Tag Archives: artificial intelligence

Fiction, Future Tech, Future War

Grail War 2050, Last Stand at Battle Site One

Leave a comment

This piece by Dave Shunk is part of our Future Military Fiction Week for the New Year. The week topic was chosen as a prize by one of our Kickstarter supporters.

The nation state had decided not to invest in robotic armies. Autonomous killing machines were beyond their ethics. However, the enemy had no problem building autonomous robotic killing machines.

The enemy robotic land assault caught the nation state by surprise. The enemy forces especially sought to destroy the nation state’s treasure nicknamed “The Grail Project.”  The enemy’s battle plan sought to overcome the human defenders at the various Grail Project sites by overwhelming swarms.

The tactical fight went badly against the solely human forces defending the outlying Grail Project sites. The horde of enemy robotics on land, sea and air were the perfect attrition strategy.  Soul less killers, mass produced, networked together and built cheaply with advanced 3D printers in secret production facilities were deadly.

The nation state had not pursued the robotic armies but went a different route. HAL and Major Wittmann were the first experimental AI/Human team at training site “One” adjacent to one of the remaining Grail Project sites.  They were a prototype weapon – human and AI bonded together as a weapon system team within the tank with a shared neural network. However, this tank was unlike early 21st century tanks. This tank had advanced weapon systems – a tank on technology steroids.

HAL (Human Armor Liaison) is the artificial intelligence (AI) that controls the tank, the weapon systems, and communications. HAL is incorporated and encased into the advanced nanotechnology shell of the tank.  HAL has self repairing armor and neural circuits woven into the structure of the tank.  HAL also monitors the physical and mental health of Lt Wittmann via the neural connection with nanobot sensors throughout his body and bloodstream.

Major Wittmann has twelve years of service. He is a combat veteran, tank commander and human crew of one.  With genetic, physical and mental screening beginning in preschool, Major Wittmann began his military training early. He had the mental and intellectual capability for the nation state’s Human Performance Enhancement program. During his initial military training he received the neural implant for direct communication with advanced computer AIs. He also received nanotechnology enhancements in the form of nanobots in his blood stream to enhance and accelerate his cognitive and physical attributes.

HAL and Major Wittmann had trained as a team for two weeks. Due to the neural implant and nanobots, the bonding program progressed much quicker than human to human bonding. Days of training became the equivalent of months or years of purely human to human bonding. As the first AI/Human armored team they would chart the course for the fight against purely robotic forces. The speed of warfare had overtaken purely human skills due to AI and robotic technology.  At the same time science and technology opened new doors such as AI/human teaming, enhancing both warriors.

Orders came down to protect the Grail Project adjacent to HALS/Major Wittmann’s position at all costs. HAL monitored the battle flow from the network and Major Wittmann correctly anticipated the enemy tactical attack plan.  Within .01 seconds HAL detected the inbound swarm of enemy hypersonic missiles meant for the Grail Project.  HAL countered within .001 seconds by launching a counterstrike of steel flechettes which intercepted, detonated or deflected the inbound hypersonic missiles.  Inside the tank, observing from his 360 degree visual hologram of the battle, Major Wittmann thanked HAL via the neural network for his quick and decisive action to protect the Grail Project and them.

HAL and Major Wittmann knew if the enemy held to his doctrine, the robotic tanks would be next on the scene and attempt to destroy the sole AI/human tank. The twenty enemy robotic tanks announced their arrival by firing their laser cannon main weapons. Within .002 seconds of their firing HAL modified the external nanotechnology armor to disperse the energy along the entire hull and recharge the backup energy grid.

Before the last laser impacted the hull, HAL counter targeted the enemy robotic tanks. HAL fired the multiple barrel railgun and destroyed or severely damaged the robotic force. Fifteen burning hulks remained stationary and would move no more. Five other damaged tanks attempted to retreat. In .003 seconds HAL targeted the five with miniature hypersonic anti-tank missiles turning them into molten scrap. The enemy robotic scout force had been destroyed.

HAL knew they would need reinforcements to defeat the upcoming main robotic assault force. Major Wittmann came up with the “Improvise, Adapt, Overcome” solution.  On the training grounds in an underground warehouse were ten more experimental tanks – with AI’s on board but no human team member.  Due to neural limits Major Wittmann could not directly control another 10 AIs  – but HAL could.

