Navy awards Applied Intuition $25M to deploy AI sensor fusion at sea

The U.S. Navy awarded Applied Intuition a $25 million contract to begin deploying its Intelligent Perception System, or IPS. The first deployment will place the technology on five Arleigh Burke-class destroyers. Its purpose is to help sailors understand their surroundings more quickly and clearly.

IPS gathers information from different cameras and infrared sensors. Many of those sensors were not designed to work together. The system combines their feeds into one coherent picture and can issue automated alerts, including warnings about possible collisions with other vessels.

IPS also records sensor data and can send it ashore when bandwidth and security conditions allow. Land-based systems can use that data to improve AI models for manned and unmanned vessels. Ships can also receive software updates while at sea, potentially allowing several improvement cycles during one deployment.

AI-powered sensor fusion means using software to combine separate sensor feeds into a single understanding of the surrounding environment. IPS is designed for cameras and infrared sensors that often were never built to work together. That matters because sailors otherwise may need to interpret multiple, disconnected streams of information.

In near real time, IPS pulls data from those sensors and synthesizes one coherent picture. It can also provide automated alerts, such as warning when another vessel appears to be approaching on a collision course. The article does not describe the exact algorithms or alert thresholds.

The result is a shared operational picture that can support faster human decisions aboard ship. IPS also records the underlying sensor data. When bandwidth is available and security conditions permit, that data can move to shore for analysis and future model training.

IPS will initially be deployed on five Arleigh Burke-class destroyers. That is the specific starting point under the Navy’s $25 million contract with Applied Intuition. The deployment is therefore a limited operational beginning, not an immediate fleet-wide rollout.

The system is intended to process data from shipboard cameras and infrared sensors. It can combine those feeds, present a coherent picture, and provide automated alerts. The first five ships will give the Navy a setting to use the capability under real naval operating conditions.

The possible end state is much larger. Applied Intuition said the system could expand “across the fleet,” but the article gives no timetable, ship count, or commitment for that expansion. Its broader potential includes both manned and unmanned vessels, because shore-based analysis can train improved AI models for each.

IPS could change how sailors respond to nearby traffic by giving them earlier, more organized information. Instead of relying on separate camera and infrared feeds, they would receive one synthesized picture of the surroundings. That could make a developing hazard easier to recognize.

One concrete feature is an automated alert when another vessel is approaching on a collision course. IPS gathers sensor data in near real time, compares the information, and presents the result to sailors. The article does not claim that the system takes control of the ship or replaces human judgment.

This capability is especially notable after two Navy collision incidents in 2017. The USS Fitzgerald and USS McCain collisions killed 17 sailors and severely damaged both ships. The article says sailors aboard those vessels could have benefited from such a capability, while emphasizing IPS’s potential rather than proving it would have prevented those incidents.

IPS creates a feedback loop between ships and shore. While operating, ships record sensor data and send it back when bandwidth is available and security conditions permit. Shore-based systems can then study those real-world observations and use them to train improved AI models.

The mechanism works in both directions. Data travels from ship to shore for analysis and model training. Updated software can then travel over the air back to ships at sea. This avoids waiting for maintenance or modernization in port, according to Applied Intuition.

The result could be multiple refresh cycles during one deployment. The article connects this rapid adaptation to the drone war in Ukraine, Defense Secretary Pete Hegseth’s push for fast upgrades, and the Pentagon’s interest in commercial technology. It does not provide a fixed update schedule or guarantee that every deployment will receive several updates.

The USS Fitzgerald and USS McCain collisions are an important historical example because both incidents occurred in 2017 and caused serious harm. Together, they killed 17 sailors and severely damaged both Arleigh Burke-class ships. Those consequences show the stakes of recognizing nearby vessel movements correctly.

IPS is designed to combine camera and infrared feeds into a coherent picture. It can also issue an automated warning when another vessel is approaching on a collision course. The article therefore links the system’s intended function to a kind of danger seen in those earlier incidents.

The article does not establish that IPS would have prevented either collision. It says sailors aboard the Fitzgerald and McCain definitely could have benefited from such a capability. Their cases provide the historical context for why better onboard perception and timely alerts matter to the Navy’s deployment.

An autonomous system must first perceive its environment. It needs information about surrounding objects, movement, and conditions before deciding what to do. Without that understanding, later decisions about routes, actions, or responses would lack the necessary input.

Applied Intuition describes the core sequence as perception feeding planning and control algorithms. IPS follows that pattern by collecting information from cameras and infrared sensors, combining it, and creating a coherent picture. Automated alerts are one example of how processed perception can support action by people aboard a ship.

The article connects this approach to Applied Intuition’s self-driving-car business. The company says gathering data for driverless cars involves operating environments and data challenges similar to naval work. It applies the same data-engine approach to naval operations, while the article notes that details vary between commercial vehicles and Navy vessels.