An Examination of UAS Incidents: Characteristics and Safety Considerations

Abstract

This paper examines the characteristics and implications of reported Unmanned Aircraft Systems (UAS) incidents in the National Aeronautics and Space Administration (NASA) Aviation Safety Reporting System (ASRS) database for UAS incidents operated by remote pilots licensed under Part 107. Characteristics examined include seasonal patterns of incidents, operational mission, and number of contributing factors, as well as crew configuration, timing of incident detection, and airspace class. Results are compared with previous research and with incident data for recreational users. The narratives for each incident are assessed to provide a greater context for the incidents and to determine how the incidents vary in different classes of airspace. Findings reveal that UAS incidents often involve multiple contributing factors, including environmental, human, equipment, and policy issues; there is an increasing prevalence of human-related issues over equipment problems compared to previous research; this reflects historic safety trends in crewed aviation. Near-miss incidents with crewed aircraft are a very real concern, particularly in Class D airspace, which often includes general aviation (GA) and helicopter operations. This research underscores the need for timely communication during urgent nighttime UAS operations as well as enhanced safety culture at both operator and organizational levels.

1. Introduction

Small Unmanned Aerial Systems (sUAS, also known as UAS or drones) have grown dramatically in popularity over the last decade. These systems can be quickly deployed and gather data more efficiently than traditional methods for many applications, particularly for “dull, dirty, dangerous, and demanding” tasks. UAS are widely used in photography, engineering, forestry, agriculture, archaeology, rescue missions, and infrastructure inspections [1,2]. As applications continue to expand, the U.S. commercial UAS market is expected to grow by almost 10% per year and reach revenue of $9.78 billion by 2030 [3]. While UAS provide significant social and economic benefits, they also pose potential threats to the National Airspace System (NAS). To reduce the likelihood of conflicts, UAS are restricted to altitudes below 400 ft and are not permitted to operate on or near airports without approval. Potential threats include collisions with crewed aircraft or being drawn into an aircraft engine, which could have catastrophic consequences such as bird strikes. These risks are exacerbated by the small size of UAS, making them difficult to detect and avoid, especially at high aircraft speeds [4].

While existing rules for UAS have been effective, with no major accidents involving commercial flights in the more than eight years since the publication of Part 107 in 2016, it is useful to examine UAS incidents to better understand the risks and safety context for UAS operations. This philosophy is supported by Heinrich’s Iceberg Theory, which suggests that catastrophic accidents often stem from seemingly minor errors and safety hazards that may go unnoticed or unaddressed. Proactively identifying and mitigating these minor errors is critical for preventing accidents, particularly in the emerging UAS sector [5,6,7]. Using UAS incident reports from NASA’s Aviation Safety Reporting System (ASRS) database, this study explores current UAS safety considerations, challenges, and areas for improvement. The main research objectives are as follows:

• To provide insights into the safety characteristics of reported incidents for Part 107 UAS operations;
• To compare the characteristics of current Part 107 operations with those of recreational UAS operations;
• To identify trends in UAS operations by comparing characteristics of current Part 107 operations with previous studies
• To investigate incident characteristics in different classes of airspace.

2. Literature Review

A thorough examination of UAS incidents requires an understanding of the current regulations and safety requirements for UAS, as well as an appreciation of the historical context. This section explores four key areas: the regulatory development for UAS in the United States, UAS safety training requirements and outreach engagement efforts, safety frameworks from crewed aviation that are applicable to UAS operations, and prior research on UAS safety. By addressing these themes, this review establishes a foundation for analyzing current UAS safety challenges and identifying opportunities for improvement.

2.1. UAS Regulatory Development

The FAA Modernization and Reform Act of 2012 required the establishment of commercial drone regulations [8], and in response, the FAA published Part 107 in 2016; this was subsequently amended to enhance safety and operational capabilities and now affects the activities of nearly 1.3 million Part 107 and recreational pilots [9,10]. The FAA continues to explore innovative methods to integrate UAS into the national airspace system (NAS) by leveraging existing regulations and also through the development of performance-based rules [11]. In this context, Part 107 applies to most small unmanned aircraft operations (under 55 pounds), and example provisions include:

• Always avoid manned aircraft
• Maintain Visual Line of Sight (VLOS);
• Do not exceed a ground speed of 100 miles per hour (87 knots);
• Fly no higher than 400 feet above ground level unless within 400 feet of a structure;
• Obtain authorization for operations in Class B, C, D, and E airspace;
• Obtain FAA waivers for operations that do not conform to these provisions.

UAS regulations have continually evolved in response to UAS growth, new technologies, and FAA objectives to develop performance-based rules [12,13]. For instance, in 2021, the FAA amended several rules in Part 107, which now permit routine nighttime operations and allows UAS flights over people and moving vehicles under specific conditions [14]. This rulemaking highlights the FAA’s proactive regulatory approach to accommodate advancing UAS technology and the growing demand for flexibility in UAS operations.

In addition to developing a regulatory framework, the FAA has employed an innovative regulatory approach to grant access to the ever-increasing demand for routine, complex UAS operations while ensuring safety risks remain “As low as reasonably practicable” (ALARP) [15]. Key components such as Registration and Marking (a requirement for UAS to be registered with the FAA and marked with a unique identifier number that is registered with the FAA), Remote Identification (a requirement for an ADS-B out transponder that emits a signal identifying the unmanned aircraft or UA), Low Altitude Authorization, and Notification Capability (LAANC, an internet-based system that provides immediate authorization) serve as fundamental pillars of unmanned traffic management (UTM) [16]. Much like the function of an automobile license, the Remote ID of UAS in-flight provides identification information transmitted via broadcast signals, which can be received by other aircraft and air traffic control, facilitating enforcement by public safety [17]. Collectively, these capabilities and regulations provide the safety and security that enable more complex drone operations.

LAANC, a collaboration between the FAA and industry, plays a crucial role in managing the growing volume of UAS operations and provides:

• Access to controlled airspace at or below 400 feet;
• Awareness of safe operational areas;
• Information and updates on temporary flight restrictions (TFRs), the Special Use Airspace (SUA) schedule
• Visibility and transparency regarding locations and times of drone operations.

As a foundational element of the UTM ecosystem, LAANC is accessible to commercially certified operators, recreational flyers, and other NAS users. While some users report issues with interface design, overlapping authorization zones, and inconsistent approval times, LAANC has transformed airspace authorization requests from a labor-intensive manual process into a streamlined, near-real-time system, greatly benefiting both UAS operators and the FAA [18].

While the FAA has made significant progress in ensuring the safe and efficient integration of UAS, regulations remain relatively immature due to the sector’s infancy [19,20,21,22]. For instance, challenges include delays in UAS emergency authorizations, a regulatory gap for UAS over 55 pounds [23], and the absence of rules for Beyond Visual Line of Sight (BVLOS) operations [24]. Furthermore, the lack of requirements and recommendations for appropriate crew configurations and shift schedules for UAS operators highlights another critical gap [25]. As the UAS sector continues to grow and evolve, addressing these regulatory challenges will be essential for ensuring the safe and efficient integration of UAS into the NAS.

