Unmanned Aerial System Operations for Retail
Guido Manfredi, Elgiz Baskaya, Jim Sharples, Yannick Jestin
To cite this version:
Guido Manfredi, Elgiz Baskaya, Jim Sharples, Yannick Jestin. Unmanned Aerial System Operations
for Retail. ICAS 2018 : 14th International Conference on Autonomic and Autonomous Systems,
IASFR: Intelligent Autonomous Systems for the Future of Retail, May 2018, Nice, France. pp.ISBN:
978-1-61208-634-7. hal-01859054
HAL Id: hal-01859054
https://enac.hal.science/hal-01859054
Submitted on 23 Aug 2018
HAL is a multi-disciplinary open access
archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est
destinée au dépôt et à la diffusion de documents
scientifiques de niveau recherche, publiés ou non,
émanant des établissements d’enseignement et de
recherche français ou étrangers, des laboratoires
publics ou privés.
Unmanned Aerial System Operations for Retail
Guido Manfredi, Elgiz Baskaya, Jim Sharples and Yannick Jestin
ENGIE Ineo - Groupe ADP - SAFRAN RPAS Chair
Universite de Toulouse, Ecole Nationale d’Aviation Civile
Toulouse, France
Email: guido.manfredi@enac.fr
Abstract—The number of Unmanned Aerial System (UAS) applications is quickly increasing as technology, standards and
regulation allow them. With each new application, more industrial
sectors get affected, and the retail sector is already being
impacted. This paper presents five UAS applications that will
impact the retail sector: freight, monitoring, guiding, delivery, and
advertisement. For each application, concepts of operation are
provided along with the associated technological, standard and
regulatory locks. These operations are then organized along time,
from earliest to latest accessible, with accompanying explanation
as to why and when. It is shown that the applications currently
most publicized are not the ones that will come first. Finally, a
discussion regarding the accuracy of our forecast is proposed and
leads to support the enabling of drones, in the retail sector, are
provided.
Keywords–UAS; RPAS; drones; retail; mass distribution; integration; drone operations.
I.
I NTRODUCTION
The retail sector has been quickly evolving in the last
decades, as it aptly integrated emerging technologies. The
advent of the Internet enabled big data and machine learning approaches, e.g., recommender systems based on customers browsing history to target advertisement [1]; automation changed the way general merchandise stores worked with
self-checkout machines and automated drive supermarkets [2].
Smart sensors changed the way online retail stores present
their goods, with the development of virtual dressing rooms
[3]. Even the most complex technologies have been integrated
in the retail sector, with warehouse robots allowing fast and
efficient content management [4] and robotic guides improving
the customers’ experience [5]. The successive technologies
transformed the industry in many ways by changing how
goods are stored and managed, reducing the number of intermediates, automating dull and dirty tasks, and changing the
customers’ habits and experience. Now, a new technology is
about to become fully available: Unmanned Aerial Systems
(UAS). A well-known example is the Amazon prototype [6]
for autonomous package delivery. However, there are lesser
known applications, which could have more impact on the
retail industry and sooner than one might think.
This paper studies five applications in which UAS operations will impact the retail sector. For each application, different concepts of operation are considered. For each operation
type, the key enablers are listed. Exemples of such operations,
and ongoing research on the topic, are also provided. As the
reader may not be familiar with UASs, Section II introduces
their architectures, their environment, and their regulation.
With this information in mind, Section III presents the five applications along with the technologies and regulation required
to enable them. Tentative dates at which these applications
will become possible are provided in Section IV. Finally, the
conclusion provides future leads to explore in order to enable
UAS operations for retail.
II. T HE UAS WORLD
There are numerous words to designate drones, each with
its own meaningful nuance. For the sake of clarity, in the rest
of this paper, we use indifferently the terms UAS and drones,
with the definitions provided hereafter.
A. UAS Architecture
The S in UAS stands for ”System”, a UAS is composed
of three elements: an Unmanned Aerial Vehicle (UAV), a
Command and Control link (C2 link), and a Ground Control
Station (GCS). We provide some general information regarding
each one of these elements, as it will be useful to understand
the limitations in different applications.
