The Human Challenge with Complex Systems

15 November 2010

The importance of the human-machine interface increases as systems become more complex – capable of formerly ­impossible operations. Understanding human factors (HF) issues becomes critical, not only operational capability, but also to survival.

In many cases, modern automated systems intended to reduce high levels of operator workload have created environments in which the workload is so low under normal conditions that the human finds it difficult to remain actively engaged at all times. While this can increase performance under normal conditions, it makes system failures even more dangerous as recovery is extraordinarily difficult for an out-of-the-loop operator. This issue, and other challenges associated with complex systems, are often exacerbated by the capability of new systems to deliver vast amounts of data at a rate that exceeds the human ability to process and comprehend. Such problems are especially acute in the provision of situational awareness information intended to aid pilots operating in very low visibility conditions.

Over the past several decades, Canada’s Air Force has successfully accomplished a wide variety of difficult missions in challenging environments – both at home and worldwide. As with any air force, or even any civilian aircraft operator, the key element in this success has been the personnel, from the pilots to the maintainer crews that ensure the aircraft are safe and ready to fly.

To keep aircrews safe and performing optimally, military and civil aircraft have been steadily updated with advanced systems designed to support pilot performance. While systems like GPS navigation and integrated ‘glass’ cockpits have already been operationally proven, and new developments, such as enhanced vision systems (EVS) and synthetic vision systems (SVS), are capable of providing further performance benefits, the design, implementation, and deployment of these complex systems must be done carefully to ensure that potential human system integration risks are identified and mitigated.

Since the early days of flight, operations in conditions of limited visibility, such as night and instrument flying, have been acknowledged as being significantly more challenging and dangerous than operations in favourable conditions. This is not ­surprising, as the perception of the world outside the aircraft – including terrain, landmarks, and other aircraft – is crucial to an accurate situation awareness picture.

The effort to maintain or improve flight safety during poor visibility conditions has been a key driving factor for advances in aerospace technology in the past several decades. Technologies such as advanced autopilot systems, GPS navigation, and night vision imaging systems (NVIS), enable military and civilian aircrews to operate more safely in these conditions than ever before. However, while intended to increase the margin of safety, these improvements have often had the opposite effect as they allow flight operators to push the envelope into more and more difficult conditions. It is not uncommon for operations to occur when visibilities drop to “0”.

The current Canadian mission in Afghanistan and U.S. military experience in Iraq have demonstrated the dangers of ­helicopter operations in sandy terrain, with the ‘brownout’ effect being cited as a contributing factor in several fatal accidents involving both Canadian and U.S. helicopters. In addition, with Canada’s size and varied climate, domestic operations face an incredibly wide variety of weather conditions including dense fog, intense thunderstorms, and, of course, snow. In fact, the sandy ‘brownout’ effect that has proved so dangerous in Afghanistan and Iraq is similar to the ‘whiteout’ effect that can occur in snowy conditions – both have contributed to accidents.

In the civil world, the need to support the natural resources sector, off-shore and in the north, creates a demand for reliable flight operations in areas known for unpredictable weather. As the changing climate in the north results in expanding natural resource operations and an associated push to assert and defend Canada’s sovereignty, the demand for flight operations in these challenging conditions will only increase in the future.

Complex New Systems
Effectively supporting aircrew situational awareness is another critical element of maintaining flight safety in low visibility conditions. Previously, this support came in the form of navigation systems that allowed pilots to know their position with ever increasing precision using advanced radio navigation aids being supported by inertial and, eventually, GPS navigation ­systems. At the same time, cockpit instruments were steadily being improved, moving from separate analog instruments to integrated digital displays, allowing information from the aircraft systems to be displayed in a more usable form.

The development of sensor systems, such as NVIS and infrared (IR) sensors, allowed some of the “lost” visual information to be restored to aircrews. NVIS have been in service with military aircrews for over 25 years and are gaining popularity in civil use as well, particularly for critical missions such as medical transport. IR sensors have also been used in military operations for over 30 years, and FAA/EASA/TC certified, IR-based enhanced vision systems have been available on the civilian market since 2001, allowing aircraft operators to fly Category I approaches to Category II minimums.

