Drones

New mission simulator tests potential drone operators based on personality info

Prospective drone operators could now be asked to prove their skills on a new search-and-rescue mission simulator program. The program appears like any other interactive video game, but behind-the-scenes, next-generation algorithms are calculating scores using personality, biographical, and simulation data.

“There’s a gap right now in the technology. The selection testing that’s being used [for drone operators] is borrowing from manned aviation, which is a completely different animal,” said Dr. Phillip Mangos of Adaptive Immersion (AI) Technologies at the 2017 Future Naval Forces S&T Expo.

“It takes a completely different set of skills to fly something remotely when you’re not in it. Not just in terms of the station orientation, but the ability to plan ahead with the mission and to know the effects of your actions and reactions,” he explained.

To determine whether a person meets the Unmanned Aerial Vehicle (UAV) operating requirements, the assessment is more than just a simulator It uses data from the simulation test, but combines it with personality data and biographical history, and uses unique scoring algorithms to aggregate the data into comprehensive personal assessment results.

The program is “looking at elements like trust in automation, disengagement, stress resilience…and different aspects of a person’s personal history that could be predictive of performance…for example, there is some evidence of early interest in gaming, robotics, and automation is a good predictor of performance,” said Mangos. “All of this data is optimized using a set of scoring algorithms that minimize testing time and give the right content at the right time to get the best measurements.”

According to a report by the Navy Medical Research Unit, trust in automation is key to effective UAV operations. If operators lose trust in the UAV they are operating, they are much more likely to begin to manually override or question the actions of the automated system. Under these circumstances, the UAS will be flying much less effectively. The algorithms of the UAV must be straightforward enough for operator comprehension, said the report, but a pre-disposed operator tendency to trust technology is also important.  

“The operator has to go and find a friendly, stranded individual, who is surrounded by these weapon engagement zones,” explained Mangos. “There are rules of engagement the operator needs to observe based on how long the individual has been stranded, how many resources he or she has left, and how close he or she is to the weapon engagement zones.”

The search-and-rescue simulation, developed by Adaptive Immersion Technologies, begins with a bird’s eye view of the mission theater, which is configured using a scenario generation algorithm. Once the mission starts, the operators must find and enter authentication codes in order to survive the weapon engagement zones and identify themselves to the individual being rescued.

In the top right corner of the screen, colorful rectangles indicate remaining health and fuel percentages.
One challenge the operators face is completing the mission according to the rules of engagement without running out of fuel, according to Mangos. The system throws real-life curve balls at the operators, such as bad weather or a random loss of connection between the operator and the UAV.

UAV operators face decision-making challenges like the ones in the simulation constantly. They can be due to the environment, poor feedback loops, time stress, and multiple operators, reported the Navy Medical Research Unit. The scoring algorithms of the AI system are designed to calibrate the decision-making skills of the operator and factor them into the final score.

“The current way of doing testing is basically asking [prospective operators] questions about their knowledge of the airspace, for example: When do you talk to air traffic control? What are the limits for flying a drone in the civilian world over a crowded area?” said Mangos. “This system is going to predict the best operators, in terms of their ability not only to conduct the missions, but also to handle the work stress, job-related and operational stress of conducting remote missions.”

The fuel and health monitors, and the rules of engagement, are designed to assist the operator to make key decisions throughout the UAV flight, but deviations inevitably must occur. Decision making is further complicated because UAV operators lack the same tactile, real-time, reference points and situational awareness of manned system pilots.

However, this also means that simulation testing and training is particularly relevant and effective for UAS operators. The assessment simulation software keeps track of some 25 different measurements, according to Mangos. At the end of the game, the measurement data is input into the scoring algorithm, which determines the operators suitability to become an official UAV operator.