A report issued by the U.S. Defense shows that the army intends testing robots that “think – look – move – talk and work.” Figure 24 summarizes the Army’s vision for these five problem domains, barriers to achieving its vision, and work to be done to advance toward the vision.
The Robotic Collaborative Technology Alliance (RCTA) plans a Capstone Experiment in during 2014. The U.S. Defense robot provides the following example as shown in Figure 25.
The Capstone Experiment is centered around a notional cordon-and-search operation: during urban transit by a small unit (i.e., four to five soldiers), a fugitive is reported to have entered a building the unit is approaching. A man-transportable robot is instructed to “cover the back door” of the building by the unit commander because he cannot safely split up his limited resources. The robot must understand and acknowledge the order, associate the order with its perceived environment, move safely and securely to an appropriate vantage point, observe activity behind the building, and report any salient events to the unit commander. As needed, it enters the building and negotiates stairs or other mobility obstacles. It then returns to its unit, maintaining situational awareness, and is ready for another assignment. While this narrative occurs in the context of a cordon-and-search operation, its underlying capabilities support a broad range of potential operational missions.
Effectively the Military consider that a truly useful robot should have the ability to learn on its own from interactions with the physical and social environment. It should not rely on a human programmer once it is purchased. It must be trainable.
Unfortunately, the report goes further by indicating that work will continue to move to fully autonomous robots, something I am personally very much against:
Similar to the other Services, middle- and long-term work by the RCTA will continue to evolve and improve capabilities to increase the level of autonomy in systems from the current, remotely operated systems to autonomous systems.