ASCI 638, Module 6: Automatic Takeoff and Landing Comparison

Comparison of Lockheed Martin F-35C and Aurduplane UAS Firmware

The Lockheed Martin F-35C Lightning II is a fifth-generation Joint Strike Fighter aircraft designed for operation from aircraft carriers. The C model differs from the A model (delivered to the U.S. Air Force) and the B model (the short-takeoff vertical landing version used by the U.S. Marine Corps) by adding avionics that assist the pilot with approaches to a pitching and rolling carrier deck. 

Figure 1. F-35C Lightning II sea trials.

Additionally, the F-35C is constructed with titanium-reinforced airframe components, making the aircraft 5,500 pounds heavier than the F-35A (Lockheed Martin, 2015). The catapult-launched takeoff roll is automated, because the extremely high acceleration and short runway distance are too short for a pilot to react to any changes that might occur. The violent acceleration is also very physically demanding for the pilot. The F-35C (and many other Naval aircraft) utilize heading hold and auto-throttle during the catapult launch. The landing is a different story. There are several technological advances that have made carrier landings easier for pilots, but have not removed the human from the loop entirely. The Automatic Carrier Landing System (ACLS) uses a ship-based radar and aircraft datalink to guide an aircraft autopilot to a very small spot on the carrier deck. The pilot’s job is to monitor the approach and adjust throttle as necessary to maintain the glide slope. However, this system is only used by very experienced pilots, as conducting a manual approach is much more difficult, and is a perishable skill (Hibbetts, 2016).

However, a new feature of the F-35C is a system known as Delta Flight Path, an automated glide path tracking system that greatly reduces the amount of pilot input to maintain an approach path. Early test results show that land-based practice approaches can be reduced from as many as 18 approaches to just 6. Current carrier landing qualification requires ten good traps, but could be reduced to six in the future (Seck, 2016).

The limitations of the F-35C automated takeoff and landing systems are primarily due to integration and operability issues that are often common to newly-fielded equipment. In a recent F-35A exercise, five out of six aircraft were unable to launch due to software stability issues on the ground (Dillow, 2016). Another issue with the nose gear has led to oscillations during the takeoff roll resulting in pilot discomfort and autopilot control issues (LaGrone, 2017).

Figure 2. Example fixed-wing aircraft utilizing Ardupilot flight controller firmware.

The open-source Arduplane firmware is a popular platform for integration in hobby and commercial UAS applications. The Pixhawk autopilot is one of many compatible flight control systems, and can be configured to operate an airplane, rotorcraft, or rover using branches of the Ardupilot program. Arduplane is designed for fixed-wing propeller-driven electric aircraft and features options for automatic takeoff and landing. During automatic takeoff, the autopilot will command full throttle and pitch up at a set angle, maintain a bank angle within a set angle, and climb to a pre-set altitude before continuing on an automatic flight plan. Several parameters must be set to account for different launch methods, in particular a parameter called TKOFF_THR_DELAY that will delay starting an electric motor for a set time after an acceleration is detected. This will prevent injury when hand launching, or aircraft damage when bungee or catapult launching. There are also options for ground steering during a takeoff roll for aircraft with landing gear (Ardupilot, 2016).

An automatic landing may also be incorporated into an automated flight plan for Arduplane. A NAV_LAND command must be added to the end of the mission, indicating the GPS coordinates and altitude of the touchdown point relative to the takeoff point. The autopilot uses inputs from an airspeed sensor to calculate a glideslope that will provide the necessary touchdown speed and altitude. Additional parameters can be modified to provide an extended flare, or a pre-flare airspeed target that leads to a deep-stall or reverse-thrust landing. The autopilot will fly the approach while navigating to the touchdown waypoint, then enter a “flare” mode that maintains a set pitch angle and wings less than 5 degree bank angle. Several landing abort modes are available that permit autonomous go-arounds or pilot-on-the-loop-controlled landings using the same pitch and bank limits as the automated landing (Ardupilot, 2016).

