Miniaturization, Networking Pervade Future Unmanned Systems

It’s not a bird or plane. It’s an unmanned aircraft.

Today’s unmanned aerial vehicles look like fighter aircraft, but the next generation of aircraft will more likely resemble brainy birds. By taking advantage of miniaturization, researchers and engineers are exploring ways to put the power of intelligence, surveillance and reconnaissance collection directly into the hands of warfighters. In the not-too-distant future, these systems will employ networking technologies to give commanders ubiquitous situational awareness.

Industry analysts predict continued growth in the unmanned aerial vehicle (UAV) domain. According to researchers at Frost & Sullivan, San José, California, the market, which totaled $1.4 billion in 2002, is expected to climb to more than $1.7 billion by 2007.

While the market for the current generation of UAVs increases, scientists are developing and testing unmanned systems that will populate battlefields during the coming years. Organizations such as the Defense Advanced Research Projects Agency (DARPA) and the Office of Naval Research (ONR), both in Arlington, Virginia, are investing in the research of micro vehicles that can share and relay critical information.

Sam Wilson, program manager, tactical technology and organic air vehicles program, DARPA, relates that work in advanced concept technology demonstrations (ACTDs) will get the technology into the hands of the military quickly. DARPA researchers are working with Allied Aerospace, Newport News, Virginia, on a micro air vehicle (MAV). It is 9 inches in diameter with a 7.5-inch propeller and can lift between six and seven pounds of payload.

The MAV ACTD focuses on a ducted-fan design. “So you have a vehicle that can hover but can tilt and can go fast. The current demonstrator has a projected speed of 68 knots,” he says.

The MAVs’ size will allow warfighters to carry them in a backpack and launch the vehicles themselves. “What we’ve been told by the [U.S.] Army’s small units is that they don’t have pilots at that level, but they don’t want pilots. They want a vehicle that just collects data and sends it back to them,” Wilson relates.

To deploy an MAV, geospatial positioning system (GPS) way points are programmed into the vehicle. “One of the capabilities that we think is very important is what we call ‘perch and stare’—the idea of it landing and watching an area,” he explains. The aircraft would be programmed to fly to the edge of a building, to land, then to observe the activity below.

In terms of size, DARPA’s MAV is comparable to the U.S. Marine Corps’ Dragon Eye (SIGNAL, April 2002, page 49). Dragon Eye, which is battery-powered, can fly up to three times longer than an MAV; however, because DARPA’s vehicle lands, collects data and relays only changes in the environment, it realizes an order of magnitude increase in time on station.

Wilson explains that the electric version is estimated to expend 1.5 kilowatt hours of power to fly for approximately 25 minutes. “If I take that same amount of power and use it to run all the computers, sensors and transmitter, which are running during flight, but not fly, I’m down to expending approximately 20 watts. I have dropped from 1.5 kilowatts to 20 watts. So I have a huge reduction there,” he says.

Because the vehicle collects data while stationary, the power that is not used for flying can be used to gather and relay information. Wilson estimates that one MAV could continuously collect and send information for as long as a week. However, because only changes are transmitted, the MAV could continue on station for as long as a month effectively, he adds.

“The utility there is that, while it can fly for only 10 or 15 minutes, once it lands, the operator has hours of watching capability. And once it lands, it can do change detection. If nothing changes pixel to pixel, then it doesn’t send anything back,” Wilson relates.

For instance, if a person sticks his head out of a window, the camera will record that data and notify the operator that something has changed. “The robot doesn’t know what it saw, but it says, ‘Humans, look at what I found.’ Then the human does what he does best—knows it’s a bad guy with a gun in that window and knows not to walk in front of that window. So we reduce the bandwidth; we reduce the operator time; and we think that’s a very powerful tool,” Wilson explains. One possible interface with the operator would be the Land Warrior system, he adds.

