Journeys to the past

Imagine trying to solve a 3-D trigonometry problem on the fly, when you haven't had any sleep in the past 48 hours and a wrong answer could have serious consequences.

That's a fairly accurate description of the kind of constant pressure an ocean engineering team taking part in a deep-sea archaeological expedition last summer was up against. They were conducting field trials of a new sonar-based navigation/control system for a remotely controlled submersible robot called Jason.

Funded by the U.S. Navy, National Geographic, the Institute for Exploration and private donors, the expedition to the Mediterranean Sea in search of ancient shipwrecks is characteristic of the fieldwork being carried out by a bold new generation of researchers: They're searching for clues to our past by deliberately seeking out uncharted territory under the sea and in outer space.

Until recently, deepwater sites such as the 20-square-mile area where the remains of eight, two-thousand-year-old Roman trading ships were discovered have been virtually inaccessible. Since the 1950s, deep-diving submersibles have existed. But more scientists now have an opportunity to penetrate these alien landscapes, thanks to a growing interest in advancing not only the science but also the technological tools behind it.

One scientist at the forefront is Dr. Robert Ballard, founder and head of the Institute for Exploration at Mystic Marinelife Aquarium (Mystic, CT). An expert in deep-ocean scientific research, Ballard's triumphs include the discovery of the Titanic and the German battleship Bismarck.

Pushing the envelope. "He's different than a lot of other scientists from the standpoint that he is very much interested in the technology itself,"says Dana R. Yoerger, an associate scientist at the Woods Hole Oceanographic Institution (WHOI) and head of the ocean engineering team during the recent Mediterranean cruise. "But although he strongly encourages us to be innovative and push the envelope, we have to be able to convince him that what we have is more than just some wacko new technology."

A remotely operated robot named Jason is one example of the successful partnership between the scientist and the engineer. A submersible, 2,200-lb vehicle developed by the Woods Hole Oceanographic Institution primarily for the oceanographic sciences, Jason was deployed by the ocean engineering team in collaboration with an international team of archaeologists headed by Dr. Anna McCann of Boston University. Using Jason, the team first precisely mapped the Roman shipwreck site with sonars and underwater cameras, then retrieved selected artifacts under direction of the archaeologists.

The surveying and sampling activities performed by the Volkswagen-sized robot require a high degree of precision and maneuverability. Before Jason and his descendants, including the Autonomous Benthic Explorer, the idea of using a computer-controlled vehicle to survey archaeological sites and handle priceless objects was considered pie-in-the-sky. "The archaeologists were skeptical that we could design a robot that could pick up an artifact without damaging it," says Yoerger.

Engineers overcame those doubts by demonstrating they could achieve very precise control over robot operations. In fact, a foundation of their efforts has been the development of the tools required to precisely navigate and map archaeological sites before the recovery of any objects.

But it has been no small task, and one that is evolutionary. Depth, altitude, heading, and attitude are relatively simple to track using today's sensor technology. But achieving fine maneuvering and X-Y positioning underwater is extremely difficult. One problem is that few techniques are available for reliable 3-D underwater navigation.

The standard approach is 12 kHz long baseline (LBL) acoustic navigation, which works by measuring the time it takes an acoustic signal to travel the straight-line distance between two points. The technique, which requires an unobstructed line-of-sight, uses fixed-position transponders as points of reference.

But limited range is a major shortcoming of any acoustic navigational technique, due to the swift attenuation of high-frequency (300 kHz and up) sounds in water. LBL has a maximum range of only about 10 km, with a tracking precision of approximately 1 to 10 m. Accuracy is proportional to the size of the network. Errors could be corrected, but the frequency of fixes is on the order of only about 0.1 Hz.

Louis Whitcomb, an assistant professor of mechanical engineering at Johns Hopkins University (Baltimore, MD) and control engineering on the ocean engineering team on the recent Mediterranean expedition, points out that the accuracy of LBL is insufficient for conducting precise surveying and retrieval activities. "When the position sensor is only precise to 1.0 m, it is impossible to precisely control the vehicle to within a tracking error of 0.1 m," says Whitcomb.

Hoping to gain an order of magnitude improvement in navigational precision and update rates over the 10 kHz LBL system, the researchers recently tested a new 1,200 kHz bottom-lock doppler sonar navigation system in the Mediterranean. The advantage of doppler sonar is that it precisely measures the vehicle's velocity with respect to the ocean floor, avoiding the error in standard velocity estimation techniques.

What did the researchers learn? A hybrid system incorporating both doppler and LBL provided the best tracking performance. The combination decreased errors to approximately 0.1 to 0.2 m and increased update rates to several hertz.

