Keep 'em Flying: The Future of UAV Safety

Commander William Earl Fannin, Class of 1945, Capstone Essay Contest

A significant competition for airspace between manned and unmanned aircraft will begin within a few short years. By that time, the safety of UAVs must match that of their manned counterparts.

Unmanned Aerial Vehicles (UAVs), despite many limitations, have significant tactical and technological advantages when compared to conventional manned aircraft. Two types-autonomous and remotely operated (teleoperated)-are in use today. Autonomous vehicles only require an operator to program waypoints, launch, and recover the aircraft. Once airborne, the autonomous aircraft tracks the predetermined path and completes its mission without human control. Most operational UAVs are teleoperated, which require a human pilot, and are flown much the same as a manned aircraft, except the pilot is not on board the craft and can be as far away as the other side of the world.

UAVs excel at what are commonly referred to as the Three Ds of unmanned systems: dirty, dull, and dangerous missions.¹ Examples include nuclear, biological, and chemical detection, airborne reconnaissance, and deep strike/ low-level reconnaissance. Unmanned systems are almost always better at completing such missions, however, conventional aircraft currentl^{y outperform UAVs in almost every other mission. Conventional aircraft also have a significantly better safety record.2}

The military's goal, set by the Floyd D. Spence National Defense Authorization Act of 2001, is to have one third of all operational aerial vehicles unmanned by 2015.³ With many UAVs slated to enter ever more crowded airspace, the vehicles' safety must also be enhanced. To operate them safely in civilian, as well as military airspace, the military and the FAA must address three areas in which UAV technology and doctrine is lacking: pilot training, base station ergonomie design, and procedure standardization.

Training, Training, Training

Effective UAV pilot training is the single most important factor in increasing vehicle safety. A March 2005 Air Force study found human factors-pilot error-to be present in more than 60 percent of UAV mishaps.⁴ In the Air Force, large UAVs are flown by experienced pilots, while in the other services small vehicles are typically flown by enlisted personnel with no pilot training. Interestingly, the study also found that pilots experienced in manned flight had no substantial advantages over pilots who were taught to fly using UAVs. In fact, many experienced conventional pilots had to unlearn certain skills such as reliance on peripheral vision during landing. Essentially, the playing field is level in UAV piloting.

Training and doctrine seems to be falling short when it comes to safety, therefore, the military needs to implement several crucial changes. A separate UAV community is needed within each service that jointly cooperates for research, development, acquisition, operations, and deployment. According to an Air Force study, UAV operators are currently reassigned or separated from the military at the height of their proficiency.⁵ This could be resolved by having dedicated, highly trained professionals in a separate aviation sub-community. Because pilots trained in manned flight have little apparent advantage over unskilled pilots, it would be most cost effective and efficient to train new personnel in UAV piloting and keep them in that community for the rest of their careers. The aviation community has a precedent, pilots either fly cargo planes, helicopters, bombers, or fighters; they do not change type with duty assignments. This precedent needs to be applied to the community, keeping experienced unmanned-aircraft pilots flying UAVs.

A common pilot error and cause of many UAV accidents is the tendency of an unskilled or inexperienced pilot to overstress the airframe while performing abrupt maneuvers. This is difficult to solve using training alone. A pilot overtaxing a manned aircraft receives a number of physical cues-vibrations and sounds-that are unavailable to a UAV pilot. Manned flight doctrine has procedures to limit airframe stress and pilots can always "listen" to the aircraft to avoid catastrophic damage. UAVs provide no sensory feedback other than vision. Effects of this deficiency can be moderated by use of a flight envelope protection algorithm. This limits the amount of control input available to a pilot, or autopilot, so the aircraft's outputs do not exceed safe parameters. Flight envelope protection is in regular use on manned fly-by-wire aircraft such as the F-16 and F/A-18. These aircraft will not allow the pilot to make maneuvers that will overstress the airframe.

Design for the Pilot

One major weak point in UAV design that contributes to high mishap rates is the lack of ergonomically designed base-station controls. Only 10 percent of spending on UAVs is going to base-station equipment, while most accidents are attributed to pilot error and machinepilot miscues.⁶ Research has been primarily focused on improving the design and capabilities of the aircraft, not the design and capabilities of the machine-pilot interface. One improvement to the interface would be to use existing flight simulator technology. The simulators should be used to not only train UAV and conventional pilots but also employed in operational UAV control. Integrating simulators would place the pilot back in the air virtually, solving many of the problems associated with taking him out of the aircraft while not losing its benefits.