 

Major Hartmann use his command emergency authority to over ride HAL’s protocol and programming limits. These limits stated that HAL could not control other AI tanks – a limit set by the nation state in peacetime.  But this was war and the Grail Project must survive.

HAL reached out to the ten tanks in warehouse by their AI battle network. Within .001 seconds the AIs received the mission, the situation, enemy order of battle, and threats. With the AI’s knowledge of military history, one other AI suggested that they form a laager around the Grail Project .

The Boers, like American wagon trains in the 19th century, formed mobile defensive laagers. The laager consisted of vehicles forming a defensive perimeter in whatever shape needed. The eleven AI tanks and one human formed a formidable interlinked mobile defensive perimeter around the Grail Project.

The battle ended quickly. The massed mobile firepower of the tanks overwhelmed the robotic attack force, but at a high cost. Tanks 1, 3 and 5 suffered catastrophic laser burn through on the armor plating destroying the AIs. Tanks 2, 4 and 8 suffered massive missile hits which destroyed various armaments reducing their offensive effectiveness to near zero.  The burning remains of the robotic army demonstrated they had fallen short of destroying the Grail Project at Site One.  In the classic struggle of over whelming force against determined defense, the combined AI/human teaming had turned the tide.

 

HAL watched the unfolding scene with curiosity as Major Wittmann exited the tank. The Grail Project at Site One had survived without loss. As the doors of the Grail Project opened, Major Wittmann, age 22, reached down and picked up his four year old son and gave a silent prayer of thanks as he held him once more.

 

His son had just been admitted with other select four year olds to the AI/Enhanced Human Performance Military Academy (The Grail Project). Eighteen years ago Major Wittmann had been in the first class of the Grail Project in 2032.

 

Article motivation for Grail War 2050, Last Stand at Battle Site One

The paper is meant as a wakeup that technology is changing warfare in a unique way. The era of human on human war is almost over. With artificial intelligence (AI) and robotics the speed of warfare will increase beyond human ability to react or intervene. The paper presents one possible solution.

 

This idea of human warfare nearing an end was presented in:

Future Warfare and the Decline of Human Decisionmaking by Thomas K. Adams

http://strategicstudiesinstitute.army.mil/pubs/parameters/articles/01winter/adams.pdf

This article was first published in the Winter 2001-02 issue of Parameters.

 

“Warfare has begun to leave “human space.” … In short, the military systems (including weapons) now on the horizon will be too fast, too small, too numerous, and will create an environment too complex for humans to direct. Furthermore, the proliferation of information-based systems will produce a data overload that will make it difficult or impossible for humans to directly intervene in decisionmaking. This is not a consideration for the remote science-fiction future.”

 

Other ideas in the paper:

  • AI/Human teaming and bonding
  • Robotic armies used with attrition strategy against human armies
  • AI controlling other AI vehicles with human oversight
  • Nanotechnology adaptable armor with embedded AI neural links
  • Human neural implants for AI link
  • Human nanobot implants
  • Multi-barrel Rail Gun for armor vehicles
  • Laser weapons for armor vehicles
  • Fletchette weapon as counter missile weapon
  • Hypersonic anti-tank missiles
  • Early military screening for youth (Ender’s Game influence)
  • Early military training for youth (Ender’s Game influence)

 

The second intent of the paper is a tribute to the military science fiction of Keith Laumer and his creation of Bolos – tanks with AI and teamed with military officers. His writings in the 1960s and 1970s were not really about just Bolos but about duty, honor and a tribute to the warriors. I read Last Command in the late sixties and devoured all the Bolo stories.

 

Last Command can be found here: (with preface by David Drake, Vietnam Vet and Author of many military science fiction books)

http://hell.pl/szymon/Baen/The%20best%20of%20Jim%20Baens%20Universe/The%20World%20Turned%20Upside%20Down/0743498747__14.htm

 

Dave Shunk is a retired USAF Colonel, B-52G pilot, and Desert Storm combat veteran whose last military assignment was as the B-2 Vice Wing Commander of the 509th Bomb Wing, Whitman AFB, MO. Currently, he is a researcher/writer and DA civilian working in Army Capabilities Integration Center (ARCIC), Future Warfare Division, Fort Eustis, Virginia.

