America Might Soon Have a New Way To Kill Russian or North Korean ICBMs

September 13, 2019 Topic: Security Blog Brand: The Buzz Tags: F-22F-35MilitaryTechnologyAIICBM

America Might Soon Have a New Way To Kill Russian or North Korean ICBMs

AI could be a game-changer. 

(Washington, D.C.) If a nuclear-armed enemy Intercontinental Ballistic Missile were speeding its way through space towards a heavily populated U.S. target, commanders in charge of defending the homeland would at most have a mere 20-to-30 minutes to destroy the incoming weapon.

With lives dangling upon a precipice of total devastation, and the earth’s future potentially in jeopardy, U.S. defenders would be tasked with finding, tracking and destroying the attacking nuclear missile.

If a Space-Based Infrared satellite detected a launch, it would quickly send sensitive data signals to an Air Force command center where the information would be evaluated by computers. Then, the command center would send a wide-sweeping “alert through the ballistic missile defense system,” a Senior Pentagon official told Warrior.

“When a satellite recognizes something that looks like a launch, it has a mathematical formula in a computer that tracks the heat source,” the official said.

“Then we start turning on radar”... he added.

Response decisions, which could involve the immediate deployment of a well-armed and ready U.S. nuclear triad…. hinge almost entirely upon .. time. In effect, the faster tracking, targeting and countermeasure analysis can take place, the greater the chance lives - or even the entire country - could be saved. The speed and accuracy with which crucial intelligence information can be gathered, organized and fused for decision-makers, will determine the outcome.

“Once you see the rocket, it is an easy mathematical formula that can calculate the speed and trajectory. You can usually tell where it is going to land. If you have enough radars you know exactly how fast it is going and where it is going to be -- so you know where to send the GBI(Ground Based Interceptor),” the Pentagon official said.

This scenario is, in reality, as dramatic and serious as it sounds; it is tough to imagine a more serious warfare situation. These circumstances form the conceptual inspiration for why the Pentagon and industry are vigorously exploring ways to better “fuse” and transmit space-based information. Booz Allen Hamilton is, along with other major industry players such as Northrop Grumman, among the industry players now working on building integrated data systems and command and control nodes within an open architecture framework -- to network, harden and upgrade missile-defense oriented communications and data sharing.

“If you walk into an AOC (Air Operations Center) when you are running an air war, you will see lots of visualization of the battlespace. You will see two dimensional data showing the range of ground threats. The commander can see that because the data is sensed and fused. We do not have a lot of that from space. It is difficult to visualize what is going on in space. If a commander has battlespace awareness in near real-time, he or she can see a missile being launched and create tactics and procedures to help protect assets,” Retired Lt. Gen. Christopher Bogdan, Senior Vice President, Aerospace, Booz Allen Hamilton, told Warrior in an interview last month. (Bogdan previously served as the Program Executive Officer for the Pentagon's F-35 program)

All this being the case, just what might be needed to streamline time-sensitive data in a secure, yet streamlined fashion across multiple nodes in real-time? Instead of operating with disparate or stove-piped radar and command and control centers, the Pentagon’s Missile Defense Agency and its industry partners such as Booz Allen Hamilton, are working to safeguard and integrate communications systems necessary to share information in crucial warfare circumstances such as a nuclear attack.

For instance, as one Pentagon official described it, the idea is to align multiple radars on a single unified network as a way to quickly pinpoint the trajectory of an approaching missile from multiple angles -- using multiple different radar systems.

“We need systems to talk to each other. It only works if they are on a common communications hub. You have 20 to 30 minutes at most to figure out what you are going to do, If you have multiple systems that are not coordinated they may think there are 12 missiles...but they are looking at the same missile from different angles,” the Pentagon official said.

SBIR - Space-Based Infrared Satellite

Achieving this not only involves hardening satellite, radio and computer networks, but engineering an integrated “mesh” of secure connected nodes able to safely transmit data, thus circumventing the problems posed by losing time relaying data from one single system to another, Bogdan explained.

