Available online at www.sciencedirect.

com

ScienceDirect
Procedia Manufacturing 3 (2015) 2937 – 2944

6th International Conference on Applied Human Factors and Ergonomics (AHFE 2015) and the
Affiliated Conferences, AHFE 2015

An investigation of the harbor pilot concept for single pilot

operations
Martin T. Koltza, Zachary S. Robertsa, John Sweeta, Henri Battistea, James Cunninghama,
Vernol Battisteb, Kim-Phuong L. Vua, Thomas Z. Strybela
a
California State University Long Beach
b
San Jose State University and NASA Ames Research Center

Abstract

Commercial aviation is expected to see an average growth of 3.7% per year over the next five years. At the same time, regional
airlines are expressing concerns that they are unable to find a sufficient number of pilots who meet the minimum number of
required flight hours and certifications. One option for handling the increased demand for, and reduced supply of, pilots is to
better distribute the pilots already employed. This could be accomplished by reducing the number of required crewmembers on
the flight deck from two to one. However, a significant challenge to the implementation of a single pilot crew is the increase in
workload a single pilot would face especially in complex airspace. The harbor pilot is one concept of operations that has been
proposed to deal with the potentially serious problem. In this concept, the harbor pilot is a ground operator whose job it is to help
arriving single pilot aircraft navigate the complex terminal area airspace. The purpose of the present study was to investigate the
feasibility of the harbor pilot concept by measuring the workload and performance of harbor pilots during a series of back-to-
back simulated arrival procedures. Preliminary results show that harbor pilots rate their workload generally low and indicate
that the maximum consecutive normal approaches that they are able to complete in a row is about four to six, provided they are
not emergency situations.

©
© 2015
2015 The
TheAuthors. Published by
Authors.Published byElsevier
ElsevierB.V.
B.V.This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of AHFE Conference.
Peer-review under responsibility of AHFE Conference
Keywords:Single pilot operations; Workload; Harbor pilot

2351-9789 © 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of AHFE Conference
doi:10.1016/j.promfg.2015.07.948
2938 Martin T. Koltz et al. / Procedia Manufacturing 3 (2015) 2937 – 2944

1. Introduction

Commercial aviation is expected to see an average growth of 3.7% per year over the next five years [1]. At the
same time, regional airlines are expressing concern that they are unable to find a sufficient number of pilots to hire
as first officers who meet the minimum number of required flight hours and certifications [2]. However, according
to a 2014 Government Accountability Office (GAO) report, the lack of supply may not be a symptom of the number
of pilots who are completing the required flight training. Rather, the GAO study argues that the inadequate supply
of candidates is due to the lack of qualified pilots who are willing to work for the low entry-level wages typically
offered to new first officers. The average first officer starting salary is only $21,600 annually [2]. This
compounded with the fact that the Federal Aviation Administration (FAA) recently extended the mandatory
retirement age from 60 to 65 making it difficult for new hires to move to higher paying positions means that more
pilots are turning to military or corporate jobs. As demand for commercial air travel increases over the coming
years, the inability for regional airlines to provide competitive compensation in order to attract new hires will
become a serious issue.
One option for increasing the starting salaries of new hires is to eliminate the second crewmember and pass the
savings on to the remaining crewmember. It has been estimated that as much as 35% of the total operating cost for
flights travelling less than 200 nautical miles and carrying less than fifty passengers is due to the cost of the crew
[3]. It is apparent then that regional carriers, which typically operate these types of routes, could benefit
substantially from reducing the number of required crewmembers from two to one. While a reduction in crew by a
factor two may not reduce crew costs by the same margin, it is reasonable to expect a significant savings.
In the past, a combination of new technology and human centered design made it possible to reduce the once
crowded cockpit from five crewmembers to the now standard compliment of two crewmembers (see, e.g.,[3], [4],
[5]). The next obvious step then is to further reduce the flight crew to a single pilot. This will help reduce operating
costs, increase scheduling flexibility, and help to spread the current pool of qualified pilots more efficiently.
Several human centered challenges to single pilot operations have been identified including single pilot crew
incapacitation, fatigue, and problems associated with increasing reliance on complex automated systems [6]. A
major challenge to the implementation of single pilot operations however, is the increase in pilot workload.That is,
tasks that would normally be distributed between two crewmembers would need to be managed by one. The
increase in workload may be especially problematic during the most complex phases of flight such as in the airspace
surrounding the terminal area. The harbor pilot is one concept of operations that has been proposed to deal with this
potentially serious problem.
In this concept, the harbor pilot is a ground operator whose job it is to help arriving single pilot aircraft navigate
the complex terminal area airspace. This concept would minimally require an open microphone and shared displays
allow the harbor pilot to act as a remote co-pilot. However, unlike a current day co-pilot, the harbor pilot has more
detailed knowledge of the current traffic flow, weather, and other procedures within the specific terminal area
airspace. Because of the increased local knowledge, the harbor pilot would be able to more easily anticipate the
needs of the crew and air traffic control, a significant advantage in terms of overall crew workload reduction.
The purpose of the present study was to investigate the feasibility of the harbor pilot concept by measuring the
workload and performance of participants performing the harbor pilot role during a series of back-to-back simulated
arrival procedures. Additionally, the present study seeks to investigate the number of back-to-back arrivals that a
harbor pilot may be expected to accomplish. Participants acted as the harbor pilot responsible for assisting arriving
single piloted aircraft into a simulated version of the Louisville terminal area. An open microphone and shared
displays allowed the participant and an experimental confederate “onboard” pilot work together as a traditional two
pilot crew. Harbor pilots were assigned the role of pilot not flying or pilot flying depending on the condition.
Task load was manipulated by requiring deviations for weather, switching the arrival runway, and varying the
participant’s crew resource management role between pilot flying and pilot not flying. Workload was measured
using the NASA Task Load Index (TLX) and real time subjective workload ratings. Additionally, qualitative post
simulation interviews were conducted to assess the feasibility of the concept from the perspective of active pilots.
 Martin T. Koltz et al. / Procedia Manufacturing 3 (2015) 2937 – 2944 2939