Drones, Future Tech

Leading the Blind: Teaching UCAV to See

3 Comments

In “A Scandal in Bohemia”, Sherlock Holmes laments, “You [Watson] see, but you do not observe. The distinction is clear.” Such is the current lament of America’s fleet of UCAVs, UGV’s, and other assorted U_V’s: they have neither concept nor recognition of the world around them. To pass from remote drones living on the edges of combat to automated systems at the front, drones must cross the Rubicon of recognition.

To See

Still can't see a thing.
Help!

The UCAV is the best place to start, as the skies are the cleanest canvas upon which drones could cast their prying eyes. As with any surveillance system, the best ones are multi-faceted. Humans use their five senses and a good portion of deduction.  Touch is a bit too close for UCAV, smell and hearing would be both useless and uncomfortable at high speed, and taste would be awkward. Without that creative deductive spark, drones will need a bit more than a Mk 1 Eyeball. Along with radar, good examples for how a drone might literally “see” besides a basic radar picture are the likes of the layered optics of the ENVG (Enhanced Night Vision) or the RLS (Artillery Rocket Launch Spotter).

Operators for typical optical systems switch between different modes to understand a picture. A USN Mk38 Mod-2 24MM Bushmaster has a camera system with an Electro-Optical System (EOS), Forward Looking Infrared (FLIR), and a laser range-finder. While a Mod-2 operator switches between the EOS and FLIR, in the ENVG, both modes are combined to create an NVG difficult to blind. For a drone, digital combination isn’t necessary, all inputs can be perceived by a computer at one time. Optical systems can also be put on multiple locations on the UCAV to aid in creating a 3D composite of the contact being viewed. Using an array of both EOS and FLIR systems simultaneously could allow drones to “see” targets in more varied and specific aspect than the human eye.

For the deployment of these sensors, the RLS is a good example of how sensors can “pass” targets to one another. In RLS, after target data is collected amongst audio and IR sensors, flagged threats are passed to the higher-grade FLIR for further designation and potential fire control solution. A UCAV outfitted with multiple camera systems could, in coordination with radar, pass detected targets within a certain parameter “up” to better sensors. Targets viewed in wide-angle scans (such as stealth aircraft only seen) can be passed “down” to radar with further scrutiny based on bearing. UCAV must be given a suite of sensors that would not merely serve a remote human operator, but for the specific utility of the UCAV itself that could take advantage of the broad-access of computer capabilities.

And Observe

In-game models for real-life comparison.
In-game models for real-life comparison.

However, this vast suite of ISR equipment still leaves a UCAV high-and-dry when it comes to target identification. Another officer suggested to me that, “for a computer to identify an air target, it has to have an infinite number of pictures of every angle and possibility.” With 3-D rendered models of desired aircraft, UCAV could have that infinite supply of pictures with varying sets of weapons and angles of light. If a UCAV can identify an aircraft’s course and speed, it would decrease that “range” of comparison to other aircraft or a missiles by orienting that contact’s shape and all comparative models along that true motion axis. Whereas programs like facial recognition software build models from front-on pictures, we have the specifications on most if not all global aircraft. Just as searching the internet for this article, typing “Leading” into the search bar eliminates all returns without the word. In the same way, a UCAV could eliminate all fighter aircraft when looking at a Boeing 747. 3-D modeled comparisons sharpened by target-angle perspective comparisons could identify an airborne contact from any angle.

A UCAV also need not positively identify every single airborne target. A UCAV could be loaded with a set of parameters as well as a database limited to those aircraft of concern in the operating area. AEGIS flags threats by speed, trajectory, and other factors; so too could a UCAV gauge its interest level in a contact based on target angle and speed in relation to the Carrier Strike Group (CSG). Further, loading every conceivable aircraft into an onboard database is as sensible as training a pilot to recognize the make and model of every commercial aircraft on the planet. A scope of parameters for “non-military” could be loaded into a UCAV along with the specific models of regional aircraft-of-interest. The end-around of strapping external weapons to commercial aircraft or using those aircraft as weapons could be defeated by the previously noted course/speed parameters, as well as a database of weapons models.

Breaking Open the Black Box

The musings of an intrigued amateur will not solve these problems; our purpose here is to break open the black box of drone operations and start thinking about our next step. We take for granted the remote connections that allow our unmanned operations abroad, but leave a hideously soft underbelly for our drones to be compromised, destroyed, or surveilled at the slightest resistance. Success isn’t as simple as building the airframe and programming it to fly. For a truly successful UCAV, autonomy must be a central goal. A whole bevy of internal processes must be mastered, in particular the ability of the UCAV to conceive and understand the world around it. The more we parse out the problem, the more ideas we may provide to those who can execute them. I’m often told that, “if they could do this, they would have done it”… but there’s always the first time.

