RADAR SENSORS on the AWACS sentry aircraft flying oval patterns over the Persian Gulf last May detected the unknown aircraft around 7:55 p.m. The crew designated the blip "Track 2202" and alerted U.S. ships in the area through the Navy Tactical Data System. There was no immediate cause for alarm.

As the aircraft headed farther south, AWACS updates on Track 2202 became more frequent, and radars and other electronic apparatus on ships confirmed the blip to be a single Iraqi F1 Mirage fighter flying about 3000 feet above the water. Around 9 p.m., the lone Iraqi pilot switched on his Cyrano IV fire-control radar to search for targets abetting Iran. Crewmen of the USS Stark, hunched over luminescent screens in the darkened combat information center, picked up the emissions and realized that the fighter was within striking distance.

The rest is known all too well. Through a series of blunders, the Stark failed to ward off two radar-guided Exocet missiles fired by the Iraqi plane. Remarkably, the launchings of the missiles apparently went undetected by the ship's various radar operators.

The Navy's report on the incident was released last month. The version sanitized for the public masks out all sections involving the performance of the Stark's radars, leading some readers to conclude that human error was entirely to blame. But reading between the lines makes it apparent that glitches in the ship's radars and electronics may have contributed to the disaster, as Stark's former captain, Glenn R. Brindel, contends. The missile "wasn't seen on any of the ship's radars," Brindel says. "If the sensors would have divulged the things they should have, then I'm sure my TAO {tactical action officer} would have taken additional measures."

The Stark controversy not only shows how dependent ships, aircraft and missiles are on radar sensors but also points to deficiencies in some of today's systems. Most search radars like those on the Stark, which belongs to the Navy's newest class of frigate, rely on mechanically steered antennas which re-scan an area every few seconds -- a relatively long time in many battle situations. The slow scan-time problem is exacerbated when radar is called upon to do more chores, from navigation to tracking of hostile and friendly forces to directing missiles to their targets. Is the stalwart sensor, relied on since World War II, still up to the job?

Radar (Radio Detection And Ranging) is often preferred to "passive" infrared or optical sensors because it measures distance accurately and can therefore be used to distinguish target range. {See Box.} It has the added benefit of long-range detection and the ability to see no matter what the weather. But its strength is also a military liability. By emitting signals, radar often gives away its identity because it has a distinctive "signature" based on its transmission parameters.

Passive sensors -- those that rely solely on receiving emissions from potential targets -- are often good for identifying targets but cannot measure range well: Without knowing the emitter's strength, it is impossible to determine how its signal was affected by distance. The U.S. military now favors a mix of "active" radar and passive sensors to monitor its adversaries.

New radar systems range from colossal over-the-horizon backscatter radar which bounces waves off the ionosphere to detect aircraft as far as 2,000 miles away to "quiet" radars that generate a rapid fire of beams transmitted in random sequence with constantly changing frequency. The goal here is to baffle enemy receivers -- devices similar in principle to commercial "fuzz-busters" -- with a signature that appears to be no more than random noise.

Other recent developments include: 1) phased-array radar for the B1B bomber and the Aegis cruiser that adds new dimensions to radars' dual use as a detector and an abettor in guiding weaponry; 2) Doppler radar to distinguish moving targets from stationary ground clutter and, for civilian air transport, to detect hazardous wind shear; 3) synthetic-aperture radar, using antennas miles long in effect, to peer through foliage and several meters into the ground; and 4) use of smart computers to analyze radar signals.

Expanding Horizons

The phased-array radar, instead of using a typical mechanically steered antenna, employs a fixed flat plate that inside looks like a huge honeycomb, with its concatenation of cells. Each small cell in the array usually has its own antenna. Adjacent antennas radiate energy at the same frequencies. These signals intermingle and reinforce each other to produce one large beam.

The unique characteristic of the phased array is its ability to electronically steer the beam in millionths of a second -- even though the antenna face rests immobile. When all the individual antennas send signals timed precisely in phase the beam will be directed straight ahead. But by electronically orchestrating small delays across the face of the array, the beam can be shifted at great angles. Each delay causes a signal to lag a fraction of a wavelength behind the signal from a neighboring element. The delays increase successively in a motion like rows of dominoes collapsing. The result is a change in the beam in the direction of the increasing delay (or where the last row of dominoes would fall). By varying the magnitude of the time delay, the angle of the beam is controlled.

Phased-array radar is the linchpin of the Aegis cruiser, the Navy's latest warship, as well as the B1B bomber -- the first operational aircraft to employ it. Because the system steers the beam electronically, it can be directed much faster than conventional radar. The B1B's beam can be shifted in 150 millionths of a second (theoretically more than 6600 times a second), compared to 1-to-2-second scans in mechanically steered systems. The speed makes it possible to "interleave" various functions, such as mapping the ground, following terrain and delivering weapons. In effect, the phased array can simultaneously track many targets while it searches for new ones.

Phased-arrays require large amounts of computing power to send precisely timed commands to the thousands of cells across the array and assess the returns. This is one reason why only recently they are being put on ships and large planes. But with the inexorable miniaturization of computers, proposals are being made to put the phased array on small fighter aircraft like the F16.

