a few minutes and could therefore ensure timely responses in rural regions (Pulver et al., 2016; Cleasson et al., 2017). In order to facilitate the use of drones in OHCA treatment, there is a need to implement an AED drone network (Pulver et al., 2016). A network is defined as a collection of network elements that have interdependent interactions (Möller and Halinen, 1999). The main elements are the operational processes, the resources that support these processes and a control structure that provides guidance regarding how these processes and resources should interact.

Decision-making in a network is typically affected by regional characteristics, including stakeholders, weather conditions and regulations (Möller and Halinen, 1999). Stakeholders may, for example, require certain elements to be included in a network. This means that every network will be unique to a specific region. Designing AED drone networks therefore also entails making decisions concerning all of the main network elements while also taking into account region-specific characteristics.

(8)

8

no viable guidance for designing AED drone networks. This gap in the literature stresses the need for a comprehensive approach that seeks to identify the key network elements in alternative settings and provide support for the decisions associated with designing an AED drone network. Therefore, the research objective of this study is as follows: To develop and evaluate a framework for designing a

responsive AED drone network for the use in rural regions facilitating treatment of out-of-hospital cardiac arrests.

To develop this framework, the design science method outlined by Wieringa (2014) is adopted. This framework is intended to provide support in understanding how AED drone networks should be designed. On the basis of insights provided by the literature, observations of documents and interviews with domain experts, AED drone networks are first characterized in terms of their key elements, i.e., their operational processes, network resources and control structure. Thereafter, building on these findings, a framework for the design of such networks is developed. The proposed framework is then applied and evaluated in a pilot study concerning a region in which the EMS provider UMCG Ambulancezorg operates. As such, this EMS provider serves as the research vehicle.

(9)

9 2. Research design

This chapter is intended to provide an overview of the research design employed in this study. Section 2.1 will describe the problem background. Thereafter, the research objective and research deliverables are identified in Section 2.2. Section 2.3 introduces and explains the conceptual model intended to facilitate the delivery of AED services. Finally, an outline of this research is provided in Section 2.4.

2.1 Problem background

Currently, the survival rates in rural areas for out-of-hospital cardiac arrest patients are relatively low when compared to urban areas1_{(Fredman et al., 2016). This is largely due to the high response times of} EMS providers, which occur as a result of the lengthy distances that must be travelled to reach OHCA locations and the difficulty of reaching certain areas (Fredman et al., 2016). Given the chain of survival, early CPR and defibrillation are critical to promoting survival probabilities (Perkings et al., 2016). Therefore, EMS providers strive to reduce response times by implementing AED and volunteer networks intended to facilitate early CPR and defibrillation (Fredman et al., 2016). An AED network includes publicly accessible AEDs or AEDs that are stored at volunteers’ homes. In the Netherlands, registered volunteers can be activated using an app in order to facilitate rapid CPR and defibrillation (HartslagNu, 2017). Despite these initiatives, however, the responsiveness of individuals who deliver AED services remains unacceptable, as publicly accessible AEDs are subject to the opening hours of buildings or are not accessible due to the distances involved, particularly in rural areas (Van de Voorde et al., 2017). An increase in AED responsiveness would therefore increase survival rates (Larsen et al., 1993); therefore, alternative approaches to AED facilitation should be considered in order to increase the responsiveness of such services.

Drones represent a new approach to facilitating the delivery of AED services. Studies have shown that AED drones are capable of delivering an attached AED to an OHCA location within minutes (Pulver et al., 2016). As a result, bystanders are enabled to perform early defibrillation. The concept of AED drones emerged in 2014, when Alec Momont developed a prototype of an ambulance drone (Van de Voorde et al., 2017). From that point, many stakeholders became increasingly interested in AED drones as their medical relevance to society was revealed (Cleasson et al., 2016). However, no practical AED drone operations have been identified. For the practical integration of AED drones intended to facilitate the delivery of AED services by EMS providers, the implementation of an AED drone network is crucial. A network is a collection of operational processes, network resources and a control structure that have interdependent interactions (Möller and Halinen, 1999). Implementing an AED drone network would therefore entail making decisions concerning these various network elements. Decision-making is driven by regional characteristics, including stakeholders, weather conditions and regulations (Möller and Halinen, 1999). Region-specific characteristics will likely ensure that each AED drone network is unique.

As stated previously, the existing literature offers little support for designing AED drone network; existing studies only support the final design step by addressing the selection of launch site locations from which the AED drones fly to the OHCA locations throughout a region (Pulver et al., 2016; Cleasson et al., 2017; Cleasson et al., 2016). In doing so, previous studies neglected all of the relevant design steps that should be considered prior to the selection of locations. For example, these studies offer no details concerning the types of drones that their authors selected or why they have made certain operational choices.

(10)

10 2.2 Research objective

The research objective of this study is as follows:

To develop and evaluate a framework for designing a responsive AED drone network for the use in rural regions facilitating treatment of Out-of-Hospital Cardiac Arrests.

Herein, a framework is defined as a structure that provides support for the understanding of or building on a certain concept (Wieringa, 2014). The framework developed in this study is a stepwise structure for designing AED drone networks.

Responsiveness, or response time, is the time elapsed from an emergency call to AED arrival on the scene of an OHCA (Larsen et al., 1993). Other times are not taken into account, as doing so would be beyond the scope of this study. According to Larsen et al. (1993), response time is directly linked to patient outcomes: When it comes to performing defibrillation, a difference of one minute can result in a 10 percent lower survival probability. This topic is elaborated upon in Section 3.1.3.

