Autonomous flight systems developer Reliable Robotics is teaming with machine-learning-based avionics specialist Daedalean to develop advanced navigation and situation awareness systems for commercial aircraft operations. According to the partners, the proprietary flight-automation systems they are working on will support both onboard and remote pilots in making faster, better-informed operational decisions based on data from advanced sensors.

Last month, Reliable Robotics completed a series of test flights controlled by a ground-based pilot at its headquarters in Mountain View, California. From a distance of over 50 miles, the pilot instructed a Cessna 208 Grand Caravan to taxi, takeoff, and fly over a populated area, and then land, while communicating with air traffic controllers via the aircraft’s onboard radios.

According to Reliable Robotics, it expects to take another two years to achieve certification for its remote-pilot technology. Initially, the company is focused on converting Cessna Caravans mainly for freight operations, but it sees potential to do the same for other aircraft types and for other applications.

The company maintains that automation will give cargo operators, in particular, greater flexibility in opening new routes and allow these to be launched more quickly in response to shifting demand. It would not comment on reports that express delivery giant Fedex has shown interest in using the system.

The business plan calls for increased service to smaller regional airports in underserved markets. The company claims that turnarounds between flights will be quicker and more efficient because they won’t depend on the availability of local flight crew in these locations.

“Pilots will no longer be limited to flying specific types of aircraft, the same routes, or only one or two flights a day,” it explained in a statement about the latest phase of flight testing. “Geographically distributed control centers will enable constant operation across time zones, giving remote operators the ability to fly more frequently than traditional airlines.”

Reliable Robotics was founded in 2017 by Robert Rose, who formerly led the development of flight software for SpaceX and the autopilot system for Tesla. It has attracted $33.5 million in funding, mainly from venture capital groups, and is pursuing a business model that would have it operating automated aircraft as a service.

Switzerland-based Daedalean is developing machine-learning applications based on situational awareness technology that will meet safety standards now being defined by both FAA and EASA. In joint work with the European agency, it has developed high-performance machine-learning algorithms for safety-critical applications that formed the basis for "The Concepts of Design Assurance for Neural Networks," a report published in March 2020. The company is already partnered with avionics manufacturers Avidyne and Honeywell Aerospace, with the latter having invested in it last year.

“Our team has developed advanced machine learning that can adapt to the inherent uncertainties in airspace and increasing levels of onboard autonomy,” said Daedalean founder and CEO Luuk van Dijk. “Bringing our core competencies together was a logical next step to jointly develop a solution set that makes aircraft safer.”

Reliable Robotics started remotely piloted test flights in a smaller Cessna 172 Skyhawk aircraft in 2019. It began testing its system on the Cessna 208 Caravan in June 2020.

The test flights have been conducted with a remote pilot based in a newly established control center, working in tandem with an onboard safety pilot who monitors flight performance. The company’s goal is to achieve incremental certification to support operations using gradually increasing levels of autonomy.

The Reliable Robotics team has been supported by colleagues from SpaceX’s Falcon 9 and Dragon 2 programs in developing the human interface for the control system and also cybersecurity protection. The company believes that successful commercial flights will depend on highly trained pilots who will transition into remote operations roles that will include communicating with ground and air traffic controllers so that the converted aircraft can use existing airspace and airport infrastructure.

From the workstations, the remote pilots are expected to maintain situational awareness and manage flight plans similarly to how an onboard pilot would complete the task. The converted aircraft will feature a proprietary sensing and computing platform that will handle fault management in conjunction with added levels of mechanical redundancy.

Meanwhile, rival autonomous flight technology pioneer Xwing is working on what appears to be a similar plan for converting the Cessna Caravan single-engine turboprop aircraft. In December, the company started piloted commercial cargo flights after receiving FAA approval for its Part 135 charter operation in late November. It is using these flights as a means to continue development work on its Autoflight System.

The California-based start-up has been conducting test flights since July 2020 as it prepares to seek approval for operations with its autonomous (non-piloted) Autoflight System, for which work started in 2017. It aims to secure FAA approval to begin autonomous commercial flights in 2022.

AIN Products

SUBSCRIPTIONS

RESOURCES

ABOUT AIN