At this point in development, the MAV uses only electro-optic/infrared cameras; however, the goal is to move to acoustics, so any noise can be tracked and relayed to the operator. In addition, current thinking is to use wide-angle lenses and transmit only the part of the image that changes, Wilson offers. Digital images from various MAVs could then be stitched together so that the pan-tilt-zoom feature is electronic rather than mechanical.

The next iteration of MAV will use diesel fuel. Wilson estimates that the vehicle would use a pint of fuel to run for one hour. During the next 18 months, DARPA researchers plan to develop an electric-diesel hybrid version of MAV. The electric motor would start the engine, and the engine would run the motor as a generator to recharge the batteries.

If these MAVs are to be used in large numbers, Wilson points out that the price per unit must be acceptable. At least part of this primary goal of the ACTD was achieved when the price of the airframe alone went from more than $20,000 for the first unit to $325 per unit by the time the first experiments were held earlier this year.

Although the price of the airframe has been reduced to an acceptable level, once loaded with avionics and sensors Wilson estimates that each MAV will cost approximately $10,000. “We want to be at a place where they are affordably expendable,” he says.

Wilson predicts that MAVs could be transitioning into the force in the 2004 to 2006 timeframe. The Army’s Future Combat Systems program may purchase several to equip the first units soon after that time.

In addition to the MAV project, Wilson is involved in the organic air vehicle (OAV) program. Organic vehicles would be attached to the squad, platoon or even a company. This aircraft would be larger than the MAV—approximately 19 inches in diameter and weigh 30 to 40 pounds—but would feature a larger payload. The OAV would include the perch and stare capability, although it would need a larger landing area.

Wilson’s third program involves UAV autonomous collision avoidance systems. The first year of the program will focus on flying in forests. Research the second year will involve flying in and around buildings, and during the third year of work research will focus on tunnels and caves, he shares. To reach this goal, DARPA researchers will examine small radars, laser-range finders, acoustics, stereo vision and mono vision. Initially, the vehicles would travel at half a mile an hour, comparable to soldier movement. A superior capability would be to fly at two miles per hour, Wilson explains.

Wilson’s vision for the future closely resembles work now being conducted by ONR’s Dr. Allen Moshfegh, program manager, autonomous intelligent network and systems (AINS) initiative, as unmanned air and ground vehicles are networked to work as a team. They would relay information without soldiers knowing or caring how the data is acquired or transmitted.

“My goal is to design intelligent systems, infrastructure, architecture, technology, sensor processing and information fusion, networking and autonomy concepts. So when you have all these intelligent modules working in concert you can wrap a casing around it, and you can call it a smart tank, smart car, smart airplane, smart bomb. But before you invent another air platform, you really have to understand how you can fly these in a safe manner reliably over a populated area,” Moshfegh says. These intelligent systems must be able to interact with humans at the Pentagon, on a ship or at any other command and control stations where people make the decision about battlefield targets, he adds.

The program has three focus areas. First, researchers are exploring intelligent autonomy, which is planning, reasoning and decision making by the human and the machine together. Operators would program teams of unmanned platforms with high-level tasks, then the vehicles would carry out the mission. The second element of the program is the wireless networking of platforms in a dynamic environment where no infrastructure exists. Intelligent control of the unmanned nodes of the network by humans is the final goal of the program.

“This is a totally integrated concept. That’s how this program is different than any other research program,” Moshfegh relates. The work includes both scientific reports and hardware and software technology demonstrations, he adds.

Although today’s UAVs are effective, they require a large staff to carry out a single mission. Moshfegh contends that the best way to deploy multiple platforms in a cost-effective and viable manner is by using autonomous intelligent agents that can communicate and interact. Humans only intervene to assign specific tasks.

“Humans have to be at very high levels, supervising the infrastructure and providing direction to the nodes. These nodes would take the low-level mission on their own and interact with each other, conduct data fusion and low-level decision making amongst themselves versus having humans directing every minute task for each node,” Moshfegh shares.

