Analysis of Drone Flight Stability for Building a Korean Urban Air Traffic (K-UAM) Delivery System

Abstract

The Ministry of Land, Infrastructure, and Transport conducted a demonstration project targeting pilot areas to commercialize drone delivery services in urban areas and to present a standard model. In this study, flight data on drone delivery routes in Ulju and drone hovering in Yeosu were collected and analyzed for flight safety. Since there are no domestic or international regulations on the stability of drone flight, we were given the task of analyzing whether drone path flight should be maintained within a 10 m error range from the planned path line by the Korea Transportation Safety Authority and whether hovering works while satisfying the left and right radius errors and altitude errors within 3 m. Accordingly, the drone flight path data analyzed in Ulju met the criteria of up to 1.07%, and the hovering data analyzed in Yeosu met the criteria of less than 3% for the entire section data. Therefore, the drone flight stability evaluation analyzed in this paper is considered to have been passed. Based on the results of this study, is the data are expected to serve as a cornerstone for establishing guidelines for drone delivery flight data analysis regulations.

1. Introduction

Currently, the concentration of population in the metropolitan area of South Korea is accelerating. The metropolitan area currently accounts for 11.8% of the country’s land, while 49.5%, or about half of the population of South Korea, resides there. To solve this problem, various transportation methods such as shared electric kickboards, self-driving taxis, demand-responsive buses, and transportation services such as MaaS (Mobility as a Service) and Seamless Service are emerging, but ground transportation congestion is expected to persist due to the concentration of population in the metropolitan area. As an alternative, UAM (Urban Air Mobility), a three-dimensional transportation method using eVTOL (Electric Vertical Take Off and Landing, electric vertical take-off and landing aircraft or airframe), is attracting attention. UAM was developed as an air mobility platform that can improve urban traffic and environmental problems due to the limitations of ground transportation caused by global population growth and overcrowding in large cities. UAM is an air transportation ecosystem that utilizes an urban three-dimensional aerial transportation system and is expected to be composed of entities such as aircraft developers/manufacturers, transportation operators, traffic management service providers, vertiport operators, and value-added information service providers. The National Aeronautics and Space Administration (NASA) first named this concept UAM, and it has been used in related industries since then. Recently, NASA and the Federal Aviation Administration (FAA) have been expanding UAM to include cargo transportation, called Advanced Air Mobility (AAM) [1,2,3,4,5,6].

Accordingly, the government recently launched UAM Team Korea, which consists of about 40 key figures from industry, academia, research, and government in the urban air traffic sector, centered around the Ministry of Land, Infrastructure, and Transport, and proactive measures are being taken such as surveying domestic related organizations on their willingness to participate in the development of core technologies in six fields based on the Korean Urban Air Traffic Roadmap (K-UAM roadmap) [7,8,9]. Research and development of core flight-related source technologies targeting existing drones is being actively conducted, centered around academia, expanding to UAM. In addition, the research and development needs mentioned in the roadmap are being promoted step by step through public hearings, etc. [4]. Meanwhile, as the demand for non-face-to-face delivery and emergency goods delivery in urban and mountainous island areas has increased significantly due to COVID-19, much research and development is being conducted to commercialize drone-based logistics delivery services in urban/non-urban airspace [10,11,12]. In addition, research and development on drone taxis and flight safety are actively being conducted [13,14]. However, commercialization has not yet been achieved due to the lack of unification of policies, procedures, and standards, and it is judged that research is necessary in various fields [14,15,16]. To this end, the need for establishing a traffic management system for the safe operation of drones is emerging. Accordingly, research is actively being conducted to analyze and establish countermeasures for dangerous situations that drones may encounter during flight based on simulations [17,18]. To take these situations into account, there have been research cases that explored the flight performance of drones under various weather conditions and hovering situations [19].