1) UAV: The UAV is the aerial part of the UAS. There
are numerous ways to categorize UAVs. For clarity we will
consider that they can belong to one of three families: rotorcraft, fixed wind, and airship. The most represented type
of UAV is the rotorcraft. It does not have wings and entirely
relies on the thrust from its propellers to fly. Easy to pilot,
especially for take-off and landing, it allows the pilot precise
positioning. Its drawback is the energy needed to remain aloft,
as well as the noise and wind created by the propulsion. With
low endurance, it remains the most popular form of UAV. The
fixed wing UAV flies thanks to the lift produced by the flow
of air above its wings, like most manned airplanes. For this
reason, it must remain inside a given flight envelope and is thus
less maneuverable than rotorcrafts. As the lift comes from the
air reaction and not purely from the propellers’ thrust, it has
high endurance and can fly for long periods of time. Plus, it
can carry high payload relative to its own weight. It needs
an initial speed for takeoff, and some distance for landing,
though it is possible to design Vertical Take Off and Landing
(VTOL) fixed wing aircraft, greatly simplifying the beginning
and end of flight. Its aerodynamic shape allows it to reach high
speeds. For a airship type UAV, the majority of its lift comes
from a lighter-than-air gas inside its structure allowing it to
stay aloft with energy spent only for motion. Consequently, it
has the highest endurance of the three types. However, having
to store gas onboard gives it a bulky shape which reduces
the maximum speed/maneuverability and makes it sensitive
to external perturbations (e.g., wind). The UAV also carries
the equipage, i.e., the onboard equipment needed for the UAS
operation (e.g., flight, rules compliance) but not necessary to
perform its assigned task. The equipage comprises the sensors
TABLE I. P ERFORMANCES OF DIFFERENT UAV TYPES .
UAV type
Endurance
Speed
Manoeuvrability
Payload
Cost
Rotorcraft
Low
Medium
Very high
Medium
Medium
Fixed-wing
High
High
Medium
High
Very low
Airship
Very high
Low
Low
Low
Low
used to navigate (e.g., GPS, cameras), the communication
systems, and the equipment required by regulation (e.g., lights,
sound). A brief summary of the differences between these
types of aircraft is provided in Table I.
2) C2 link: The C2 link is the medium used to uplink/downlink Command and Control (C2) data to/from the
UAV. From an operation point of view, C2 links can belong to
one of three categories: direct link, terrestrial network, satellite
network. The direct link allows short range (few kilometers)
control over a UAV. It has the benefit of not relying on a service
provider. For long range operations, the C2 link can be carried
over a terrestrial network. Due to the cost of such installations,
a service provider is required. The delay in such type of
communication is low enough to be comparable to direct link.
Similarly, satellite networks allow very long range operations
and operations where no ground network is available (e.g.,
over mountains, water). But it has longer delay than direct or
terrestrial networks. This type of C2 link may not be useable
for certain types of operations, e.g., indoor operations. Again,
due to the cost of such structures, it requires a service provider.
3) Ground Control Station: The GCS can be a simple
remote control, a cockpit sized control room, or even bigger.
Its size and complexity is related to the type of UAV and the
type of operation. However, as the GCS is not considered a
limiting factor, we will not take into account the different types
of GCS in the rest of this paper.
The capabilities of a UAS change greatly depending on
its type and equipage. These need to be adapted to the type
of operation to ensure a safe flight while allowing the UAS to
perform its mission. Note that, for a same UAS, the regulatory
constraints vary depending on the environment and type of
operation.
B. UAS Environment
The environment of an aircraft can be classified based on
three elements: the available services, the systems required on
the aircraft and the applicable procedures. Services encompass all external support to the operation, like aeronautical
meteorology or conflicts management. System requirements
designate onboard equipment needed to interact with services,
other aircraft, and to ensure proper following of procedures.
Procedures are pre-defined behaviors that an aircraft needs to
follow in certain airspaces or situations. When considering
manned aviation, these three elements form the Air Traffic
Management (ATM) system; these same elements, when specific to drone operation, form the U-Space system.
1) ATM: The ATM has a strong safety record due to its
numerous services, high system requirements and numerous
procedures. As a consequence, integrating in such an environment is complex. Plus, the fact that manned aviation is
currently flying in these environments asks for drones to be
integrated seamlessly, i.e., a drone should be able to behave
like a manned aircraft in every aspect. Drones capable of
integrating in the ATM system are likely to be large ones,
with performances close to these of existing manned aviation.
2) U-Space: The notion of U-Space encompasses the services and procedures offered to drones to allow their integration. The U-Space is planned to be rolled out in four phases:
U1, U2, U3, and U4 [7]. Phase U1will bring e-registration, eidentification, geofencing (don’t go in this area) and geocaging
(don’t leave this area). Phase U2 will set up management of
drone operations through flight planning, flight approval, tracking, airspace dynamic information, and procedural interfaces
with ATM. In Phase U3, operations in dense areas, capacity
management, assistance for conflict detection will be rolled
out. Finally Phase U4 will provide integrated services with
manned aviation, high levels of automation for all services.
When reaching U4, all types of operations, even the most
complex ones, will be supported by the U-Space services and
procedures. U-Space will provide traffic management for small
UAS but also drone specific services for large UAS.
Depending on the type of operation, the drones will have
to integrate with: neither, one, or both of these environments.
However, the complexity of integrating a drone not only comes
from the environment but also from the risk of the operation,
as will be explained in the next section.
C. UAS Regulation
In order to provide an unified answer to the problem of
regulating UASs, national regulators and industries gathered to
create the Joint Authority for Rulemaking on Unmanned Systems (JARUS). This entity, building on the European Aviation
Safety Agency (EASA) rules, proposed to split UAS operations
based on risk levels. The risk is mainly defined in terms of
probabilities to hit a human on the ground, another airspace
user or a critical infrastructure. This led to the definition of
three operations categories (A, B and C) described below.