Matt Hipple is a surface warfare officer in the U.S. Navy.  The opinions and views expressed in this post are his alone and are presented in his personal capacity.  They do not necessarily represent the views of U.S. Department of Defense or the U.S. Navy.

Drones, Future War

Video Game AI and the Future UCAV Top Gun

2 Comments
My brother in flight school should be glad we played so much Ace Combat 4.
Alright, Roomba, now start sweeping for enemy units.

A Roomba is useful because it can sweep up regular messes without constant intervention, not because it can exit and enter its docking station independently. Although the Navy’s new X-47B Unmanned Combat Air Vehicle (UCAV) has, by landing on a carrier, executed an astounding feat even for humans, this ability only means our weapons have matured past their typical one-way trips. The real challenge will be getting a UCAV to defend units while sweeping up the enemy without remote guidance (i.e. autonomously). The answer is as close as the games running on your Xbox console.

 

 

Player One: Insert Coin

ace-combat-5-the-unsung-war-20041024114416859
Simulated fighters are UCAVs having an out-of body experience.

Considering the challenge of how an air-to-air UCAV might be programmed, recall that multiple generations of America’s youth have already fought untold legions of advanced UCAV’s. Developers have created artificial “intelligences” designed to combat a human opponent in operational and tactical scenarios with imperfect information; video games have paved the way for unmanned tactical computers.

A loose application of video game intelligence (VGI) would work because VGI is designed to operate in the constrained informational environment in which a real-life UCAV platform would operate. Good (i.e. fun) video game AI exists in the same fog of war constraints as their human opponents; the same radar, visual queues, and alerts are provided to the computer and human players. The tools to lift that veil for computer and human are the same. Often, difficulty levels in video games are not just based on the durability and damage of an enemy, but on the governors installed by programmers on a VGI to make competition fair with a human opponent. This is especially evident in Real Time Strategy (RTS), where the light-speed all-encompassing force management and resource calculations of a VGI can more often than not overwhelm the subtler, but slower, finesse of the human mind within the confines of the game. Those who wonder when humans will go to war using autonomous computers fail to see the virtual test-bed in which we already have, billions of times.

This Ain’t Galaga

No extra lives, and forget about memorizing the level's flight patterns.
No extra lives, and forget about memorizing the level’s flight patterns.

Those uninitiated must understand how VGI has progressed by leaps and bounds from the pre-programmed paths of games such as the early 1980’s arcade shooter Galaga; computer opponents hunt, take cover, maneuver evasively, and change tactics based on opportunities or a sudden states of peril. The 2000’s Half-Life and HALO game series were especially lauded for their revolutions in AI – creating opponents that seemed rational, adapting to a player’s tactics. For the particular case of UCAV air-to-air engagements, since the number of flight combat simulators is innumerable, from Fighter Pilot on the Commodore 64 in 1984 to the Ace Combat series. Computers have been executing pursuit curves, displacement rolls, and defensive spirals against their human opponents since before I was born.

However, despite its utility, VGI is still augmented with many “illusions” of intelligence, mere pre-planned responses (PPR); the real prize is a true problem-solving VGI to drive a UCAV. That requires special programming and far more processing power. In a real UCAV, these VGI would be installed into a suite far more advanced than a single Pentium i7 or an Xbox. To initiate a learning and adapting problem-solving tactical computer, the DARPA SyNAPSE program offers new possibilities, especially when short-term analog reasoning is coordinated with messier evolutionary algorithms. Eventually, as different programs learn and succeed, they can be downloaded and replace the lesser adaptations on other UCAVs.

I’ve Got the Need, The Need For Speed

Unlike Maverick, drones will never have to go through motorcycle safety training.
Unlike Maverick, drones will never have to go through motorcycle safety training.

