Laser vision for robots

Rebecca C. Jernigan, rebecca.jernigan@laurin.com

Modern-day soldiers are trained in urban warfare. They know how to maneuver the streets of a strange city, where the enemy is likely to hide, and how to avoid booby traps. Many technological advances have focused on making soldiers’ jobs easier and safer, including development of the PackBot from iRobot Corp. of Bedford, Mass. This tough robot, designed for use in small spaces or hazardous situations, currently is being used by American troops in Iraqi cities.

Researchers at the US Army Research Laboratory (ARL) are working on a sensor system upgrade for the robot – as well as for other small, unmanned ground vehicles – that consists of a low-cost, compact and low-power ladar (laser detection and ranging) imager. The research stems from the difficulties encountered when attempting to use two-dimensional imagery from visible and IR cameras to guide autonomous vehicle operation. These problems result mainly from variations in environmental conditions and distances as well as from unpredictable lighting.

Ladar systems provide their own laser light and reliable three-dimensional imagery, calculating the distance to each pixel in their view to offer a sense of depth that is lacking in 2-D imagers. Unfortunately, most of the ladar systems tested previously had flaws, including a short detection range, poor range resolution, high power consumption or large size.

The ladar system being developed by the US Army Research Laboratory delivers an image that is different from what a human eye would see. Compare the photograph (left) with the ladar image (right). Images courtesy of Barry Stann.

ARL’s experimental system consists of a Perseus TS36 PS36L 1541 pulsed fiber laser from Aculight Corp. (now Lockheed Martin Aculight Corp.) of Bothell, Wash., and a two-axis scanning mirror from Mirrorcle Technologies Inc. of Albany, Calif. To reduce the system’s price, the laser eventually may be replaced with a 1.55-μm fiber-coupled diode laser from QPC Lasers Inc. of Sylmar, Calif., although that would require the latter device to be heavily modified. The substitution would decrease the overall cost of the illuminator system by a factor of ~10.

The researchers are working on final integration of the baseline scanned ladar system, having completed the testing of the electronic boards. After they receive feedback about the image quality, they will work with the laboratory’s robotics group to incorporate the system into the PackBot. The integration should be completed this fiscal year.

Because of the small size of the imaging system, the transmitter and receiver modules will fit inside the robot’s sensor head, while the rest of the components are carried inside the robot’s body.