Breaking down radar and communications stovepipes by using multiple radar tracks at the same time is, as a Pentagon official explained, referred to as operating in “stereo.”

“You have to protect command and control. That is where redundancy comes in. It creates a self-adapting system which is hard for the adversary to take out. Your adversary is confronted with a tough dilemma,” Bogdan explained.

AI for Missile Defense

Pursuing faster, more secure technical modernization of data sharing systems and command and control are, not surprisingly, greatly fortified by emerging applications of Artificial Intelligence, Bogdan explained. AI, which works by comparing and organizing new information against a vast or seemingly limitless database to organize data, solve problems and perform procedural functions using processing speed well beyond human capability.

Extending this reasoning, it seems self-evident that AI could massively help enable the "stereo" strategy to operate and align multiple radar tracking systems upon a single approaching ICBM. Advanced algorithms could instantly integrate multiple radar returns at one time, presenting a comprehensive, multi-faceted unified picture for decision-makers. While human cognition is, of course, indispensable to decision-making, AI-enabled machine learning programs can organize and distribute this kind of data with mathematical accuracy at lightning speed. Advanced iterations of AI can also determine context by analyzing multiple interconnected variables at one time. For instance, an AI system could be able to distinguish the meaning of “ball” as in “football” from “ball” as in dance by analyzing the surrounding words to discern overall meaning. This concept is analogous to how AI is helping satellite command and control make sense of and organize otherwise disparate pools of information.

The more advanced the AI system, the faster is absorbs, learns and integrates new information. Emerging AI technologies, such as one or two now being developed by DARPA, are looking at engineering a kind of real-time machine learning process wherein computers do not have to wait for new information to bounce off an existing database but rather perform analytics even faster - in near real-time as information is flowing in.

Satellites themselves, Bogdan offered, could be programmed with AI such that they can organize, analyze and distribute sensor data while in space out on the far edges of combat.

This, naturally, could only further streamline an already time-sensitive command and control process. “You don’t have to take that data back to a control center, you can do it at the edge with a smart satellite,” Bogdan said. “That is the forward-looking piece of AI. It learns when it is faced with something it may not have seen before.”

This kind of data synergy, which gathers and organizes information for human decision-makers, often comparing new information against existing data, is not unlike the well-known “sensor-fusion” used in the F-35. Multiple sensors, including targeting, navigation and EO-IR sensors are all synchronized on a single screen for the pilot. In addition, comparing newly gathered sensor information against an existing database is entirely consistent with the F-35s Mission Data Files, which identify enemy aircraft in certain global geographic areas by analyzing new sensor information.​

Booz Allen Hamilton is, according to Bogdan, developing these kinds of AI-enabled systems for satellites and missile defense through simulations able to gather and catalog new data for machine-learning systems.

“Through simulation we can create synthetic data to help the machines learn better. With the information that has already been cataloged, machines can help discern the “best way to counter something and reduce your vulnerability,” Bogdan said.

This process, he added, can reduce some of the need for human involvement in command and control - while of course still needing human decision-making faculties. Humans are, among other things, better able to make sense of a wide range of more subjective variables less likely to be captured by algorithms or mathematical formulas.

Nevertheless, this kind of computer-enabled real-time machine learning also finds applications in what Bogdan described as “adaptive EW.” Satellites, and even possibly ballistic missile interceptors, could defend against certain kinds of “jamming” or electronic warfare attacks by virtue of drawing upon AI-methods of comparing and identifying enemy signals. An Adaptive EW system, Bogdan explained, “can tell what countermeasures were used in the past and create a signal to go against them. AI can learn new signals and absorb a new kind of RF, depending upon what the enemy is using. If you have a radio that can sense the environment in real-time, you can understand what that environment looks like and create a waveform that can be changed,”

Low Earth Orbit Satellites

The Pentagon and industry are working on engineering large numbers of smaller, Very Low Earth Orbit satellites. Connecting a large volume of lower, smaller interconnected groups of LEO satellites can, as Bogdan explained it, help foster the construction of the sought after “networked” command and control apparatus.