Scattershot approach
Cameras fired from a rifle? DARPA's ELASTIC effort is showing how
By Ross Stapleton-Gray
Monday 3rd September, 2007 (UPI)

A combat team entering hostile urban terrain faces the same challenges as a rat in a dangerous maze — open to ambush at every turn, without the ability to conduct reconnaissance and broaden its awareness beyond its limited line of sight. Technological advances, however, have made it possible to bombard the future battlefield — including the complex urban environment — with sensors, affording the military eyes and ears that can be lobbed into a window or fired onto rooftops. The Defense Advanced Research Projects Agency (DARPA) launched its Expendable Local Area Sensors in a Tactically Interconnected Cluster (ELASTIC) initiative in 2005 to promote research and development in this area. DARPA envisions a set of small, ballistically distributed optical imaging sensors able to form ad hoc wireless networks. These sensor nodes should also be able to support hierarchical networks for aggregating and up-linking high-bandwidth image and other sensor data directly to personnel on the scene. The sensor network would then report back to local nodes, where the information could be immediately used to guide tactical decisions. DARPA's efforts build on dramatic gains across a number of technologies already in battlefield use. Radio technology, for example, has evolved from large — the SCR-536 “handie-talkie” introduced in World War II weighed more than five pounds — and expensive, to become consumables, like ammunition and rations. It's now possible to condense the communications capability of that bulky field radio into a package a few centimeters across, and mass produce radio-equipped sensors at a price where they can be cost-effectively scattered to construct defensive perimeters and leave-behind sensor networks, or to project digital eyes and ears into unknown terrain in advance of the troops.
Similar gains have been made in other areas. Camera technology has plummeted in both size and price, to the point where more than half of all cell phones in use in the U.S. are camera phones. The ELASTIC solicitation also envisioned power consumption ranging from no more than several hundred nanowatts, to manage a network of sensors communicating sporadically and hibernating when possible, to a mere milliwatt to deliver compressed full-motion video, allowing for battery-powered sensors to survive for a considerable time. DARPA and others are also researching technologies for power scavenging, to allow remote sensors to collect energy from their environment and recharge batteries indefinitely. DARPA's initial ELASTIC solicitation was made under the Small Business Technology Transfer (STTR) program. STTR grants, capped at $150,000 for initial Phase I awards, are made to collaborating teams of a small business and a nonprofit research or academic organization, most commonly a university. The premise of the STTR program, established in 1982 by the Small Business Act, is that it will accelerate the commercialization of items developed in laboratories, and encourage the growth of small technology companies.

FOCUS ON IMAGING
Phase I awardees under the ELASTIC program were CDM Optics, teamed with the University of Arizona; Nova Research, teamed with Innovative Wireless Technologies and Johns Hopkins University; TechFinity, teamed with the University of Wisconsin, Madison; and Toyon Research Corp., teamed with NASA's Jet Propulsion Laboratory (JPL). Each team was tasked to develop a feasibility study of its selected system. Bob Cormack was the principal investigator on the ELASTIC work performed by CDM Optics, based in Boulder, Colo. Cormack said CDM's principal area of expertise was in computational imaging. “This is a broad field, but one of the things that our technology enables is the production of very small, very cheap cameras, with surprisingly good performance,” he said. “We came up with some interesting results; from sensor designs that could survive terminal velocity drops — distributed from a high-altitude aircraft, for example, or delivered by ‘shotgun' on the ground — to new methods of fusing, compressing and moving redundant data over a band-limited and power-limited ad hoc network.” At the University of Arizona, work connected to the ELASTIC program fell to the Optical Computing and Processing Laboratory within the Electrical and Computer Engineering Department. Its main concern was with combining multiple images to achieve improved resolution, designing custom optics to overcome sensor pixel-limited resolution, and refining the communications protocol (including compression) to increase network lifetime. By encouraging a diversity of approaches across a number of competing investigators, programs like the STTR and its better-known Small Business Innovation Research counterpart generate lots of ideas. Some prove feasible, others less so. “One thing we briefly looked at was delivering an image sensor by bullet,” Cormack said. “The bullet would be designed to decelerate the sensor on impact [by crushing], and leave the camera stuck to a wall, say, by a spike. It's kind of mind-boggling to imagine setting up a sensor net by machine gun.” To the TechFinity/University of Wisconsin team, wireless sensor networks appear to offer the most potential, when combined with collaborative image processing and object tracking. In this application, network nodes would lie dormant until awakened by “beacon” nodes, which would locate each other through the use of GPS. Of the four teams funded for a Phase I effort, the Toyon/JPL team was selected to receive follow-on funding to take its work to the point of a functional prototype. Toyon, founded in 1980 and based in Goleta, Calif., has a long history of high-tech R&D for federal agencies, including in areas of antenna and other radio-frequency devices, and C4ISR systems. Its partner, JPL, is a federally funded research and development center run by the California Institute of Technology. The STTR program allows small businesses to partner with a range of nonprofit research and academic entities, aiming to encourage the transfer of expertise developed in the research setting into products for customers in both commercial and government markets.