Finally, rural areas are defined as all populations, housing, and territory not included within a human settlement with a high population density (i.e. an urban area). Whatever is not urban is considered rural (Health Resources and Services Administration, 2017).

The research deliverables are as follows:

- An evaluated and validated characterization of AED drone networks; and

- An evaluated and validated framework design for designing AED drone networks

2.3 Conceptual model

Figure 1 presents a conceptual model of a typical AED facilitation system. The input is the call for help from an OHCA patient or a bystander/family member. An AED should then be made available at the

(11)

11

scene of an OHCA. This study focuses on the AED facilitation of AED services by means of a drone network. As mentioned previously, such a network is comprised of operational processes, network resources and a control structure. In the context of this research, operational processes include the dispatching of AED drones or the delivery of an AED. Network resources are, for example, the launch sites from where AED drones fly to OHCA locations. The control structure provides guidelines on how network resources should be used and operational decisions determined. These network elements are externally affected by region-specific characteristics, such as the landscape, stakeholders and regulations. Region-specific characteristics will affect decisions concerning all network elements. For example, an AED drone may not be permitted to fly over certain areas as a result of regulations or environmental issues. This would affect decisions concerning both operational processes and network resources, since it may be necessary to relocate launch sites.

Other means of facilitating AED delivery are the EMS networks, volunteers and public accessible AEDs. The EMS network typically consists of ambulances that are equipped with a defibrillate function; these are sent to OHCA locations by a dispatch and control centre. Secondly, volunteers may have AEDs and be capable of transporting them to OHCA scenes. Finally, public accessible AEDs are mainly present at public locations such as train stations, public libraries and offices. A dispatch and control centre must decide which AED facilitation option will have the lowest response time when responding to an OHCA report. Therefore, the four AED facilitation options should always be linked to each other. As mentioned in Section 2.2, performance of AED facilitation is typically measured in terms of responsiveness; ultimately, this will determine a patient’s outcome.

2.4 Research outline

In order to achieve its objective, this study adopts the design science method as described by Wieringa (2014). This approach is appropriate because it results in the development of an artefact that can provide support when designing AED drone networks. The design science approach for this study consists of five phases: a review of literature, network characterization, framework design, the application and evaluation of the proposed framework and a discussion of the theoretical contribution.

The phases and their corresponding thesis chapters are presented in Table 1. The first phase reviews the existing literature concerning the design of AED drone networks and evaluates the gaps in the literature. Thereafter, the second phase characterizes the key elements of AED drone networks. Building on the work done in the second phase, the third phase develops a framework for designing AED drone networks. In the fourth phase, the framework design is applied and evaluated in a pilot study. Finally, on the basis of the framework design and the evaluation thereof, the theory that underlies it is generalized. As a result, the research deliverable are produced. By generalizing the findings, the framework is rendered applicable to other EMS providers. The five phases are described in more detail below Table 1.

Research phase Chapter

1. Literature 3. Literature review

2. Network characterization 4. Characterization of AED drone networks

3. Framework design 5. Framework design

4. Application and evaluation of the framework 5. Framework design 6. Evaluation of framework

5. Theoretical contribution 7. Discussion

(12)

12

Phase 1: Literature review

In this phase, the existing literature is reviewed in order to identify the basic elements of the OHCA treatment, drones and AED drone network design. Furthermore, this review also investigates the existing gaps in the literature. The following main sources are consulted:

- Scientific articles;

- Relevant documents from aviation authorities; and - Relevant documents from health associations.

Phase 2: Network characterization

The second phase characterizes the key elements of AED drone networks, describing them in detail. This phase provides the basis for the actual design of the framework. The following main sources are consulted:

- The literature review;

- Non-structured interviews with domain experts in the provision of EMSs, representatives of the Dutch Heart Association and experts in emergency helicopter operations, drone operations and the Dutch regulatory aviation authority; and

- Relevant documents from the (unmanned) aviation authorities

Phase 3: Framework design

This phase focuses on developing a framework for designing AED drone networks, with the previous phase serving as the input for the actual framework design. The relevant elements are placed in a stepwise order, thus allowing that AED drone network can be designed logically. The framework is validated continuously in an iterative process until no further improvements can be made. The main sources consulted are as follows:

- The literature review;

- Non-structured interviews with domain experts in the provision of EMSs, representatives of the Dutch Heart Association and experts in emergency helicopter operations, drone operations and the Dutch regulatory aviation authority; and

- Relevant documents from the (unmanned) aviation authorities

Phase 4: Application and evaluation of the framework

In this phase, the framework design is put into practice in a pilot study. The EMS provider in the region of Drenthe in the Netherlands, UMCG Ambulancezorg, serves as the research vehicle; the pilot study is therefore conducted in the region in which UMCG Ambulancezorg operates. The main sources consulted are as follows:

- Non-structured interviews with domain experts in the provisions of EMSs and emergency helicopter operations

The evaluation is limited, as a full evaluation could not be performed due to time constraints. The evaluation consists of a basic check for completeness and an illustration of how the framework functions. As a result, only a limited number of experts were involved in validating the decisions made.

