Sharon Weinberger, David Hambling andBill Sweetman Washington and London

One key advantage of unmanned vehicles is that their persistence is not limited by the human operator. This also means their size is set by the payloads and sensors that they carry, not by the need to accommodate a pilot or crew. But in many cases, this now means that unmanned systems run into an energy limit, both in terms of endurance and their ability to deliver power to radars, lasers and communications links.

Unmanned ground vehicles and small hand-launched UAVs continue to benefit from the fact that their electrical power requirements are not vastly different from those of consumer electronic devices, so they have been able to ride on a multibillion-dollar wave of investment. Larger systems face different challenges.

A number of efforts to develop multiday-plus endurance UAVs have foundered in the past year or so. The Defence Advanced Research Projects Agency's (Darpa) Vulture project for a huge solar-powered UAV was canceled earlier this year. Lockheed Martin's HALE-D solar-powered airship impaled itself on Pennsylvania treetops last July, on its first flight attempt. The U.S. Air Force let the gas out of the MAV6 Blue Devil 2 project with a stop-work order. AeroVironment's Global Observer was canceled after the first vehicle crashed at Edwards AFB, Calif., and the other U.S. hydrogen-powered UAV, Boeing's Phantom Eye, is under repair after a rough first-flight landing. The surviving airship program, the Northrop Grumman Long Endurance Multi-Intelligence Vehicle (LEM-V) is running late, to the surprise of almost no one.

The art and science of lighter-than-air control is not well understood, and even with hydrogen fuel, heavier-than-air ultra-long endurance requires light, long wings, unusual propulsion arrangements and aggressive weight-saving. Phantom Eye takes off from a trolley and lands on a skid-and-wheel combination, the latter being the source of its problems.

Power issues were one reason for the USAF's attempt (being opposed in Congress) to retire the RQ-4B Block 30 Global Hawk while retaining the U-2S. The Rolls-Royce AE3007H meets the RQ-4B's needs for efficient propulsion, but such engines are sensitive to generator loads at the edges of their envelope (high altitudes and low speeds) and cannot run the kinds of sensors that the U-2's fighter-based F118 can support.

One propulsion engineer notes that standard military specifications are part of the problem, requiring the engine to tolerate very rapid fluctuations in generator load—necessary in combat, less so in a typical UAV mission. But the problems involved are severe—and explain why the notion of a nuclear-powered UAV, explored recently in a Sandia-Northrop Grumman study, still gets serious consideration.

At the lower end of the UAV size scale, attempts to develop heavy-fuel engines (spark ignition or diesel) to replace gasoline-fueled reciprocating or rotary engines have been underway for decades. Advantages in efficiency and logistics (gasoline has disappeared from ground forces and is detested on ships) are undoubted, but the challenge of developing an HFE that is light, reliable and low in vibration (including torque vibrations that stress reduction gears) and that is supportable, is still daunting.