The race is on to develop AI applications that can track and down hypersonic missiles. Image: Twitter

Chinese researchers claim to have developed an artificial intelligence (AI) application that can predict the course of hypersonic missiles, potentially allowing Beijing’s existing air defenses to intercept such threats.

A hypersonic glide weapon can enter and exit the atmosphere akin to a stone skipping on water and perform random evasive maneuvers while flying at five times the speed of sound, making such weapons nearly impossible to intercept with current air defense technologies. That could change, however, if China’s AI claim is accurate.

In a paper published in April in the Journal of Astronautics, a peer-reviewed publication run by the Chinese Society of Astronautics, Chinese researchers from the Air Force Early Warning Academy in Wuhan said they created an AI that can calculate the most probable course of a hypersonic weapon during its terminal stage, using data gleaned from early stages of hypersonic flight.

This is made possible by the fact that any missile, no matter how fast or modern, is confined by fixed physical laws, and as such the high-speed projectile’s every move can be analyzed to provide clues about its design, capabilities and mission.

China’s new AI is reportedly more sophisticated than any previous used for hypersonic trajectory prediction and can produce results within 15 seconds on a laptop computer. Simulated tests suggested that it remains effective against hypersonic weapons flying up to 12 times the speed of sound.

The researchers claim that an air defense system powered by their AI can estimate the potential kill trajectory of an incoming hypersonic weapon and initiate a counter-response within a three-minute lead time.

However, they also noted that raw data collected by missile early warning systems contain so much noise that they could confuse the AI and overwhelm the computers running the application.

To resolve these problems, the researchers developed a unique deep learning algorithm that automatically removes noise from detected signals and imitates the human brain by focusing only on the latest, most important data.

To be sure, China is not alone in exploring the role of AI in missile defense. The US is also investigating AI applications against missile threats such as more maneuverable and precision-guided ballistic missiles, low-flying cruise missiles as well as hypersonic weapons. 

China’s DF-17 hypersonic weapons on parade. Image: Twitter

Vice Admiral Jon Hill, director of the US Missile Defense Agency, stressed that the speed and pace of contemporary warfare and missile defense is beyond the ability of the human mind to comprehend and analyze, leaving AI as the only feasible solution to accelerate response times and reduce the probability of human error.

Conventional manned missile defense systems such as the US Aegis have auto-response systems which human operators can activate in case they are overwhelmed by a missile attack. Human operators pre-program threat parameters into the system such as inclination angle, altitude or speed of contact. When the system detects a contact that is consistent with these parameters, it engages the target autonomously.

However, using these semi-autonomous, pre-programmed defense systems against unpredictable threats such as maneuvering hypersonics is insufficient and requires an operator within the system loop, either a human being, AI or man-machine tandem.

AI affords several advantages over human operators. Most basically, AI can be on watch 24 hours a day without the need for rest and nutrition. AI also retains effectiveness over long periods of time and becomes even more effective as more data is fed into it, in contrast to human operators who need constant practice to be effective and whose skills atrophy with disuse.

AI also reduces the element of human error, where fear, fatigue and anxiety can often cause human operators to make poor or wrong decisions with potentially disastrous consequences. AI also has a faster reaction time than any potential team of human operators and can give instantaneous feedback of system performance.

AI-powered missile defenses could also be networked with each other to maintain high readiness and protect critical areas in a potential conflict scenario, significantly at a time of rising geopolitical tensions.