Drones operated by remote control or autonomous flight are used in various fields such as aerial photography, pesticide spraying, surveillance, and reconnaissance. In particular, logistics companies such as Amazon in the U.S. are promoting cargo delivery using drones, and in Korea, they are also trying to utilize drones for delivery of small cargo and food. The Ministry of Land, Infrastructure, and Transport has established the ‘K-drone delivery promotion team’ to commercialize delivery services using drones in urban areas and to present a standard model, and has announced ‘Drone Delivery Guidelines’ to conduct pilot projects targeting some local governments. The above ‘Drone Delivery Guidelines’ stipulate that drones used for delivery services must conduct flight tests to ensure that they maintain their flight path and altitude as part of safety measures to prevent damage caused by collisions with ground personnel or obstacles in urban areas, and that a decision is made on whether or not to deploy them for delivery work based on the results of the tests.

Therefore, in this study, we collected and analyzed drone flight data from selected demonstration cities as part of the K-drone delivery commercialization project. The drone flight was measured for flight and hovering along the delivery route, and an error analysis on the stability of the flight was performed for submission to a national certification agency. The structure of this paper is as follows: Section 2 describes the K-drone delivery demonstration project. Section 3 introduces coordinate transformations to facilitate flight data calculations. Section 4 details the path flight of the drone, and Section 5 provides an in-depth description of the analysis procedures and error analysis related to hovering.

2. K-Drone Delivery Demonstration City Construction Project

The K-drone delivery commercialization project is a sub-task of the Korean Urban Air Transport Core Technology R&D Project. This is a drone delivery service implemented in islands, ports, and parks by applying the K-drone delivery commercialization standard model to discover drone utilization models suited to regional characteristics and apply them to public services, etc. The K-drone delivery project plans to begin building drone delivery infrastructure, including 41 drone delivery bases and 183 drone delivery points, by the first half of 2024, and to launch full-scale drone delivery of daily necessities and groceries in 38 island regions from the second half of the year. Figure 1a shows the delivery of parcels, convenience store items, chicken, etc., between inhabited islands and the mainland and reverse delivery of agricultural and marine products (3 or more bases, 10 to 20 delivery sections). Figure 1b refers to the delivery of convenience store items, chicken, kimbap, etc., to urban parks, campgrounds, beaches, tourist destinations, etc., in the form of park delivery (3 or more bases, 10 to 20 delivery sections). Figure 1c is a concept of port delivery, which includes delivery of food, ship supplies, fishing and daily necessities to anchored ships and marine leisure areas around ports (one to three bases, delivery area near the port). In Figure 2, 14 local governments are selected for the drone demonstration city construction project. In this study, flight data of drones for delivery were acquired from 3 out of 14 local governments and their stability was analyzed.

3. Flight Data Coordinate System

The coordinate system for analyzing drone flight data is generally a combination of the Global Coordinate System and the Body Coordinate System. This allows for accurate interpretation of the drone’s position, attitude (Euler angles), speed, etc. The Global Coordinate System is a coordinate system that considers the Earth as a three-dimensional ellipsoid and indicates the location on the Earth. The Earth is considered an ellipsoid and a point on the Earth’s surface is represented by latitude and longitude, and the unit is degree. The drone’s GPS data are expressed as latitude, longitude, and altitude, and the definitions of each term are as follows:

-

Latitude: The angle formed by the vertical line of a given point on the Earth’s surface and the equator.

-

Longitude: The angle formed by the prime meridian and the meridian drawn from the South Pole to the North Pole through a given point on the Earth’s surface.

-

Latitude: A line connecting points of the same latitude.

-

Longitude: A vertical line on the Earth’s surface that connects the North Pole and the South Pole at the shortest distance, also called a meridian.

-

Prime meridian: A line with a longitude of 0° that serves as the standard for longitude. Passes through the Greenwich Observatory in England.

The GPS receives latitude, longitude, and altitude information from satellites. Since the Earth is round, we can easily express this on a map by converting it to the UTM (Universal Transverse Mercator) coordinate system and expressing where it is on the map. The UTM coordinate system is one of the geographic coordinate systems that converts the Earth’s surface into a flat surface. The UTM system divides the Earth into several zones and applies a flat coordinate system to each zone, providing high accuracy. It is mainly used in fields such as map making, navigation systems, and GPS. In Figure 3, the UTM coordinate system divides the Earth into 60 vertical zones. Each zone covers a 6-degree longitude range and separates the Northern and Southern Hemispheres. For example, Zone 1 covers longitudes from 180°W to 174°W, and this zone is divided into the Northern and Southern hemispheres. The zone numbers increase eastward. Each UTM zone is defined by the central meridian of that zone. The central meridian is not a meridian with longitude 0, but rather the central longitude of each zone.