1) Category A: Category A encompasses low risk operations, like most of leisure flights and some professional
activities. Operations falling in this category do not require an
explicit authorization from civil aviation authorities. But, they
are subject to strict operational limitations (e.g., no proximity
to people, traffic, infrastructures; no dangerous items; one pilot
per UAS; no item dropping). These operational limitations are
sufficient to mitigate the low risk. Though some professional
activities fit in this category, the main goal is to regulate leisure
types of operations.
2) Category B: Category B regroups medium risk operations, such as operation beyond visual line of sight (i.e.,
no visual contact between pilot and UAV during flight). To
facilitate the regulation process, a set of scenarios with specific operational limitations are designed. To operate within a
scenario, the UAS needs to comply with a list of requirements.
For operations outside the scope of the scenarios, a risk
analysis must be carried out to show that the existing risks
are properly mitigated. To facilitate this risk analysis process,
JARUS developed a framework called the Specific Operations
Risk Assessment (SORA). The SORA considers threats, which
contribute to the risk, and barriers, which mitigate the risk,
to evaluate the actual risk of an operation and decide if the
resulting mitigated risk is low enough to allow the operation.
The risk analysis needs to be validated by the authorities in
order to authorize an operation not included in the standard
scenarios.
3) Category C: Operations with risk that cannot be mitigated in Category B are evaluated in Category C. These
represent high risk operations, for example large cargo delivery
in urban areas. These are likely to be operations with a
risk close to current manned aviation’s one. As such, the
aircraft, avionics, pilot/crew and operator will need to be
certified in order to fly these operations. The fact that a
certification process is involved increases the complexity of
the introduction of UAS in these types of operations. Indeed,
before being able to certify a piece of equipment, a standard
must be developed for this equipment. Standardization can be
understood as the process of defining details and minimum
requirements for safe and uniform operation across a diverse
range of implementations. However, as of today, not all the
parts required to fly a UAS have corresponding standards, e.g.,
Detect And Avoid (DAA) systems, C2 links, pilot training. So,
enabling this category of operation requires extra effort and
time for the industry to agree on standards. For Category C,
harmonization at the International Civil Aviation Organization
(ICAO) is planned to allow international operations.
4) The special case of indoors operations: Indoor operations are not regulated by civil aviation authorities, so the
above categories do not apply. In fact, indoor operations fall
under the responsibility of the operator and usually no risk
analysis or certification is required by the authorities.
With a knowledge of the existing technologies, environments, and regulations, next section introduces UAS operations
that will impact the retail sector in the near to long term.
III. UAS A PPLICATIONS FOR R ETAIL
In the following, we consider five domains where the
UASs will impact the retail industry and examine the prerequisites for integration, as well as the obstacles faced on the
technological, standardization and regulatory levels.
A. Freight
Freight operations involve the transportation of large quantities of goods, which represents the input of most retail stores.
Depending on the way suppliers operate, opportunities for the
retail sector can change. Thus, an evolution of the freight
can have an impact on the retail sectors. In the following we
analyze two freight operations: cargo flights transporting goods
on long distances and cargo delivery to urban areas.
1) LUCA: Large Unmanned Cargo Aircraft (LUCA) offer
economic opportunities, as demonstrated by Collins [8], especially when it comes to long haul point to point, delivery
to/from remote places and high value cargo over distances
requiring multiple piloting crews. Removing the cockpit will
cut costs related to having a crew onboard and the associated
maintenance. Plus, the UAV will be freed of dynamical constraints related to the presence of humans onboard (e.g., limited
acceleration, turning angles, braking force). This kind of
operation will ask for integration in the ATM system. Aircraft
performing these missions will fly along pre-determined routes,
will have to access airport environments, and to follow specific
ATM procedures. Because it will fly in the middle of the traffic,
this type of operation is classified as high risk, Category C,
and will ask for certification. From a technical point of view,
transforming large manned aircraft into UAVs is possible with
existing technology; the UAV having the same capabilities as
existing aircraft, it will integrate seamlessly with the ATM
environment. But it will require to agree on some standards.
Notably for the C2 link, DAA and Automated Take Off and
Landing (ATOL) equipment. Indeed, long range operations ask
for a terrestrial or satellite communication network, collision
avoidance is a requirement to fly with the rest of the traffic,
and operating in airport environments remotely requires ATOL
and auto-taxi (from gate to runway) capabilities. The rest of the
equipment remains the same so no additional standardization
effort is required. Once these hurdles are removed, this type
of operation will be at hand. Using LUCA could even open
new opportunities. Indeed, in the context of the SESAR2020
projects [9], experiments have been carried out to determine
drone specific trajectories which could facilitate their landing
on airports without disturbing the existing traffic. It has been
determined that UAVs can use steeper descent angles and
sharper turning angles. Added to the fact that there is no human
onboard, so less airport services are required, this could open
access to more options in terms of airports and available supply
routes.