MILSPEC SYSTEM
Toyon's first ELASTIC sensor prototype is slightly larger than a quarter. Follow-on development includes an electronically steered imager array; 2.1 megapixel imagers; onboard, hardware-based image compression; and a single data link, in Defense Department-specific frequency bands, for the transport of telemetry and compressed image data. According to Richard Cagley, a Toyon engineer, these features will give the Toyon sensors some distinct advantages over current systems, such as the Remington Eye Ball R1, a baseball-sized wireless imaging device using a digital command-and-control link as well as an analog video link. Popular with law enforcement agencies, Eye Ball can be mounted on a stick, poked around corners and left in out-of-the-way areas. It is still a civil system, however, built to different standards than military hardware, according to Toyon. “While something like the Eye Ball is great from an operational perspective, there are mechanical and systems issues we felt needed to be addressed,” Cagley said. Cagley noted Toyon's use of an array of six imagers that allows slewing of the imager's field of view electronically, rather than mechanically, reducing points of failure and providing fewer entry points for sand, rain and debris. In addition, Toyon's sensor employs a single digital link for command and control as well as image data transport for multi-hop routing, relaying and reach-back capability. Also of interest to military customers, according to Cagley, is that Toyon's system operates in U.S. Defense Department-specific frequency bands. Remington's Eye Ball transmits its video in the 2.4 GHz band, which is increasingly popular for a broad range of consumer applications, such as wireless LANs, cordless telephones and baby monitors. The 2.4GHz band is also used as an unlicensed industrial, scientific and medical (ISM) channel, and is therefore crowded. Cagley also raised the issue of cost. “As camera modules are used, we can leverage the economics of scale from the cell phone market,” he said. “The all-digital link fed back to a PCMCIA card allows us to use commodity computing hardware, including PDAs, toughbooks and tablet personal computers. A soldier can use a COTS [commercial-off-the-shelf] PC, grab an image, and then send it electronically to someone else in the unit. That potential has the Army very excited.” Toyon is developing both hand-placed and thrown sensors, and the optical package for a ballistically-delivered system under development by the Army at Picatinny Arsenal, N.J., under ELASTIC and for another contract vehicle. Where the World War II-era SCR-536 was based on fragile vacuum tubes, the technologies incorporated into the newest sensors are to be sufficiently rugged to allow them to actually be fired onto the battlefield, using a unit's own ordnance, such as a light mortar or M203 grenade launcher. The ability to deploy a sensor ballistically requires that it be able to withstand considerable G-forces of firing, and, unlike a shell, to survive the impact. Mark Mellini, with the U.S. Army Armament Research, Development and Engineering Center Acoustic and Network Sensors Office at Picatinny Arsenal, said Toyon's work complements an existing U.S. Army program, SMADSnet (Small Arms Deployed Sensor Network). That effort had already been developing other sensor types — acoustic, magnetic — within the form factor of a 40mm shell able to be fired from an M203 grenade launcher, which would allow sensors to be deployed by soldiers using their standard small arms. The M203, commonly carried by the infantry as an attachment to the M16 rifle, has an effective range of 350 meters for area targets. Mellini characterized the state of R&D as technology readiness level 5 — in development and approaching technology demonstration. •