Recently I attended Day One of an industry workshop sponsored by CASA and the University of NSW on “Human Factors and the Automated Flight Deck.”
Day One consisted of a series of highly qualified and experienced speakers across a range of regulatory based automation management subjects, and Day Two we moved into a series of workshops on specific areas with guided discussions.
Quite apart from the content, the opportunity to meet and mix with a cross section of airline representatives from most aspects of the Australian airline and regulatory environment is fascinating and unique for a line driver. It’s only when you have this opportunity that you really see how small aviation – particularly Australian aviation – really is, with familiar faces going back 5, 10, 20 … 30 years.
Owing to the nature of the content, this post won’t go into the details of what I learned over these days – but I’ll provide links for further reading and post some of the aspects of what I saw that struck me the most.
The primary focus of the presentations and workshops was?the 2013 release of the work of 4-6 years of the?FAA CAST (Commercial Airline Safety Team) and PARC (Performance Based Aviation Rule-making Committee) teams who jointly formed the Flight Deck Automation Working Group (FDAWG) to review accident data and produce the “Operational Use of Flight Path Management Systems” published in September 2013.
This team of dedicated professionals – some of whom were present today – were in part updating the work of a decade before on “The Interfaces between Flightcrews and Modern Flight Deck Systems” – the seminal 1996 document which identified key features of automation and flight path management safety issues after in the move to FMS, EFIS, Electronic Warning and Checklist Systems of the 80’s and 90’s. The introduction of technology into aviation has at times been a case of two (or more) steps forward with the odd step back – the original 1996 report reviewed years of data (including accident/incident statistics and investigations) associated with the introduction of advanced technologies into the flight deck, identified a number of shortcomings in system design, flight crew training and flight deck procedures, focussing on industry and regulatory solutions. Both the 1996 and 2013 reports are well worth a read if you operate glass flight deck aircraft, particularly if you are looking for a more in depth understanding.
Dr Abbott reviewed the background behind the report, including the genesis in the 1996 document, and the focus of the group across the various data sources used to collate meaningful conclusions and recommendations which included ASR’s, Accident/Incident reports and LOSA data. One aspect touched on was areas identified in the 1996 report that had changed/improved, or had still remained an area of concern today.
I noted the following points from Dr. Abbotts remarks. Note many of these remarks are based around specific findings of the report – I encourage you to read these directly,
- Pilots mitigate safety and operational risks – and often do so very well – and the aviation system is designed to rely on that mitigation.
- Significant vulnerabilities exist in pilot knowledge and skills for manual flight operations, and the “development and retention of such skills” – Manual Flight is much more than just stick and rudder skills.
- In managing automation malfunctions, there can be insufficient systems knowledge and over reliance on inappropriate procedures.
- An over reliance on automated systems was found, with a reluctance to intervene, along with mode confusion and FMS usage/programming errors. One area of focus was the automation of information processing tasks on the flight deck, which has received more focus in this report.
- Pilot to Pilot, Pilot to ATC Communication/Co-ordination issues were highlighted and reviewed.
- Data Entry / Cross Check errors are still a feature in accidents and LOSA.
- Technology (Alerting Systems) are effective as a backup “safety net” – but increasingly these safety nets are ?becoming a primary line of defence, which may never have been envisaged by the certification processes.
- Training has progressed in pilot skills for flightpath management – but the regulatory minimum standards have not.
- Much has been added to the syllabus in relation to expanding procedural and technical competencies required of pilots; very little has been taken away. The knowledge expected of an initial training program is much more that was was expected 20 years ago – but has the training footprint expanded accordingly.
On the back of Dr. Abbotts review, I look forward to working through the report in more detail over the coming weeks, particularly in light of the workshops tomorrow.
Presentation 2 : Flight Path Management Policies by Captain David McKenny, ALPA/IFALPA
Captain McKenny reviewed the need for an adaptive and comprehensive policy on flight path management in airline documentation. Of particular focus was the point that perhaps too much focus has been placed on the automation itself – rather than on the need to monitor, predict and control aircraft flight path through the various appropriate levels of automation available. Essentially the Tools have been the focus, rather than the Task. Policies need to stress that the responsibility for flight path management remains with the pilot.
Meanwhile policies should identify opportunities for manual flight. Automation must be recognised as a tool – and distinguish between modes of Guidance vs Control. Separating out the two facilitates transitioning between the various levels of automation.
Useful guidance can and should be provided on when pilots should say to ATC “Unable.”
Information Automation – EFB, FMS, etc – how to use them without them becoming a distraction in the management of flight path.
Flight Safety Foundation recently published a document on Pilot Monitoring. A review of how we define, document, train and check pilot monitoring is long overdue.
The issues of Manual Flight and the maintenance of the skills – both cognitive and manipulative – receives special focus in the report and by these excellent presenters. As a long haul pilot, where 50% of my flights are conducted from the jump seat on the aircraft in the role of a checker – this is a particular interest for me as well. The following were noted as areas of concern.