Phase 5: Theoretical contribution

(13)

13

Figure 2 – System overview of OHCA treatment ent

3. Literature review

This chapter conducts a review of the existing literature on OHCA treatment and how an AED drone network fits into such a OHCA treatment system. The chapter begins with Section 3.1, which describes a typical OHCA treatment system. Section 3.2 introduces the use of drones as AED carriers by characterizing their components and modes of operation and identifying relevant regulations; it also provides an initial review of the use of AED drones. Thereafter, the existing literature concerning AED drone networks is discussed in Section 3.3. Finally, the main findings of the literature review are presented in Section 3.4.

3.1 System description of OHCA treatment

This paragraph provides a system description of OHCA treatment. An OHCA is defined as a cessation of cardiac mechanical activity, as confirmed by the absence of signs of blood circulation, that occurs outside of a hospital setting (Roger and Lloyd-Jones, 2011; Jacobs et al., 2004). A cardiac arrest is often confused with a heart attack; however, a cardiac arrest differs from a heart attack in that the former refers to the cessation of blood circulation, while a heart attack refers to an impaired flow of blood to the heart muscle itself (Roger and Lloyd-Jones, 2011; Jacobs et al., 2004). A typical OHCA treatment system is characterized by the input of an OHCA patient, the treatment itself (which is subdivided into the OHCA care services and the OHCA resources) and finally results in an output, the treated OHCA patients. The core processes and resources are schematically depicted in Figure 2. The AED network is marked in bold, as this study is built upon this OHCA resource. This discussion continues with Section 3.1.1, which elaborates on OHCA patients. Thereafter, the chain of survival is discussed in Section 3.1.2. The survival probability curves that are used to evaluate the performances of the care services in the chain of survival are addressed in Section 3.2.3. Finally, Section 3.2.4 discusses the OHCA pathway.

3.1.1 OHCA patients

(14)

14 3.1.2 Chain of survival

The chain of survival encompasses the core care services required for a successful resuscitation. The chain of survival consists of the following four vital links: early recognition and call for help, early bystander CPR, early defibrillation and early Advanced Life Support (ALS). The chain of survival is depicted in Figure 4, and each link is explained below.

I. Early recognition and call for help

Once a collapse occurs as a result of a cardiac arrest, early recognition is critical in order to allow rapid activation of an EMS provider and prompt initiation of bystander CPR. Key observations for recognition are unresponsiveness and a lack of normal breathing (Perkings et al., 2015).

II. Early bystander CPR

Following recognition, early bystander CPR should be performed. Cardiopulmonary resuscitation is defined as administering chest compressions together with ventilations (Perkings et al., 2015), which ensure that blood circulation to the body will continue. In this manner, vital organs will be provided with the oxygen that they require, thus minimizing their deterioration (Perkings et al., 2015). This creates time for the initiation of defibrillation or the arrival of ALS. Cardiopulmonary resuscitation can therefore quadruple survival probability from cardiac arrests (Valenzuela et al., 1997; Waalewijn et al., 2001).

III. Early defibrillation

Next, defibrillation has to be performed as soon as possible, as doing so results in the highest survival probability (Larsen et al., 1993). Early defibrillation can be performed using public accessible or onsite AEDs. An AED consist of defibrillation pads, which should be attached to the patient’s chest. An AED operates semi- or fully automatically; these modes differ in the manner the shock is administered. For a semi automatic AED, a bystander administers the shock by pressing a button; a fully automatic AED does this automatically (Perkings et al., 2015). An AED measures the initial heart rhythm and determines whether or not it is shockable. When shockable, a shock is recommended in order to restore the heart rhythm to a steady state from which the heart can restore its normal heart rhythm (Perkings et al., 2015). If the rhythm is non-shockable, one should administering CPR to the OHCA patient, as only ALS can possibly transform such a rhythm into one that can be shocked.

IV. Early Advanced Life support

Finally, ALS should be performed. Advanced life support (ALS) refers to the operation performed by an EMS provider, such as intubation, the provision of medicines, infusion and transportation to a hospital (Perkings et al., 2015).

Cardiac arrest Cardiac cause (80%) Initial shockable rhythm (60%) Treatable by AED Initial non-shockable rhythm (20%) Non-treatable by AED Non-cardiac cause (20%) Non-treatable by AED I. Early recognition and call for help II. Early bystander CPR III. Early de9ibrillation IV. Early Advanced Life Support

Figure 3 – Categorization of OHCA patients for the treatment of an AED

(15)

15 3.1.3 Survival probability curves

Given the chain of survival, it is important to determine the survival probability of an OHCA patient at the point in time at which, for example, defibrillation by means of an AED is to be performed (Larsen et al., 1993). The survival probability curve estimates this survival probability for a certain period of time after the patient has collapsed. This survival probability curve is based on the relationship between the response times of medical interventions and the survival rate of OHCA patients (Larsen et al., 1993). By applying the curve, one can determine what a patient’s probability of survival is at a certain point in time with respect to the response times of medical interventions. This includes the time elapsed since a collapse occurred and the initiation of CPR or defibrillation. Survival probability curves have been established in the literature for the most common group of OHCAs, namely those with a cardiac cause and a shockable rhythm.

Jansma (2014) and Steenkamer (2015) studied the wide body of literature concerned with survival probability curves. These curves are all based on response time in minutes and define survival as a patient being discharged from a hospital after suffering an OHCA. The authors determined that the studies conducted by the American Heart Association (2000) and Larsen et al. (1993) are most relevant to the OHCA treatment. The two two survival probability curves developed in these studies are explained below.

