UCAV Mishaps: Role of Human Factors
There is substantial history of unpiloted, powered aircraft operations, by many of the world's militaries, following there appearance in the first decades of the 20th Century. Unquestionably there has been a rapid rise in the use of Uninhabited Aerial Vehicles [UAV’s] but along with the perceptions of greater use and versatility come hard facts of high mishap rates.
The accident rate of unmanned flight is far greater than for manned aircraft. Certainly remote controllers have a dramatic disconnect to the touch feel and situational awareness of the pilot manning unique flight. High accident rates can and will compromise mission availability. Office of the Secretary of Defense’s UAV Reliability Study did the most comprehensive review of UAV mishaps to date reporting human error induced mishaps at 17% [Predator HF factors appear higher] though providing no breakdown of these human factors.
Pentagon unhappy about drone aircraft reliability: rising mishap rates of unmanned vehicles attributed to rushed deployments.</a>
Early Accident Rate
In 2003 the Department of Defense (DoD) reported the period 1986-2002 accidents per 100,000 hours. The accident rate of the Predator has fallen since this report was commissioned, but remains approximately 10 times the rate of General Aviation.
|1986-||-2002||Mishaps per 100,000 hrs.|
Image 1 : A group photo of aerial demonstrators at the 2005 Naval Unmanned Aerial Vehicle Air Demo.
It was hoped a pooled analysis of studies might provide Department of Defense (DoD) greater perspective of human factors in UAV mishaps but this was not possible due variety of human factor taxonomies employed. A purpose of this topic is to discuss a quantitative analysis [Military Unmanned Aerial Vehicle Mishaps, 2005] with forum of further discussion of the role and patterns of active and embedded human failures in UAV mishaps within the U.S. military services using standardized human factors taxonomy. However it is intended to provide statistics and comment on the research beyond.
N.B. emphasis on unmanned rather fully automated.
Study protocol Tvaryanas, Thompson, and Constable (2006) was approved by the Brooks City-Base Institutional Review Board in accordance with 32 CFR 219 and AFI 40-402. The study design was a 10-year cross sectional quantitative analysis of UAV mishaps using the DoD Human Factors Analysis and Classification System (DoD HFACS) version 5.7 taxonomy.
Inclusion criteria was U.S. Air Force, Army, or Navy/Marine UAV Class A, B, or C severity mishap occurring during and including financial years 1994-2003. DoD made available records and databases pertaining, of which 221 UAV mishaps were submitted for further analyses, each double rated independently and classified all human causal factors using the DoD HFACS. Cases of disagreement were reconciled to form consensus classification, embedded within study database.
'Unmanned Aircraft' is standardized to refer to the airborne element of the system whereas the term 'Unmanned Aircraft System' [UAS] will be used to reference the entire system, that is inclusive of Ground-Based Components be it a GCS (Ground Control System), communication links and support equipment.
How much full automation [not just remotely controlled] do we allow increasingly sophisticated, faster, more lethal unmanned aircraft (or submarine) that holds sufficient artificial intelligence to self interrogate and self correct?
Such debate must continue with such statistics telling us to date unmanned military Ariel mishaps are occurring at alarmingly higher rates.
Overall a little over 60% of mishaps were judged to involve Human Factors. Below Table 2 indicates latent error rates detected with associated nanocodes for the USAF.
Table 2. : Root categories of latent error and associated nanocodes of USAF.
|Model Variables||Associated Nanocodes||Human-Factor Mishaps|
|Technological Environment (P = 0.001)||47.1%|
|Instrumentation & Sensory Feedback Systems||26.5%|
|Cognitive Factors (P = 0.009)||26.5%|
The military services operate their respective UCAV's in distinctly different variations. The machine operator for the Airforce will view flight from a centralized camera in the nose. Whereas Army and Navy view the unmanned aircraft externally as though from the runway. This factor alone can grossly disproportion the appointment criteria of mishaps and whether attributable to human factors? Without standardization it made incidents harder to compare between services.
It is noteworthy the Army’s UCAV investigations into mishap policy prior to 2003 referred to them as ground not aviation incidents and simply involved checking 'appropriate' boxes on forms provided. This has likely caused an under representation of true frequency.
A major finding of the study was that organizational factors were present in two-thirds of Air Force UCAV mishaps followed by one-half of Navy/Marine mishaps, mainly involving acquisition policies and processes. These organizational factors were present in but one-quarter of Army mishaps a possible distortion as above.
It was established the UCAV operator has limited sensory perception compared to the pilot on a manned aircraft being a distinct disadvantage. NASA tested placing a microphone within the UCAV with a feed to the ground operator which was enormously helpful.
Given the Predator flight simulator as at 2004 (reported by USAF Safety Center Predator mishap report) does not accurately reproduce handling characteristics of the actual vehicle so it recommended acquiring a virtual simulator with high-fidelity to mirror vehicle handling characteristics and thus increase operator proficiency.
The alternative was to automate landing phase of flight to eliminate the need for proficiency in the landing skill set and thus decrease landing mishaps. Again this alters the criteria for attributing further mishaps as human factor related. General Atomics Predator UAV's cost approx. $nz3.2 million though the ground mission control unit will cost far more.
Teleoperation here refers to the control of the UCAV by the UAS with a remotely located human operator. Teleoperation decouples operator safety from the fate of the remote aircraft. Freed of the constraints of this shared fate operators or their commanders may unintentionally pose survival risks not normally contemplated by an onboard pilot. Ground controllers may in extreme circumstances intentionally use the aircraft as a weapon or otherwise intentionally destroy the remote craft. Most UAS possess 'flight termination systems' often an engine kill switch though other methods can bring the aircraft down quickly. Of note the first loss of a Teledyne Ryan Global Hawk, introduced following flight tests in 1998 with approx. cost of $nz10million for the UAV and a further $nz20-25 million for the ground mission control unit, occurred when it's flight termination system was inadvertently activated.
Image 2 : Teledyne Ryan Global Hawk
The study commissioned of military UCAV mishaps identified recurring factors being organizational, supervisory, and pre-conditioned levels all needing to be addressed in order to make UCAVs a greater success and improve reliability and sustain usability.
Predator’s, first procured in 1995, propulsion subsystem has caused the vast majority of the system losses that were not combat losses, findings from the Defense Science Board.
A major finding of the study was that organizational factors were present in two-thirds of Air Force UCAV mishaps
However although organizational factors were largely at fault it was not directly from the unreliable units but identified as further up the chain at acquisitions level the problems originating in vendor suppliers not been given scope of reliability.
Lastly the study identifies a new frontier surrounding aviation human factor issues involving UCAVs.
Anthony P. Tvaryanas, M. M. (2005 , March). (M. Bill T. Thompson, Ed.) U.S. Military Unmanned Aerial Vehicle Mishaps: Assessment of the Role of Human Factors , 15. Alan Hobbs Ph.D San Jose State University, Human Systems Integration Division., Chapter 16 Human Factors in Aviation 2nd Ed.. 2010.
Contributors to this page
MADDEN, Keith student Massey University