June 5, 2014
As I wrote last week, DARPA’s war on costs is an effort in technology, funding potentially “powerful new approaches” in armaments, sensors, navigation and communications, aimed at radically decreasing the cost of fielding the next generation of weapon systems. What may be less obvious is just how commercially disruptive those technological developments could be. After all, DARPA’s intended effect fits Clay Christensen’s classic description of a market disruption: entry by non-traditional suppliers with a technology whose performance is initially inferior to the state of the art, but far less expensive than anything on offer. For exhibit one, I will cite again the ACTUV program, the agency”s effort with SAIC to build a "Anti-Submarine Warfare Continuous Trail Unmanned Vehicle” without a crew.

The appeal of robotics to budget-contrained military customers is clear: war is a labor-intensive business, and Americans make for expensive labor. Americans on the other side of the planet make for particularly expensive labor, so they shouldn't travel there hundreds or thousands at a time, when a few sharp eyes can accomplish the same task. The technological solution to that problem may lie in autonomy, which as Mike Nemeth of Zyvex Technologies has put it, takes systems beyond the limits of physics and human physiology. As I wrote in May, this makes new missions possible, but it also helps accomplish existing missions at radically lower costs.

More specifically, as I wrote with some colleagues in ‘Robots in the Age of Pirates” (Proceedings of the USNI, December 2011), a small drone ship offers a compelling solution to part of this problem. As I noted last week, DARPA’s ship

     could track hostile diesel submarines for longer than they can stay underwater. If the submarine skipper gets angry about the pinging on his hull, torpedoing the drone won’t be a good idea—the loss of an unmanned, hundred-ton boat will just produce a flaming datum for the fleet. And if a legion of American geeks are right, it won’t ever get that far: DARPA crowd-sourced the tactics with a video game.

In 2010, DARPA awarded SAIC a contract to build its 150-ton trimaran. As the project was getting started, The Register opined that the ships should be deemed the Mary Celeste-class of robo-frigates, for ACTUV is intended to undertake missions of up to 90 days without a crew. Indeed, in the long run, what’s really different may be the implied reliability of her onboard systems. If engines and sensors can run without fail for 90 days, then why embark crew just for maintenance on any small vessel? The US Navy already suffers from this problem in a big and comparative way. Consider how American submarines have crews of over 100 sailors, while Swedish and German boats get by on fewer than thirty. Every additional body is a lot of expensive space inside the hull. Eventually, taxpayers fund those sailors’ impressive and long-lasting retirement benefits. If the crews and their support staffs ashore can be reduced without an undue loss in combat effectiveness, the long-term savings could be huge.

Indeed, in a commercial case in point, Rolls-Royce thinks that crews on cargo vessels are just pointless, and is working on designing completely unmanned ships. Bloomberg’s story on the subject cites a study by the accountancy Moore Stephens that even a small crew, and its accompanying accommodations, contribute to 5 percent of the weight, 12 percent of the fuel burn, and fully 44 percent of the operating costs of a large container ship. That’s a lot of money that the people doing the shipping could save by not paying for people actually on the ships.

Noted naval thinker B.J. Armstrong has argued in a counterpoint that we may be overselling drones. And as evidenced by that treaty convention assembled last month in Geneva, much of the world wants to ban truly robotic weapons. So perhaps we should not oversell the concept in all aspects. Instead, as General James Cartwright has repeatedly argued, what’s really needed may be a better man-machine interface, a better way to leverage the capital of mechanical labor when that is far less expensive, and at least nearly as effective, as a human presence. 

Perhaps then, in contrast to Rolls Royce’s view, a large robotic military vessel makes less sense than a small one. With a mere boat, as my Proceedings co-author Commodore James Soon has asserted, "the compromise of optional manning may itself significantly compromise sensor quality and craft performance, simply because the boat needs to be designed for human carriage as a norm, rather than optimized as unmanned from the start.” But with a ship of corvette size, a modicum of habitation may not be such a handicap, and the local talents of just a few wrench-turning humans may be very valuable. The question is whether it can be made sufficiently comfortable to merit a ninety-day trip.

Moreover, if the crew is truly small, their safety may be addressed with a different paradigm. Think about the difference between a warship and a warplane: the latter is not supposed to repair damage during the fight. It returns to base, if it can, where it is patched up and sent back. If the damage is too great, the aircrew attempt to either bail out, punch out, or crash-land. This is rather what former Director of Naval Research Jay Cohen had in mind over ten years ago when he suggested an equivalent to ejection seats on small surface combatants.

Put these ideas together, and you have an alternative for hunting submarines. Today’s approach requires a frigate with a crew of a more than a hundred, and a cost in the hundreds of millions. Tomorrow’s approach may manage with a flotilla of drones, or lightly manned corvettes, costing in the few tens of millions. The diffusion of assets makes the entire fleet more robust by limiting the cost of any single loss. It nearly upends the cost-equation for offense and defense, where today sometimes billion-dollar warships are attractive targets for multi-million dollar missiles. Would any single target of ACTUV size merit that expenditure? If DARPA’s plan works sufficiently well, American leaders may eventually find themselves asking not just where are the carriers, but where are the corvettes, for these small ships may very effectively blanket operating areas with overwatch rather like Reaper drones do today.

The effect on the shipbuilding industry could be dramatic. Control software for autonomous systems would grow in importance, of course. But the real change would come at the scale and rate of production: building 20 small ships for $50 million each is a far different undertaking than building single ships for $1 billion each. The capitalization of the yards is very different: the latter requires cranes and graving docks; the former is almost like an aircraft factory. The production and engineering skills may prove considerably different, opening an opportunity for aircraft manufacturers or systems engineering outfits to enter the business, if they chose. After all, the first of these robo-corvettes is being designed and built by SAIC—at heart, an IT firm.

James Hasik is a senior fellow at the Brent Scowcroft Center on International Security.

RELATED CONTENT