- The American Heart Association (2000) constructed the following well-known, frequently cited and commonly used survival probability curve:

-

!!"#$%$&',!!!"#$%&' ! = 100% − 10! , with t = time since collapse with defibrillation

This curve shows the survival probability of an OHCA patient based on the time interval between collapse and defibrillation (Figure 5). The curve suggests that the initial survival probability is 100 percent and is reduced each minute with 10 percent.

- Larsen et al. (1993) constructed a survival probability curve that estimates the impact of CPR, defibrillation and ALS on survival (Figure 6):

!_{!"#$%$&',!!!"#$%&'} ! = 67% − 2,3!_!"#− 1,1!_!"#− 2,1!_!"# , with !!"# = time since collapse, with CPR being provided

!!"# = time since collapse, with defibrillation being provided

!_!"# = time since collages, with ALS being provided

The curve suggests that the initial survival probability is 67 percent and that the final survival probability is determined by !!"#, !!"# and !!"#.

(16)

16 3.1.4 The OHCA pathway

An OHCA pathway is the physical architecture of the network that an EMS provider has designed in order to provide the care services and resources necessary to facilitate timely treatment of OHCAs (Aringhieri et al., 2017; Perkings et al., 2015). These three primary care services are typically telephonic care, pre-hospital care and in-hospital care. These care services are supported by resources such as the operators working in dispatch and control centres, volunteers, EMS networks and hospitals (Campbell et al., 2017). Figure 7 presents the OHCA care services and resources schematically. In addition, it also depicts each relevant event in OHCA treatment such as the OHCA, the telephone call and arrival at the hospital.

When a dispatch and control centre operator receives an emergency call from a bystander, telephonic care will be initiated. This entails completing a triage protocol. A triage protocol provides guidelines for assessing patients’ situations and assigning priorities to them (Dami et al., 2015). During the administration of telephonic care, the ambulances in the EMS network will be dispatched and a volunteer network will be activated if available. Thereafter, pre-hospital care is provided. CPR can be administered by EMS staff who arrive by ambulance within the EMS network, volunteers from the volunteer network or bystanders. These three potential performers of CPR can also operate an AED in order to perform early defibrillation. An AED can be retrieved by an AED network. ALS will eventually be administered by the EMS provider. When the OHCA patient arrives at the hospital, in-hospital care will be provided, and his or her final outcome will be determined, be it discharged from hospital or death. In this thesis, the focus is on the first two care services, namely telephonic care and pre-hospital care. In-hospital care is not addressed in this study, with the exception of the determination of final OHCA patient outcome. In the OHCA pathway, there are many links with the chain of survival. Early recognition and calling for help are addressed in telephonic care since the OHCA is recognized and reported to the dispatch and control centre. Early CPR, early defibrillation and early ALS, however, are typically performed in the pre-hospital care phase, if possible.

Figure 7 – Schematic overview of OHCA pathway

(17)

17 3.2 Drones as AED carriers

This section describes the use of drones as carrier of AEDs. An Unmanned Aerial Vehicle (UAV) or drone is defined as a flying vehicle that does not carry a human operator, flies remotely or autonomously and can carry payloads (Gupta et al., 2013). Drones are first characterized in Section 3.2.1; thereafter, the literature on AED drones is discussed in Section 3.2.2.

3.2.1 Characterizing drones

Drones are first described in order to provide an overview of all their relevant aspects. These aspects are their physical components and modes of operation; relevant regulations are also discussed.

Drones’ physical components

Drones are widely categorized and are divided into many types, sizes and weights, depending on their intended applications (Hassanalian and Abdelkefi, 2017). Essentially, drones can be considered as capable of performing a vertical take-off and landing (VTOL), horizontal take-off and landing (HTOL) and hybrid forms (Hassanalian and Abdelkefi, 2017). Based on these categories, several types of drones have been identified and analysed with respect to healthcare applications. Three main drones types have been identified: fixed-wing, rotary-wing and tilt-wing/tilt-rotor drones. A fixed-wing drone is characterized by its rigid wings and uses a motor and propeller as its propulsion system (Mueller, 1999), whereas a rotary-wing drone is characterized by its multiple rotors and the ability to perform a VTOL and hovering flight (Austin, 2011). Finally, a tilt-wing/tilt-rotor drone is characterized by its ability to perform both VTOL and HTOL; thus, it combines the capabilities of fixed-wing and rotary-wing drones. The ability of performing both VTOL and HTOL can be achieved by a mechanism that allows the angle of the rotor or wing to be tilted (Jin and Shim, 2014).

Turning to the interiors of drones, there are multiple components that are relevant for any drone application. Guidance, navigation and communication systems are the most relevant ones, as they affect a drone’s operation. Guidance systems are concerned with the command and control or behaviour of drones. A drone can be commanded and controlled from a ground control station by a pilot, by automation systems, by autonomous systems or by a combination of these options (Chao et al., 2010). A system is autonomous if it can decide by itself if a certain action should be performed; otherwise it will be considered to be an automation system, as commands will still be provided by a pilot (Chao et al., 2010). Navigation systems consist of location estimation sensors such as a global positioning system (GPS)2 or an inertial navigation system (INS)3 and other sensors that are capable of detecting a Drone’s location and assisting in planning a route (Hassanalian and Abdelkefi, 2017). Communication systems consist of communication links such as a C2 data link4_{that ensures that a} drone can send and receive data and commands from a pilot or an air traffic management system (Gupta et al., 2013).

Drones are controlled from a ground control station. Essential aspects of control include monitoring, flight command and control and situational awareness systems (Gupta et al., 2013). A ground control station transmits command signals to the drone and ensures that it behaves accordingly. These commands are usually issued by the drone pilot who operates a ground control station, but they can also be issued autonomously, without the interventions of a pilot (Gupta et al., 2013).