2. Method

2.1. Design

The present study utilized a 2 (runway switching vs not switching) x 2 (weather deviation vs no deviation) x 2
(CRM role pilot flying vs pilot not flying) within subjects design. Participants completed four training scenarios
followed by 16 experimental scenarios in blocks of four over the course of two days. The runway and CRM role
variables were presented using a Latin square counterbalanced order such that each participant experienced every
unique runway and CRM combination four times and the weather condition was switched every four scenarios.

2.2. Participants

For this preliminary report, four participants participated in the present study. All four participants were air
transport pilots who had extensive experience with modern glass cockpit displays. The minimum number of
reported flight hours was 8,000 hours and the maximum was 23,000 hours. Three of the four participants reported
being captains while the remaining participant reported being a first officer.

2.3. Apparatus

The Multiple Aircraft Control Simulation Software (MACS) was used to simulate air traffic within the Louisville
International airport terminal area. Participants and experimental confederates sat in three separate rooms of the lab.
Graduate student research assistants were trained to act as air traffic control and confederate pilots. They were
responsible for managing the air traffic within the sector and acting as the pilots of all aircraft in the sector other
than the aircraft piloted by the single pilot harbor pilot crew. Expert confederates, who were current instrument
rated pilots and had glass cockpit experience acted as the single pilot operator who utilized the services of the harbor
pilot during the simulated arrival procedure.
The participant and all experimental confederates were able to communicate through a push to talk microphone
system that simulated current day very high frequency (VHF) radio communications used by pilots and air traffic
control. An additional open microphone channel was established between the harbor pilot and the single pilot
operator using a separate microphone that was always active. This allowed the single pilot and harbor pilot to
communicate without using a push to talk button and without any other parties listening in. The harbor pilot and the
single pilot confederate also used shared screens so that they could see exactly what the other was doing and they
could share control of the simulated cockpit.

2.4. Procedure

Participants completed four experimental blocks each consisting of four consecutive scenarios lasting
approximately 25 minutes each. Each scenario involved a simulated arrival procedure into the Louisville
International airport. One of two Standard Terminal Arrival (STAR) procedures was used followed by an
instrument landing system (ILS) approach. The sequence of events for each approach is shown in Figure 1. Shortly
after each scenario began the single pilot confederate requested the services of the participant who was acting as the
harbor pilot. At this time, a short briefing was conducted between the participant and the single pilot in order to
communicate the current status of the aircraft and the expected arrival runway, traffic flows, and terminal area
weather.The single pilotalso assigned the pre-determinedCRM role for which the harbor pilot would be responsible.
In half of the scenarios the harbor pilot acted as pilot flying. In this role the participant was responsible for all of the
primary flight controls, calling for checklists, and aircraft spacing. On the other half of the scenarios the participant
was assigned the role of pilot not flying. In this role, the participant was responsible for communicating with air
traffic control, setting up secondary aircraft systems, and reading the required checklists. Shortly after the initial
briefing, ATC would request that the single pilot aircraft maintain 10 miles spacing behind a specified lead aircraft.
Spacing was done manually and was the responsibility of the pilot flying. Roughly one third of the way into each
2940 Martin T. Koltz et al. / Procedia Manufacturing 3 (2015) 2937 – 2944