Drexel University's Autonomous System Laboratory is leading research in UAV technology investigating the safety issue. Their solution to the machine-pilot interface problems is to give the pilot a simulated feeling of flying. This approach should work with the Ak Force's large UAVs but will probably be impractical with smaller, more mobile systems. The Marines simply cannot afford to control a Dragon-Eye or Dragon-Warrior with a simulator that costs ten times more than the aircraft. There are also tactical advantages to having a UAV pilot in theater as close to the front line as possible. Small, portable UAV base stations with virtual-immersion technologies must be created.

Lockheed Martin Aeronautics' Ship-Air Integration Lab uses virtual technology to integrate the design of the F-35 with Navy aircraft carriers. Virtual-reality technologies would be the perfect solution to machine-pilot interface problems. Other-than-visual virtual technologies are needed to augment current vision-based base stations. Technologies including ambient sound and haptics should be included to make UAV piloting more realistic. Haptics is the technology of force-feedback, which uses computerdriven devices to produce realistic physical interaction in a virtual environment. According to an Air Force Performance Enhancement Directorate study, many UAV mishaps are caused by the pilot's attention being channeled to one aspect of flying. The pilot becomes so focused on one task that he neglects the others needed for flight.⁷

Although UAV pilots have some visual cues, they are essentially flying aircraft remotely, using instrument flight rules. Giving these pilots other-than-visual cues would put control closer to visual flight rules. Coupling these technologies with visual immersion equipment such as helmet-mounted displays would provide a UAV pilot with a realistic sensation of flight. Instead of flying the vehicle based on instruments and a video screen, pilots would be able to see the (simulated) horizon, hear the (simulated) sounds, and feel the (simulated) vibrations of the aircraft. The combination of cues with peripheral vision would reduce pilot fatigue and help prevent attention tunneling.

A Balancing Act

Building completely autonomous UAVs would solve the problem of pilot error contributing to mishaps. Technology, however, is only now maturing enough to allow completely autonomous flight from take-off to landing. This technology is still very limited and not yet ready for most military applications. Completely autonomous UAVs also raise ethical questions, especially because some are armed. Neither the public nor the military are ready to grant totally independent machines the power to take life. A human needs to be in the decision-making process. Some level of automation, however, is needed in UAV flight, because they were developed for dirty, dull, and dangerous missions.

Not surprisingly, remotely flying a UAV is turning into a dull mission. With long periods of relative inactivity followed by short intervals of intense and frantic work, mishaps are certain to occur.⁸ The solution for this is to find the right balance between automation and human control. This balance will be found in an autonomous UAV that receives only its mission plan and releases ordnance on a target with human intervention. This will ensure that the UAV does not crash because of human factors and does not launch ordnance without a human operator weighing all the consequences.

An additional measure needed to decrease mishap rates is the development of standardized procedures. Peter van Blyenburgh, president of Unmanned Vehicle Systems International noted,

There is a real mix of UAVs in our skies and there are more being developed all the time by an increasing number of companies in an increasing number of countries . . . but amazingly no one, on an international level, is regulating these UAVs-there are no international standards, no certification norms, no airworthiness norms, and no air traffic control regulations.⁹

The result of the absence of control is that UAVs are being designed to complete missions, not to complete missions safely within busy airspace. This oversight could diminish their value in the future. The FAA is tackling the problem of UAV integration within civilian airspace and has hired Lockheed Martin to develop a road map for unmanned system introduction into national airspace.¹⁰ This will include how UAVs will interact with conventional aircraft, and the FAA will use this to develop flight rules and regulations for unmanned systems. The Air Force is also developing flight rules and regulations for its UAVs in conjunction with the FAA and other military services. The shortcoming of these standardization actions is that they are not being coordinated with other countries and international organizations. This may happen in the future, but it would be advantageous for all to have an international summit to standardize procedures as soon as possible. Developing a single international standard would dramatically decrease UAV mishaps and facilitate their faster and safer integration into civilian airspace.

Once a UAV accident occurs, flight mishap investigation procedures must be in place to investigate the cause and ensure steps are taken to prevent its repetition. Once again, these procedures will be borrowed and modified from manned-flight practices. The Air Force stated that, "accident investigation rules will be the same . . . [but] what is undecided is how we will categorize [UAVs]."¹¹ The current system of manned-flight accident investigation procedures will be suitable for UAV mishaps, but one change must be made. Instead of just aviators within the military having access to investigation reports, they must be made available to UAV manufacturers and designers as well. This will ensure that lessons learned will not only be gained by teleoperated UAV pilots but also incorporated in future autopilots and vehicle control algorithms.

Mr. van Blyenburgh declared, "UAVs are going to be big, and there will be more and more of them and without regulation . . . they could be a real danger in the skies."¹² This can be avoided by implementing proper training, technology, and procedures. UAVs can certainly match, if not exceed, the safety of manned aircraft in the near future.