Seeing the Wind

To distinguish moving targets from stationary ground clutter and tell how fast objects are moving, the military employs Doppler radar. It is also of interest to radar specialists at the Federal Aviation Administration who hope to install Doppler systems at selected airports to prevent accidents caused by wind shear. Two years ago a sudden downspout of air from high altitude caused the crash of a jetliner in Dallas, killing 136 people. Since 1970, crashes due to wind shear are the No. 1 cause of air travel fatalities, according to the National Transportation Safety Board.

Many people are familiar with the Doppler effect by hearing the changing pitch of a passing train's whistle. As the train approaches, its pitch rises to a higher frequency, then drops as it leaves. This same principle is used to relate the reflected microwave energy to the speed of the wind or other targets. (When a Doppler radar is monitoring the winds, it actually measures the velocities of dust, raindrops or insects being blown about, or lacking these, on minute changes in air density attributable to humidity variations.) The echo returns at an altered frequency proportional to the speed of the objects. Unlike conventional radars, Doppler radars measure this frequency shift of the pulses bounced from objects. For example, each mile per hour of the object's speed produces a frequency shift of about 31 Hertz (cycles per second) with a radar in the X-band frequency of 10.5 gigaHertz (giga denotes a thousand million), or an alteration of 72 Hz with the K-band frequency of 24 gHz.

Some bugs remain to be worked out. One problem is that radar sensitive enough to measure humidity differentials picks up many extraneous signals that need to be suppressed. And such systems are expensive. The Terminal Doppler Weather Radars are estimated to cost about $4.5 million per installation, and are expected to be installed starting in 1992.

Another development underway is the stationing of radar satellites in orbit. The incentive is great: Important parts of the Soviet arsenal, particularly in Eastern Europe, are often obscured by cloud cover. But radar can generate appropriate wavelengths not only to penetrate clouds but also to peer through foliage and even look beneath earth several meters deep to discover old river beds or buried mines.

For the same reason that one can listen to radio inside a house, the wavelengths of radar do not degrade as fast when passing through certain media. Hence they can penetrate dry soil. (Performance is drastically reduced when the ground is moist.)

The potential of more precise radar satellite surveillance has been demonstrated by such civilian projects as the Seasat and Shuttle Imaging Radar B experiments, funded by NASA. The key to each of these systems is the synthetic-aperture radar technique. This method takes advantage of a satellite's speed to make a small antenna work like one that is miles long. While the transmitter sends signals, its antenna gathers echoes from points along the ground. A computer selectively combines these echoes, based on time intervals and the Doppler frequency shift of the signal relative to the moving spacecraft. The effective "length" of the synthetic antenna is equal to the distance that the satellite moves during the time a signal is sent and received.

In 1978, Seasat pioneered these high-resolution developments. From an orbit 800 km. high, it could resolve some objects of less than 10 meters. The length of its synthetic antenna was 15 km. When its images were processed, the acuity of furrows from ship wakes astonished many people. The shuttle imaging radar launched in 1984 took the Seasat approach a step further. Unlike Seasat, the shuttle's 11-by-2-meter antenna could be moved at different angles. This enabled three-dimensional views. A series of images of Mt. Shasta in California obtained from a spacecraft 225 km. high looked as if they had been taken by a person circling the base of the mountain and pausing every 1000 meters or so to snap a polaroid.

There is much room for improvement, however. The use of multiple frequencies for radar imaging is a near-term possibility which Charles Elachi, project scientist of the Shuttle Imaging Radar at the Jet Propulsion Laboratory (JPL), Pasadena, Calif., likens to going from "black and white to color." Each wavelength emitted would be reflected differently according to the composition of the material. Both JPL and the Environmental Research Institute of Michigan in Ann Arbor have experimented with aircraft versions of such "color" radar for civilian and military uses. The first space version -- expected in the early '90s -- is expected to have three frequencies.

Too Much Too Fast

Radars often collect data at such rates that humans cannot analyze it fast enough. Studies in the '60s and '70s by the Applied Physics Laboratory at Johns Hopkins University showed that radar operators often missed targets even when they were detectable. The error rate, says a radar specialist, "caused a panic" and prompted Navy work into computer-aided detection.

Many Navy ships employ automatic detection systems to alert radar operators -- who may be tracking numerous blips on their screens -- to new targets. The problem with many of them is that they cannot distinguish spurious targets (reflections from birds, ocean, ground or clouds) from real ones. Consequently, radar operators often prefer to switch off the automatic-detection mode.

But even if fully operative, systems such as those used on the Stark may have trouble spotting small, low-flying objects. Capt. Brindel says that his radars were supposed to be able to detect Exocets "well beyond the visual horizon. They did not." And, he says, the first warnings came "almost simultaneously with the impact of the first missile." (An Exocet fired from 20 miles away reaches its target in about 120 seconds.)

The Stark's failure to defend itself was the result of numerous causes, including the crew's state of readiness. But the Navy's October report (Vol 1: Summary of Damage and Assessment) conceded that the French-made Exocet can pose a "challenging threat" to such ships' combat systems, especially when less than optimal attack angles and other conditions cause "degradation factors" which reduce performance "in some cases to zero."

Newer ship systems seek to spot missiles faster while reducing the number of false alarms. One way is to meld data from various types of radars aboard the ship. Such an integrated system can produce multiple detections within the scan time of a single radar. Similarly, by using integrated radars with different characteristics, authentic targets can be winnowed from spurious signals because a spurious signal on one radar is often not spurious in the electronic eye of another.