$2.5 million NASA project will develop and test safety management for ‘air taxis’

Multi-university team tackles safety systems for autonomous eVTOLs

Vanderbilt engineers are part of a NASA-funded, multi-institution effort to develop safety systems for a mode of transportation that doesn’t exist yet—small, commercial, autonomous planes that move people by air between locations in large, crowded cities.

The task is a formidable one with machine learning at its core. Autonomous, or self-piloted, airplanes must communicate with each other. They must respond to hazards, from weather to equipment malfunction to “uncooperative” other aircraft to prevent collisions and crashes. And all this must unfold in real time, in defined corridors separate from existing air traffic routes but without continuous air control support on the ground.

Gautam Biswas

Commercial, pilotless “air taxis” are perhaps a decade away, maybe less. With this and similar projects, NASA, with the Federal Aviation Administration, wants to stay ahead of that curve and have air traffic management systems (called UTMs) in place when commuters take to the skies.

To that end, the $2.5 million, three-year project, will develop and test the foundations of safety management for advanced urban air mobility. It is hoped these eVTOLs, or electric takeoff and landing aircraft, will reduce fossil fuel consumption and traffic congestion.

“This is a very exciting task,” said Cornelius Vanderbilt Professor of Engineering Gautam Biswas, who heads the Vanderbilt effort. “A machine learning algorithm is not like a person–it can only do what it has been trained for. It can analyze a situation better than a human, but it doesn’t have the intuition to deal with unusual situations.”

The team includes more than a dozen researchers from George Washington University, which is project lead, the University of Texas-Austin and MIT’s Lincoln Lab, as well as Vanderbilt. The Vanderbilt team also includes Marcos Quinones-Grueiro, research scientist at the Institute for Software Integrated Systems. The government partner is NASA’s Aeronautics Research Mission Directorate.

The research team will fly the Tarot T-18 octocopter at NASA Langley for for safety demonstration experiments. The application will be package deliveries with the vehicles flying at low altitudes.

These vehicles resemble a cross between a helicopter and a small airplane. They don’t need long runways of passenger jets or even smaller planes, but their lighter, pilotless profile makes them more susceptible to certain types of risk.

The project tackles three types of hazards: adverse convective weather, winds and fog; corridor incursion by non-cooperative aircraft; and vehicle and component level degradation and faults. Vanderbilt engineers are focusing on the latter.

The Vanderbilt researchers are expert in tracking performance of components and monitoring degradation, and a key innovation they bring to this work is the idea of using reinforcement learning algorithms for online fault tolerant control. “What happens with a fault occurs? How can aircraft keep flying by adjusting its controller? Should it continue or alter its route, or is the situation so bad the vehicle must find a place to land immediately?” Biswas said.

“All of this decision-making has to happen on board,” he said.

The approach, which also supports condition-based maintenance for safe flying, will require a cloud infrastructure as well as to support prognostics, risk evaluation and hazard response functions.

Automated aircraft control is not new; autonomous systems handle more and more functions in commercial and military flights each year with software trained by system models. The difference with the eVTOL safety management project is that it will be data-driven.

The research team will use an Airbus Vahana UAM vehicle for its taxi studies.

NASA has been testing such aircraft at its research center in Hampton, Virginia. Combined with its Advanced Air Mobility Campaign, the agency is exploring how different aircraft technologies and configurations will perform in an urban environment, and researchers will use this data. With the information trove, the team will tackle not only what and how a “machine” learns, but also how to make the leap into “learning as you go” using reinforcement learning.

“All of this will be done in a data-driven manner,” Biswas said. “That is the interesting thing and the challenge–how rich is your data? If you don’t consider the limitations of your data, you will fail.”

In a related effort, he and the larger team are advising NASA on extending and operationalizing a Concept of Operations, or ConOps, for advanced air mobility. Included will be separate local and regional corridors for autonomous flights, takeoff and landing requirements on the ground, and designated areas for passenger transfer.