UAS Safety Training Requirements and Outreach Engagement

The FAA provides two types of certificates for pilots who operate UAS in the NAS. The remote pilot certificate is issued for commercial use under Part 107, and the Recreational UAS Safety Test (TRUST) is designed for purely recreational operations [26]. To obtain a remote pilot certificate, applicants must be at least 16 years old, proficient in English, physically and mentally capable of safely operating a drone, achieve a minimum score of 70% on a 60 min aeronautical knowledge test, and complete the TSA security background check. For people who hold a Part 61 Private Pilot Certificate, this knowledge test is substituted with the Part 107 Small UAS Initial (ALC-451) online training course, as these individuals are considered to have a solid understanding of general aeronautical principles [27]. To become an authorized recreational flyer, applicants must complete TRUST, a free online test of aeronautical and safety information. As of October 2024, the FAA issued 415,635 Part 107 certificates and 883,094 TRUST certificates [28]. Public UAS operators who utilize UAS for safety- or security-related government functions may operate under Part 107 or obtain a Certificate of Waiver or Authorization (COA), which is submitted by the public agency. The COA allows public agencies to design their own training and certification programs for UAS crews; these programs must align with the terms in the COA application and reference Federal Aviation Regulations (FAR) Parts 61, 91, and 107 [29,30].

The UAS pilot certification and training process has seen ongoing improvements, and related research is required to ensure completeness and effectiveness. For example, the scope of the exam, including several knowledge test sections (e.g., reading runway numbers, understanding crewed aircraft procedures, and interpreting METARs), remains a topic of debate by UAS operators [31]. Additionally, there are no formal training or certification requirements for visual observers, who play a critical role and work under the direction of Part 107 pilots to maintain the aircraft within line of sight and ensure safe UAS operations [32]. The rapid evolution of UAS technology often outpaces regulatory updates and emphasizes the need to revise exam requirements to address emerging technologies [33].

In addition to safety training and certification, the FAA is committed to outreach to the UAS community, especially recreational flyers who may have less understanding of operational guidelines [34,35]. Initiatives such as the FAA UAS Symposium, national drone safety awareness week, and drone webinar series aim to foster robust safety awareness among commercial and recreational operators. Programs such as the Law Enforcement Assistance Program (LEAP) support public safety agencies in managing drone operations, while UAS STEM education and the UAS Collegiate Training Initiative (UAS-CTI) prepare students for careers in the drone workforce [36,37]. These efforts have demonstrated a positive impact, as evidenced by FAA drone webinars garnering over 244,500 views on YouTube alone [38].

2.2. Safety Development in the Crewed Aircraft Sector

While UAS safety is still evolving, given the fact that Part 107 was published less than a decade ago, it builds on the foundation of safety for crewed aircraft. In 2011, ICAO declared that UAS should follow the same safety standards as crewed aircraft [39]. Although the UAS sector differs in a number of ways from crewed flight operations, most notably in that there are no risks to souls on board the aircraft, lessons learned from crewed aircraft safety practices can provide valuable guidance for developing UAS safety standards [15,23].

Generally, the progression of safety in commercial aviation can be divided into three eras: the initial focus was on technical aspects, followed by a focus on human factors, and most recently, safety has evolved to emphasize organizational factors [39]. In the early days of flight, most safety deficiencies in aviation were related to technical aspects and equipment failures, which reflected the immaturity of aircraft technology. By the 1950s, significant technological improvements to aircraft helped make aviation a safer mode of transportation; however, the increase in air traffic resulted in issues related to air traffic control. A number of dramatic aircraft crashes, such as the midair collision over the Grand Canyon, led to increased funding for air traffic control improvements [40]. By the early 1970s, technologies were more reliable, and safety efforts shifted to address human factors, which were frequently cited as contributing to incidents and accidents. One framework for human factors is the “SHELL” model, which was developed by Hawkins in 1975 and provides a structured approach to analyzing the contributing factors in incidents and accidents, including:

• Software (procedures and documentation);
• Hardware (displays and functional systems);
• Environment (social, economic, and natural conditions)
• Liveware (human operators within the system) [41].

Starting in the 1990s, a more holistic approach to safety emerged, incorporating broader considerations, including organizational culture, external conditions, and operational policies alongside human and technical factors. Since incidents and accidents in commercial aviation are often caused by multiple factors, various safety models have been developed to illustrate and analyze the causal relationships. One of the most recognized models is the “Swiss Cheese” model developed by James Reason, which illustrates how accidents in complex systems such as aviation have a number of contributing factors and illustrates the importance of implementing multiple layers of defense to mitigate variations in human performance and address latent external conditions that may contribute to safety issues [5,42].

The continuous improvements in safety management practices have contributed to aviation becoming one of the safest modes of transportation today. While adaptations are necessary to apply these crewed aircraft safety frameworks to UAS operations, they establish a strong foundation for developing UAS safety protocols [15,43].

2.3. UAS Safety Considerations and Prior Research

UAS safety encompasses more than just the flying drone; it involves an integrated system of equipment, personnel, procedures, missions, and environmental factors [44]. Compared with crewed aircraft, the emerging UAS sector faces unique safety challenges due to the absence of onboard crews, rapid technological evolution, and immature regulatory frameworks [15,45].

Human factors considerations, such as crew resource management, situational awareness, and fatigue, represent some of the most significant safety hazards in UAS operations [46,47,48]. For instance, single pilots monitoring multiple UAS over extended periods are more prone to fatigue [49,50]. Additionally, remote operators may take more significant risks since the absence of personal physical danger lowers their perceived stakes [44]. These human factors are further complicated by reliability issues with UAS equipment and the operational environment. Equipment failures, including GPS malfunctions, lost links, and propulsion issues, are significantly more frequent in UAS operations than in crewed aviation, with failure rates of 1 in 1000 flight hours—far higher than in commercial aviation [51,52]. Understanding the interaction of the UAS with the environment is also critical. Without being onboard, operators may struggle to navigate hazards such as inclement weather, obstructions, or terrain [44]. Moreover, gaps in organizational safety policies and culture may compound these risks and are essential areas to address for the effective development of UAS SMS [53,54,55].

Analyzing past occurrences is one of the most effective strategies for understanding safety risks and identifying patterns in UAS operations [5]. Early research primarily focused on military UAS incidents, revealing that human factors contributed to 60.2% of 221 recorded cases [56]. These early findings also underscored the need for regulators to prioritize airworthiness requirements, which remain underdeveloped for civil UAS [13]. For civil UAS operations, Wild et al. (2016) [57] analyzed 152 unsafe events, identifying equipment issues as the most prevalent contributing factor, accounting for over 60% of cases. Unsafe events during landing were frequently attributed to human factors; these findings are consistent with Huh and Shim’s (2010) [58] findings that human factors are the leading contributors to incidents during the landing phase for both unmanned and crewed aircraft.

Building on this, Wang and Hubbard (2021) [59] analyzed 6544 UAS sighting reports from 2016 to 2019, and identified seasonal and temporal patterns, with higher sightings in summer and during daylight hours. These findings align with Gettinger and Michel’s (2015) [60] observation that UAS sightings are more frequent during the day. Similarly, Pitcher (2022) [61] examined over 9000 UAS sighting reports from 2015 to 2019 and found that incidents were most frequent in densely populated areas during summer weather. Collectively, these studies highlight critical factors that influence UAS safety and support the concept that examining UAS incident data is an effective approach for deriving meaningful insights into UAS safety development.

3. Methods

This study examines characteristics of UAS incidents reported in the National Aeronautics and Space Administration (NASA) Aviation Safety Reporting System (ASRS) database from 2019 to June 2024, including 187 reports for UAS operating under Part 107 and 70 reports for recreational UAS activities. The NASA ASRS was established by the FAA and NASA in 1976 to assist NAS stakeholders in identifying safety hazards and deficiencies. Since its inception almost 50 years ago, ASRS has compiled over 1.7 million reports, making it a vital resource for aviation and airspace safety research [62]. Recognizing the rapid growth of the UAS sector, ASRS introduced a safety reporting system for UAS in 2019. The ASRS system is voluntary, confidential, and non-punitive. Stakeholders are encouraged to report close calls, hazards, violations, and other safety-related incidents using the tailored reporting form shown in Figure 1. Information about the reported UAS incidents is publicly available and accessible online. The reports in ASRS are self-reported and non-punitive. This means that remote pilots file a report when they have violated regulations or may have violated regulations, and there will not be a penalty against the remote pilot who reports a violation.

