Thursday, February 5, 2009

Virtual Warfare -- PEO STRI

I went this week to a demonstration/ exhibition put on for Members of Congress and their staffs by the U.S. Army's PEO STRI office. The purpose of the event was to give us "the chance to experience interactive military simulations and training devices that currently prepare Warfighters for their missions." I came away impressed by what our military is already doing with virtualization and the potential for virtual worlds.

One demonstration that impressed me used a 3-D camera and Second Life (plus 3 large LCD panels). The 3-D camera translates the movement of a human being into movement by an avatar. In the demonstration I saw, there was a virtual punching bag that existed solely in Second Life. When the user throws a punch in the real world, the system reacts by showing the virtual punching bag being hit. Users can also move their avatars around Second Life simply by leaning forward. Very cool! Update: As Tami notes in her comment to this post, the punching bag is not actually in Second Life. It must therefore, I think, be in a construct created with the SL code, but not on the SL grid (kinda like the sparring program in The Matrix?).

I learned from talking with folks at the exhibition that the Army has begun using Second Life to a significant degree. While Forterra's Olive software is being used in parts of the military, Second Life has the appeal of being open source, so end users (such as PEO STRI) can easily develop their own scenarios.

Another impressive demonstration was Virtual Battlespace 2 (VBS2). VBS2 allows soldiers and their commanders to conduct virtual training in environments designed by the soldiers themselves. Thus, a platoon could map out a detailed, realistic simulation of an upcoming mission. They could then use it to identify potential choke points along the route, learn to identify key markers in the dark, and scope out enemy sniper positions. VBS2 further offers the capability for some soldiers to play as the enemy (or civilians), such as a sniper or roadside bomber, thus adding an element of surprise, intelligent response, and immersiveness. VBS2 could also be used to help train a National Guard unit that is about to deploy to Iraq or Afghanistan. A unit in the field could play the role of the opposing force, and the new unit could practice how to identify IEDs and locally-unique aspects of their deployment.

One of the best features is that soldiers and officers can design and modify VBS2 scenarios, which then go into a communal library of sorts. One field commander might design a sim to train his soldiers on convoy protection near Kohat. Another commander could then use that same scenario to develop a sim for interacting with local tribesmen, or he might tweak the sim for a night-time raid when there's 3 feet of snow on the ground.

VBS2 is clearly an exciting technology, and is already being used by the Army. A basic VBS2 deployment consists of around 50 networked computers, corresponding to the size of 1 platoon. Additional groups can be assembled allowing for upwards of 250 soldiers. Fort Lewis, KY is the primary deployment site for VBS2 thus far, where it has been well-received.

Talking with some of the members of the Army, it was clear to me that the adoption of such virtualization technologies complements nicely the fact many soldiers grew up playing, and continue to play, warfighting games. Many of these soldiers are already familiar with Call of Duty and other games, so they take to these virtual trainers quite enthusiastically.

From an economics perspective, the benefits are clearly the cost-efficiency of virtual training. Giving a trainee an extra hour on a simulated construction vehicle costs far less than going through all the regulations required to identity land for digging, wear and tear on the machine, and damage to an environment (not to mention the added bonus of letting them train in a variety of settings -- jungle, mountains, etc. -- for no additional cost). VBS2 allows for soldiers in the field to transfer hands-on street knowledge to incoming units, and the ability to work through mission simulations will hopefully cut down on casualties. All these benefits, and yet the marginal cost for an additional hour on the virtual trainer/sim is next to nothing. Although development and deployment costs can be significant, it seems abundantly clear that in the long (and even in the short) run, the benefits of virtualization far outweigh the costs.