Hundreds of nodes would be networked to carry out complex tasks in an environment that would be unstructured, such as Afghanistan. “We were lucky in Afghanistan that they did not have technology to jam our GPS infrastructure. Without GPS, none of these platforms can navigate, and UAVs like Predator and Global Hawk would be rendered useless. The goal in the AINS concept is to create our own pseudo-GPS signaling so that we’re not going to rely on our present GPS satellite infrastructure. It would provide navigational information so that these nodes can fly without having infrastructure that can be jammed or brought down or destroyed,” he says.

During the past six years, the team has developed the dynamic multiagent network, demonstrating all of the protocols. Fully autonomous docking between two platforms has been demonstrated, so a UAV can carry and deliver munitions to troops in a treed area or on a ship without human intervention. This is accomplished through passive sensing, which Moshfegh says involves a video camera on the platform that searches for the landing area, finds it then autonomously lands. Because the systems do not use active or semiactive sensing such as radar, operations can take place in a passive manner, impeding detection.

In addition, the team has demonstrated that nine helicopters can be navigated simultaneously, safely and autonomously in a precise manner in limited space.

Moshfegh believes that these types of technologies can be used in the field in the next six to eight years; however, some experts predict that it will be 15 to 20 years before they are incorporated into operations because of doctrinal issues that need to be resolved, he says.

Biologically inspired concepts of multisensor fusion also are being examined. The team is capitalizing on how the human brain coordinates input from sensory organs. One goal is to determine how to transfer this functionality to a technical platform.

Industry is well aware that the future is likely to include multiple types of unmanned vehicles. Charlie Guthrie, director of rapid prototyping and advanced concepts, unmanned systems, The Boeing Company, Seattle, describes that firm’s work in this field as broad-based. “A lot of companies have taken a focused look at being a niche player. We looked at it from being a full-spectrum provider that best addresses the needs of our customer base,” Guthrie says.

Although the company is known primarily for its involvement in the unmanned combat aerial vehicle program, it also is investing in other types of unmanned vehicles. Guthrie believes that, although the military is the primary customer today, in the future these aircraft will have homeland security and commercial applications.

To develop some of these systems, Boeing is working with other companies. For example, work on ScanEagle, one of the company’s newest UAVs, is being conducted with The Insitu Group, Bingen, Washington, a small entrepreneurial company. The four-foot-long vehicle, which has a 10-foot wingspan and can fly at speeds of up to 68 knots, is classified as a long-endurance vehicle, operating for up to 60 hours. It could support network-centric warfare by acting as a communications relay station.

“The applications for ScanEagle initially are in the different aspects of reconnaissance and surveillance. We are addressing service needs in that respect and are trying to establish where and how it can support homeland security,” he says.

The Buster, an even smaller UAV that can fit in a backpack, could provide base security as well as allow commanders to survey areas prior to establishing a forward base. This vehicle also has civil security applications, Guthrie adds.

Working with DARPA, Boeing is designing a fuel-cell-based propulsion system for an ultra long-endurance UAV. The project, called Ultra Leap, involves a vehicle that can operate for two weeks and fly at altitudes of up to 60,000 feet. Boeing plans to incorporate currently available automotive fuel cell technology into the propulsion system.

In addition to developing UAVs, Boeing is working on other unmanned craft, including ground and underwater vehicles. These systems can be useful in military, civilian and commercial applications, Guthrie says.

While the company’s efforts support current UAV development, much of the work will support systems that the military will use in the next five to 10 years. To take advantage of the benefits of UAVs on the commercial side, several regulatory issues need to be resolved. Industry and the Federal Aviation Administration are working together to address these issues with a goal of flying the systems with the same level of restrictions that currently govern manned systems, Guthrie states.

Additional information on future unmanned aerial vehicle work is available on the World Wide Web at  http://www.darpa.mil/tto/programs/fcs_oav.html .