This contrasts with the framework for reports published by the FAA in the UAS sightings report. Reports published in the FAA sightings reports can be made by anyone who sees a potentially dangerous or illegal UAS operation [63]. A fundamental difference between the reported incidents in ASRS and the FAA sightings reports lies in the source and purpose of the reports: ASRS prioritizes safety improvements through voluntary self-reporting, whereas UAS sighting reports prioritize documentation of public safety concerns and regulatory violations through a public reporting system. FAA sightings reports may be made by remote pilots, pilots of manned aircraft, air traffic controllers, airport operations workers, and even members of the general public. The FAA will investigate the reports, and there may be penalties against a pilot if the FAA documents a violation. Since ASRS is non-punitive, a remote pilot may file a report in ASRS to protect themselves from a penalty if they believe there was a violation.

As a result, the ASRS database may include more information about reported events than the UAS sightings report, since the ASRS report was filed by the remote pilot. ASRS reports are more likely to be filed by a remote pilot since there are no associated risks of penalties to the UAS pilot regarding a violation. The self-reported nature of ASRS reports may include subjective evaluations or opinions from the reporters. However, the value and accuracy of the ASRS reports are enhanced by the first-person perspective, safety awareness, and the motivation to provide accurate and meaningful insights to improve aviation safety. The non-punitive nature of the platform further encourages reporters to share their mistakes and perspectives openly, with minimal concern for potential consequences.

Figure 1. Overview of the UAS reporting form [64].

Figure 1. Overview of the UAS reporting form [64].

Characteristics of the reported UAS incidents were categorized and analyzed based on variables such as the time of occurrence, operational mission type, crew configuration, as well as contributing factors. The categories for each variable are defined such that there is no overlap between categories. Descriptive statistics of the incidents are presented, and comparisons are given for UAS operations under different circumstances.

The classification of the variables time of day, month, operational mission, crew size, airspace, and time of detection follows industry standards, regulatory requirements, and generally established practices, as presented in

Table 1. Classification criteria for variables used.

Variable	Categories
Time of Day	Dawn: 30 min before the official sunrise Daylight: From sunrise to sunset Dusk: 30 min after the official sun Nightlight: From dusk to dawn
Month	January to December: By calendar month
Operational Mission	Top five most frequent operation missions in the database
Crew Size	Single operator: Only one operator is involved Multiple operators: More than one operator is involved
Airspace	High-density controlled airspace: Class B and Class C Moderate-density controlled airspace: Class D and Class E Uncontrolled Airspace: Class G Special Use Airspace: Restricted and temporarily restricted airspace
Time of detection	Pre-flight: Violations/incidents noticed before the operation In-flight: Violations/incidents noticed during the operation Post-flight: Violations/incidents noticed after the operation
Contributing Factors
Human Factors	Fatigue Complacency Lack of knowledge Distractions Impaired situational awareness Hazardous attitude
Environment	Weather-related (e.g., wind shear, low viabilities) Non-weather-related (e.g., terrain, obstructions, lighting, fire)
Equipment Issues	Components and sensor failure Lost link Frequency interference
Policy Issues	Unclear company policy Deficiencies in safety culture Deficiencies in operation procedure

. The time of detection variable indicates when the breach of regulations or incident was noticed by the remote pilot: before the UAS operation began (pre-flight), during the UAS operation (in-flight), or after the UAS operation (post-flight). In some cases, a remote pilot may have realized that there would be a violation before the flight but chose to proceed anyway, perhaps due to the urgency of the mission or a perception that the safety risk was low. Additionally, a remote pilot may have realized there was a violation during the flight when there was a gust of wind or equipment issue that caused the UAS to go into an airspace region that was unauthorized. In some cases, the remote pilot may have realized that there was a potential airspace violation after the flight during a post-flight review.

The classification of the variables for contributing factors was based on the criteria and assumptions shown in

Table 2. Classification method for contributing factors.

Contributing Factors	Associated Keywords	Associated Issues	Assumptions
Human Factors	Human factors Staffing	Relates to operator behavior, decision-making, and mental state, including fatigue, complacency, distractions, impaired situational awareness, and hazardous attitudes.	The keywords ‘Human Factors’ and ‘Staffing’ are always categorized under human factors.
Environment	Environment non-weather-related Weather	Includes weather-related factors (e.g., wind shear, low visibility) and non-weather-related factors (e.g., terrain, obstructions)	The keywords ‘Environmental—Non-Weather Related’ and ‘Weather’ are always categorized under environment.
Equipment Issues	Aircraft Equipment tooling Software and automation	Relates to hardware and software failures, including component and sensor malfunctions, lost link, and frequency interference.	The keywords ‘Aircraft’, ‘Equipment Tooling’, and ‘Software and Automation’ are always categorized under equipment issues.
Policy Issues	Company policy Procedure Manuals Airspace structure Airport	Relates to organizational safety policy issues, including unclear company policies, deficiencies in safety culture, and deficiencies in operational procedures especially regarding airport and airspace structure.	The keywords ‘Company Policy’, ‘Procedure,’ Manuals’, ‘Airspace Structure,’ and ‘Airport’ are always categorized under policy issues.

. The four categories of contributing factors are human factors, equipment, environment, and policy; these categories are consistent with methodologies used in previous studies by Boyd in 2015 [65] and Wild et al. in 2016 [57]. Boyd’s research [65] included contributing factors such as human error, weather, mechanical failures, and procedures; these categories are analogous to human factors, environmental issues, equipment issues, and policy issues in this study. Wild et al.’s research [57] included contributing factors as human factors, environmental issues, equipment problems, and organizational issues; these categories are almost identical to those in this study, with Wild’s organizational issues analogous to policy issues in this study. Although the specific terms in the three studies differ slightly, the general regimes of the four contributing factors in all three studies align closely.

The use of human factors, environmental factors, equipment issues, and policy issues as contributing factors aligns closely with contributing factors commonly recognized and used in aviation, such as those described by the SHELL model as defined in the ICAO Safety Management Manual (Doc 9859, 2017) [5]. The SHELL model reflects Software (i.e., policy issues in Table 1 or the software component of equipment issues), Hardware (i.e., equipment issues in Table 1), the Environment, and Liveware (i.e., the people involved in the operation and the human factors considerations in Table 1). Policy issues that could contribute to a UAS incident may include unclear company policies or procedures (e.g., a lack of pre-flight checklists) and/or an immature safety culture; this would align with the safety policy component of SMS in commercial aviation.

The classification of contributing factors for each reported incident was based on keywords provided in the “Assessment” column of the database. These keywords were used to identify the contributing factors, as described in Table 2. To ensure an accurate representation of each keyword’s meaning in the context of the reported incident, the narratives were carefully examined and considered. The combination of the keywords interpreted in the context of the narrative description enabled the classification of contributing factors for each reported incident and ensured that each keyword was assigned to only one category.

Contributing factors were based on the associated keywords for each reported incident. Most reported incidents had multiple contributing factors (as reflected by the multiple keywords in the assessment section). This is not surprising; it is consistent with the error chain philosophy in crewed aviation, which acknowledges there are usually multiple factors that contribute to an aircraft accident or near miss.

To illustrate the classification method for contributing factors, consider the following examples.