- Definition of errors to include both flight path and energy management; cognitive as well as motors skills; and a lack of monitoring as well as maintaining automation and flight path.
- Go-around, especially all engine from well above minima is of particular interest. LOSA data indicates that 98% of all go-arounds from unstable approaches result in procedures that include errors made by pilots.
Note: The BEA report on Aeroplane State Awareness During Go-Around is particularly germane and illuminating reading here). The videos (even if your French is as rusty as mine) are also worth a look.
- Returning to nominal flight path after a flight path deviation.
- Knowing and Executing corrective actions (Stall Recovery)
- Knowing when/how to intervene when automation is not behaving as expected.
- Degraded manual flight handling skills due lack of practice owing to high reliance on automation.
- Once competent – motor skills degrade very quickly, even with limited practice. Meanwhile Cognitive skills degrade faster, regardless of the level of expertise – unless the pilot can practice those skills.
- Automation policies that encourage/required maximum use of automation reduce opportunities and incentives to practice the knowledge/skills.
- A lack of confidence in knowledge and skills for manual flight was observed.
- There may be a strong perception that automation is safer and more efficient that is not always justified. Automation may well lower workload, but may not actually make the task completion easier.
- Transition from automated to manual operations is not trained/practised.
- The cognitive aspect of manual flight are not addressed.
- Some manual handling skills commonly practised at lower altitudes do/may not transfer to high altitude. ?Separate exposure to high altitude handling – both inside and outside the normal operations flight envelope – may be required.
Presentation 4 : What Monitoring and Pilot Intervention Skills do Pilots Need? by Dr/Captain Simon Henderson
Simon spoke at length on this area and provide much food for thought with a very practical approach to the implementation of the report.
- While flight deck presentation, tools and automation have changed significantly – the requirement for the pilot to manage the flight path has not.
- Systems on board are now highly integrated – but there is a large amount of information that is seldom referred by by pilots until a unusual event happens – at which point the lack of familiarity with this information and it’s presentation can detract from analysis.
- Critical phases of flight may not permit diagnosis of events but instead demand that the pilot make a flight path control response.
- Events are often trained as canned sequences and are advised beforehand to produce perceived better training results. However subsequent training events are not advertised and include a significant requirement for recognition/analysis – this is when pilots are failing in the simulator and resulting in accidents in the aircraft.
- Pilot Flying (PF) skills are highly technical in nature; Meanwhile Pilot Monitoring (PM) skills required are Non-Technical (CRM) in nature including the ability to establish intent and challenge the actions and behaviour of the PF.
- In fact these two skills overlap. You can’t be an effective PM unless you have the skills to be PF.?PMs need Technical (Systems/Path/Acft-ATC/Modes/Rules of Thumb/Manual Flight/Recovery/Malfunction) skills. Worthy of review is Flight Safety Foundation’s “A Practical Guide to Improving Flight Path Monitoring.“
- Intent : LOSA has shown us that the most important thing the PF can do is state intent to the PM.
- Challenge : It’s hard for the PM to correctly/adequately challenge the PF if you aren’t sure what the other pilot’s Intent is.
- Help :?People dont like to state that they dont know what is going on. Airline Check and Training systems often don’t support this mode ? pilots under check/training dont state they are unsure of whats going on.
- Unable : Its intrinsic for pilots to comply with instructions/requests.
- PF-PM Workload – When the workload is Low or High, what can we do to ensure intervention when it’s required?
- Manual and Cognitive skills are both perishable and frangible. A recent study indicated what has become obvious over time to Ultra/Long Haul pilots : Manual Flight Handling skills do degrade through lack of regular use – but are quickly regained with practice. But the cognitive skills otherwise required by pilots also degrade; often degrade as quickly (if not more so) than manual flight skills; and are harder to regain through simple practice. In addition the cognitive load that results from a lack of regular practice in the normal day to day operational environment is a significant detractor from safe, efficient operations.
- Anticipation seems to be missing; Statements of intent; Briefings based on Threat and Error Management theory/practice.
- Recognition : Recognising when things are going wrong is no longer a trivial skill. Flight Path Management algorythms are complicated and it can be difficult to recognise when things are going awry. Know what to expect in order to be able you recognise a deviation.
- Reacting : Check and Training systems tend to encourage pilots to react too quickly. How do you Challenge? What are you going to challenge? How to Recover ? Higher/Lower/No automation? How to reset automation.
- Review : Critical Review skills to analyse whether an action taken is working. This rests on some key foundations ? Knowledge / Critical Appraisal / Be prepared to Change.
- Knowledge underpins all this. LOSA indicates that if a problem had an underlying foundation of a knowledge gap ? it was far more likely to result in an undesirable aircraft state.