2

_A

_{GPS is a satellite navigation system that uses a radio receiver to collect signals from orbiting}

satellites to determine position, speed, and time

3_{AN INS is a system that includes gyroscopes and accelerometers which are used to calculate the} position and orientation of the drones and is able to support GPS signal when lost

(18)

18 Modes of operation

According to Eurocontrol (2017), there are three airspaces and several modes of drone operations: 1. Very low level (VLL) refers to the airspace below instrumental flight rules (IFR)5_{and visual}

flight rules (VFR)6 altitudes. No IFR and VFR flights are allowed in this airspace and there is no controlled manned aviation. There are four approaches to operating within this airspace: - Visual line of sight (VLOS): This refers to exactly keeping a drone within visual line of sight

at all times, using the pilot’s unaided vision.

- Extended visual line of sight (EVLOS): This refers to a pilot being supported by one or more remote observers in order to maintain visual contact with the drone

- Beyond Visual Line of Sight (BVLOS): This refers to a drone being operated without the pilot or observers having visual contact with it. This form of operation additional technological support in order to maintain a safe operation.

- Fully autonomous: This refers to a drone that is able to fly fully autonomously, without the intervention of a pilot;

2. IFR or VFR in Radio Line of Sight (RLOS) refers to the operation of drones in a non-segregated and/or controlled airspace in which manned aviation and air traffic control are present. This requires additional communication and autonomous systems such as detect and avoid and a positioning link. These operations can only be conducted in the BVLOS or fully autonomous modes; and

3. IFR or VFR Beyond Radio Line of Sight (BRLOS) refer to the operation of a drone when it is no longer be in direct radio contact with air traffic control. This requires more complex communication and autonomous systems. These operations can only be conducted in the BVLOS or fully autonomous modes.

Drone use - regulations

Drones operations are subject to several regulations; these regulations vary by the country or specific region in which an operation is being conducted. Essentially, drone regulations can be divided into two categories: those that exist and those that are in the process of being refined. Globally, existing regulations prioritize minimized risk, and they place little emphasis on the further integration of drones within aviation airspace (Rijksoverheid, 2017). Hence, it can be concluded that the existing regulations lead to many obstacles to the implementation of (AED) drone operations. These obstacles, such as a ban on flying BVLOS or significant restrictions on the transmission power of the C2 data link (Agentschap Telecom, 2018), will ensure that (AED) drone implementations are highly likely to remain unfeasible in the short term.

However, based on an evaluation of documents published by EASA (2017), EASA (2015) and JARUS (2017), there is also a global process of defining future regulations. This process is driven by the potential social importance of and benefits offered by drone operations. These future regulations will focus on ensuring safe drone operations that are characterized by minimal risk (EASA, 2017). According to the above-mentioned documents, these regulatory priorities mainly focus on risk mitigation. In other words, in order to prevent harm (i.e. damage to infrastructure or fatal damage to people) from occurring, risks should be mitigated to an acceptable level in order to ensure safe drone operations. Thus, given the social importance of and benefits offered by drones, certain risk levels are likely to be tolerated.

(19)

19

In order to be able to assess the risk of associated with drone operations, JARUS (2017) developed the specific operations risk assessment (SORA), which focuses on risk assessment specifically in the context of drone operations. The SORA is widely accepted by aviation authorities, as many regulatory authorities were involved in its development. As it has not been widely tested in practice, it remains only a proposed approach to risk assessment for drone operations. This is currently the most relevant document that focuses on the risks associated with of drones; therefore, this document forms the basis for the assessment of the risks associated with drone operations in this thesis (Chapter 5). The risk assessment is conducted with reference to the airspace and ground characteristics that affect drone operations. If a drone flies in controlled airspace, a higher risk will be assigned to its operation, as the chance of collision with other aircrafts will be higher. Risk mitigation approaches include technological mitigations and emergency procedures (JARUS, 2017). If risk is minimized to an acceptable level and the importance of and social benefits offered by drones are recognized, then it is likely that regulatory authorities will allow a certain drone operations.

These two approaches to regulating drones are presented in Table 2, below.

Characteristics

Current regulations - Focus on minimal risk

- No focus on further integration of drones with airspace - Social importance or benefits of drones are not recognised - Many regulatory obstacles

- Highly likely that AED drone operations will not be allowed Future regulations - Focus on mitigating risk to an acceptable level

- Based on risk assessment specific to drone operations - Social importance or benefits are recognised

- AED drone operation allowed under strict conditions

3.2.2 AED drones

Using a drone to transport an AED to the site of an OHCA is a concept that was pioneered by Alec Momont7 in 2014, when he created a prototype of an ambulance drone that enabled the rapid delivery of an AED to an OHCA location.

Currently, there are four main articles that address the use of AED drones based on this concept (Pulver et al., 2016;Van de Voorde et al., 2017; Cleasson et al., 2017; Cleasson et al., 2016). Van de Voorde et al. (2017) elaborate on Momont’s ambulance drone concept and call for future research into AED drones. The other studies mainly contribute to this concept by making unsupported decisions on the type of drone and region of choice and the selection of launch site locations. Finally, these other studies tested the performance of their prototype AED drone network. Launch sites are strategic locations within a certain region from where AED drones fly to OHCA scenes (Pulver et al., 2016). When choosing launch sites, one must take into account the required response time, the speed and endurance of the AED drone being used. Given these parameters, each launch site can only cover a certain area; therefore multiple launch sites may be necessary in order to cover an entire region. The test results, however, only indicate the potential of AED drones since these tests do not adequately represent reality; this is due to the fact that they neglect certain relevant factors, such as regulations and decisions concerning operational processes, network resources and the control structure, that will likely affect the performance of an AED drone network.

