Military Robots in the Field

Of course these are not robots. Journalists just keep on using this misleading term to generate excitement. The devices filmed by NewsHour are simply small radio-controlled vehicles armed with a light weapon, a seventy year old technology. Radio-controlled tanks were first deployed by the USSR in 1939 in the Russo-Finnish War--where they had no effect in staving off the disastrous defeat of the Red Army.

Where I seriously disagree with you is on your point that no one dies when flawed weapons systems fail in combat (or, sticking to your narrower point, no one dies when flawed remote control weapons fail in combat).

Any time you send untested weapons into a real combat operation and they fail -- whether they're remotely controlled or not -- the costs are high: in casualties, in failure to achieve the presumably-important military objective of the operation, and in far less accurate data on the failure than you would have gained from a well-designed field test back in the U.S. Consider this scenario: you send a SWORD machine gun-armed, radio-controlled vehicle out as the point man on a short range reconnaissance patrol in wooded terrain (this means really short range, because the radio link probably won't even yield 250 m range through the woods). Say the noise of the SWORD's tracks gives away the patrol's presence 50 m earlier than with a quiet-walking infantryman on the point. That's an extra 35 to 50 seconds warning to the enemy waiting in the woods to ambush the patrol. A patrol that had a good chance of seeing the enemy first and bringing home critical tactical information instead gets torn up because of the test vehicle on point, suffers 50 percent casualties in the ambush, comes home with no information on the enemy's numbers, weapons and dispositions, AND no one knows exactly why the patrol got ambushed because the enemy isn't likely to announce "We nailed you because your silly noisy vehicle gave us an extra 50 seconds warning."

We did this kind of thing over and over in Vietnam, with the Defense Advanced Research Project Agency sending untested, gee-whiz prototypes (of hopeless ideas like reconnaissance balloons and man-portable tunnel radars) into combat with so much priority that soldiers and airmen had to risk their lives shepherding these prototypes through (or over) jungle terrain harboring hostile forces. Sending weapons that haven't been thoroughly field tested into combat to see whether they work is a terrible, reprehensible thing to do to soldiers, sailors and airmen -- and is a disservice to the designers and builders of the weapons.

A teleoperated device doesn't have the ability to move any of its parts by itself. In other words, all movements taken by the robot are at the expressed command of the human operator usually using a joy-stick or gamepad controller. Today's battlefield robots are teleoperated but they have semi-autonomous functions or sub-routines that can be selected by the operator allowing movements independent of direct teleoperation. Pre-programed manipulator arm, gripper, wrist and camera movements currently exist in today's teleoperated robots. Additionally, QinetiQ North America is well on its way to providing voice activated track movements, GPS waypoint navigation, vision-based sensors with associated software algorithms that includes collision avoidance and path planning as a product improvement. These embedded semi-autonomous navigation and locomotion product improvements are in the process of being fielded with full backward compatibility to the current fleet of over 2,500 military robots.

Weaponized robots will likely have semi-autonomous locomotion, manipulation and camera scanning capabilities but will always require the soldier to use teleoperation (direct control by the soldier) when remotely engaging the weapon. So, to that extent I agree with Mr. Holt that the most important decision, to engage a target with lethal force, remains with the operator.

The main factors influencing the expanded use of these systems will include efficiency of operations that is largely impacted by the availability of semi-autonomous navigation and locomotion combined with effective concepts of operations and tactics, training and procedures developed by the soldiers and marines.

A purely teleoperated device wouldn't have the necessary efficiency of operations because the soldier or marine wouldn't be able to both teleoperate the robot and use his weapon for personal protection.

Maintaining robust secure telemetry for data, video, audio, and networking with other battlefield assets is of course necessary but over time it will undoubtedly expand the operational footprint from today's thousands of meters of secured encrypted communications to tens of miles in the future.

Durability, ease of use and the ability for the soldier to easily tailor payloads to the mission will drive design considerations much more than unit costs. Cost reduction will come as a function of increased volume from hundreds of military robots we are producing each per year today at costs in excess of $100,000 to hundreds per week in the future at costs that are substantially lower, as a result of affordable automation due to higher and sustainable volumes.