Off the chart. Navigation is just one aspect of underwater vehicle performance that engineers are working to improve. To assist Jason in its surveying and retrieval efforts, David Mindell, assistant professor in the program on Science, Technology, and Society at MIT, co-principal investigator on the expedition and technology historian, developed a sonar navigation system that is used to create exceptionally accurate, digital photomosaic maps of underwater archaeological sites.

In deep water, long-range, low frequency sonar can be used to map the general area of a shipwreck. However, the size of the smallest resolvable object is on the order of a large building. Mindell's strategy for detailed acoustic mapping consists of transmitting high frequency (300 kHz) sound waves over short distances. The centimeter-scale mapping precision he achieves is sufficient for the safe, efficient surveying and retrieval of small objects at an archaeological site.

Mindell is currently focusing on strategies for taking prototype systems to the production stage. "It may take four PhDs to operate the prototype, but that's not cost-effective," says Mindell. "We know we are at the production stage when anon-electrical engineer can operate and maintain the system as a standard part of the operational suite. (For the record, a PhD-free team runs all routine operations.)

The fact that Jason is not merely a research curiosity, but is an actual operational system poses a challenge for the research team. More often than not, the first opportunity to test out new prototypes is out in the field, when no one can afford a true calamity. That means problems must be corrected on the fly. "If you think you are going to go out there and have everything run perfectly, you are going to be disappointed," says Yoerger. "I've been exhausted to the point of pain trying to work out math problems while everyone else stands there and waits. But you know, I wouldn't trade the experience for anything."

Mining Martian fossils. At the turn of the Century, an unmanned robotic rover will roam a very different terrain--on Mars, investigating the role water may have played in the planet's past and searching for evidence of ancient life. The mission, to be launched in 2001, is part of NASA's long-term systematic exploration of the red planet. A science team from the Jet Propulsion Lab and Lockheed Martin Aeronautics and researchers from Cornell University will be heading up a series of expeditions to Mars.

One objective is to retrieve rocks and soil, bringing them back to Earth to study for evidence of ancient Martian environmental conditions and possible life activity. Other goals include identifying the elements of the rocky Martian landscape and providing color stereo imaging of the planet's surface.

The rover, dubbed the Mars 2001 Lander/Rover, is designed to travel longer and further than its predecessor, the Pathfinder Sojourner. During its year-long excursion, it will travel 100 km. It will feature an integrated suite of instruments for conducting in-situ scientific analyses of surface materials.

Three types of spectrometers will be used to help gather mineralogical data, including a thermal emission spectrometer. Operating within the thermal infrared (IR) range of the electromagnetic spectrum, it will study the mineral composition of surface features from afar. One major advantage of using thermal IR is its ability to penetrate though Mars' dusty landscape.

A microscopic imager mounted on an instrument arm at the front of the rover will be used to produce extreme close-up views of surface features such as rocks and soil. These images will aid the analysis of sedimentary rocks that formed in water and yield data on the microscopic features of rocks formed by volcanic and impact forces. The imager's large-image format of 1,024 x 1,024 pixels allows even somewhat large areas (measuring up to 2 cm square) to be imaged without repositioning the arm. And its 1-cm depth of field allows the imaging of the planet's irregular terrain without complex focus control.

"I don't know what we expect to find, says Steve Squyres, professor of astronomy at Cornell University and principal investigator of the mission. "If we knew, it wouldn't be exploration."

JASON AT A GLANCE

Depth capability:6,000 m

Size:2.2 m long by 1.2 m high by 1.1 m wide

Weight: 1,200 kg (in air)

Maximum speed: 1.25 knot forward; 0.75 knots lateral; 1 knot vertical

Propulsion: 7 brushless dc thrusters providing the following thrust:

260 N (120 lbf) fore, aft

200 N (45 lbf) lateral

300 N (135 lbf) vertical

Navigation:Long base line responder or relay transmitter/receiver

Attitude sensors:Two-axis clinometer, 0.1 deg resolution

Two-axis rate sensor, 0.01 deg/sec resolution

Heading sensors:Gimballed flux-gate compass, 0.36 deg resolution

Gimballed gyro, 0.1 deg resolution

Altitude sensors:100 kHz updating at 2 Hz, 33 m range, 0.1 m resolution

Acceleration:Tri-axial fore-compensated accelerometer

Manipulator:5 deg of freedom, electric-powered, maximum lift 15 kg at .5 m

Source: Woods Hole Oceanographic Institute

Other applications for archaeological technology

- Remotely operated underwater or airbornevehicles.

- Robotics for materials handling, precision assembly, and semi-conductor operations.

- Factory automation and machine tools.

Engineering Challenges

Achieving superior maneuverability and precision positioning of remotely operated or automated vehicles by:

- Achieving reliable 3-D navigation accuracy of underwater vehicles on the order of .1 m

- Developing actuators, optical and acoustic sensors, and computer systems for precision robotics control

- Bringing prototype technologies to production stage