• Reported incident ACN 1721078 contained three keywords: ‘Procedure’, ‘Company Policy’, and ‘Human Factors’. In this case, ‘Procedure’ and ‘Company Policy’ were classified under ‘Policy Issues’, as they reflect inadequacies in the company’s safety procedures, while ‘Human Factors’ was classified under ‘Human Factors’ category.
• Reported incident ACN 1840720 included the keywords ‘Environment—Non-Weather Related’, ‘Human Factors’, ‘Aircraft’, and ‘Software and Automation’. In this case, ‘Environment—Non-Weather Related’ referred to obstruction interference and was classified under ‘Environmental Problems’. ‘Aircraft’ and ‘Software and Automation’ were associated with connection loss and categorized under ‘Equipment Issues’. ‘Human Factors’ reflected the operator’s impaired situational awareness and was directly linked to the ‘Human Factors’ category.

A chi-square goodness-of-fit test (with a 5% significance level) was used to compare the distributions (e.g., the incident characteristics concerning variables such as contributing factors, crew size, detection times, airspace class, and operational framework of Part 107 commercial operations vs. recreational operations). The chi-square test was also used to compare the incident reports with previous findings by the time of year, as investigated by Wang and Hubbard (2021) [59] based on data in the FAA UAS sightings report and by contributing factors as investigated by Wild et al. (2016) [57].

Moreover, examining the narrative reports and associated summaries provides a more robust understanding of the incident circumstances and a context for some representative incident causes. The incidents were sorted by airspace, and the narratives were reviewed to assess the prevalent circumstances of each airspace class. Representative cases are presented to highlight key findings and provide a better understanding of the safety considerations and operational challenges.

4. Results

The results are presented in four sections which relate to the research objectives and present information regarding the characteristics of reported UAS incidents under Part 107 UAS; a comparison of reported incidents for Part 107 operations with recreational UAS operations; a discussion of how the Part 107 findings relate to previous studies; and the results of an investigation of the characteristics of incidents in different classes of airspace.

4.1. Characteristics of UAS Incidents Under Part 107 Operations

There were 187 UAS incident reports for operations under Part 107 filed between January 2019 and June 2024, as depicted in Figure 2a. The average number of reports has increased over time, from six per year during the initial two years (2019 and 2020) to an average of 51 reports per year since 2021. This upward trend reflects the increasing prevalence of UAS applications and the growing maturity of reporting practices within the UAS sector. Figure 2b depicts the monthly distribution of UAS incident reports from 2019 to 2023 (reports in 2024 were not included in this chart since there was not a full year of data). A peak in reporting is observed during the summer, with June and July collectively accounting for over 23% of the total reports, which likely reflects increased UAS activity during favorable weather conditions and may correspond with UAS being used to document summer construction and other seasonal operations. Interestingly, a higher-than-expected proportion of reports is recorded from October to December, a trend that will be explored in greater detail later in this paper.

Most reported incidents occurred during daylight hours, as shown in Figure 2c. Six of the seven nighttime cases (86%) were related to public safety missions, while only about 5% of daytime operations were related to such missions. The higher concentration of public safety missions during nighttime operations is not surprising since only time-critical and essential activities are likely to occur at night.

An examination of the narratives for these nighttime cases indicates that in some instances, UAS operational regulations were knowingly violated due to the urgency of the mission and because there may have been insufficient time to obtain the required authorization, as illustrated in

Table 3. Narratives for incident reports (NASA ASRS, 2024).

Nighttime Public Safety Missions—Flights Undertaken with Known Regulation Violations
Association Number (ACNs)	Operational Mission	Time of Day	Incident Summary
(a) 2062688	Public Safety/Pursuit	Night	In response to a request from the local fire department to search for a potential accident victim in controlled airspace, the operator attempted to contact ATC but was unable to do so due to off-hours. Prioritizing the urgency of the mission, the operator proceeded with the flight, employing visual observers and safety measures to mitigate risks.
(b) 2057627	Public Safety/Pursuit	Night	In response to a search for a suspect who fled a stolen vehicle, the operator exceeded airspace ceiling limits during a nighttime operation. Safety measures (visual observer and thermal display monitoring) were employed. The violation was acknowledged and attributed to the urgency of the mission rather than intentional noncompliance.
Incident Detection during Pre-flight May Reflect “Invulnerability Attitude”
(c) 1918662	Photo Shoot/Video	Daytime	The operator operated without obtaining the required LAANC authorization to avoid a lengthy return trip. Citing prior difficulties in requesting approval for the same airport, they chose to fly at a low altitude above the tree line, completing the mission in approximately five minutes. Despite being aware of the regulations, the operator prioritized operational convenience and assumed the operation would not pose significant safety risks.
(d) 1853802	Public Safety/Pursuit	Night	Operating in controlled airspace without obtaining the required LAANC authorization, the operator proceeded with the flights, maintaining visual line of sight (VLOS) and staying below 400 feet AGL while citing adherence to other safety measures. However, prioritizing mission completion over postponing for proper authorization suggests a potential assumption that the operation posed minimal risks.

(rows a and b).

While public safety is an important operational mission, it is not the most common one. Figure 3 illustrates the top 5 operational missions based on the incidents reported in the database. Photography was the operational mission associated with the highest number of incident reports (about 43%); this underscores the popularity of using UAS for photography, and it may also highlight a potential gap in safety culture for this user group. This gap in safety culture is illustrated in the narrative in Table 3 (row c), in which a photographer realized during pre-flight that authorization was needed but the choice was made to fly without LAANC authorization to save time and for convenience. While the public safety mission (row d) was driven by an urgent operational need, the photographer’s mission lacked this urgency. These two narratives (Table 3, rows c and d) suggest a possible “invulnerability attitude” among some UAS operators, who decide to proceed with an operation for convenience despite lacking authorization; this reflects a belief that accidents are unlikely to happen to them [44,66].

There were very few cases in which the incident was detected during pre-flight (examples were described in Table 3, rows c and d), as shown in Figure 4. Most incidents were discovered during flight (57%), while about 42% were discovered post-flight; only 3 (<1%) were found pre-flight, and in all three cases, the flight was conducted without obtaining the required LAANC authorization despite being aware of this requirement (only the case described in row d of Table 3 was driven by an urgent operational need). Conducting a UAS operation despite knowing it is not authorized may reflect a lack of training, a poor safety culture, and/or an inadequate organizational policy that forbids such an operation.

While a lack of authorization is one factor that may contribute to an incident, there are a variety of other factors that may be involved. In most cases, there is not a single factor that can be implicated in the incident, but rather, there are multiple factors that contribute to the incident, as shown in Figure 5a. An example of a single contributing factor is when the UAS accidentally enters unauthorized airspace because the remote pilot is fatigued (human factors). An example of multiple contributing factors is when the UAS accidentally enters unauthorized airspace because the remote pilot is fatigued (human factors) and there is an unpredictable gust of wind (environment). The involvement of multiple contributing factors is analogous to trends observed in crewed aircraft operations and may reflect the growing complexity of UAS operations. The involvement of multiple contributing factors may also highlight the value of comprehensive safety analysis and the detailed assessment of contributing factors. Acknowledgment of multiple contributing factors may also reflect a maturing safety framework.

Figure 5b shows the distribution of contributing factors. Human factors are the most prevalent contributing factor (37%), followed by equipment issues (27%) and policy issues (24%). Environmental problems were the least common, comprising only 12% of the contributing factors; this may reflect the fact that remote operators are well aware of environmental risks such as wind, bad weather, and terrain and are able to manage these risks effectively most of the time.

