Analysis of Civil Aviation Accidents to Avert Probable Aviation Tragedies

Abstract 

A recently reported collision at Heathrow between a civil drone and an airliner – Airbus 320 owned by the British Airways brought to the fore the need to comprehend incidents and accidents associated with civil remotely manned aircraft systems (RMAS). Such and understanding will go a long way in enabling improvements in aviation safety by ensuring that all initiatives are focused on reducing extreme risks. 152 RMAS incidents and accidents were analyzed. The data used was for a ten-year period spanning from 2006 to 2015. The findings indicate that, as opposed to commercial air service (CAS), RMAS accidents and incidents have a significantly different distribution when they are classified based on the type of occurrence, safety issue, as well as flight phase. In particular, this analysis found that operations of RMAS are very susceptible to events such as in-flight loss of control, accidents and incidents during takeoff, as well as when in cruise, and problems with equipment. It was shown that issues of technology were, as opposed to human factors, are the major contributors of RMAS accidents and incidents. This is an important finding, since it differs from the industry view, which for a long time has considered human factors as being the key contributor to aviation accidents and incidents – this however, remains the case in CAS. In this regard, industry regulators need to look at technologies as well, rather than focusing only on operators.

Analysis of Civil Aviation Accidents to Avert Probable Aviation Tragedies

Introduction 

In April 2016, it was reported that an Airbus 360 owned by the British Airways was struck by a drone when it attempted to land at Heathrow Airport at around noon. According to Clothier & Walker (2010) , drones are a risk to human life because they can collide with other aircrafts or with people on the ground. The reported event at Heathrow highlights the drone risk of colliding with other aircrafts. These remotely piloted aircrafts or drones, have been referred to in the industry by many names including unmanned aerial vehicle, unmanned aircraft system, amongst many others. Recently, however, the International Civil Aviation Organization adopted the name Remotely Manned Aircraft System (RMAS) (ICAO, 2015) .

Consequently, ICAO (2015) also noted that civil RMAS market space is rapidly growing to the extent that it is no longer limited to those flying radio-controlled aerial vehicles as a hobby. In fact, its applications have increased to include activities such as the monitoring of traffic, oceanography, monitoring of wildlife, firefighting, as well as volcanography (Valavanis & Vachtsevanos, 2014) . Furthermore, RMAS are increasingly being applied, as has been extensively discussed in academic journals and news media, in package delivery such as what Amazon is doing with its Prime Air, as well as for the general delivery of mail (Boršˇcová & Draganová, 2014) . For this reason, the volume of traffic, especially of RMAS, in densely populated cities and other urban areas has significantly grown and is likely to rapidly increase in the years to come. Thus, an understanding of the risks that will accrue as a result of RMAS operations is critical to ensure that safety is improved and can be guaranteed. It is unfortunate that extensive research that has been done on RMAS application in the military has not been in any way matched to civilian application of RMAS despite the apparent growth. In fact, civilian application of RMAS has remained almost unnoticed in academic literature (Klauser & Pedrozo, 2015) .

The purpose of this study is to assist in ensuring that civilian RMAS operations safe not just for the aviation industry, but also the entire general population because operations of RMAS take place in and around citizen populations. A comprehension of the differences between accidents and incidents in CAS and RMAS will go a long ay in helping the aviation industry reduce the number of accidents and incidents associated with RMAS. The main research question that this study attempts to answer is to compare the distribution of common factors between RMAS and CAS accidents and incidents over a period of ten years, with regards to the type of occurrence, issues of safety, as well as the stages of flight such as take-off and cruise.

Literature Review 

Previous studies such as that of Clothier & Walker (2014) have highlighted the need and the importance of reactive strategies in helping improve safety in the aviation industry. In fact, according to the study by Clothier and Walker in which they examined a sample of accidents associated with and classification of military RMAS from Wiegmann & Shappell (2011) , the authors found that about 61% of the 221 cases they reviewed had human factor as a major contributor. The aim of their study was to identify risks and suggests how they could be alleviated prior to them actually occurring.

Similarly, in another study by Boyd which used a post-incident explorative research, risk factors and causes of deadly accidents in private double-engine aircrafts were examined. In this study, the author found that there was a potential dearth in major instruction areas relating to multi-engine assessment training curricula. Due to the recommendations in this study by Boyd, regulators are today able to disseminate safety information in order for flight training agents can relook at their materials and approaches.

There are very many reporting bodies in the aviation industry such as the International Air Transport Association, airline manufactures such as Airbus and Boeing, as well as local and national agencies; for instance, the US Federal Aviation Administration. These bodies gather and report data for accident and incident from various aviation segments and for different classifications. They also report statistics related to safety. In fact, the findings form their data have reported that more than 70% of all aviation accidents and incidents are in one way or another associated with human factors (Clothier & Walker, 2014) .