7_{https://www.tudelft.nl/en/ide/research/research-labs/applied-labs/ambulance-drone/}

(20)

20 3.3 Designing AED drone networks

There are two studies relevant to designing an AED drone network, that of Pulver et al. (2016) and Kim et al. (2017). Both studies focus on determining optimal launch site locations for drones through the use of deterministic models. Pulver et al. (2016) specifically concentrate on determining and measuring the performance of launch sites locations for AED drones. In their study, the authors make several decisions, including the use of a rotary-wing drone and a region of operation. However, the authors do not provide any support with regard to making these decisions. Building on those decisions, the authors created a model that can determine optimal launch site locations for responding to OHCAs by using a response time standard. This approach is based on the maximum coverage location problem model developed by Church and Revelle (1974). Finally, Pulver et al. (2016) evaluated their model using three scenarios:

- Scenario 1: Using existing EMS Stations as potential drone launch sites;

- Scenario 2: Using new locations and EMS stations as potential drone launch sites; and - Scenario 3: Using only new locations as potential drone launch sites.

The study conducted by Kim et al. (2017) also focuses on determining optimal launch site locations, but differs from that of Pulver et al. (2016) in its nature, as this study primarily addressed designing a drone network for delivering medicines to patients with chronic illnesses. Kim et al. (2017) developed two models, of which the strategic planning (SP) model is relevant to designing AED drone networks. Similarly to the model of Pulver et al. (2016), this model determines the optimal launch sites by building on several decisions without providing support with regard to how they should be made. The SP model differs from the model of Pulver et al. (2016) in that it also addresses investment costs.

3.4 Summary of findings

The literature has shown that AED responsiveness is low in rural regions, which thus affects a critical link in the chain of survival. Existing AED networks are therefore frequently not capable of delivering AEDs to OHCA sites within an acceptable timeframe. Such networks include publicly accessible AEDs, AEDs stored at volunteers’ homes or ambulances with a defibrillator function.

AED drones offer a potentially viable alternative or addition to existing AED networks. Studies have shown that AED drones are capable of overcoming the lengthy AED response times in rural areas by delivering an AED attached to them to the OHCA sites within a few minutes. However, current regulations create a number of obstacles that make it very likely that AED drone operations will not be permitted in the short term; however, opportunities are foreseen in the future. Building on these opportunities, the facilitation of AED drones in the OHCA pathway entails designing an AED drone network that consists of interdependent interactions between operational processes, network resources and a control structure. In addition, region-specific characteristics, such as landscapes and regulations, will affect decisions regarding the design parameters of such networks.

(21)

21 4. Characterization of AED drone networks

This chapter is intended to clarify the focus of the framework developed in this paper by characterizing the key elements of AED drone networks and identifying the key decisions to be made concerning each element. Section 4.1 provides a system overview. Thereafter, the main elements of AED drone networks are discussed. These elements are the operational processes (Section 4.2), the network resources (Section 4.3) and a control structure (Section 4.4). These elements are identified and elaborated upon with reference to the literature, the interviews with experts and documents from the field of drones and aviation.

4.1 System overview

Figure 8 depicts the system of an AED drone network. It is assumed that future regulations will apply globally, as stated in Section 3.2.1, since it is highly likely that AED drone operations will not be permitted under the current regulations.

The system input is the call for help that triggers the operation of an AED drone. As already mentioned previously, an AED drone network is divided into three elements: the operational processes, network resources and the control structure. The system output is the facilitation of an AED and the re-instalment of both an AED and an AED drone in order to be prepared to respond for the next call for help. Figure 8 identifies the section in which each element is discussed. Each network element consists of several sub-elements, such as the AED, launch sites and the ground control station for the network resources. Herein, each sub element invloves certain parameters that must be decided upon. Parameters are the relevant characteristics of a sub-element in which several alternatives, the so-called key

decisions, exist. For the launch sites, for example, it is for instance relevant to consider the potential

alternatives in terms of physical locations.

4.2 Operational processes

The operational processes are all of the procedures involved from recognition to the re-instalment of a drone at the launch site. Figure 9 depicts the main aspects of the operational processes of AED drones as retrieved from the literature and interviews. All processes are elaborated upon below the figure; in addition, the key decisions associated with all operational process parameters are listed in Table 3. For the operational processes, it is assumed that there is always a bystander present with the OHCA patient and that this bystander is capable of operating the AED. In addition, the AED drone is assumed

to

always be based at

its originating launch site. These assumptions are made as a result of the fact that existing alternative AED networks function similarly.

(22)

22 4.2.1 Recognition

In the recognition phase, the bystander witnesses a cardiac arrest and contacts the dispatch and control centre. Multiple contact options exist: The bystander can contact the dispatch and control centre by means of a telephone call or using an app that enables him or her to immediately initiate the AED facilitation process. A GPS tracker (in a smartphone) can prove helpful in locating an OHCA patient. A combination of the proposed options is also possible. (These options were identified during the interviews.)