When pilots assert that they are more intuitive than computer programs, they are right; this is, however, like saying the amateur huntsman with an AR-15 is lesser trained than an Austrian Arabesquer. The advantage is not in the quality of tactical thought, but in the problem solving rate-of-fire and speed of physical action. A VGI executing air-to-air tactics in a UCAV can execute the OODA loop encompassing the whole of inputs much faster than the human mind, where humans may be faster or more intuitive in solving particular, focused problems due to creativity and intuition. Even with the new advanced HUD system in their helmets, a human being cannot integrate input from all sensors at an instant in time (let alone control other drones). Human pilots are also limited in their physical ability to maneuver. G-suits exist because our 4th and 5th generation fighters have abilities far in excess of what the human body is capable. This artificially lowers aircraft tactical performance to prevent the death or severe damage of the pilot inside.

Pinball Wizard: I Can’t See!

VGI doesn’t have a problem with the how, it’s the who that will be the greatest challenge when the lessons of VGI are integrated into a UCAV. In a video-game, the VGI is blessed with instant recognition; its enemy is automatically identified when units are revealed, their typology is provided instantly to both human and VGI. A UCAV unable to differentiate between different radar contacts or identify units via its sensors is at a disadvantage to its human comrades or enemies. Humans still dominate the field of integrating immediate quality analysis with ISR within the skull’s OODA loop. Even during the landing sequence, the UCAV cheated in a way by being fed certain amounts of positional data from the carrier.

We’ve passed the tutorial level of unmanned warfare; we’ve created the unmanned platforms capable of navigating the skies and a vast array of programs designed to drive tactical problems against human opponents. Before we pat ourselves on the back, we need to effectively integrate those capabilities into an independent platform.

Matt Hipple is a surface warfare officer in the U.S. Navy.  The opinions and views expressed in this post are his alone and are presented in his personal capacity.  They do not necessarily represent the views of U.S. Department of Defense or the U.S. Navy.

Future Tech

Thinking Weapons are Closer Than We Think

Leave a comment

This piece also at USNI News.

The Defense Advanced Research Projects Agency (DARPA) has constructed a neuromorphic device—the functioning structure of a mammalian brain—out of artificial materials. DARPA’s project, SyNAPSE (Systems of Neuromorphic Adaptive Plastic Scalable Electronics) signals a new level for biomimicry in engineering. The project team included IBM, HRL, and their subcontracted universities.

Biomimicry is not new. The most recent example is the undulating “robojelly” developed by the Universirty of Texas at Dallas and Virginia Tech. This new drone swims through the sea like a jellyfish, collecting energy from the oxygen in the water, as does any breathing organism. There is also the graceful Pesto SmartBird, an aerial drone that mimics the shape and physical flight of birds. A knockoff was found crashed in Pakistan. If not the shape, at least the actions are often mimicked, as shown by UPenn’s quadrotors being programmed to use crane claws like predatory birds rather than construction cranes. However, these examples of biomimicry only cover the external actions of an animal. SyNAPSE goes deeper, building a synthetic version of the mind that develops these actions.

In the quest for autonomous machines, the suggestions have been either-or: machines programmed to be like brains or the integration of biological processors to provide that processing flexibility. DARPA has found the “middle path” in constructing a series of synthetic synapses out of nano-scale wire. This takes the physical form of those biological processors and constructs them from the base material of conventional computers. According to James Gimzewski at UCLA, the device manages information through a method of self-organization, a key trait of autonomous action and learning, “Rather than move information from memory to processor, like conventional computers, this device processes information in a totally new way.” Moving past the surface mimicking of physical shape and function, SyNAPSE will mimic living organisms’ basic way of processing information.

However, as the possibility for real autonomy approaches, the legal challenge becomes more urgent. An article in Defense News summarizes the catalogue of problems quite well, from accidental breaches of airspace/territorial waters, to breaches in navigational rules, to accidental deaths all caused by machines not having a direct operator to hold responsible. However, as the director of naval intelligence Vice Admiral Kendall Card noted, “Unmanned systems are not a luxury; they are absolutely imperative to the future of our Navy.” Like the CIA’s armed predator program, someone will eventually open Pandora’s box and take responsibility for their new machines to gain the operational edge. DARPA’s SyNAPSE project is that next step toward an autonomous reality.

A DARPA scale of the make-up of a neuromorphic circuit and their biological equivalents.
A DARPA scale of the make-up of a neuromorphic circuit and their biological equivalents.

Matt Hipple is a surface warfare officer in the U.S. Navy.  The opinions and views expressed in this post are his alone and are presented in his personal capacity.  They do not necessarily represent the views of U.S. Department of Defense or the U.S. Navy.