The coordinate system is set based on this longitude, minimizing distortion at that longitude. The UTM coordinate system is expressed as Easting and Northing, and Easting is a value from 0 to 1,000,000, and the reference starts at the Central meridian. The Central meridian is usually set to 500,000 m, so that it is expressed only in positive values. Northing is calculated starting at 0 and increasing towards the North pole, and in the Southern hemisphere it increases towards the South pole. In the Northern hemisphere it starts at 0, and in the Southern hemisphere it starts at 10,000,000 m and increases towards the South. However, since the projection calculation is performed in the same way as the UTM system but with the reference point at the equator, a lot of distortion occurs in South Korea. The UTM-K coordinate system was created as a countermeasure for this, and it is a case where only the origin and added values are applied differently from the UTM coordinate system, as shown in Figure 4 and

Table 1. Korean UTM-K coordinate system using WGS84 ellipsoid.

Designation	Latitude and Longitude of Origin	Area of Application
UTM-K	Longitude: 127° 30′ 00.0000″ E Latitude: 38° 00′ 00.0000″ N	All over the Korean peninsula

. The reference origin is set as the intersection of longitude 127.5° and latitude 38°, and the coordinates are expressed by adding a value of 1,000,000 m in the X-axis and 2,000,000 m in the Y-axis directions. The previous TM coordinate system had four origins, making it difficult to unify the coordinate system, but the UTM-K coordinate system uses one origin and unifies the entire country into one coordinate system. Therefore, the Korean UTM coordinate system uses the WGS84 ellipsoid and uses the UTM-K coordinate system according to the 2004 Ministry of Construction and Transportation notice.

4. Drone Flight Data Analysis—Route

In this study, the drone flight areas of Ulju, Yeosu, and Tongyeong among the local governments participating in the K-drone in Figure 2 are briefly shown in Figure 5.

A drone delivery flight route was planned passing through each delivery point and returning to a previously designated demonstration city with regional bases. In this paper, we introduce a drone delivery route that passes through specific points and returns to the city of Ulju, using the Operations Office and Jinha Public Parking Lot as two bases. All drone flight path courses in Ulju are listed in detail in

Table 2. Drone flight path course in Ulju.

Base Point	Course
Operations Office (Office)	(a) Office → Ulju Sports Complex → Office
	(b) Office → Complex Welcome Center → Office
	(c) Office → Moonlight Campground → Office
	(d) Office → Starlight Campground → Office
Jinha Public Parking Lot (Jinha)	(e) Jinha → Marine Sports Center → Jinha
	(f) Jinha → Ganjeolgot Sports Park → Jinha
	(g) Jinha → Solgae Park → Jinha
	(h) Jinha → Myeongseon Bridge → Jinha

, Figure 6, and Figure 7. In Figure 6 and Table 2, the Operations Office is abbreviated as Office. In Figure 7 and Table 2, the Jinha Public Parking Lot is abbreviated as Jinha.

Figure 6 and Figure 7 show the planned drone delivery route for the demonstration project and the actual measured drone flight route on the map. The planned drone flight path is a line connecting the points that the drone must pass through, marked in red, and the blue line is the actual flight route of the drone. Visually and on the map, the planned route and the actual drone flight route appear to match. However, there are rules that must be followed in order to write a drone delivery flight test report as the goal of this project. First, the drone must maintain flight within a margin of error of 10 m from the planned route line. Second, the drone’s flight altitude must also be maintained within 3 to 10 m of the planned flight altitude. For this purpose, it is necessary to perform calculations comparing the planned route with the drone’s flight data for the entire section.

