When it comes to modern military operations, robotic technology provides a tremendous tactical advantage. Drones, ground robots and autonomous vehicles are routinely used for missions such as intelligence-gathering, surveillance and reconnaissance (ISR), allowing military personnel to conduct operations from a safe distance. And yet, despite these technological advances, the vast majority of vehicles in use by the military — whether in the air or on land — still require a human operator.

The Highly Dexterous Manipulation System (HDMS) designed by RE2 Robotics. (Photo: RE2 Robotics)

Undoubtedly, the use of unmanned systems by the military will continue to surge. Still, it would be both impractical and exorbitantly expensive for the military to replace all of its current and still-reliable vehicles with autonomous ones. The Air Force has invested in more than 40 distinct types of aircraft that cost anywhere between several hundred thousand and hundreds of millions of dollars to build. On land, articulated front-loaders, such as those deployed to clear airfields after an attack, cost upwards of $250,000 each.

These costs, of course, don't take into account the expenses and amount of time required to train military operators and pilots. According to the Department of Defense, it takes 1½ to 2½ years for a pilot to become fully operational and mission-ready. The cost of basic flight training was estimated by the DOD in 1999 at $1-million per pilot, while it can cost more than $9-million to train a pilot for specialized missions — and, at last count, the Air Force has more than 12,500 pilots on active duty.

Information is passed between system components using interfaces and communication standards defined in the CARNAC architecture (top). The CARNAC system builds on technologies from industry leaders to create a fully autonomous flight control system to fly a small general aviation aircraft after initial development with a flight simulator (bottom).

To help the military bridge the technological and economic gaps between traditionally manned vehicles and autonomous ones, robotics companies are researching and developing “drop-in” applique kits that allow legacy vehicles to be rapidly converted into autonomous, robotically controlled ones. Since these retrofit kits do not require modifications to existing vehicles, they can enhance the system performance of existing platforms, reduce costs, and enable new operations, including missions that dictate a high level of risk for military personnel and, therefore, are compatible with the implementation of highly adaptable robotic systems developed at lower costs than that of new robotic vehicles.

One applique kit in development is the use of autonomous robotics systems to control existing aircraft. Pittsburgh-based RE2 Robotics, a Carnegie Mellon University spinoff, using a Direct to Phase Two Small Business Innovative Research contract, is working with the Air Force Research Laboratory's (AFRL) Center for Rapid Innovation on a robotic system named “Common Aircraft Retrofit for Novel Autonomous Control (CARNAC),” in which a robot would replace a seat in a normally manned aircraft's cockpit. By expanding upon existing autopilot technology, the system is designed to fully operate an aircraft, from takeoff, through performance of a mission, to landing. Through the use of human-oid-like robotic manipulation capabilities, dedicated custom actuation, vision-based flight-status recognition, and cognitive architecture-based decision making, the system will interface with the same physical controls that a human pilot uses, including a cockpit's rudder pedals and yoke. Dedicated actuators and robotic arms that can mimic human dexterity will manipulate controls and respond to standard onboard gauges. Camera, range, and tactile sensors will be mounted to the robotic arm's end effector, allowing the system to perceive and process the state of controls and modify that state when necessary. The system would allow the aircraft to be completely unmanned, with no crew aboard.

RE2 Robotics CEO, Jorgen Pedersen (left), and CTO, Keith Gunnett, display one of the manipulator arms used in their innovative CARNAC robotic applique kits. (Photo: RE2 Robotics)

While autonomous flight systems and unmanned air vehicles (UAVs) are already being developed for the Air Force, those systems typically require either a completely new build, or extensive and irreversible modifications to an aircraft. CARNAC will be the first truly “plug-and-play” system in which a robotic pilot could be quickly installed into the cockpit of an existing airplane without making any modifications to the aircraft, relying upon onboard power, or connections to a fly-by-wire system. The drop-in pilot can be just as easily removed, allowing for quick conversion back to a traditional, manned aircraft. While the technology is being tested on a simulated Piper Seminole, the overarching goal is to design an applique kit that could easily be configured for use in a variety of aircraft, both civilian and military.

“Development of robotic applique kit pilots will enable a whole range of new missions for the military that can leverage the significant number of manned aircraft available, both civilian and military,” said CARNAC principal investigator Dr. Andrew Mor of RE2 Robotics. “These missions will cover the range from extremely long-duration missions that are not viable with a pilot in the cockpit to missions that are too dangerous to execute with a manned aircraft. And since these missions will be performed with existing aircraft, the cost/benefit ratio is improved significantly, especially in comparison to the development of a purpose-built autonomous aircraft.”

