Using new trajectory-planning software they have developed, the researchers say spacecraft will dip into a planet's atmosphere, harness aerodynamic forces to fly like an aeroplane, and then zip back into space, having steered into a new orbital trajectory.

The idea is to save precious fuel, says Bruce Conway and his colleague Kazuhiro Horie of the University of Illinois at Urbana-Champaign. Today, says Conway, spacecraft must carry enough propellant to carry out most of the position and orbit changes they might be expected to make during their lifetime.

The fuel's weight cuts the craft's useful payload, so space scientists have long pondered the possibility of building craft that can use a planet's atmosphere to steer onto a new course.

Until now, space probes have only used atmospheres for "aerobraking"--where the craft skims the outer edge of an atmosphere to slow down and gradually change the shape of its orbit. The NASA Mars Global Surveyor probe did just this in 1997.

"In aerobraking, you're only interested in getting drag, reducing the amount of energy that the spacecraft has," Conway says. But the "aero-assisted" vehicles he and Horie envisage would use aerodynamic lift to take a probe into a new orbit rather than merely slow it down.

"If you have some control surfaces, you can steer the craft," he says. An articulated solar panel might do the trick. "It would be like moving the ailerons on an airplane," Conway says. Alternatively the craft itself--or its heat shield--could have an aerodynamic "lifting body" shape.

The Illinois work was focused on trajectories for the interception of one satellite by another, determining the best route for an aero-assisted spacecraft to take in order to fly safely through an atmosphere and then exit to a given orbit, depending on the craft's lift and drag capabilities, and the characteristics of the atmosphere.

"The program finds the trajectory from a given initial position that enables the vehicle to intercept the target in either the minimal time or with minimal fuel," Conway says (see "Up a bit...", below).

Crucially, it does so while taking into account how hot the craft can safely get. "The heat it can stand determines how low in the atmosphere it can fly," he says. "That's the critical factor."

Charles Whetsel, chief project engineer for Mars Global Surveyor at NASA's Jet Propulsion Laboratory in Pasadena, California, says of the Illinois project: "It's definitely valuable work that will help us launch more efficient missions that use less fuel and which can carry larger scientific payloads."

To move a craft from one orbit to another, there are many possible trajectories but only one "optimal" trajectory that uses either the least fuel or takes the least time. The conventional way to find this, says Bruce Conway, is called "shooting".

A computer starts from guessed initial conditions and uses equations of motion to work out where the craft will end up after, say, an aerodynamic pass through a planet's atmosphere.

If the craft does not end up where it is wanted, the guessed initial conditions are adjusted and the process is repeated--"like adjusting the aim of a cannon to hit a target", says Conway. But this "iterative" process is very slow, and may never find an optimal solution at all.

So Conway and Kazuhiro Horie have worked out how to divide the trajectory into a curve and then chop it up into many discrete, mathematically malleable chunks. Conway says these can be adjusted to give "a correct optimal trajectory even from a very bad guess".