- Trust ? Trust in The System; Trust in Themselves. Boeing recently did a study in pilot confidence in themselves on their ability to use the FMC after checkout. 60% felt it took 12 months before they achieve a high level of competency
Presentation 5 : “Operational Policies/Training” by by Captain David McKenny, ALPA/IFALPA and?Dr/Captain Simon Henderson
This briefing covered the need for an automation and manual flight path management strategy in airline policy documentation; the need for that policy to encourage appropriate automation use as well as manual flight exposure.
- Aviation is Very Safe. Pilots (and others) mitigate operational risk; improvements have been made – but there are still vulnerable areas including:
- Manual Handling
- Interaction with Automated Systems
- Pilot Training/Qualification
- Operational Policies/Procedures
- Integration with ATC/Airspace
- No common terminology (FCU/MCP, CDU/MCDU, Philosophy/Policy across Manufacturers)
- Wide variations found in operational policy on automation use
- SOP Compliance is Emphasised – but Pilots don’t always follow : Operational match; Workload; Too prescriptive;
- The introduction of technology to our flight decks have resulted in an overal?increase in the knowledge/skill requirements of our pilots
- Operational Policy has to adapt for the exigencies of the operation, including -?Equipment Type; Operator; Long/Short Haul; Pilot Culture; Pilot Experience Levels; Environment; etc.
- Operational Policy recommendations from the Report include:
- Focus on the Task, not the Automation.
- Clarify and Identify appropriate opportunities for Manual Flight
- Identify Automation as a Tool – a Means to an End; the End being Flight Path Management.
- Policy needs to be consistent with Training and Practice
- Guidance provided in the Selection to drive Flight Path Control – Who / Waht is controlling.
- Policy should reflect Operator / Manufacturer / Training. Review (Dynamic Policy)
- Information Management : Meaning, Terms, Guidelines.
- Cross checking :
Background – The 1996 Report
Primarily the seminars are focussed on the results of 4-6 years of hard work by a team of dedicated professionals on ?This document updates the work done over “. The 1996 document was the seminal work based on observations that while progress in flight technologies (FMS, EFIS, Warning Systems and the like) had introduced higher levels of safety and efficiency in the flight deck – it had also been found to be a contributory factor in a number of high profile accidents. ?as follows (Note – This is a very short and subjective summary – I recommend reviewing the first few pages of the 1996 report itself …)
- There was a lack of understanding of the capabilities, limitations modes and operating principles and techniques of automation – this was codified into the now historical “What’s it doing now”; followed by “Why’s it doing that” and the “What will it do next?” triumvirate.
- Concern over the different levels of automation and the appropriate use of automation levels (including no automation) in unusual and non-normal situations. This included the old chestnut of “throwing the aircraft at the autopilot”.
- Flightcrew automation/mode awareness
- Flightpath awareness, including insufficient terrain awareness, sometimes involving loss of control or CFIT – and energy awareness, especially low state energy.
- Insufficient communication/co-ordination – between pilots, between organizations, including?ATC, designers, regulators and operators.
- Design process that inadequately addressed human performance issues.
- Insufficient criteria, methods and tools for design, training and evaluation.
- A need for better knowledge and skills in certain areas related to human performance.
- Cultural differences in design, training, operations and evaluation. Culture here is used in the widest meaning to span nationalities, language, organizational, regulatory philosophy, and others.
Interestingly the report stressed the need for slow, carefully implemented change, and in a manner to avoid detracting from existing safety practices. How nice is that to read? Thirty six recommendations across XX key subject areas were documented, summarised as follows, commencing with the imperative to invest in People (designers, users, evaluators and researchers) – including a refocussing of flightcrew training investment towards automation issues.
Management of Automation
- FAA to ensure a uniform is conveyed to flightcrew regarding manufacturer’s automation philosophies
- Operator documentation and recurrent training programs to provide clear/concise guidance on automation issues such as appropriate times for AP engagement/disengagement; or used in a mode with greater/lesser authority; conditions under which the AP will not engage or disengage or revert to another mode; and appropriate combinations of automatic and manual flight path control (such as Autothrottle engagement during manual flight).
- FAA to review the AP’s on all airliners to identify the potential to produce hazardous energy states; excessive pitch/bank; subtle departures from intended flightpath; or other undesirable manoeuvres. Results should initiate design improvements, flight manual revisions, operating limitations and/or changes in training programs and operational procedures.
- Better analysis of why flightcrew deviation from procedures, especially when the deviation contributes to causing or preventing an accident or incident
- Industry to take the lead in developing design guidelines for the next generation of flight management systems.
Flightcrew Situation Awareness
Operators to increase flightcrews’ unders
- Better flightcrew understanding of mode and energy awareness issues associated with autoflight systems.
- Better position awareness with respect to intended flight path and proximity to terrain, obstacles or traffic.
The most recent treatise on the subject of Automation includes in the title