Therefore, using the UTM-K coordinate system described in Section 3, the latitude and longitude values were projected into x-y rectangular coordinates, and the altitude was used as the z coordinate to perform calculations and analyze the drone’s flight data. In Figure 8, the drone delivery route of the Jinha—Myeongjin Bridge—Jinha course in Figure 7d is converted into x, y, and z coordinates using the UTM-K coordinate system conversion, and (a) represents the entire xyz plane, (b) represents the xy plane, (c) represents the xz plane, and (d) represents the yz plane.

In particular, analysis of drone delivery route data converted to xy coordinates in Figure 8b is important. In Figure 8b, a MATLAB (2022b) code was developed to calculate in real time the distance (d) difference between the two adjacent blue points along the planned delivery route that the drone passes through at each moment on the actual path that the drone passes through, using the distance formula between the point and the straight line in Figure 9. The altitude z-coordinate was also calculated as the difference between the planned altitude and the drone’s real-time flight altitude. Therefore, Figure 10a is the result of calculating the distance deviation from the planned blue path in the Jinha—Myeongjin Bridge—Jinha route of Figure 8b to the red point, which is the actual drone delivery movement coordinate. The x-axis is expressed as the number of data loaded instead of time. As with the visual observation in Figure 8b, the calculated numerical distance error in Figure 10a also shows that the drone is flying along a very stable path within 10 m. Next, the rule is applied to analyze the distance deviation of the z-coordinate, which is the altitude in Figure 10b, in the order of 10 m, 5 m, and 3 m. The error rate is expressed as the percentage of the number of data exceeding the distance error compared to the total number of data in Equation (1). When the excess data analysis was performed with the most conservative rule of 3 m or less, the altitude error of the Jinha—Myeongjin Bridge—Jinha drone delivery route was analyzed to be 0.11762%. According to the regulations set by the Korea Transportation Safety Authority, the error rate must be within 3%.

$$Error (\%)={\displaystyle \frac{\mathrm{Data} \mathrm{exceeding} \mathrm{the} \mathrm{distance} \mathrm{error} \left({\mathrm{N}}_{\mathrm{exceeding}}\right)}{\mathrm{Total} \mathrm{number} \mathrm{of} \mathrm{data} \left({\mathrm{N}}_{\mathrm{total}}\right)}}\times 100$$

(1)

The results of analyzing the path error and altitude error of the drone delivery routes in the Ulju area summarized in Table 2 are shown in

Table 3. Analysis of drone flight path course data errors in Ulju.

Course	Distance Error (%)	Altitude Error (%)
(a) Office—Ulju Sports Complex—Office	0	0.0068946
(b) Office—Complex Welcome Center—Office	1.0691	0.10097
(c) Office—Moonlight Campground—Office	0	0.11727
(d) Office—Starlight Campground—Office	0	0.14045
(e) Jinha—Marine Sports Center—Jinha	0	0.66181
(f) Jinha—Ganjeolgot Sports Park—Jinha	0	0
(g) Jinha—Solgae Park—Jinha	0	0
(h) Jinha—Myeongseon Bridge—Jinha	0	0.11762

. It can be seen that all drone delivery flight paths are flying very stably within the planned route and regulations.

5. Drone Flight Data Analysis—Hovering

Unlike fixed-wing aircraft, drones are aircraft that can take off and land vertically. Therefore, the task of testing vertical takeoff and landing is given for analyzing the stability of the drone’s body. Hovering is a state in which a drone is launched into the air and remains in that position without moving. Among aircraft capable of vertical takeoff and landing, helicopters, multicopters (drones), and VTOLs can implement hovering. Aircraft that do not fall into this category (typically fixed-wing aircraft and autogyros) cannot hover because they stall and crash if they stop in midair. The hovering process consists of three steps:

• Step (1) Selection of flight location,
• Step (2) Ascent and descent of the aircraft in place,
• Step (3) Landing at desired location.

In the study for this project, the hovering scenario was defined as in

Table 4. Hovering Scenario.