It is not surprising that environmental problems are more commonly detected in-flight rather than post-flight, as shown in Figure 6. Equipment issues are also more frequently detected during flight (26%) compared to post-operation inspections (10%). In contrast, human factors and policy issues are more likely to be associated with incidents detected post-flight; human factors and policy issues collectively comprise about 85% of the factors identified post-flight.

4.2. Compare Part 107 Operations with Recreational Operations

The previous section presented data for incidents in which the remote pilot was licensed under Part 107. The database also includes incidents reported by recreational pilots who do not hold a remote pilot certificate. This section compares the characteristics of these two groups of operators, as seen in Figure 7. The chi-square test is used to determine if the distribution of characteristics is different for incidents under Part 107 than for recreational operations.

As shown in Figure 7a, there is no significant difference between the contributing factors for incidents in recreational operations compared to those observed in Part 107 operations (χ² = 2.75, df = 3, p = 0.45). Among the 70 recreational cases analyzed, human factors accounted for 43% of the total contributing factors, followed by equipment issues (23%), policy issues (20%), and environmental problems (14%).

As shown in Figure 7b, approximately 65% of the reported recreational incidents were detected during flight, while the remaining reported detections occurred post-operation. No pre-flight cases were recorded for incidents by recreational pilots. This distribution is similar to that observed in Part 107 operations, with a chi-square test indicating no significant difference (χ² = 0.68, df = 1, p = 0.41).

In recreational operations, approximately 90% of the incidents reported reflect operations conducted by a single operator. This highlights a difference from Part 107 incidents, in which only 59% of incidents were conducted by a single operator. This difference in crew size for recreational pilots compared to Part 107 pilots is statistically significant ($\chi 2=19.23, df=1, p<0.01)$ as illustrated in Figure 7c. Interestingly, the larger crew size typically observed in Part 107 operations does not apparently correspond to a higher proportion of incidents despite the fact that human factors sometimes increase when there are more people involved.

4.3. Comparing Part 107 Operations with Previous Studies

It is also interesting to compare the findings of this study regarding UAS incidents reported in the ASRS database with previous research reporting on UAS incidents, as shown in Figure 8.

4.3.1. Comparison with Previous Research Regarding Temporal Distribution of UAS Incidents

The temporal distribution of incidents in this research was compared with the findings reported by Wang and Hubbard (2021) [59], as illustrated in Figure 8a. Both studies identified a peak in UAS incidents and sightings during the summer months, with June recording the highest number of reports. However, the two distributions were found to differ significantly when all 12 months of the year were considered $({\chi }^2=23.107, df=11, p=0.017$).

The monthly distribution of UAS incidents for the data presented in this study shows greater variability than the results of the previous study, and there is a higher proportion of reports from October to December. This difference may indicate a trend toward more consistent UAS usage year-round (perhaps newer batteries can accommodate colder weather), or it may be an artifact of the smaller sample in this research study. The dataset used by Wang and Hubbard (2021) [59] is approximately 35 times larger than the dataset analyzed in this study. More consistent trends regarding seasonal variations in UAS incidents may emerge as additional data is collected. Furthermore, since the research by Wang and Hubbard (2021) [59] reflected data in the FAA sightings report, and these incidents can be reported by pilots, air traffic control, airport workers, or members of the general public (though members of the general public are probably less likely to be knowledgeable about how to file a report). Anyone who sees a UAS that may be dangerous or in violation of the rules may file a report. As a result, the data in the FAA sightings reports may reflect not only the likelihood of a UAS incident but also the likelihood that someone will see and report it. In the summer months, there may be more people outside enjoying nice weather and thus in a position to report a UAS; for example, crewed aircraft operations such as GA operations typically increase during the summer months. Conversely, most reports analyzed in this study were self-reported by UAS pilots, making the fluctuations in crewed aircraft operations throughout the year less relevant.

4.3.2. Comparison with Previous Research Regarding Contributing Factors

The distribution of factors that contributed to Part 107 UAS incidents reported in the ASRS database from this research was compared with the results of previous research conducted by Wild et al. in 2016 [57], as shown in Figure 8b. While it is important to be very careful when making comparisons with previous research due to the potential differences in characteristics of the user groups and the way the data is collected and analyzed, there may be some value in a preliminary comparison to explore potential similarities and differences. The distribution of the four contributing factors (equipment issues, environmental issues, human factors, and policy issues) was found to be statistically different ($\chi 2=74.44, df=3, p<0.01)$, suggesting that the nature of factors that contribute to UAS incidents may have changed over time. In this study, human factors were the most prevalent contributing factor, whereas the research by Wild et al. (2016) [57] found that equipment failures were the most frequently reported issue. This shift may reflect UAS technology advancements, as equipment has become more reliable and less prone to technical failures. The findings of this research also indicate that human factors and policy-related challenges have become more common. The findings of a statistically significant difference may also be affected by the fact that Wild et al. analyzed a sample of 152 sUAS-related accidents and incidents from Remotely Piloted Aircraft Systems (RPAS), the Australian accident and incident database, and there may be differences due to the operation in Australia under a different regulatory framework.

However, the evidence suggesting an evolution of contributing factors from technical issues to human factors is a compelling one and one that aligns with the evolution of safety in crewed aircraft, as mentioned earlier (ICAO, 2017) [5]. In crewed aviation, the most prevalent contributing factors to incidents and accidents have shifted from technical aspects to human-oriented errors as technology has advanced over the last century, as shown in Figure 9. The findings of this research, compared with previous research suggest that as UAS technology continues to advance, the UAS sector may be experiencing a shift with more human factors and policy-related issues becoming more prevalent as contributing factors to incidents.

4.4. Investigation of Incident Characteristics in Different Classes of Airspace

UAS operations are restricted in some classes of airspace. As shown in Figure 10a, most UAS incidents were reported in Class D (39 cases) and Class G airspace (38 cases). It is not surprising that many incidents occurred in Class G airspace since it is commonly used by UAS operators and does not require any additional authorization. The significant number of incidents in Class D airspace is concerning, however, since this is controlled airspace that serves crewed air traffic, especially GA. Class B and Class C airspace each have 23 reported incidents, which suggests the potential for a considerable safety risk due to the concentration of air traffic serving commercial crewed aircraft operations at the airports that have Class B and Class C airspace. There were 15 incidents reported in Class E airspace, which is frequently associated with non-towered airports, and 12 incidents reported in special use airspace, the majority of which were under Temporary Flight Restrictions (TFRs).

Figure 10b illustrates the distribution of contributing factors in the various classes of airspace. Human factors are the most prevalent contributing factor in nearly all airspace classes, except for Class D, where equipment issues are the most prevalent factor. Pairwise comparisons were conducted to determine if there were differences between the distribution of contributing factors in the different airspace categories. As shown in

Table 4. Chi-square results for pairwise comparisons of contributing factors by airspace (no significant difference for any pairwise comparison).

Airspace Classes	Class B, C	Class D, E	Class G	Special Use Airspace
Class B, C	N/A	p 1 = 0.14	p = 0.84	p = 0.73
Class D, E	p = 0.14	N/A	p = 0.11	p = 0.58
Class G	p = 0.84	p = 0.11	N/A	p = 0.89
Special Use Airspace	p = 0.73	p = 0.58	p = 0.89	N/A

¹ p-values above 0.05 indicate no statistically significant difference in all cases.

, the chi-square test results suggest that the distribution of contributing factors does not vary in different classes of airspace. The factors contributing to UAS incidents may be affected more by shared UAS safety challenges that transcend the airspace classification than by characteristics specific to the airspace classifications. It is reasonable to conclude that the lack of significant difference may also be due to the relatively small sample size, which limits the ability to detect differences in the distribution of contributing factors across different airspace classes.