2) Urban freight: Urban is currently defined in some civil
aviation instances as an area with a density of population
higher than 1295 people per square kilometer (plus a 0.5NM
buffer around it). Nowadays, delivery of goods to urban outlets
is mainly done with trucks. This brings problems related to
traffic, and increases the cities’ complexity. Using air transport
can simplify the way deliveries are carried out. It would still
require some urban planning for the landing sites and air
routes, but could remove the impact and dependency on road
traffic. Moreover, little infrastructure would be required for
operations contrary to trucks, which require at least roads,
facilitating delivery to peculiar places (e.g., islands, old cities).
Comparatively to what can be done today with helicopters,
using a UAV would allow saving space and weight by removing the cockpit. Plus, today’s helicopters operations over
urban areas are mostly exceptional. Developing an appropriate
regulation for UAS will allow them to operate routinely. The
urban freight operation is complex due to the fact that the UAV
is likely to have to go both into integrated airspace (cruise)
and low level (take-off and landing), so these UAVs must both
integrate with ATM and U-Space. In both cases, flying above
urban areas will be classified as a category C operation. A
benefit of flying above urban areas is that manned aviation is
not allowed near buildings, simplifying operations in terms
of DAA for traffic. However, the density of infrastructures
requires a DAA for fixed obstacles. Plus, these operations are
likely to use stable flight routes, with deconfliction services
required considering the operational risk. This asks for USpace services from phase U3. Due to the size of these UAVs,
and to reduce nuisance to the population, it is likely that
some areas will be forbidden. Because of the large payloads,
delivery from the air or dropping seems unlikely, so the freight
UAV will have to land on some properly defined spaces. The
limited amount of space will ask for rotorcrafts and/or possibly
VTOL UAVs. Urban areas being well covered in terms of
communication networks, ground communication will be the
preferred solution.
Enabling these operations will take time, as there is no
equivalent in manned aviation, so there is little experience.
Having a UAV transition from integrated airspace to low level
will ask for equipage to deal with both environments. In
integrated airspace, a DAA for traffic will be required, while in
urban environment (take-off and landing) a DAA for obstacles
will be needed. Developing both of these pieces of equipment
for medium sized aircraft is a significant challenge and will ask
for research to be done, especially on sensor miniaturization.
The sensors will also be crucial for the localization part,
as some urban areas can be GPS denied, asking for an
independent localization mean. From a procedure point of
view, insertion of a relatively slow UAV in the traffic will
be challenging. At lower levels, dynamical flight planning
will be needed to adapt to a future busy airspace around
cities. Procedure for helipad use will also need to be defined,
though this could benefit from regulations of cities with high
helicopter traffic. A solution to the miniaturization problem
can come from the automotive industry. Indeed, hardware and
software developed for autonomous cars could answer the
requirements for low level flight. The envisioned UAV being
on a scale similar to cars, the Size Weight and Power (SWaP)
requirements could match. From a procedure perspective, this
type of operation will benefit from the advances brought by the
current exploratory research projects on U-Space (see current
SESAR2020 ERC projects). Alternatively, in the future, a
drone could stay aloft and be used as a warehouse while being
supplied by ground-to-air freight: Amazons floating warehouse
[10].The idea is to have a medium altitude floating warehouse
to which a swarm of drones have access. This might serve
as a fast delivery platform, assuming the regulatory hurdles
behind would be solved in the future. The company patented
the design, with details about its fuel efficiency solutions for
the route from airship to the delivery destination.
Freight related flights are likely to be among the first high
impact operations to be enabled since they already exist and
only require modifying the transport medium (from aircraft to
UAV). Plus, they don’t involve the transportation of humans,
greatly reducing the risk. However, operations in urban areas
will still ask for complex procedures and systems to be set up
before they can take place.
B. Warehouses, Distribution Centers and Outlets Monitoring
Monitoring in those large buildings offers some challenges
which can be eased with the help of drones. A wide range
of applications can be envisioned like security protocol improvement, customer habits studies, process improvement, and
inventorying [11]. Todays solutions can be expensive and lack
efficiency. For example, inventory management of distribution
centers is most of the time done by human workers via handheld devices scanning through stocked pallets of goods in
labyrinth like aisles. UASs abilities to move quickly and in
3-D will facilitate these tasks with an automated, flexible and
cost-efficient solution. Fleets of UAVs going around a building
with appropriate sensors can reduce the number of required
sensors as well as their cost (closer range, so less costly).
