Human-machine teaming is key to the future of aerospace

By JOHN-PAUL CLARKE|December 2020

The Human-Machine Teaming Technical Committee fosters the development of methodologies and technologies that enable safe, trusted and effective integration of humans and complex machines in aerospace and related domains.

There is little debate that the future of warfare will include greater integration of humans and machines, so it should come as no surprise that many of the accomplishments in human-machine teaming this year occurred in the military domain.

In January, the Kratos XQ-58A Valkyrie test drone completed its fourth test flight only four months after suffering damage in a crash in October 2019. The Valkyrie is a testbed for the U.S. Air Force Research Laboratory’s Skyborg “loyal wingman” technology that the Air Force envisions will enable formations of autonomous, low-cost and relatively expendable drones that would accompany and collaborate with F-15EX and F-35 fighter jets in the same way as crewed aircraft.

In July, Boeing Australia conducted a series of flight tests involving three uncrewed aircraft that took off, achieved and departed from their required formations and landed autonomously. The flight tests were an important step toward the development of Boeing’s Airpower Teaming System, an uncrewed loyal wingman aircraft in development in Australia.

In August, the DARPA Competency-Aware Machine Learning project received BAE Systems’ MindfuL software program designed to increase transparency in machine learning and artificial intelligence systems by auditing them to provide insights about how these systems reached their decisions, and thereby enable autonomous systems to assess their own competency and strategy and express both in a form understandable to humans.

There were also significant events in the civil domain — especially with reduced crew operations and single pilot operations. In July, Airbus concluded its Autonomous Taxi, Take-Off and Landing project to explore how autonomous technologies could help pilots focus less on aircraft operations and more on strategic decision-making and mission management. Over two years, Airbus conducted 500 test flights, culminating in a series of six test flights in which A350-1000 aircraft taxied, took off and landed autonomously using fully automatic vision-based onboard image recognition technology.

In August, Airbus subsidiary Acubed followed up on the work in the Autonomous Taxi, Take-Off and Landing project by starting flights in California to collect data and advance autonomous technology that will make the next clean-sheet narrow-body aircraft capable of single-pilot operation. Also in August, Xwing, a San Francisco-area autonomous systems startup, conducted numerous fully autonomous passenger flights in a Cessna 208B Grand Caravan.

Events outside aerospace also indicated a future with greater interaction between humans and machines. In July, researchers with the U.S. Army Combat Capabilities Development Command’s Army Research Laboratory, in collaboration with researchers from the University of Southern California, announced that they had developed the Joint Understanding and Dialogue Interface capability, which enables bidirectional conversational interactions between soldiers and autonomous systems, thereby allowing soldiers to verbally interact with machines.

Given all the activity, human-machine teaming is poised to become a key area of research and development in aerospace, especially considering the increasing capabilities of machines as illustrated in August during DARPA’s third and final AlphaDogfight Trials, an AI air combat competition that pitted AI against humans.