Episci develops AI to detect hypersonic missiles from space

Five years on from suggesting that AI be given autonomous control over certain US defense activities the Space Development Agency selected California-based EpiSci to develop a software tool capable of detecting hypersonic missiles in flight from satellite data, a challenging task given the extreme speeds of these weapons.

Hypersonic missiles travel at speeds of at least Mach 5, which is five times the speed of sound, posing a significant challenge for current defense systems. The Space Development Agency (SDA), an organisation within the US Space Force, is building a network of satellites in low Earth orbit with the goal to provide global indications, warning, tracking, and targeting of advanced missile threats.

As many as 100 missile-tracking satellites are projected to be part of this network.  But in addition to having satellites in orbit, the SDA needs advanced software that can analyse the data collected by these sensors and identify targets in the clutter of objects in the atmosphere.

That’s what EpiSci hopes to accomplish under the $1.6 million two-year Small Business Innovation Research Phase 2 contract from SDA, announced earlier this year. It will test its AI-powered software against data collected by low Earth orbit sensors to identify and track hypersonic threats.

Samuel Hess, technical director at EpiSci, said the project is not without its hurdles. Maintaining “custody” of these fast-moving targets across vast distances requires collaboration between multiple satellites and precise tracking algorithms. For this project the company is partnering with Raytheon Technologies, a large defense contractor with expertise in missile defense simulations and data analysis, that also is an investor in EpiSci.

“As the hypersonic target manoeuvres in flight, it’s moving through different camera images, so how do you communicate that across the whole realm of satellites?” he said. “That’s something that we need to work out.”

EpiSci specialises in AI for autonomous drones and to enhance combat pilots’ performance. Some of these tools, for example, enable pilots to respond to threats faster and to “team” with unmanned aircraft.

One of the company’s strategic investors is Top Aces, a tactical aviation training company that uses EpiSci’s technology to generate complex scenarios for pilots in training.

To demonstrate hypersonic missile detection for SDA, the company will use Raytheon’s missile defense simulators, Hess said. It will start with data from just one satellite and gradually add more.

“Raytheon’s simulation is extremely powerful. So they can actually simulate multiple satellites, and provide us video feeds from what that would look like.”

The challenge for EpiSci is to develop the right software algorithms, he said, to “actually detect these small targets over this large swath and also maintain custody of the targets in an area where other objects are flying, such as commercial aircraft.”