The autonomy level required for such tasks is to be able to
navigate and avoid collisions with static obstacles and humans
(assuming there are no other flying object). Automation will
allow the routes to be changed dynamically by the operator
providing flexibility and usage of a same fleet for different
tasks. The fleet will operate for long periods of times with
automated UAV replacement and charging. In this section, we
limit ourselves to indoor operations. These applications don’t
ask for high speed, large payload or maneuverability, the main
requirement will be endurance, plus it is indoors operations
in a known environment subject to little perturbations. For
such tasks, airships are the best pick as they provide high
endurance and stability for onboard sensors. The two main
tasks of navigation and collision avoidance ask for a precise
localization system, which will most likely be based on Simultaneous Localization and Mapping (SLAM) methods which
have been extensively studied. Existing sensors, even lowcost ones, like monocular cameras, would allow to perform
this task. As there is no interaction with traffic, ATM, or
U-Space, regulatory constraints will be low. The two main
obstacles for this application are the high level of automation
and the operation above the public, which asks for Category B
regulatory requirements. Though it will not be the case for all
applications, if the UASs operate above the public, they will
require mitigation means to prevent any incident developing
into injuries on ground. Again, airship provide some de-facto
mitigation means because they are lighter than air. For the level
of automation, operations with a pilot per UAV, or even one
pilot for multiple UAV, does not seem like an economically
viable solution. These applications will most likely ask for
supervision only from one human who could perform other
tasks.
Though not existing yet on the market, all the pieces for
these applications are on the shelves. There are no strong
regulatory or technological barriers to the introduction of UAS
in this context.
C. Improving Customer Experience
According to 2018 retail predictions, the organizations have
to review their structure to set the experience of the customers
as their priority [12]. With the data based methods, such as
machine learning, offering powerful solutions for customer
data, it is easy to have a direct link to the habits of individuals
which would help to offer preferable solutions to customers.
This customer-centric approach would be likely to involve
interaction between the customer and the organization both
online and in stores, where a drone can be a part of this link.
While the concerns of the public about drones is already on
the rise before they have started to populate the sky, maybe
indoors applications can serve as a warming environment to
start with human-machine interactions. A scenario will be to
have drone guides that will show you the path to the item of
preference, saving time during the supermarket item searches.
Their ability to move in 3-D already will allow them to point
items in higher shelves where it might not be easy with other
solutions. They can even calculate the best route to follow if
a list of items is provided. They can offer alternatives for the
missing items, shuffling through the database. Another possible
application will be drones for assisting with car parking in
malls. For big cities, finding a spot in a mall parking might
be quite challenging. Multiple drones can work in harmony
to guide the customers to the closest parking lot, and win
the hearts of the public in time. Yet another application of
drones might be to merge with virtual goggles, displaying the
stream through a camera on the drone to wander in shops
to see the items or environment without the need to actually
be there. For example, this would help customers choose the
less crowded time to go shopping, naturally improving the
customer flow. In this type of application, two operations
have to be distinguished: indoors and outdoors. As mentioned
earlier, indoors operations are not concerned with civil aviation
rules. However, contact with the public asks for risk mitigation
measures. To reduce safety concerns, one solution could be to
use very small drones (<800g). A rotorcraft or airship of this
size would have sufficient autonomy to guide a customer while
being small enough not to be a hazard. Since they will be
flying in large aisles, avoiding obstacles is not a concern and
avoiding the other drones will be the main challenge. Now, for
the outdoor case, things get complicated. Mitigation measures
to ensure that drones won’t crash onto cars will be required.
Indeed, a failure to the drone could lead to a car crash. The
resulting risk is likely to classify this application in Category
C. Moreover, at least U1 U-Space services will be required to
ensure tracking and geocaging of the drones in the mall area.
In both cases, the fleet management and autonomy aspects
will be the main concerns, as they will determine the overall
efficiency of the system. Human-machine interaction will also
be crucial to allow easy communication between drones and
customers. As far as guiding humans with drones is concerned,
feasibility studies showed that drones can be used to guide
blind runners [13]. Similarly, prototypes of flying street lights
have been designed to guide people through cities[14]. Still,
guiding simultaneously a group of persons is a challenge yet
to be undertaken. For guiding cars, manufacturers like ford are
currently exploring the solution for a wider application [15];
guiding cars through a parking could be a first step, yet a
challenging one.
When it comes to customer experience, customers are used
to have what they need, as they need it, at the time they want,
most probably as fast as possible thanks to interactions with
computers, internet and mobile networks. Thus, the companies
are in a rush to keep up with the race of expectations, otherwise
they face to lose the customers quite fast as well. Faster
deliveries is of importance and how it might be achieved is
described in the following section.
D. Delivery
As mentioned earlier, the advent of internet shopping has
profoundly altered the logistics of retail. According to the BBC
[16], parcel volumes surged almost 50% globally between
2014-2016 and they are on track to increase at rates of 17-28%
annually up to 2021. This has bolstered the need to rethink
logistics, particularly in the delivery part. Currently, delivery
is mostly done with delivery vans and trucks. Because of the
sheer volume of parcels, this is causing problems in big cities.
According to [17], in Paris, one moving vehicle in five is a
delivery vehicle, generating 25% of urban CO2 emissions.