A Review of Methods Used 

An investigative research design was utilized in investigating the spread of factors involved in accidents and incidents associated with civil application of RMAS. This multifaceted approach began with a qualitative stage where data for civil RMAS events was collected. This was then followed by an evaluation of the data using content analysis, which helped in determining the prevalent themes and trends (Leedy & Ormrod, 2013) . The qualitative stage ended with the data coding which involved classifying all the factors of concern, particularly, types of occurrence, issues of safety, as well as stages of flight. This was then followed by the quantification and examination of the classified data to determine prevalent causal factors, as well as to find out whether or not any statistical significance existed.

The quantitative stage included the statistical analysis of data using the Pearson’s Chi Square test for Goodness of Fit (GoF). The data for RMAS was designated as the observed data (R) while the CAS data represented the expected data (C) (EASA, 2014) . The tested hypotheses were given as:

(H 0 ) - Null hypothesis: P RMAS, n = P CAS, n 

(H A ) - Alternative hypothesis: P RMAS, n ≠ P CAS, n 

In which P referred to the proportion of the n th category for both RMAS and CAS. In this regard, (H 0 ) - null hypothesis can be stated as:

(H 0 ) - Null hypothesis: The proportion of RMAS accidents and incidents equals the proportion of CAS accidents and incidences, for all the different classes 

On the other hand (H A ) - Alternative hypothesis can be stated as:

(H A ) - Alternative hypothesis: All the proportions of RMAS accidents and incidents are not equal to the proportion of CAS accidents and incidences, for all the classes. 

Thus, the chi square test for GoF, also denoted as x 2 , is (Berman & Wang, 2011) :

(1)

Where there are n class (4 for stages of flight, 4 for issues of safety, and 7 for the types of occurrences). In order to determine the available Degrees of Freedom (DoF), often denoted as v, for every test, n-1 is used. Then, to get the Critical Value (CV) from the DoF, the chi square statistic table was used with a 95% confidence level. In the end, where the CV was greater than, the null hypothesis was accepted, if not, (H 0 ) was not accepted.

For every subset of C and R – the expected and the observed, the differences in comparative percentage, also denoted as deltas (∆) were determined using:

(2)

In order to get the i th class percentage, the data points of R were divided by the summation of all values of R; the result was then multiplied by 100%. Henceforth, the delta ∆ value results in a direct error for the percentage of R. Therefore, this enabled a side-by-side evaluation of the observed data for RMAS for every class and what would be expected for a random sample of CAS. This means that when there is a positive ∆, a conclusion that an RMAS accident or incident is possible than a CAS accident or incident in that class can be drawn. On the other hand, a negative ∆ points to a conclusion that an RMAS accident or incident is less probable compared to a CAS accident or incident in that class.

Results 

Figure 1 represents a breakdown of all the 152 cases gathered. Comparing the accidents and incidents for RMAS, it is clear that 24% were accidents, while 76% were incidents, which almost similar to what EASA reported as the breakdown for CAS accidents and incidents (EASA, 2014) . For both the classes of occurrence types and issues of safety, there is a clear difference between the proportions of CAS and RMAS. Regarding the stages of flight, a small increase in events related to cruise and takeoff is apparent for RMAS accidents and incidents.

Ten chi squared tests were carried out in total. The first test compared the amount of reported accidents and incidents for RMAS to those of CAS. The equivalent chi squared value was determined as 0.52, with two classes which results in a singled DoF, and 3.84 as the CV. Thus, at a 95% level of significance, (H 0 ) - null hypothesis is accepted; in other words, the proportion of accidents and incidents with respect to RMAS equals that for CAS.

Nine more chi squared tests were carried out using subsets of data, classified as either accident, incident, as well as combined total. The classes that were subsequently investigated included safety factor, occurrence, as well as the stages of flight.

Figure 1 :Accidents and Incidents distribution 

The results of the nine statistical tests are shown in Table 1. The table contains the results from the chi squared test, including the DoF and the associated CVs. For easy comparison, a conclusion is also included. A vertical examination of e table shows that for the class, “type of occurrence”, each of the tests results is greater than the CV. Therefore, (H 0 ) is rejected; thus, a conclusion is arrived at that at 95% level of significance, the proportions of RMAS incidents and accidents differs from the proportions of CAS. With regards to stage of flight, the test result is below the CV leading to a conclusion that the number of incidents throughout various stages of flight are the same for CAS and for RMAS, at the 95% SF. In contrast, however, the test results for combined total and accidents are both more than the CV. Therefore, a conclusion can be arrived at that for all the events combined and for the accidents, the number of events throughout all stages of flight for RMAs differ from those CAS at 95% level of significance. With regards to issues of safety, the statistics of each of the tests surpasses the CV. Therefore, the (H 0 ) must be rejected and a conclusion be made that for all issues of safety, the proportions of RMAS events differ form those of CAS at the 95% level of significance.