“Unmanned flight operations utilizing traditionally manned airplanes offer an increase in mission planning flexibility for a large set of missions and reduced cost while leveraging existing traditionally manned airframes,” said Dr. Alok Das, AFRL Senior Scientist and leader of the AFRL Center for Rapid Innovation. “Non-invasive approaches to robotically piloted aircraft using existing commercial technology and components offer the benefits of unmanned operations without the complexity and upfront cost associated with the development of new unmanned vehicles. Unmanned, low cost cargo transportation, resupply, refueling, and ISR missions are envisioned applications of this technology.”

Enhancing Safety During Rapid Airfield Recovery

The Piper Seminole flight simulator from Redbird Flight Simulations that is being used to develop and test RE2's CARNAC robotic applique kit. (Photo: RE2 Robotics)

Like UAVs, unmanned ground vehicles have a long history with the U.S. military, and many are already in use by the military to deliver supplies, complete dangerous missions or conduct reconnaissance. As autonomy and artificial intelligence technology progress, the military will continue to research and develop systems that serve to protect troops and reduce their cognitive and physical burdens. Another of these systems currently in development at RE2 is the Rapid Airfield Damage Recovery-Teleoperated (RADR-T) program, which will enable the Air Force to use existing construction ground vehicles as robotic vehicles during clean-up efforts after an airfield strike.

Unexploded ordnances (UXOs) create a significant hazard during cleanup of a stricken airfield. According to a report published by the Air Land Sea Application Center in 2001, U.S. military personnel have been killed or injured by UXOs in almost every conflict in which the United States has participated. The clearing of UXOs is one of the most hazardous occupations in the military, and yet doing so quickly and efficiently is critically important, as it affects the continued execution of operations.

Following an enemy strike, the goal of the U.S. Air Force is to repair an airfield in 8 hours or less. This timeframe includes the clearing of UXOs, which can be found in spalls and cracks or under debris, and which could detonate unexpectedly during the clean-up process. Typically, trained explosive ordnance disposal personnel handle the clearance and safe detonation of UXOs. However, following an attack, a large number of airbase personnel perform RADR duties by hand, including the identification of possible UXOs, as well as the positioning of them for neutralization.

Typically, the Air Force deploys manually driven front-loaders to help clear debris from the airfield. Taking into account an average “dud rate” of 5-percent for UXOs, this is an extraordinarily dangerous job. For instance, a B-52 that drops a full load of 45 cluster bombs, each of which produces 650 submunitions, could produce approximately 1,462 unexploded bombs that must be cleared quickly by Air Force personnel. While technology components, such as the laser-driven Recovery of Airbases Denied by Ordnance, or RADBO, exist to assist military personnel in this process, there is no system that is 100 percent autonomous.

The RADR-T program would allow the Air Force to leverage existing assets to reduce the risk of severe injury or death to personnel during damage assessment and cleanup. Similar to the CARNAC system, RADR-T will employ a drop-in robotic driver that can provide real-time switching between manned and unmanned operation. The objective is to develop a viable in-cab telepresence solution that will rapidly adapt to a variety of operational requirements, depending upon the type of vehicle. As with CARNAC, RADR-T will require no special modifications to existing vehicles.

“By creating the capability to easily turn construction vehicles into robotic systems without losing the integrity of the original man-drive vehicle, the personnel responsible for conducting the dangerous mission of airfield damage recovery will soon have the ability to perform their jobs at a safe distance, when necessary,” says Jorgen Pedersen, CEO of RE2 Robotics.

Leverage Existing Assets

Moving forward, the development of drop-in robotic kits will allow the U.S. military to take advantage of advancements in unmanned systems while maintaining its fleet of existing vehicles and leveraging an even larger fleet of existing commercial vehicles. Applique kits such as those in development at RE2 Robotics will initially provide military personnel with the opportunity to utilize existing assets to conduct missions from safe, remote locations. Eventually, these retrofit kits will offer greater autonomy, including the ability to self-direct high-level tasks, such as driving a ground vehicle or flying an aircraft.

As artificial intelligence and machine-learning methods advance, fully autonomous, unmanned systems are expected to become an integral aspect of military training exercises and operations. By enhancing operational capabilities, advances in system autonomy will help to ensure that the U.S. military maintains its strategic advantage during global conflicts and warfare.

This article was written by Jennifer Brozak, Marketing Communications Manager, RE2 Robotics (Pittsburgh, PA). For more information, visit here .