Last mile delivery is also costly: it represents 20% of the
cost of the whole delivery chain [18]. This is where drones
may come in handy. With their low cost, automation, high
speed and flexibility, they could become the future of delivery.
This section focuses on the last leg of the journey, more
specifically on the last mile delivery, between the warehouse
and the customer or a pick-up point. This application can be
declined in three operational scenarios: delivery to a remote
location, to an urban pickup point, to a house or person in an
urban area. Each of these scenarios has its own constraints
that will be detailed in the following. For all applications,
the main technological capacities needed are endurance, automation, maneuverability, payload carrying ability and DAA.
The most important aspect is ensuring the integrity of the
data used to automate the flight. If this data is corrupted then
the flight might be jeopardized. This means several onboard
sensors must be used and fused. Global Navigation Satellite
System (GNSS) is to be avoided in the cities (too imprecise).
And malicious corruption of the information source must be
mitigated. Parcel delivery operations will likely fall within
Category B. If parcel delivery falls outside of the set of preestablished specific scenarios, a risk analysis must be carried
out and an authorization must be requested. Since delivery
drones might be numerous, and certainly will not be the only
drones in the urban sky, services from U-Space will need to
be at least in phase U2.
1) To a remote location: This is the simplest scenario,
flying to a remote location avoiding urban areas greatly reduces
the risk to humans. This type of long range mission is best
suited to fixed wing or VTOL UAV, with a connection through
terrestrial or satellite network. Depending on the remote place
to reach, flight can be performed at low levels, taking advantage of U-Space services, or at medium altitude which would
require interactions with ATM, thus adding requirements on
the operation. Assuming that the flight goes through low
density air traffic, because of the remote location, constraints
on DAA and deconfliction services are low. However, such
long-range operations are likely to ask for a stable flight plan.
Thus, a U-Space in phase U2 would be sufficient for this type
of operation.
2) To a pickup point: Flying to a pickup point has various
benefits. Firstly, the objective is fixed so a stable flight plan
can be used. Secondly, a landing facility will be available at
the pickup point. However, due to these operations being in
urban areas, the landing facility is likely to be small, like a
helipad, thus VTOL and rotorcraft will be the preferred UAVs.
This type of operation will have to be integrated in the middle
of a complex airspace above cities, asking for a U-Space in
phase U2 to U3. Protection areas around cities will prevent
interactions with manned aviation, so no interaction with ATM
would be required in normal operations.
3) To a house/person: This is the most complex case, as
each delivery has a different destination which can potentially
move (moving person as target), asking for maneuverable
aircraft like rotorcraft. From a procedural perspective, this
asks for services capable of dynamic flight planning, dynamic
geofencing and deconfliction to ensure that the drone is able to
reach the person/house. Such a high level of dynamic airspace
management asks for a U-Space in phase 4. On top of the
equipage required to be integrated in the U-Space, the UAV
will need to carry identification means to ensure that the parcel
is delivered to the correct customer, increasing the payload. It
is still unclear how delivery could be done: dropping the parcel
would be dangerous for the customer, and the parcel; attaching
it to a rope would jeopardize the flight and be dangerous to
the population below; landing for the delivery would expose
the drone to hazards like animals, or malicious humans.
As a summary, delivery drones must be agile, enduring,
and loaded with sensors; this means that the maximum parcel
weight they can carry might be reduced. This also means that
the number of drones delivering parcels will be high; and since
other drones will be flying for other missions (police, etc.)
a high level of airspace management will be required, thus
delaying the integration of these operations.
body for ATM research and development. Lack of standardization is addressed by EUROCAE, the main aeronautical
standardisation body in Europe, especially the working group
WG-105 dedicated to UAS. In terms of regulation, the EASA
is developing a regulation for drone operations.
In terms of obstacles, it appears from the analysis of the
previous operations that the biggest obstacles when introducing
UAS are likely to be:
Figure 1. The European Union timelines for integration of drones with ATM
(green), and with U-Space (orange).
•
•
E. Advertisement
Advertisement has evolved recently with the appearance
of dynamic ads and recommender systems. Instead of advertising goods to an as large as possible audience, tailoring
advertisement to the customer’s needs is now possible. The
integration of UASs is an opportunity to develop both approaches. For mass advertisement, UASs could be equipped
with banners or signs, depending on their size. For targeted
advertisement, displays on a UAV could allow displaying the
relevant information to the right customer. For example, one
could envision a delivery drone asking for the customer to
watch an advertisement (just like on some web videos) before
releasing the parcel. This would imply some design challenges.
Indeed, UASs are systems optimized for a given operation,
with all design factors: Cost, Size, Weight And Power (CSWAP), being pushed to the limit. Any additional element not
directly related to the mission or the UAS might jeopardize the
optimality of the design. For example, adding advertisement
structures would add significant aerodynamic drag asking for
more power or a different structural design. Similarly, adding
displays would increase the weight and require more power.