Step	Scenario
(1)	Takeoff and ascent 30 m
(2)	Maintain altitude at 30 m
(3)	Descending altitude 10 m
(4)	Maintain altitude at 10 m
(5)	Descending altitude 5 m
(6)	Maintain altitude at 5 m
(7)	Descent and landing

.

Figure 11 shows the planned step-by-step hovering scenarios listed in Table 4 and the analysis of the drone body’s hovering data. The drone’s hovering data were conducted in Yeosu, and the stability analysis of the drone for hovering presented by the Korea.

Transportation Safety Authority looks at two things: the left and right radius error and the altitude error when the drone takes off and lands. The hovering test of the drone body was specifically conducted in the areas of Gaedo and Jageum in Yeosu, as shown in Figure 12. The initial position before the drone took off was used as the origin, and the relative positions of the drone were calculated in real time and compared while the drone hovered according to the scenario. During the drone hovering performed in the Yeosu Gaedo in Figure 13a, the drone body shows a good example of safety in terms of shaking from side to side. The relative positions of the drones are relatively close to the origin, and the numerical error is also very small, with a difference of about 50 cm, as shown in Figure 13b. On the other hand, the drone hovering data collected in Yeosu, as shown in Figure 14, show a relatively bad example compared to the hovering performed in Gaedo. Figure 14a shows that the left and right movement of the drone becomes more severe during hovering compared to the initial position of the drone in Figure 13a. The numerical error in Figure 14b also increases to 2 m compared to the maximum value of 50 cm in Figure 13b. Nevertheless, the drone’s hovering radius error satisfied all requirements within 3 m.

The second rule for the hovering test is that the altitude error must be within 3 m. The hovering maintenance altitude measurement data collected at Yeosu Gaedo show that it is almost identical to the test procedure process in Figure 15, and the actual numerical error is also stably satisfactory within 3 m. On the other hand, the hovering data collected in Yeosu Jakgeum in Figure 16 show that the drone’s altitude is maintained at an average of 2 m or lower, and sometimes even momentarily exceeds the numerical error limit of 3 m. Fortunately, the error rate analysis of Equation (1) is 1.2195%, which does not exceed 3%, so it is considered to have passed the gas stability category.

Finally, the results of analyzing the hovering left/right errors and altitude errors in the Yeosu area are as shown in

Table 5. Analysis of drone hovering data errors in Yeosu.

Region	Distance Error (%)	Altitude Error (%)	Evaluation
(a) Gaedo	0	0	Stable (Pass)
(b) Jakgeum	0	1.2195

. It can be seen that the hovering test, which evaluates the stability of the drone according to regulations, was passed.

6. Summary and Discussions

The Ministry of Land, Infrastructure, and Transport of South Korea established the ‘K-drone delivery promotion team’ to commercialize drone delivery services in urban areas and present a standard model, and announced the ‘Drone Delivery Guidelines’ to promote pilot projects targeting some local governments. It stipulates that drones used for delivery services must conduct flight tests to maintain flight paths and altitudes as part of safety measures to prevent damage caused by collisions with ground personnel or obstacles in urban areas, and that decisions on whether to deploy them for delivery work be made based on the test results. In this paper, we collected and analyzed the flight data of the drone body. The drone flight was measured along the delivery path and hovering, and an error analysis on the flight stability was performed to confirm whether it met the regulations. The relevant regulations are as follows:

(a)

Drone Flight Data Analysis—Route

-

The drone maintains flight within a 10 m margin of error from the planned route line.

-

The drone’s flight altitude is also maintained within 3 to 10 m from the planned flight altitude.

(b)

Drone Flight Data Analysis—Hovering

-

The left and right radius error during drone takeoff and landing must be within 3 m.

-

The altitude error for maintaining the hovering must be within 3 m.

As a result of analyzing the drone flight path data conducted in Ulju and the hovering data conducted in Yeosu, there were moments when the distance within the regulation was temporarily exceeded, but the error rate of the entire section data satisfied the criterion of within 3%. Therefore, the drone’s body stability evaluation analyzed in this paper is considered to have been passed. Based on the results of this study, the data are expected to serve as a cornerstone for establishing guidelines for drone delivery flight data analysis regulations.