top of page
  • Nancy Baldwin

Drones Earn Slot In Accident Investigators’ Toolkit

When air safety investigators with Taiwan’s Aviation Safety Council (ASC) mobilized to the island of Hoko on July 24, 2014, to find out why a TransAsia Airways ATR 72-500 crashed into residences near Magong Airport, they brought along a new tool—a ScientificAerospace quadcopter with a digital camera and inertial and altitude sensors.

Investigators with the U.K.’s Air Accident Investigation Branch (AAIB), who have also been deploying with unmanned aerial vehicles (UAV) the past two years, have built a flight department that includes three DJI quadcopters and battle-tested standard operating procedures. Other nations are building or at least contemplating similar uses.

The common thread among all the organizations is the newfound utility the compact, lightweight quadcopters bring, both at the scene of an accident and during the data analysis that follows.

Investigators used a quadcopter to map the area under the final flightpath of TransAsia Airways Flight 222 several days after the crash, finishing the survey in less than 90 min. Credit: Taiwanese Aviation Safety Council

At Magong, investigators used their cyberQuad Maxi to survey and map the path of the aircraft during its final seconds of flight after going off course and initiating a missed approach. The data was later used to compute height measurements of the terrain and obstacles struck by the aircraft, matching the altimeter and sound measurements downloaded from the flight data recorders.

The quadcopter data is included in the ASC’s recently released final report on TransAsia Airways Flight 222, which faulted the pilots for descending below the minimum-descent altitude during a nonprecision approach in heavy rain without having visual contact with the runway. The crash killed 48 of 58 on board, including all four crewmembers, and injured five persons on the ground.

Brian Kuo, an ASC investigator, says the preprogrammed UAV survey of the crash zone (see map) took approximately 90 min. to complete, a more efficient and less costly proposition than the typical solution—hiring and equipping a helicopter for the operation.

Martin Chen, an engineer with the ASC, operated the ScientificAerospace cyberQuad Maxi quadcopter for the survey after gaining approval from the Taiwanese Civil Aeronautics Administration (CAA), a process that took five days. He says the CAA typically requires 14 days to process a request to fly a UAV in the region; officials in this case accelerated authorization. At Magong, he started the survey mission at 5:30 a.m. and ended it before 7:00 a.m., when scheduled service starts. “We had time pressure,” he says.

The ASC, which had been studying the idea of using a UAV for investigations for several years, originally planned to use the Maxi for site surveys of accidents in difficult to access remote or mountainous areas, a task it regularly practices in joint drills with Singapore, Japan and other East Asian countries. Many of those countries are also moving ahead with plans for using drones, says Kuo.


Costco Offers Customers Private Aviation Through Wheels Up

NASA ‘Surprised’ By Hybrid Power Study Results

U.S. Navy, Marines Focus On F-18 Oxygen Problem

Although Magong Airport is neither mountainous nor remote, the ASC team upon being sent to Hoko the day after the crash opted to take the quadcopter along “to see how well we could finish the site survey task in a populous area,” says Kuo. “That was rather quick,” he says of the 90-min. survey flight time, noting that using “traditional” site survey tools—GPS position tagging of the all the various wreckage pieces by hand —takes approximately one day to complete.

Chen flew the Maxi at 100 meters (328 ft.) altitude through the survey area and collected digital images at 5-cm (2-in.) resolution along with GPS position, altitude and other variables. Using a commercial software tool afterward, he created a digital surface model—the height of the surface including the vegetation canopy, buildings and obstructions—based on stereoscopic image matching. With a digital terrain map—the height of the surface without vegetation, buildings and obstructions—obtained from the government, Chen was able to compute the height of the obstacles above the ground.

Investigators also embedded the geotagged locations of recovered components within overhead images taken by the UAV, a process that shows the actual conditions of the environment after the crash rather than an aerial map taken some time before.

Stereoscopic, georeferenced imaging from quadcopters has also become a valuable asset for AAIB investigators, who use the photos to accurately measure distances between pieces of wreckage and to create 3-D images of crash scenes.

Stuart Hawkins, a senior inspector of air accident delegates to the AAIB, recently said that UAVs had so far been used in 16 crash investigations around the U.K. He notes that in the past, images had been supplied by emergency services helicopters or by hiring commercial helicopters and UAVs, but imagery would be slow to arrive and AAIB would not necessarily get the angles or shots needed, or have control of the data.

At accident sites, the AAIB uses two UAV operators, one looking at the video feed and the other controlling the quadcopter. “We decided we should always have a second person operating the camera,” says Hawkins, “because although you can fly it and operate the camera, to get good images you need heads down, and to fly the UAV safely, you need to be heads up.”

In one accident involving a microlight aircraft, the AAIB used a quadcopter to fly near trees to examine branches broken by the crash, a view unattainable with a helicopter. After the crash of a Bell Jet Ranger into cliffs off the coast of East Yorkshire in September 2014, AAIB used a quadcopter to survey the site, where it was too dangerous for investigators to inspect.

In North America, UAVs are also a high-priority enhancement for U.S. National Transportation Safety Board investigators, but government rules currently prevent an optimal view . Whereas a company or private individual can apply to the FAA for a so-called Section 333 exemption to fly a drone, a governmental agency must obtain a Certificate of Authorization (COA) from the agency. And to obtain a COA, John DeLisi, director of the NTSB’s Office of Aviation Safety, says his agency would have to maintain a traditional flight department, with a chief pilot, director of maintenance and other roles. “We don’t operate any aircraft,” he says.

While rules sometimes change, the NTSB in the interim has briefed its investigators that commercial companies, approved through the Section 333 process, could be called in to provide a survey. “If we had a case where we felt that was a needed resource at an accident site, we could do it,” says DeLisi. However, he adds, the optimal and lowest-cost scenario would be for the NTSB to have its own drones and specially trained operators.

Preparations are also underway through back channels. One of DeLisi’s investigators has received a Section 333 exemption as a civilian and is privately developing the skills needed to do site surveys of accidents. “He’s chomping at the bit to do that for us,” says DeLisi.

bottom of page