scenario, ATC notified all aircraft that a new Automated Terminal Information System (ATIS) message was
available. At this time, the pilot flying would request that the pilot not flying change frequencies and gather the new
ATIS information. Next, on half of the scenarios, air traffic control would instruct the single pilot aircraftto switch
the arrival runway. This meant the crew had to complete a new approach briefing and the aircraft’s FMS had to be
reconfigured for the new approach. Finally, a weather cell along the arrival path would require the single pilot
aircraft to request a deviation on half of the scenarios. After each scenario, participants filled out the NASA TLX, a
subjective questionnaire intended to assess crew workload, and the Situation Awareness Rating Technique (SART)
questionnaire (see Figure 1).
The study took place over the course of two consecutive days. On the first day, participants were first briefed on
their roles and responsibilities and given instructions on how to operate their station. Then, participants completed
four training scenarios during which any questions they had were answered. Following the training scenarios,
participants completed two experimental blocks of four trials each. On the second day, participants completed
another set of two experimental blocks of four trials each. Finally, a post experiment interview was conducted in
order to elicit qualitative feedback about the overall concept.

3. Results

Because the present study was a preliminary investigation of the harbor pilot concept (and only four participants
were run) we provide descriptive results on the harbor pilot workload and situation awareness. We also summarize
pilot comments on some aspects of the simulation and their view of the harbor pilot concept.

Fig. 1.Sequence of events. Each approach began with the confederate “onboard” pilot requesting harbor pilot assistance.
 Martin T. Koltz et al. / Procedia Manufacturing 3 (2015) 2937 – 2944 2941

3.1. TLX and SART

Overall TLX workload scores were low to moderate for all conditions and pilots, (M = 34.0, SD = 17). Situation
awareness, based on SART, was consistently high (M = 6.44, SD = .72) across all experimental conditions.
Therefore, self-ratings of workload and situation awareness were at acceptable levels for all conditions in this
simulation.

3.2. Post scenario questionnaire

After each approach was flown, a post scenario questionnaire on the participant’s ability to maintain effective
communication and crew resource management techniques with the remote copilot was administered. Participants
were asked to rate five statements on a scale from 1 to 7 where 1 meant the participant strongly disagreed with the
statement and 7 meant the participant strongly agreed (see Table 1).

Table 1. Post Scenario Ratings.

Statement (1 = strongly disagree; 7 = strongly agree) Mean Rating Standard Error
It was easy to maintain proper CRM with my on-board co-pilot during this approach.
5.31 1.36

During this approach our communications were equivalent to what would be found when
4.18 1.68
both pilots are located in the cockpit.

It was easy to maintain proper spacing during this approach. 4.67 1.23

Communications between the air traffic controller and the aircraft were adequate during this
5.00 1.09
approach.

If I was the pilot flying it was easy to operate the aircraft safely. 5.51 0.90

On average, participants agreed with all of the statements (all means were above the midpoint of 4.0), suggesting
that CRM, communications, and safety were not compromised during the approach. The lowest level of agreement
was for the statement regarding crew communications; pilots indicated on some trials that the communications were
less than equivalent to what is experienced when both pilots are co-located in the cockpit.

3.3. Post experiment debriefing

At the end of the experiment, participants were given a final questionnaire to obtain feedback on the overall
realism of the simulation and the feasibility of the harbor-pilot concept. Participants were asked to rate the realism
of the scenarios, their interactions with air traffic control, and their ground station interface, where 1 meant very low
realism while 7 meant very high realism (see Table 2). In general, pilots rated the scenarios, air traffic and
confederates as realistic, with mean ratings above 4.0 in each case. The cockpit interface was rated less realistic,
which is not surprising considering that a desktop simulator was used.

Table 2.Scenario and experimental realism ratings.

