This study by Tvaryanas, Thompson and Constable considered the human factors in Remotely Piloted Aircraft (RPA)mishaps. The military use of RPA’s also know as Unmanned Aerial Vehicles (UAV) or Unmanned Aerial Systems (UAS) has increased as technology has advanced. As the use has increased so to have the occurrence of mishaps. Compared with general ‘manned’ aviation the figures are thirty to three hundred times the rate per flight hour. This high mishap rate has been identified as one of the two biggest risks to realising the full potential of RPA’s.
While there is an absence of a human in the aircraft and so the potential for injury or fatality is greatly reduced, this high mishap rate affects the feasibility of this type of aircraft because they are neither expendable or disposable.
The purpose of this study was to define the distribution and determination of operator error in RPA mishaps. It used a standardised human factors taxonomy and hierarchical model of human error. This model, know as Human Factors Analysis and Categorisation System developed form the Swiss cheese model published by Dr James Reason.
The study used data from the fiscal years 1994 – 2003 from four of the five armed services (coast guard not included). It considered only mishaps resulting in greater than US$20,000 in damage. An RPA was as defined by the US Department of Defence(DoD). The information was collated from the safety centres for each of the armed services. Navy and marine RPA are combined as one force. Target drones were not included in the study. As a result 221 mishaps were considered for the study.
The data from the safety centres was analysed by two independent raters to determine the casual factors according the DoD HFACS model. These factors are not limited to the four levels of human failure but are further resolved to root-level categories. A single mishap could have several contributing factors and each was considered.
Human factors was the sole cause of 24.0% of mishaps while mechanical failure was attributed as the sole cause of 31.7%. While in 8.1% the cause could not be identified the remaining 36.2% of mishaps it was considered to be both Human factors and mechanical failure. Of the four hierarchical levels it was the unsafe acts which produced the biggest variation.
In theory a undesirable event requires failures in all levels of of the HFACS hierarchy of human factors. From this study one specific level showed significant variation between services. In this discussion the unsafe acts detailed in the results section will be focused on. Unsafe acts can be divided into errors or violations. The difference between the two is the intent. Error are considered to be unintentional by nature while violations are wilful. These can again be divided into further sub-categories. Errors are considered to be either skill-based errors, decision errors or perceptual errors. Violations can be either routine or exceptional.
The results show a significant difference in both the skill-based errors and violations. Results for the other sub-categories were not included in the study but it was noted that there was not a great difference in the proportion of human factor mishaps involving decision making and judgemental errors. The high skill-based errors recorded by the Air Force were unexpected. in contrast to the other two services the air force uses experienced commissioned pilots while the army and navy\marine operators are trained from those with no piloting experience and given specific RPA training. This high skill-based error rate could be due to the over reliance on senses not available in RPA such as vestibular and peripheral views.
The higher rate of violations in the army mishaps over the air force and navy\marine could be contributed to a number of factors. The first that needs to be considered is the method of recording the mishaps. By nature errors have system based causal factors while violations are more individual being wilful by nature. It could be that the army reporting system tends to consider the intent of the individual in a different manner to the other services. I could also be a higher acceptance of risk by the army and willingness to violate rules and regulations.
This review of the reference study draws two conclusion
- That experience as a pilot is not a preferable precondition for RPA operations.
- The differing culture and reporting methods between the service does not allow for balance results.
TVARYANAS Anthony P, THOMPSON William T, CONSTABLE Stephan H. Human Factors in Remotely Piloted Aircraft operations: HFACS Analysis of 221 mishaps over 10 years. Aviation, Space, and Environmental Medicine, July 2006.
Contributor to this page