The dataset narratives and synopses were reviewed to provide additional insight into incidents across different airspace classes. Selected examples are shown in

Table 5. Representative cases and their contributing factors (NASA ASRS, 2024).

ACN	Reason for Incident Report	Case Summary	Contributing Factors
Class B Airspace
(a) 1966160	Near Miss	While near a busy Class B airport, an observer noticed a UAS operation during a runway shift that brought northbound departures closer to the area. Departing airline traffic posed no collision risk, but piston-powered aircraft required greater vigilance. A Cessna on a crosswind departure passed within 100 ft. of the other operator’s UAS. The Remote Pilot in Command (RPIC) of the UAS promptly initiated a descent upon alert but lacked proper situational awareness and familiarity with manned flight procedures.	Env: Obstructed visibility due to high buildings and sunlight HF: Complacency and inadequate training Policy: Lack of radio surveillance or internet-based traffic data
(b) 1893543	Airspace Violation	During a mapping operation near ZZZ airport, a DJI Matrice 210 V2.0 RTK UAS ascended to 297 ft. AGL without obtaining LAANC authorization for a 300 ft. area in Class B airspace. The violation stemmed from the PIC being distracted by a low-battery warning on the primary iPad during pre-flight setup. The oversight was discovered post-flight when notifying ZZZ ATCT. No interference with crewed aircraft operations occurred.	HF: Complacency and pre-flight distractions Policy: Pre-flight checklist gaps
Class C Airspace
(c) 2111236	Airspace Violation (Exceeding Authorized Altitude)	The crew learned from an oversight during a sUAS flight where the LAANC altitude restriction was not updated after changing the flight area. While the original plan allowed for a 300 ft. AGL flight, the revised plan limited the altitude to 200 ft. the drone was mistakenly set to 279 ft. AGL. This was noticed later when reviewing the airspace and flight data. Additionally, the reporter suggested that simplifying the LAANC interface could help reduce human error by making the authorized information more straightforward.	HF: Impaired situational awareness and inexperienced crew Policy: Pre-flight checklist oversight and complex LAANC authorization system interface
(d) 2021216	Airspace Violation (Exceeding Authorized Altitude)	At approximately XA:29, a pilot climbed to 48.2 m (158 feet) during an operation to capture the exterior of a water tower in Location A. The LAANC approval allowed a maximum altitude of 100 feet. However, the pilot misinterpreted the altitude units, believing the LAANC approval used meters instead of feet, based on the FAA Visualize It website graphic, leading to an unintentional airspace violation.	HF: Unit misinterpretation Policy: Complex LAANC interface design
Class D Airspace
(e) 1884993	Near Miss	During a dual UAS operation in Class D airspace, a crewed aircraft passed within 60 ft of UAS-1 shortly after takeoff. The drone operator attempted to descend but shifted to an ascent to avoid collision. The crewed aircraft was not detected until it was near the UAS due to environmental noise, distractions from UAS communication, and unfamiliarity with the approach path.	Env: Machinery and road traffic noise HF: Unfamiliarity of crewed aircraft approach path Policy: Inadequate pre-flight preparation
(f) 2021216	Collision	During an authorized UAS flight near a school stadium, the operator heard but could not see a crewed aircraft. In a moment of panic, the operator lost control and ascended the drone, inadvertently exceeding the 400 ft. altitude limit. The crewed aircraft was later spotted approximately one mile away, after which the operator regained control, adjusted the altitude settings, and safely completed the flight.”	HF: Impaired situational awareness Policy: Inadequate separation from ATC
Class E Airspace
(g) 2095036	Airspace Violation	A drone mission near MVN to collect flood elevation data faced challenges with restricted airspace, fragmented data management, and LAANC approval issues. Despite multiple requests, only limited approvals were granted, complicating planning. Critical information on authorized boundaries was dispersed across devices, while the drone controller’s poor interface and lack of integrated LAANC data impaired situational awareness. ADS-B and CTAF monitoring mitigated traffic risks, but systemic and technological limitations hindered operational safety.	HF: Impaired situational awareness Equip: Poor user interface on drone controllers for programming flight paths Policy: Multiple LAANC rejections without feedback
(h) 2042087	Near Miss	A drone hovering under 50 ft. while filming guests preparing for a hot air balloon ride was involved in a collision with a flying hot air balloon. The flying balloon approached within a few feet of the stationary balloon being filmed, causing the drone to lose control and strike the side of the flying balloon’s basket.	Policy: Inadequate separation from ATC
Class G Airspace
(i) 1943552	UAS Crash	On a clear, windy day, a drone launched from a park in Location A transitioned to ATTI mode after losing its GPS signal. The operator lost visual tracking and control as the wind pushed the drone eastward. Attempts to activate Return to Home failed due to the GPS loss. The drone was unrecoverable, likely due to battery depletion.	Equip: Loss of GPS signal Env: Moderate winds (15 MPH)
(j) 1924069	UAS Crash	A UAS experienced a flyaway and crashed due to suspected GPS loss while operating in an area frequented by low-altitude manned air traffic. Pilots were instructed to keep the accident secret under threat of job loss. A covert recovery mission retrieved the damaged drone. Concerns were raised about the organization’s immature safety culture.	Equip: Loss of GPS signal Policy: Extremely immature safety culture
(k) 1772563	Waiver Violation	The operator flew multiple UAS under a Part 107 waiver without issuing the required NOTAM 24 h in advance. The operator mistakenly believed that a runway closure NOTAM was sufficient. Upon re-reading the waiver, the operator realized this did not fulfill specific requirements, such as location and altitude details.	HF: Oversight in pre-operation compliance checks Policy: Misinterpretation of waiver provisions
Special Use Airspace
(l) 1629713	Airspace Violation	The operator inadvertently flew a DJI Inspire 2 into a Temporary Flight Restriction (TFR) zone despite prior awareness and briefing to avoid it. A last-minute change in location and time pressures led to oversight of the TFR.	HF: Impaired situational awareness Policy: Inadequate pre-flight checks
(m) 1910447	Airspace Violation	The operator unintentionally breached a TFR near ZZZ during an active fire. Pre-flight checks relied solely on Sky Vector and Flightradar24, which did not show the TFR. The operator later acknowledged the need for diverse official sources.	HF: Reliance on limited tools Policy: Inadequate pre-flight checks
(n) 2051533	Airspace Violation	A corporate flight was conducted without a licensed operator’s supervision, disregarding warnings about restrictions and licensing. The team exceeded the 400 ft. altitude limit and bypassed the DJI Fly Zone authorization checklist.	Policy: Organizational safety culture deficiency

. There were 23 incidents reported in Class B airspace; 18 of these were attributed to airspace violations, and operators were either certain or possibly conducting operations without the required authorization. Three cases involved UAS collisions with the ground, power lines, and a bird, respectively. The most severe incident (ACN: 1966160) was a near miss with a GA aircraft during its climb, with the UAS coming within 100 feet of the crewed aircraft. This incident had a number of contributing factors, including obstructed visibility caused by the early afternoon sun, lack of surveillance measures, and the operator’s complacency and insufficient knowledge of Instrument Departure Procedures and crewed airplane performance. Table 5 provides detailed insights into this near-miss case (row a) and a representative airspace violation (row b), which highlights the interplay of policy issues and human factors that contributed to the incident.

There were 23 incidents reported in Class C airspace, the majority of which were related to airspace violations, and a notable exception was a lost link case. Four of the airspace violations involved exceeding the authorized altitude, a violation not observed in the recorded reports for Class B airspace. Two representative cases of exceeding authorized altitude are provided in Table 5 (rows c and d).