4.2.2 Dispatching

The dispatching process essentially consists of completing a triage protocol and determining the location of the OHCA site. This enables the dispatch and control centre to assess whether or not an AED drone is required. When an AED drone is required, an initial check should be performed. This initial check will first determine if a flight is possible given factors such as the time of day (i.e. whether it is night or day), the location of the OHCA site and the availability of an AED drone. If the flight seems possible, the protocol will enable the operator to send a request to the AED drone pilot at the ground control station to initiate the AED drone flight. The key decisions in dispatching are those regarding the manner in which an AED will be delivered to the OHCA location. When completing the protocol, one can choose between multiple options for delivering an AED to an OHCA location. As mentioned in Section 2.3, an AED can be delivered by means of an EMS network, an AED drone network or volunteers; in addition, one could use a publicly accessible AED. It is necessary to determine which delivery approach will ensure that an OHCA patient receives timely AED treatment.

4.2.3 Preparation of flight

When the dispatch request is received by the ground control centre, the flight to the OHCA location must be prepared. First, the availability of the selected AED drone launch location should be checked in order to determine whether a flight is possible. When a flight is possible, the flight path and possible landing zones should be determined, taking into account factors such as weather conditions and landscape. Based on the findings of the interviews, this can either be done manually, using monitoring systems, or automatically, using a traffic management system; alternatively, a combination of both approaches can be used.

4.2.4 Activation of drone

Once preparations have been completed, the AED drone should be activated in order to allow it to take-off. Based on the findings of the interviews, this can either be done automatically or manually by means of a launch system at the launch site. Such a launch system is further described in Section 4.3 under ´Launch sites´.

(23)

23 4.2.5 Flight

The drone will fly at VLL altitudes and at a minimum safe altitude; this is due to the fact that flying above VLL altitude result in unnecessary high risks, as higher airspaces will have greater density of aerial vehicles. A high density of aircraft at this altitude can be found as a result of the activities conducted at airports, military bases and other prohibited zones (a detailed overview is provided in appendix I). Furthermore, flying at VLL altitudes ensures an efficient flight since vertical movements will be minimal. In addition, the AED drone should fly at a certain minimum safe altitude in order to overfly obstacles such as mountains or skyscrapers; alternatively, it should fly at a minimum altitude while avoiding such objects.

The flight will only be performed BVLOS. When flying with VLOS, the pilot is at a relatively short distance from the drone itself, as discussed in Section 3.2.1. Since this distance would allow the pilot to deliver the AED without the use of a drone in approximately the same period of time, VLOS is not considered to be a feasible option. No fully autonomous flight is considered, as aviation law state that at least a commander should always be present during the flight process (Wet Luchtvaart, 2017). Furthermore, a pilot is also required due to the fact that potential landing zones can vary and must be determined. In addition, given concerns regarding social acceptance and responsibility, a pilot acting at least as a commander would be most suitable. Therefore, the operation is assumed to be performed BVLOS. Thus, the AED drone will partly operate automatically and partly be remotely controlled by the pilot, as the drone is not visible to the pilot, who will be unable to optimally monitor its flight given observer delay and limits in terms of camera and audio feed. However, the pilot is always capable of intervening in a flight by performing certain flying activities or by assisting bystanders.

A flight can take the form of a straight line (Euclidian); alternatively, it may not take the form of a straight line or only do so partially (no-Euclidian). Furthermore, a flight can vary in terms of altitude and speed. The AED drone will be able to fly anywhere within its geofenced area. Geofencing is the demarcation of a certain geographical area in both horizontal and vertical aspects; it is done with the intention of identifying specific risk areas, such as airports or military zones (EASA, 2017).

4.2.6 Delivery of AED

The AED response process entails delivering an AED to the bystander. There are three main delivery options. Van de Voorde et al. (2017) have shown that delivery by landing is possible, while Cleasson et al. (2016) demonstrated that it is possible to deliver an AED to the ground using a line. The last option is the delivery of an AED by means of a parachute, as practised, for example, by Zipline8. Volunteers may possibly be able to provide support in this process by clearing the landing zone or retrieving the AED from the drone safely.

4.2.7 Return of drone and AED

This stage of the process involves the return of the AED drone to its original launch site and the AED to its designated location; these two actions can either be performed simultaneously or separately. The return of the AED can be achieved by a means of a return flight, alternative transportation methods or by an ambulance.

4.2.8 Re-installing drone and AED

This process involves the reinstallation of the drone and AED at the originating or designated launch site. The drone should be recharged and prepared for the next flight, while the used AED should be prepared (i.e. medically and physically checked) for the next flight. Since this can take some time (the defibrillation pads, for example, should be replaced), another AED should be load onto the drone. (These factors were identified during the interviews.)

(24)

24 4.3 Network resources

Network resources refer to the physical resources that support the operational processes mentioned in the previous section. For these operational processes, the physical resources are the AED drone, the AED itself, the launch sites, the ground control station and staff.

4.3.1 AED drone

For the AED drone, certain parameters should be taken into consideration. Parameters refer to the design and system components of the drone itself. The most relevant parameters with regard to its key decisions are explained below and presented in Table 4. Some parameters can be integrated in either a minimal, basic or advanced fashion. The essential difference between these three integration levels is the manner in which the risks associated with emergencies are reduced to an acceptable level. Minimal integration only addresses the most relevant risks, while basic integration addresses the most common risks, i.e., certain complex situations that may not have been prepared for. Advanced integration addresses all potential risks; the possibility of experiencing an emergency is thus negligible.