Statement (1= not realistic at all; 7 = very realistic) Pilot 1 Pilot 2 Pilot 3 Pilot 4 Mean

Scenario Realism 5 4 5 3 4.25

Realism of the cockpit interface 4 3 3 4 3.5

Realism of confederate air traffic controllers and

6 5 5 4 5
pseudopilots
2942 Martin T. Koltz et al. / Procedia Manufacturing 3 (2015) 2937 – 2944

Participants were also asked to rate and comment on the performance of the “onboard” copilot, as shown in Table
3. Note that two confederate pilots alternated runs in the simulation, so that the harbor pilot would not fly with the
same pilot on consecutive runs. Three of four participants rated the copilot very high but one participant rated the
copilot as only “somewhat adequate.”

Table 3.How well did your “onboard” co-pilot work with you in the scenarios?
Pilot Rating* Comments

6 “Good CRM worked well. Better explanation of mouse sharing would help”

“CRM- issues with one pilot. Difficulty adjusting to new event with new pilot. Too much anticipation of what
4
comes next in the scenario.”

7 “Both pilots were excellent! Standardization was great”

5 “Once the pilots learn the other traits it makes it easier to anticipate cockpit needs”

* (1 = Extremely inadequate, 7 = Extremely Adequate)

Pilots were also asked to rate and comment on the feasibility of harbor pilot concept. Overall participants rated
the concept low in terms of overall feasibility (M =2.75) as shown in Table 4. Comments made by the pilots
indicated that feasibility of harbor pilot operations would depend on having additional tools for communications,
and additional procedures for maintaining adequate workload. One pilot commented on public acceptance and
another was concerned about the impact of a harbor pilot flying repeated approaches.

Table 4. Summary of ratings and comments on harbor pilot feasibility

Pilot Rating* Comments

“The harbor pilot needs more info on aircraft systems and the aircraft pilot would have to have more
5
streamlined systems to reduce workload further”

“Possible- but unless we are looking at a backup system of flight control for future 9-11 type scenarios (or
2 Malaysian air runners, etc..), why bother? Introduces a whole new level of confusion and miscommunication
between 2 pilots, not to mention subtle incapacitation issues, etc...”
2 “Lots of issues yet. Incapacitation, costs, public acceptance, etc..”

“Pilots don’t want to sit in a box to fly. Trying to maintain separation is not easy because you have to fly the
2
plane anticipating ATCos next move. It is hard to stay focused with the repetition”

*Statement being rated: Please circle the number (1 = not feasible, 7 = feasible) that best describes how feasible you think the harbor
pilot concept is.

Participants were asked to estimate the number of approaches they could fly in succession under normal
operations. Most participants said that four to six back-to-back approaches was reasonable provided that no
emergency situations arose during those approaches. One participant suggested that it would be possible to do eight
successive approaches under normal operating conditions.
Finally, participants were asked to comment on the possible benefits of the harbor-pilot concept. Several benefits
were mentioned by the three pilots who responded to the question. In addition to the benefit of having assistance
during busy periods, two participants mentioned the benefit of having a crewmember with knowledge of the local
terminal-area conditions, for example, weather, air traffic control, and airport information.
 Martin T. Koltz et al. / Procedia Manufacturing 3 (2015) 2937 – 2944 2943

Table 5. Pilot comments on possible benefits of harbor pilot concept

Pilot Comments
Number
1 “Well, say, the single pilot concept is accepted in the work. Uh, you - it’s really good to have somebody in the really busy
stages of the approach, arrival and the approach with you working in the, in the flight deck. So, that would help
tremendously.”

2 N.A. (question not asked)

3 “The supporting role of it. I like that a lot. Even if we have to guys in the cockpit, having somebody on the ground that has
real time access to weather, conditions in the area looking at stuff. . .Whether there is one or two guys in the concept, in the
cockpit. Having a guy who is familiar with that arrival and all that who’d been watching it all day.”

4 “The concept I think is a pretty good concept. You’ve got a guy who’s familiar with that airport, those arrivals that are
working that day. That alone being familiar with what the weather conditions are and things like that. It’s really helpful. . .
To have someone that only shoots approaches into Los Angeles and knows when to expect the next turn and things like that
and can anticipate, ATC, you know because he’s been sitting there watching what this particular controller has been doing
all day, where he gets people to turn base and stuff like that. So I see that. Huge benefit . . .”