There were 39 reported incidents in Class D airspace, which represents the highest number of incidents in any airspace. Four of these were near-miss cases with crewed aircraft. The combination of these statistics suggests that the safety risks of operations in Class D airspace should be emphasized. All four near-miss cases were conducted with proper authorization, which underscores a critical concern: even when operators adhere to airspace regulatory requirements, near-miss incidents can still occur due to a mixture of factors such as unexpected, crewed aircraft movements, gaps in air traffic control (ATC) coordination, and impaired situational awareness due to unfamiliar operational environments. These concerns may be particularly pertinent in Class D airspace, which typically serves regional airports with diverse aeronautical traffic, including smaller GA aircraft and helicopters that frequently operate at lower altitudes. The various types of air traffic may increase the complexity of UAS operations in this environment. Examples of the incident reports for two near-miss cases are presented in Table 5 (rows e and f).

There were 15 incident reports recorded in Class E airspace. One case involved a UAS losing control and striking the side of a flying balloon’s basket. This near miss was partially attributed to inadequate separation between the UAS and balloon, and there is insufficient information available to determine the exact cause of the loss of control. The remaining cases were all attributed to airspace violations. Among these, one case (ACN:2095036) is particularly noteworthy, as it illustrates how multiple seemingly independent contributing factors can interact and result in an incident. This case and the hot air balloon collision are provided in Table 5 (rows g and h).

In Class G airspace, there were 20 cases attributed to equipment-related issues such as battery failures, GPS malfunctions, and lost link incidents; 10 of these resulted in UAS crashes or flyaways. In many cases, the flyaways were caused not only by technological failures but also stemmed from interactions with external factors, including strong winds, and were also caused by insufficient safety culture and impaired situational awareness. Table 5 (rows i, j, and k) highlights two representative cases illustrating crashes due to interactions of multiple contributing factors (rows i and j), along with an educational case involving a NOTAM issuance error (row k).

There were 12 incident reports in special use airspace, and all were related to airspace violations. Four out of 12 violations were identified during operations, while six were discovered post-flight. The primary causes of these violations stemmed from human factors and policy issues, such as inadequate pre-flight planning, impaired situational awareness, and over-reliance on third-party applications that occasionally failed to provide updated TFR information. One narrative highlighted a severe deficiency in the operator’s company safety culture (row n), involving a mission knowingly conducted in restricted airspace without proper authorization or a licensed operator. This case, along with two representative reports of airspace violations (rows l and m), is summarized in Table 5.

5. Discussion and Conclusions

Fortunately, there have been no catastrophic UAS accidents since Part 107 was published. However, the lack of accidents correlates with sparse information about the risks of UAS operation. Building on Heinrich’s (1972) Iceberg Theory [7], which suggests that catastrophic accidents often result from the accumulation of minor errors, violations, and hazards, it is imperative to proactively examine incident reports to identify and address latent safety deficiencies. This approach is particularly critical for the emerging UAS sector, where safety protocols and practices are still evolving. This research addresses UAS incidents and documents characteristics that are associated with these incidents and may affect UAS safety and operations.

Despite the obvious differences between UAS operations and crewed aircraft (e.g., size, speed, and agility of aircraft; (lack of) people on board; remote pilot vs. pilot on board), there are notable similarities related to safety considerations between UAS and crewed aircraft. There are also many lessons learned in the crewed aircraft sector that are relevant to UAS. Similar to crewed aircraft, most UAS incidents analyzed in this study resulted from multiple contributing factors, which can be described and assessed using established safety management models, such as the SHELL and the “Swiss Cheese” models.

A comparison of findings from this research with previous research by Wild et al. (2016) [57] suggests a shift in the prevalent contributing factors, suggesting the UAS sector may be evolving from predominantly equipment-related issues to issues related to human factors and policy-related challenges; this may reflect maturing UAS technology. The prevalence of human factors aligns with other early analyses of UAS incidents, which attributed 60% of 221 recorded cases to human factors [49]. Interestingly, this trend in UAS incidents aligns with safety trends in the crewed aircraft sector and underscores the need for increased focus on human factors (e.g., improving operator situational awareness) and policy issues (e.g., developing robust checklists and organizational safety policies) for UAS training and support [5,54,67]. Just as the FAA provides sample templates to support airport safety through regular inspections in Advisory Circulars, it may be appropriate for the FAA to provide sample templates and checklists for UAS operations.

A closer examination of human factors and policy issues suggests gaps in safety culture at both the operator and organizational levels. At the operator level, an “invulnerability attitude” was identified, with some individuals knowingly committing violations by bypassing required authorizations or making minimal efforts to obtain the proper authorizations. At the organizational level, deficiencies in safety culture were evident in reports such as ACN: 2051533, in which a corporate owner led an unauthorized operation without licensed operators despite prior warnings, and ACN: 192406, in which operators were threatened with job loss if they disclosed a UAS flyaway incident. Although these safety culture gaps were explicitly highlighted in only a few reports, they warrant attention. These behaviors may be less likely to be self-reported, especially if the report is filed by an employee to meet an employer’s requirement. Concerns related to an “invulnerability attitude” take on a different dimension in UAS operations, where violations usually do not directly endanger the operator’s personal safety [44].

When examining the timing of the incident discovery, it is not surprising that human factors and policy issues, such as impaired situational awareness and operational procedural deficiencies, are often realized during post-flight inspections. In contrast, equipment and environmental problems are more likely to be noticed immediately after occurrence during the flight. This finding underscores the importance of both routine inspections before flight and post-flight to detect intricate safety deficiencies; pre-flight and post-flight inspections are both vital components of the hazard identification process in crewed aircraft SMS [68].

Most incidents occurred during the daytime; however, seven nighttime incidents were reported, all of which followed the FAA’s authorization of night operations in 2021. Of these, six (86%) were associated with public safety missions. This suggests that outreach and education regarding the risks of nighttime flying and best practices for safe nighttime flying may be appropriate for public agencies. It also suggests that it may be appropriate to find ways to expedite emergency UAS authorization procedures (such as Special Government Interest (SGI) waivers) to address the challenges of obtaining timely approvals for urgent nighttime missions. Several incident reports highlight these authorization delays, a challenge that has persisted for years and was also emphasized by Wild et al. (2016) [57].

In terms of temporal factors related to UAS exposure, incident reports peaked during the summer months, which is consistent with seasonal patterns identified by Wang and Hubbard (2021) [59] and Pitcher (2022) [61]. However, a comparison with Wang and Hubbard (2021) [59] reveals statistical differences in monthly distributions, with this study recording a higher proportion of reports from October to December. This variation may suggest that as UAS evolves, it may become more of a year-round activity, potentially driven by expanded commercial applications and improvements to UAS battery life. Additionally, the increase in reports during October-December may be an anomaly to the limited sample size of this study rather than a trend; the dataset from the FAA UAS sightings report used by Wang and Hubbard [59] is 35 times larger than the one analyzed here. Further research is needed to confirm trends related to reported incidents by time of year and to identify whether the time of year poses different risks due to different environmental factors (e.g., snow, low temperature, etc.) or human factors (e.g., remote pilot response to extreme heat or cold).