1. Type of drone: This entails the types that were discussed in Section 3.2.1;

2. Basic design: This refers to the design of the entire drone, including the following factors:

Parameters Key decisions

1. Recognition 1. By call

2. By app 3. GPS

4. Combination

2. Dispatching 1. Send EMS network

2. Send AED drone

3. Activate volunteer network 4. Communicate public AED location 5. Combination

3. Preparation of flight 1. Manually, with aid of monitoring systems 2. Automatically determined by an air traffic system 3. Combination

4. Activation of drone 1. Automatically launched from ground control station 2. Manually launched from launch site by ground control

station at launch site

3. Manually launched by staff at launch site

5. Flight 1. Euclidian distance

2. Non-Euclidian distance 3. Altitude (variable) 4. Speed (variable) 6. Delivery of AED 1. By parachute

2. By line 3. By landing 4. Volunteer support 7. Return of drone 1. By ambulance

2. By alternative transportation method 3. By return flight

7. Return of AED 1. By ambulance

2. By external transportation 3. By return flight

8. Re-installing drone at launch site 1. Same AED drone manually 2. Same AED drone automatically 3. Other prepared AED drone 8. Re-installing AED on AED drone 1. Manual placement of new AED

(25)

25

- Industrial standards refer to the requirement that a drone be constructed of robust material; such a material should allow the drone to withstand stresses, should have a low flammability and should minimize the risk of hazards associated with the drone’s electrical properties (EASA, 2016). Exact industrial standards can vary by country, and multiple options are thus possible. - Safety for users refers to the safety of the design in terms of the users involved in OHCA

situations. Van de Voorde et al. (2017) note that a drone could have retractable rotor blades. Interviewees proposed rotor blade protectors in order to increase the safety of a drone’s design; - A weather resistant design is one that can resist certain weather conditions. Multiple options and

techniques are possible. However, no main options can be identified based on literature and interviews; and

- The payload-bay is the component in which a certain payload can be carried. A payload-bay can be integrated within the basic design of a drone or can be an external component (Van de Voorde et al., 2017; Cleasson et al. 2016);

3. Payload-bay design: This refers to the design of the payload-bay, including the following factors: - The release system is the system that facilitates the delivery of the AED to the bystander. The

AED can either be manually retrieved from the compartment in which it is stored or delivered by a line or parachute that delivers it to the ground (Van de Voorde et al., 2017; Cleasson et al. 2016).

- Situational awareness systems are devices that enable the visual and/or auditory monitoring of a flight. According to the interviewees, an AED drone can be fitted with audio and video devices that allow one- or two-way interaction with an area;

4. Type of energy source: This refers to the many possible energy sources that can power AED drones. According to the interviews, lithium polymer (Li-Po) batteries, hydrogen and fossil fuel are the three most relevant options.

5. Navigation systems: These positioning systems are used to determine a drone’s location and assist in path planning. According to Hassanalian and Abdelkefi (2017), GPS and INS are the most relevant options, but also other possibilities exist;

6. Automation systems: These are autopilot systems that control a drone automatically; however, commands will still be issued by a pilot. These systems can be integrated in a minimal, basic or advanced fashion;

7. Autonomy systems: these are systems that can decide themselves if a certain action should be performed, as mentioned in Section 3.2.1. These include a detect and avoid system that autonomously detects objects in flight and avoids them in a flight and an autopilot system that autonomously flies a drone to determined locations (EASA, 2017). These systems can also be integrated in a minimal, basic or advanced fashion;

8. Communication systems: These are systems that ensure interaction between pilot and drone; they consist of the following:

- A C2 data link, as mentioned in 3.2.1. According to the interviewees, a telecommunication network, radio signal, satellite, combination, other systems or a combination thereof can be used as C2 data link; and

(26)

26

of all of a drone’s components; once again, such systems can be integrated in a minimal, basic or advanced fashion. Furthermore, once can choose whether or not to integrate such systems; and 9. Emergency equipment: This refers to equipment that can be used to mitigate risks in certain

situations. According to the SORA (2017), there are two emergency systems: parachute/crash impact mitigation and a flight termination system that aborts flights in emergencies. These systems can be integrated in a minimal, basic or advanced fashion. However, certain systems cannot be integrated into an AED drone.

Parameter Key decisions

Type of drone: Wing type 1. Fixed-wing 2. Rotary-wing 3. Tilt-wing/rotor Basic design:

Industrial standards Multiple options possible Safety for users 1. Retractable rotor blades

2. Rotor blade protectors 3. Other

Weather resistant Multiple options possible Payload-bay 1. Integration with basic design

2. External component Payload-bay design:

Release system 1. Compartment

2. Line or parachute Situational awareness systems 1. Video/audio camera

2. Speaker device Type of energy source:

Energy source 1. Li-Po battery 2. Hydrogen 3. Fossil fuel Navigation systems: Positioning system 1. GPS driven

2. Inertial navigation system driven 3. Other Automation systems: Autopilot 1. Minimal 2. Basic 3. Advanced Autonomous systems:

Detect and avoid 1. Minimal 2. Basic 3. Advanced Autopilot

Communication systems:

C2 data link 1. Telecommunication network 2. Radio signal

3. Satellite

4. Combination/other Location tracking system 1. No integration

2. Minimal 3. Basic 4. Advanced Health-monitoring system

Emergency equipment:

Parachute/crash impact mitigation 1. No integration 2. Minimal 3. Basic 4. Advanced Flight termination system

A framework for designing AED drone networks