Automatic takeoff and landing may both be interrupted and continued manually by a pilot with an RC transmitter. The main limitation of the automatic landing mode is that it is susceptible to barometric altimeter errors due to changes in atmospheric pressure during the flight, errors due to increased temperature of the autopilot electronics, or errors due to local pressure changes from airflow around the altimeter sensor. All of these can be mitigated through the use of a low-cost laser rangefinder, which is an additional plug-and-play sensor that can be easily incorporated into the aircraft systems (Ardupilot, 2016). There is no specific training required to operate the Arduplane firmware, although a background in remote control aircraft flight is highly recommended.

References:

Ardupilot. (2016). Automatic takeoff. Retrieved from http://ardupilot.org/plane/docs/automatic-takeoff.html

Ardupilot. (2016). Automatic landing. Retrieved from http://ardupilot.org/plane/docs/automatic-landing.html

Dillow, C. (2016, August). Only one of six Air Force F-35s could actually take off during testing. Retrieved from http://fortune.com/2016/04/28/f-35-fails-testing-air-force/

Hibbetts, T. (2016, March). How effective is the automatic carrier landing system? Retrieved from https://www.quora.com/How-effective-is-the-Automatic-Carrier-Landing-System-ACLS-Do-pilots-use-it-or-do-they-prefer-manual-landings

LaGrone, S. (2017, February). Navy to test fix for F-35C catapult problem next week. Retrieved from https://news.usni.org/2017/02/16/f-35c-catapult-problem-next-week

Lockheed Martin. (2015, October). The C at sea: the F-35 aboard the USS Dwight D. Eisenhower. Retrieved from https://www.f35.com/in-depth/detail/the-c-at-sea-the-f-35-aboard-the-uss-dwight-d.-eisenhower

Seck, H.H. (2016, August). F-35’s new landing technology may simplify carrier operations. Retrieved from http://www.military.com/daily-news/2016/08/17/f-35s-new-landing-technology-may-simplify-carrier-operations.html

ASCI 638, Module 5: Medium-Altitude Long Endurance (MALE) UAS Shift Work Schedule

          For the modified shift work schedule based on the example provided of a 4-crew 24-hour MALE UAS operation, the objective is to provide the aircraft and payload operators with sufficient rest between shifts and the ability to maintain constant availability for mission readiness. The current shift of 6 days on/2 days off rotates in a clockwise fashion, i.e. each crew will work for 6 days, take 2 days off, then shift forward to the next block.

Figure 1. Original shift schedule, 8-day portion. 

            There have been several studies conducted of rotating shifts and their effect on fatigue, most notably in the air traffic control industry. There are several recommendations to add to the MALE ISR shift schedule, and several operational constraints that prevent full implementation of the recommendations.

            The overall recommendation is to shorten the number of consecutive shifts and more rapidly rotate through the shift schedule. The current clockwise shift rotation (day-swing-night) is preferable to counter-clockwise, because it is analogous to flying east-to-west during air travel. Recovering from jet lag has been shown to be significantly easier when flying east-to-west because of our 25-hour circadian rhythm. Because the time zone is earlier on arrival, the day is longer, which matches the natural circadian rhythm more closely (Lu, et. al, 2016). Rotating on a 2-day shift schedule (2-2-2) with 2 days off following the night shift has been shown to maintain a generally diurnal circadian rhythm, and is currently used by USAF air traffic controllers (Luna, French, Mitcha, & Neville, 1992).

Figure 2. Revised 2-2-2 shift schedule.