This effect can be lessened by new technologies, e.g., by providing more powerful motors, more efficient batteries, better
control laws. From a regulatory perspective, privacy concerns
might arise from the fact that targeted advertisement in a public
place can reveal personal information that a customer would
prefer to keep private. This could be solved by technological
means (e.g., displaying the advertisement on the customer’s
cellphone) but it could lead to more customer acceptability
issues, on top of the ones linked to the presence of the
advertisement itself. This type of operation could be started as
soon as drone operations will be allowed in close proximity
to humans.
This section presented the operational concepts related
to each of the five considered applications, as well as the
obstacles preventing their integration and existing projects. The
next section compares the different obstacles with the timelines
of large European efforts in order to forecast dates at which
the required services and procedures will make the operation
available.
IV. E NABLING T IMELINE IN E UROPE
Since the beginning, three types of obstacles have been
considered for the integration of drones in various types of
operations: technological, standardization and regulatory. In
the European integration effort, each of these locks is currently
tackled by an international effort. Technological locks are being identified and dealt with by SESAR2020, the coordination
•
•
Safety aspects, especially related to the presence of
humans near the operation,
The need to integrate with ATM, or U-Space services,
or both,
High automation levels,
Definition of standards for UAS sub-systems.
The obstacles and ongoing efforts deployed in Europe are
presented side-by-side for each type of operation in Table II.
When the humans are involved in the operation, the risk
is lower, and in fact such a scenario is on its way to being regulated in some countries. For this reason, drones for
monitoring in non-public places are already at hand. Now
when the humans are not part of the operations, like in
the customers monitoring and customer assistance operations,
mitigation factors are required to ensure a low overall risk
level. For both applications, the UAV design and size can be
chosen to provide safe operation (e.g., UAV in a ball, micro
UAV). For flights far from humans and traffic, the complexity
of the operation is related to the amount of interactions with
the rest of the traffic. Operating in the middle of a dense traffic
is a lot more complex than in less occupied airspaces. For this
reason, delivery to remote places will likely come fairly soon.
The second obstacle is interactions with the ATM system and
manned traffic, though this is a highly complex problem, the
fact that it is being actively investigated will enable these operations, like cargo drones, relatively soon. Now, another factor
that may delay some operations is the automation level. Indeed,
the regulation being developed envisions a pilot for each UAV
in Category C operations and possibly one pilot for multiple
UAVs for Category B operations. However, to be feasible,
some operations will require full autonomy with little or no
supervision. We believe this level of acceptance and trust will
come only after years of successful drone operations, at which
point offering drone services for customer experience outdoors
will become a realistic application. Finally, the integration of
UAS in high traffic density airspaces, like the future airspace
above cities, asks for numerous mitigation measures that large
UAV (e.g., freight UAV) will be able to carry, allowing them
to operate earlier than smaller ones. Small UAV operations,
like parcel delivery, will come with the miniaturization of
equipment and setup of UAS traffic management, but that will
ask for time.
Based on the SESAR-JU roadmap for integration of drones
in ATM and EASA’s U-Space roll-out (see Figure 1), considering that EASA regulation is planned for 2019, and in
view of EUROCAE’s ongoing work, we propose tentative
dates for the beginning of each operation. For the EUROCAE
timeline, currently studied topics (DAA traffic, C2 link, ATOL)
are expected to yield standards in three years (around 2021).
Topics not studied yet, if started this year, would still take four
to five years to complete. Note that these dates represent the
TABLE II. O PERATION TYPES , MAIN OBSTACLES TO INTEGRATION , CURRENT EFFORTS TO SOLVE THEM , AND TENTATIVE DATES FOR INTEGRATION IN
THE E UROPEAN CIVIL AVIATION CONTEXT.
Application
Freight
LUCA
Urban
Monitoring
Public
No public
Guide
Indoors
Outdoors
Delivery
to remote location
to pickup point
to house/person
Advertisement
Main obstacles
Main ongoing efforts from civil aviation communities
Prospective beginning date
ATM integration, C2 link,
DAA traffic, ATOL
ATM integration, U-Space U2
deployment, DAA obstacles,
safety (Cat. C)
SESAR2020 projects for RPAS, EUROCAE WG-105 developing
standards (DAA, C2link, ATOL).
SESAR2020 projects for RPAS, SESAR2020 projects for U-SPACE,
automotive sector developing sensors, existing helicopter procedures.
2022
High level of automation,
public safety
High level of automation
Not regulated by civil aviation authorities.
2018
Not regulated by civil aviation authorities.
2018
High level of automation,
public acceptance
High level of automation,
public acceptance, U-Space
U1 deployment, safety (Cat
C.)
Not regulated by civil aviation authorities.
2018
SESAR2020 projects for U-SPACE, EASA’s regulatory framework
2022-2023
ATM integration, U-Space U2
deployment, C2 link, DAA
traffic, safety (Cat B.)
U-Space U3 deployment,
safety (Cat B.)
U-Space U4 deployment,
safety (Cat B.)