Table 1 : Results from x 2 for the three factors (issues of safety, type of occurrence, and stage of flight) 

The results for the types of occurrences, particularly using classifications of GSIE are depicted in Figure 2. Results for the stages of flight are shown in Figure 3, while those for issues of safety are illustrated in Figure 4.

Figure 2 : Comparative % differences between (R) and (C) (accidents, incidents, and combined for each of GSIE classification. 

Figure 3 : Comparative % differences between R and C (accidents, incidents, combined total for every stage of flight) 

Figure 4 : Comparative % differences between R and C (accidents, incidents, and combined total for each issue of safety) 

Conclusion 

The objective of this analysis was to investigate the incidents and accidents occurring within the RMAS sector in the aviation industry. the motivation was the need to improve and guarantee safety in the civil operations of RMAS, thereby guaranteeing safety for the flying and the general population. The study attempted to answer the question: “does the distribution of common factors between RMAS and CAS accidents and incidents over a period of ten years, with regards to the type of occurrence, issues of safety, as well as the stages of flight such as take-off and cruise, compare?”. The findings from the quantitative data analysis suggest that all the classes for RMAS significantly differed for type of occurrence, issues of safety, and for stages of flight with regards to the number of combined total events, and particularly with regards to accidents. With regards to incidents in RMAS, issues of safety and type of occurrence were found to be statistically significant; stage of flight, on the other hand, was determined to be somewhat an unclear case.

Since the number of events were determined to be statistically momentous and very dissimilar from what is reported for CAS, major differences can be identified. For instance, it was apparent that in RMAS, in-flight loss of control and operational damages were very common. In fact, accidents and incidents in RMAS were as a result of various equipment challenges as opposed to human factor which was common in CAS. With regards to stage of flight, it became clear that most accidents and incidents in RMAS occurred during take-off and as well as when aircraft was in cruise; this was relative to the case in CAS.

The most obvious and important conclusion from these findings, therefore is that regulators in the aviation industry need to seriously consider licensing operators of RMAS, regardless of their type or form. While this might be an obvious solution with regards to improving skills, knowledge, competencies, as well as all those attributes that impact human fact, this contrasts the findings of this study. More emphasis should be put on technical issues including operation-worthiness of aircrafts, as well as on the integrity of communication channels. This, as suggested in this study, has the potential of producing greater safety dividends that only focusing on human factors.

References 

Berman, E., & Wang, X. (2011). Essential Statistics for Public Managers and Policy Analysts. Thousand Oaks, CA: SAGE Publications. 

Boršˇcová, D., & Draganová, K. (2014). Utilization possibilities of unmanned aerial systems in postal and parcel services. Acta Avion, 16 (2). 

Clothier, R., & Walker, R. (2010). Determination and evaluation of UAV safety objectives. 21st International Unmanned Air Vehicle Systems Conference , (pp. 18.11–18.16). Bristol, UK. 

Clothier, R., & Walker, R. (2014). Safety risk management of unmanned aircraft systems. In K. Valavanis, & G. Vachtsevanos, Handbook of Unmanned Aerial Vehicles (pp. 2229–2275). Dordrecht: Springer Netherlands. 

EASA. (2014, September 8). Annual Safety Review 2014. Retrieved March 9, 2019, from asndata.aviation-safety.net: http://asndata.aviation-safety.net/industry-reports/EASA-Annual-safety-review-2014.pdf 

FAA. (2013, March 14). Integration of Civil Unmanned Aircraft Systems (UAS) in the National Airspace System (NAS) Roadmap. Retrieved March 10, 2019, from FAA: https://www.faa.gov/uas/media/UAS_Roadmap_2013.pdf 

ICAO. (2015). Manual on Remotely Piloted Aircraft Systems (1 ed., Vol. 10019). Montréal, QC: International Civil Aviation Organization. 

Johnson, C., & Holloway, C. (2014). On the Over-Emphasis of Human ‘Error’ As A Cause of Aviation Accidents: ‘Systemic Failures’ and ‘Human Error’ in US NTSB and Canadian TSB Aviation Reports. NASA. Hampton, VA: NASA Langley Research Center. 

Klauser, F., & Pedrozo, S. (2015). Power and space in the drone age: A literature review and politico-geographical research agenda. Geogr. Helv., 70 (2), 285–293. 

Leedy, P., & Ormrod, J. (2013). Practical Research: Planning and Design (10 ed.). Boston, MA: Pearson Education Inc. 

Valavanis, K., & Vachtsevanos, G. (2014). Handbook of Unmanned Aerial Vehicles. Dordrecht: Springer Netherlands. 

Wiegmann, D. A., & Shappell, S. A. (2011). Human Error Analysis of Commercial Aviation Accidents: Application of the Human Factors Analysis and Classification System (HFACS). Aviation, Space, and Environmental Medicine, 6 , 1006-1016.