Human-Machine Interface
While the proliferation of electronics has allowed for the development of more complex automated aircraft and related systems, the implementation of these new and advanced systems brings additional risks associated with their integration into the existing human-machine interfaces.

In many modern aircraft, the biggest limitations in aircraft performance and flight safety are attributable, not to the ­aircraft or the aircrew individually, but to issues that arise from the integration of the two as parts of a complete system. For example, the use of NVIS in military flying has been a factor in a significant number of incidents and accidents. Investigations into many of these incidents have indicated that the narrow field of view of the aircrew NVIS, combined with the procedures for checking clearance in confined landing zones, could allow for blind spots in the visual field around the aircraft.

These incidents show that while the use of NVIS for night operations can ­provide aircrew with additional capability, the integration of a new complex system into existing operational procedures has inherent risks that must be taken into account.

As more advanced EVS and SVS are deployed, it is likely that they will be subject to some of the same human system integration issues that have been associated with other advanced cockpit technologies such as flight management systems. Experience in the civilian sector has shown that excess workload and channelized attention can occur in some situations while using flight management systems (FMS).

These issues pose a particular danger when they occur in critical phases of flight, as evidenced by accidents such as those that occurred on American Airlines Flight 965 and Crossair Flight 498. Operator interaction with an FMS was found to be a contributing factor in both of these accidents.

So What?
As the complexity and variety of aircraft systems continue to increase, it is important to ensure that new systems are designed and deployed in a way that will minimize human system integration risks, ensuring that critical missions can continue to be flown safely and successfully.

The most reliable method of reducing these risks is the use of an effective HF engineering process throughout the development and deployment of a new system. This consists of three main elements: analysis, design, and evaluation.

A thorough HF analysis at the early stages of system development may appear to be an unnecessary expense but it can make the difference between success and failure in a development program.

Ensuring that the functional and human-machine interface requirements for a system are well defined and understood early on can save significant time and expense later in the development process by reducing the probability of encountering major issues that would require late-stage system modifications or re-design.

Of equal importance is developing an understanding of the context of operations for a new system. For example, if the system under development is to become a piece of a larger overall interface (such as an EVS or SVS system in a cockpit), the system requirements must reflect the need to integrate effectively with the complete human-machine interface.

The design of an effective human-machine interface largely flows from the results of a thorough analysis. Designers need to ensure that the proposed interface designs meet the requirements identified during the analysis process and follow the well-established guidelines and heuristics for usability in interface design. It is also important to recognize that interface design is an iterative process; design concepts should be evaluated with subject matter experts and prospective operators at logical stages of progression. This means that as the level of fidelity increases, it is important to evaluate design concepts in a realistic context to ensure that they are usable not only individually, but as parts of a larger system.

HF techniques can also be used to good effect to evaluate existing system designs and identify any usability issues. The relative severity of the issues can also be assessed and used to focus mitigation efforts on the areas of greatest concern. When dealing with an existing system design, human system integration issues can be mitigated either by updating or re-designing the system, which is typically expensive, or by developing training and operational procedures that can help operators work with any existing limitations. While improving the design of the system itself is usually considered to be the best option if cost is no concern, it is often possible to obtain effective improvements through less expensive modifications to training and operational procedures.

Given the pressures on military and civilian aircraft operators to take on more difficult missions in more challenging environments while maintaining or improving flight safety, it is certain that the development and deployment of new complex aerospace systems will continue into the future because new technologies can offer definitive benefits to performance and safety. However, the complexity of these systems also brings some risks associated with their integration as part of an overall human-machine system.

As advanced systems demand more operator bandwidth, it becomes increasingly important that system design, operator training, and operational procedures be tailored to take advantage of the strengths of human capability while mitigating the limitations. In this way, effective human system integration can help to ensure that new systems support operator performance without introducing new and potentially dangerous issues.
Paul McKay obtained his M.A.Sc. degree in Systems Design Engineering from the University of Waterloo, where he specialized in advanced interface design and aviation human factors. He is currently the Manager of Human Performance Engineering for Gladstone Aerospace Corporation.
© FrontLine Defence 2010