4. Discussion

The purpose of the present study was to investigate the feasibility of the harbor pilot concept as it applies to
single pilot operations. It was also intended to gather insight from active duty pilots on the total number of back-to-
back approaches that could be considered reasonable and to understand how crew resource management changes
when the crew is not co-located. Participants acted as harbor pilots during a series of simulated arrivals into the
Louisville International Airport. Weather, traffic flow, and CRM roles were manipulated. Pilots rated their
subjective workload and situation awareness and provided quantitative feedback during post experimental
interviews and questionnaires. Overall workload was rated as being moderately low and was not affected by any of
the experimental manipulations. Situation awareness was rated high and also was not affected by any of the
experimental manipulations.
Participants rated the harbor-pilot concept low in terms of feasibility citing technological issues and the increased
potential for miscommunications as reasons. This suggests that additional communication and collaboration tools
may be necessary for the success of the concept. For example, in the present study, the harbor pilot and the single
pilot could not see each other. Thus, non-verbal cues such as pointing at a gauge or reaching for a lever were not
present. A video link between the two crew members may help fill in some of these missing cues, however,
differences in the layouts of the actual cockpit and the ground station limits the usefulness of a live video link [7].
Because the harbor pilot switched CRM roles frequently, collaboration tools such as NASA’s CRM tool may have
been more helpful than a video link in this particular application [7]. The NASA CRM tool is intended to help
remote co-pilots maintain awareness of their current CRM role. The tool also requires each crewmember to
manually acknowledge state changes in the aircraft’s configuration. This is similar to the non-verbal cues used in
today’s co-located cockpits where the second crewmember typically acknowledges state change by pointing at the
gauge or indicator that has changed.
The present study shows the challenges associated with remotely located crews and highlights the need for
additional collaboration tools. The concept of the harbor pilot as a whole while rated poorly in terms of feasibility
was seen as having potential. When asked if any benefits of the harbor pilot concept could be seen, participants
responded favorably and two pilots reported on the benefits of having a crew member who is knowledgeable about
current terminal conditions: “. . . That alone being familiar with what the weather conditions are and things like that.
It’s really helpful. . . ” ; “ To have someone that only shoots approaches into Los Angeles and knows when to
expect the next turn and things like that and can anticipate, ATC, you know because he’s been sitting there watching
2944 Martin T. Koltz et al. / Procedia Manufacturing 3 (2015) 2937 – 2944

what this particular controller has been doing all day, where he gets people to turn base and stuff like that. So I see
that. Huge benefit.. .”

Acknowledgements

This study was supported by NASA cooperative agreement NNX09AU66A, Group 5 University Research
Center: Center for Human Factors in Advanced Aeronautics Technologies (Brenda Collins, Technical Monitor).

References

[1] Federal Aviation Administation, FAA’s NextGen Implementation Plan, March 2011, Federal Aviation Administration, 2011.
[2] U.S. Government Accountability Office, Aviation Workforce: Current and Future Availability of Airline Pilots, Publication No. GAO-14-
232, 2014.
[3] Harris, D., A Human-centered Design Agenda for the Development of Single Crew Operated Commercial Aircraft, Aircraft Engineering and
Aerospace Technology, 2007, pp.518-526
[4] Federal Aviation Administration, Advisory Circular 25.1523-1, Minimum Flightcrew, U.S. Department of Transportation, 1993.
[5] Schutte, Paul C.; Goodrich, Kenneth H.; Cox, David E.; Jackson, Bruce; Palmer, Michael T.; Pope, Alan T.; Schlecht, Robin W.;
Tedjojuwono, Ken K.; Trujillo, Anna C.; Williams, Ralph A.; Kinney, J. Bryan; Barry, John S., Jr., The Naturalistic Flight Deck System:An
Integration System Concept for Improved Single-Pilot Operations, NASA, 2007
[6] Comerford, D,. Brandt, S.L., Lacter, J.B., Wu, S., Mogford, R.H., Battiste, V., Johnson, W., NASA’s Single-Pilot Operations Technical
Interchange Meeting: Proceedings and Findings, NASA/CP-2013-216513, ARC-E-DAA-TN8313, 2013.
[7] Lachter, J., Brandt, S., Battise, V., Ligda, S., Matesse, M., Johnson, W., Towards Single Pilot Operations: Developing a Ground Station,
Proceddings of the HCI-AERO 2014 Conference, Silicon Valley, California, 2014.