Public acceptance
SESAR2020 projects for RPAS, SESAR2020 projects for U-SPACE,
EUROCAE WG-105 developing standards (DAA, C2link).
2022
SESAR2020 projects for U-SPACE, EASA’s regulatory framework
2027
SESAR2020 projects for U-SPACE, EASA’s regulatory framework
2035
Not regulated by civil aviation authorities.
2018
times at which integration will be possible, however it does
not ensure that the industry will develop the applications and
that the public will be ready to accept them.
V. C ONCLUSION
This paper presented five envisioned applications, with
associated operations for UAS, that will impact the retail
sector. Each type of operation has been described, and enabling
technologies, standards, rules and services have been put
forward. This allowed to provide an estimation as to when the
different operations will become possible. One must remain
careful regarding these estimations as the retail habits are very
different from one country to another, and some countries may
want to push some applications regardless of the difficulties.
The same applies for regulation, with some countries going
faster and being more liberal.
[7]
[8]
[9]
[10]
[11]
[12]
[13]
R EFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
B. Smith and G. Linden, “Two decades of recommender systems at
amazon.com,” IEEE Internet Computing, vol. 21, no. 3, May-June
2017, pp. 12–18. [Online]. Available: doi.ieeecomputersociety.org/10.
1109/MIC.2017.72
Reuters. French retailers in Drive to win market share.
URL:
https://www.reuters.com/article/france-retail-drive/
french-retailers-in-drive-to-win-market-share-idUSL6E8INECL201207
24, [retrieved: 04, 2018]. (2012)
Econsultancy. Fashion ecommerce: are virtual fitting rooms
the silver bullet? URL: https://www.econsultancy.com/blog/
66058-fashion-ecommerce-are-virtual-fitting-rooms-the-silver-bullet,
[retrieved: 04, 2018]. (2015)
P. R. Wurman, R. D’Andrea, and M. Mountz, “Coordinating hundreds of
cooperative, autonomous vehicles in warehouses,” AI magazine, vol. 29,
no. 1, 2008, p. 9.
Softbankrobotics. Who is Pepper? URL: https://www.ald.
softbankrobotics.com/en/robots/pepper, [retrieved: 04, 2018]. (2018)
Amazon. Amazon Prime Air. URL: https://www.amazon.com/
Amazon-Prime-Air/b?ie=UTF8&node=8037720011, [retrieved: 04,
2018]. (2018)
[14]
[15]
[16]
[17]
[18]
2027
SESAR-JU. U-Space blueprint. URL: https://www.sesarju.eu/
u-space-blueprint, [retrieved: 04, 2018]. (2017)
M. P. Collins, “The future market for large unmanned cargo aircraft
in the national airspace system,” Ph.D. dissertation, Lewis University,
2017.
SESAR-JU. SESAR2020: PJ10-05. URL: https://www.atmmasterplan.
eu/data/sesar solutions/347, [retrieved: 04, 2018]. (2017)
Telegraph. Amazon’s flying warehouses to dispatch drone deliveries
from the sky. URL: http://www.telegraph.co.uk/technology/2016/12/29/
amazons-flying-warehouses-dispatch-drone-deliveries-sky/, [retrieved:
04, 2018]. (2016)
Overflo. How Drones Are Changing the Face of Warehousing. URL:
http://www.overflo.com/blog/drone-delivery.html, [retrieved: 04, 2018].
(2016)
InternetRetailing. 2018 Retail predictions. From drones to the changing customer experience. URL: http://internetretailing.net/2018/01/
customer-experience/, [retrieved: 04, 2018]. (2018)
M. Al Zayer, S. Tregillus, J. Bhandari, D. Feil-Seifer, and E. Folmer,
“Exploring the use of a drone to guide blind runners,” in Proceedings
of the 18th International ACM SIGACCESS Conference on Computers
and Accessibility. ACM, 2016, pp. 263–264.
DesignBoom. Fleetlights: a prototype team of flying torch drones to
guide the way at night. URL: https://www.designboom.com/technology/
fleetlights-direct-line-drone-app-11-18-2016/, [retrieved: 04, 2018].
(2016)
IndustryWeek. Ford Studies Drone Use to Guide SelfDriving Cars. URL: http://www.industryweek.com/automation/
ford-studies-drone-use-guide-self-driving-cars, [retrieved: 04, 2018].
(2016)
BBC. Crowded streets: Cities face a surge in online deliveries. URL:
http://www.bbc.com/news/business-42245367, [retrieved: 04, 2018].
(2018)
RungisInternational.
Transport,
la
logistique
du
dernier
kilometre (Transports, the logistics of the last mile).
URL:
https://www.rungisinternational.com/a-l-heure-du-digital/
transport-logistique-dernier-kilometre/, [retrieved: 04, 2018]. (2017)
TechnologyReview. Urban Drone Deliveries Are Finally Taking
Flight. URL: https://www.technologyreview.com/the-download/608912/
urban-drone-deliveries-are-finally-taking-flight/, [retrieved: 04, 2018].
(2017)