In the comparative analysis between Part 107 and recreational operations, the distribution of contributing factors and the timing of incident detection were similar. This suggests that the general safety considerations for UAS may supersede any other differences (e.g., purpose of operation, training requirements, etc.). It was notable that 35% of incidents in recreational operations were reported as detected post-flight, which indicates a robust and commendable safety culture among recreational operators as it relates to post-flight inspections. However, this observation may reflect a potential bias, as recreational operators who submit reports are likely those with stronger safety practices, including more frequent post-flight inspections. While Part 107 operations were more often associated with larger crew sizes (perhaps due to the increased complexity of their missions), this did not result in a higher proportion of incidents with human factors as a contributing factor, which may suggest that Part 107 operator training, safety policies, and safety culture are effective in managing some human-related errors.

This research provides insight into the characteristics of UAS incidents in different classes of airspace, which addresses a significant gap in previous UAS research. Airspace violations emerged as the most common infringement in controlled and special-use airspace; these violations were often committed inadvertently, as shown in

Table 6. Incident characteristics by airspace class.

Airspace Class	Number of Incidents	Notable Circumstances
B	23	18 airspace violations (78%) Three collisions (ground, powerlines, and bird) One near miss with crewed GA aircraft (4%)
C	23	21 airspace violations (91%) Four of these exceeded the authorized altitude One lost link
D	39	30 airspace violations (77%) Four near misses with crewed aircraft (10%) Two UAS crashes One bird strike risk
E	15	14 airspace violations (93%) One collision with balloon (UAS loss of control) (7%)
G	38	Five exceeded the authorized altitude 20 incident reports with equipment issues (e.g., battery failure, GPS malfunction, lost link) resulted in 10 UAS crashes or fly-aways One operating with an expired license One flying over people One operating without a beacon at night One single operator managing multiple UAS without NOTAM One near miss with crewed aircraft (crop duster) (3%)
Special Use	12	12 airspace violations (100%)

. These issues are attributed mainly to human factors and policy issues, including misinterpretation (of the airspace map and current location), inadequate training, and gaps in pre-flight checklists. Challenges with the LAANC applications, including their complex interface and occasional difficulties in obtaining real-time authorization, exacerbate these human factors and policy issues, potentially warranting attention from authorities and LAANC providers, as highlighted in several reports from this study (e.g., ACN 2111236 in Table 5, row c), the remote operators’ discussion forum, and prior research [69,70]. This suggestion to expand the capabilities of LAANC is made with an acknowledgment that LAANC has been a resounding success overall, enabling near real-time authorization for many UAS operations.

For uncontrolled airspace, Class G airspace presents a more diverse range of issues, including UAS flyaways, crashes, flyover standbys, night operations without beacon lights, and exceeding authorized altitude. The diversity of issues, combined with the large number of violations, suggests that UAS operation and incidents in this airspace may be worth additional study to better understand the complexities of UAS operational safety, especially given that it is the most common environment for recreational UAS operations.

Near misses with crewed aircraft are a very real concern in UAS operations, underscoring the significant challenges of safely integrating UAS into the NAS. Most near-miss incidents occurred in Class D airspace, which serves regional airports with diverse air traffic operations, including GA aircraft and helicopters frequently operating at lower altitudes. Notably, all four near-miss cases involved authorized UAS operations, emphasizing that ensuring safety in complex operational environments is a shared responsibility. This requires due diligence and coordination among remote operators, aircraft crews, and ATC, as Williams (2019) [71] discussed.

One near-miss (ACN: 1966160, in Table 5, row a) in Class B airspace highlights a UAS operator’s lack of understanding regarding crewed aircraft departure procedures; this understanding is essential for UAS operations in complex environments such as Class B airspace. This finding contributes to the ongoing debate regarding the scope of the Part 107 exam [31], as discussed earlier, and underscores the benefits of incorporating knowledge of crewed aircraft operational procedures into remote pilot training. Just as pilots hold certifications with various types of ratings, maybe UAS remote pilots licensed under Part 107 should be required to have additional type ratings (and correlating exams) for operations in some kinds of airspace; in this case, LAANC authorization could be connected to the remote pilot ID, and the LAANC authorization would consider these certifications when providing approvals.

Two primary limitations of this research stem from the nature and size of the ASRS UAS dataset. The limited sample size limits the application of statistical models (e.g., log-linear analysis and other statistical tests) to explore the impact and interaction of incident characteristics and contributing factors, including incident characteristics in different airspace classes. A larger dataset is required, particularly when examining complex or higher-order relationships. Although limited in size, this data and the associated analysis provides a lot of interesting and useful information, and the narratives lend a valuable context to the data and offer a lot of additional insight. Future research may uncover and address additional insights into risks and safety considerations, such as gaps in safety culture, which are challenging to capture through statistical models. As more reports are filed in ASRS UAS, future studies can build on this work by running this analysis with additional data and by expanding the analysis to employ more advanced methods, such as lexical analysis and machine learning. As time goes on, the ASRS database will include more UAS incident reports. This larger dataset would enable time series analyses within the ASRS UAS system across different time periods, which would allow researchers to explore trends and identify shifts in operational characteristics and safety considerations over time. These methods can help identify patterns and may provide a better understanding of the underlying causes of incidents; future research can also make significant contributions by leveraging the rich information provided in the narrative reports.

While the self-reported nature of the UAS ASRS provides a valuable resource for disseminating safety information, it may also introduce potential bias since only operators with strong safety awareness are likely to report incidents to the UAS ASRS; it is likely that many incidents and violations are unreported. In addition, the self-reported and non-punitive nature of the reporting system greatly encourages reporters to share their experiences, providing valuable insights and lessons learned. However, the inherent subjectivity of self-reporting remains a limitation. Furthermore, the lack of standardized reporting practices contributes to a lot of variability in incident severity; reported incidents range from critical near misses with crewed aircraft to minor issues such as temporary data losses during operations. Adding a case severity selection feature to the report forms could facilitate future analysis by categorizing incidents uniformly. Just as runway incursions are labeled A, B, C, and D with standard definitions, it may be appropriate to develop definitions and severity ratings for UAS incidents. This improvement would lay a strong foundation for future research in safety risk analysis and the development of UAS-tailored SMS.

UAS are a recent addition to our NAS, but they are an important user. FAA has numerous successful programs to support education and safety culture for crewed aircraft pilots and aviation maintenance professionals. These programs include WINGS (a pilot proficiency program), the Aviation Maintenance Technician Awards Program (AMT), the General Aviation Awards Program, the Wright Brothers Master Pilot Award, and the Charles Taylor Master Mechanic Award [72]. Perhaps similar programs should be initiated to raise the profile and promote safety culture for remote pilots and recreational fliers, building on the foundation of the UAS initiatives the FAA has already launched (e.g., UAS symposium, national drone safety awareness week, etc.).

Further research is recommended to explore whether analysis of reported incidents should be the same for Part 107 pilots and recreational pilots since Part 107 pilots may be held to a higher standard, which is analogous to crewed aviation. Furthermore, leveraging artificial intelligence-based solutions, such as automated compliance checks, predictive analytics for risk management, and real-time operational monitoring, may provide innovative avenues to address safety gaps effectively. In this scenario, software could integrate the operation of the UAS with questions to assure important safety checks have been completed, as is pertinent to the mission, environment, and operational framework.

Some of the findings from this research may be generalized to other countries, including considerations related to safety issues related to human factors such as fatigue, communication, and complacency; environmental considerations such as wind gusts, cold weather, and obstructions; and policy considerations such as a robust framework for safety checklists, post-flight reviews, and organizational safety culture. Other findings may be less generalizable to other countries, such as safety issues related to the use of LAANC, which is unique to UAS operations in US airspace.

UAS may be used for a wide variety of applications, and the sector is expected to continue to grow in the future. Additional research to support safety and education in UAS will pay significant dividends not only for UAS operators but also for all NAS users.