            Additional steps can be taken to reduce pilot fatigue on the various shifts. The 1992 study by Luna, French, Mitcha, and Neville (title) showed that subjective fatigue levels were higher on the night shifts in a 2-2-2 schedule. For the night shift, the main recommendation is to maintain a cool environment to encourage alertness. Another recommendation is to provide caffeine at the beginning of the shift to increase alertness, as the lowest levels of alertness occurred in the first 30 minutes of the night shift (Luna, French, Mitcha, and Neville, 1992). An additional recommendation is to provide tasty, nutritious foods around the clock so that night shift workers do not have to rely on frequent snacks for hunger satisfaction. Exercise was also shown to improve alertness on every shift, and it was recommended to provide access to a stationary bicycle or treadmill for crews in confined spaces (such as ships or small bases).

One recommendation was to lengthen the day or swing shifts and shorten the night shifts. However, most aviation units limit pilots to 8 duty hours in a 24-hour period, so that is not practical with this current crew schedule. Another consideration is the frequent shifting of mission times throughout each day and potential loss of fidelity while performing ISR tasks such as long-term observation of a target. The handoff briefs between shifts should be very thorough and specific, with a script provided for each crew to follow with a list of tasks and items that must be briefed to prevent fatigue-induced mishaps.

One additional risk to any rotating screw schedule is the potential for sickness or injury that prevents a crew from performing normal duties. 4 crews rotating on a 3-shift schedule is the minimum number needed for mission readiness; if one crew is sick for 24 hours or longer, the impact of schedule disruption will be felt by all crews. The final recommendation is to post one additional crew, which will facilitate an additional day off after the night shifts and provide backup in the event of illness or injury to a crewmember.

References

Cruz, C., Detwiler, C., Nesthus, T., and Boquet, A. (2003, June). Clockwise and counterclockwise rotating shifts: effects on sleep duration, timing, and quality. Aviation, Space, and Environmental Medicine Journal, 74(6), pp. 597-605. Oklahoma City, OK: Federal Aviation Administration.

Lu, Z., Klein-Cardeña, K., Lee, S., Antonsen, T.M., Girvan, M., & Ott, E. (2016, July). Resynchronization of circadian oscillators and the east-west asymmetry of jet lag. Chaos, 26. doi: http://dx.doi.org/10.1063/1.4954275

Luna, T.D., French, J., Mitcha, J.L., & Neville, K.J. (1992, September). Forward rapid rotation shiftwork in USAF air traffic controllers: sleep, activity, fatigue, and mood analyses. Brooks Air Force Base, TX: Air Force Materiel Command.

ASCI 638, Module 4: UAS Beyond Line of Sight Operations

The Aerosonde family of aircraft began development in 1991 in response to a growing need for accurate oceanic weather data and the great expense of using ships and manned aircraft to launch weather sondes. The implementation of small GPS units and robotic controllers led to the development of the Aerosonde system, which could conduct a flight and conduct data collection over a very large area for the same cost as a weather sounding balloon (McGeer, 1999). The aircraft first flew in 1995, and work began to plan a trans-Atlantic flight. Aerosonde Laima successfully crossed the Atlantic Ocean from Bell Island, Newfoundland to South Uist, Scotland in August 1998, burning less than 2 gallons of fuel in a 26 hour 45-minute flight.

Figure 1. Aerosonde Laima flight track, 20-21 August 1998.

          The Aerosonde uses the CloudCap Piccolo family of flight controllers. The Piccolo ground station is based on the same hardware as the airborne autopilot and provides the datalink to the aircraft, a pilot-in-the-loop interface for line of sight operations, and an interface for command and control via a PC (Vaglienti, Hoag, and Niculescu, 2005). Because the Piccolo is purpose-built for beyond-line-of-sight (BLOS) operations, a minimum crew of three was used for the 1998 flight and following operations: the pilot operating the command and control interface, an External Pilot (EP) for visual line of sight (VLOS) operation during launch and landing phases, and a ground crewmember to operate the launch and recovery system. The 1998 flight utilized a car-top launcher, while later versions of the system use a hydraulic catapult launch and net recovery system (McGeer, 1999 and AAI Corporation, 2016). Ground support equipment for crew communication systems is required, as the pilot does not typically have visual line of sight to the aircraft throughout the operation. Coordination of the launch is performed by the ground crewmember and the EP, with the pilot providing a backup role in relaying telemetry data to the EP, such as airspeed and altitude. The EP only has control of 5 channels: one each for pitch, roll, yaw, throttle, and mode switching; the pilot has control of autopilot commands and engine ignition. Following a successful launch, the EP will turn over control to the autopilot, which will follow waypoints and flight modes as input by the pilot. The autopilot will also automatically take over flight control if input from the EP is interrupted (Vaglienti, Hoag, and Niculescu, 2005).

The advantage of using a BLOS system such as the Aerosonde/Piccolo is that a much greater area can be surveyed than remaining within VLOS. An example is the long-range weather data collection test flights that were conducted by the Aerosonde system in 1999. Flights were launched from Atlantic Field near Cherry Point, NC and flown to Pamlico Sound, about 20 miles North. The flights were controlled form Norfolk Naval Station in Virginia, about 100 miles away (Holland et. al, 1999). Follow-on flights were conducted over the Atlantic Ocean and in the Arctic/Antarctic research areas. More recently, an Aerosonde was used in March 2017 to conduct BLOS power transmission lines in New Zealand (Groenestein, 2017).

The disadvantage of BLOS operations is the dependence on the datalink. The Laima crew originally intended to fly to Ireland, but were denied entry into Irish airspace because of a rule requiring trans-Atlantic aircraft to provide position reports every hour when operating within controlled airspace, something the Aerosonde crews could not reliably provide, as their datalink was limited to radio line of sight; the Laima was flying an automated flight path (McGeer, 1999). Situation awareness is also compromised by the inability to see or feel the aircraft environment; all sensory input is provided through visual interpretation of the ground station control software. A large amount of human factors research in UAS centers on this issue.

Figure 2. Aerosonde with UTC Aerospace TASE daylight camera view of forest fire.

Commercial applications of BLOS flight include long-range infrastructure inspection (as in New Zealand) and intelligence, surveillance and reconnaissance (ISR) collection for military customers. An additional potential commercial mission that has been tested by the Aerosonde is in wildland fire fighting. In 2015, a team from AAI Corporation used an Aerosonde to fly over the Teepee Spring Fire in Idaho, providing real-time imagery during the night – a period when manned firefighting aircraft are grounded for safety. Active surveillance during the night provides incident commanders on the ground with information to make decisions to place assets when daylight is available (Ridler, 2015).

References

AAI Corporation. (2016). Aerosonde SUAS. [Brochure]. Retrieved from http://www.textronsystems.com/sites/default/files/resource-files/TS%20US%20Aerosonde%20Datasheet.pdf

Groenestein, C. (2017, March). Taranaki company makes history with drone flight. Taranki Daily News. Retrieved from http://www.stuff.co.nz/taranaki-daily-news/news/90012497/taranaki-company-makes-history-with-drone-flight

Holland, G. J., Webster, P. J., Curry, J. A., Tyrell, G., Gauntlett, D., Brett, G., … & Vaglienti, W. (2001). The Aerosonde robotic aircraft: A new paradigm for environmental observations. Bulletin of the American Meteorological Society, 82(5), 889-901.

McGeer, T. (1999, February). Laima: the first Atlantic crossing by unmanned aircraft. Retrieved from http://aerovelco.com/wp-content/uploads/2015/03/Laima_the-first-Atlantic-crossing-by-unmanned-aircraft1.pdf

Ridler, K. (2015, September). Wildfire managers plan drone test over Idaho blaze. The Gazette. Retrieved from http://gazette.com/wildfire-managers-plan-drone-test-over-idaho-blaze/article/feed/272561

Vaglienti, B., Hoag, R., and Niculescu, M. (2005, April). Piccolo user’s guide version 1.3.0. CloudCap Technology: Hood River, OR.