U.S. Navy tests artificial intelligence in the Battle of submarines

Fighting the submarine (ZOP) has always been a race against a time in which the detection of a suspicious signal is only the beginning of work. During RIMPAC 2026, the U.S. Navy checked Lockheed Martin's solution, which is expected to speed up this process and allow crews to almost continuously adapt the system to the new acoustic signatures appearing during operations.

RIMPAC 2026 exercises conducted off the coast of Hawaii were attended by two ZOP helicopters MH-60R Seahawk equipped with SensorMAX using artificial intelligence and machine learning. The machines carried out the search for the submarine, placing Radio-hydroacoustic buoys (PRHA), the data of which was subsequently analysed by Lockheed Martin's system.

This does not change the very essence of the activities of the ZOP. The helicopter continues its search using acoustic sensors deployed in the selected area, and operators analyse the signals they receive. SensorMAX, on the other hand, is expected to take over some of this work by constantly tracking the incoming data and indicating those that may indicate the presence of a submarine.

Eight buoys under one system

SensorMAX analyses the energy spectrum of signals received from below the sea surface and assists the crew in detecting and classifying acoustic sources. During the trials, PRHA data were transferred to both the helicopter and the station of operators on land, so that the analysis could be conducted simultaneously.

The system interface allows the crew to track data from up to eight radio-hydroacoustic buoys simultaneously. By Lockheed Martina This means a double increase in the number of data streams that can be simultaneously supervised compared to the existing solutions.

With the activities of the ODP, this is not a purely quantitative difference. With the increase in the number of sensors deployed, there is also a rapid increase in the number of information required to be evaluated by operators, especially in an environment where the signal of the submarine being sought mixes with the acoustic background of the sea, the movement of other units and a range of natural sound sources.

This is where SensorMAX is supposed to support the crew by sifting out some of the information and paying attention to those signals that need further analysis.

The most important thing is when you detect a new signal.

The use of algorithms for the analysis of hydroacoustic data is not, however, the most interesting element of the tests carried out during RIMPAC. Much more important is the way the system can be adapted to the new signal already when performing the task.

If operators identify acoustic characteristics which SensorMAX has not previously recognized, they can mark them and use them to retrain the model. According to Lockheed Martin, one such operation takes less than five minutes.

Upon completion, the update may be sent via encrypted data link directly to the helicopter. This means that the crew do not have to wait for the task to be completed, later analyzing the material and preparing another version of the software. The newly acquired information can still be used for the same mission.

W ZOP it is not always enough to refer to previously recorded acoustic signatures, so the ability to quickly adapt the system to new data may be as important as its effectiveness in recognizing signals already known.

MH-60R continues to do the same work

SensorMAX does not replace the ZOP equipment of the helicopter or the operators themselves. MH-60R Seahawk remains a carrier of sensors and weapons, and radio-hydroacoustic buoys form a distributed submarine search system. In particular, the way information from these sensors is processed and presented to the crew changes.

During RIMPAC, both helicopters performed a series of ZOP tasks, during which they searched for a submarine and collected acoustic data using PRHA placed. The information went to SensorMAX, while operators on land were able to adapt the model on the basis of it and send its subsequent versions back to helicopters in the air.

This creates a closed circulation of information: signal detection, signal marking, model adjustment and crew reuse. Lockheed Martin reports that such a process can be repeated repeatedly during one task.

Not just for Seahawk

The manufacturer also draws attention to the possibility of using SensorMAX with other acoustic sensors. The system was not designed solely as a helicopter equipment MH-60R, but as a tool capable of analysing data from different sources.

This is a particularly interesting direction for the further development of the ZOP. The modern search for submarines is increasingly based on one sensor and one unit. Data may come from submarines, helicopters, marine patrol aircraft, radio-hydroacoustic buoys, and increasingly from unmanned vessels in the future.

Where information from such sensors can be analysed according to common rules and quickly transmitted between participants, the time needed not only for the purpose to be detected but primarily for its classification and further tracing.

However, this does not mean that during RIMPAC The US Navy has already demonstrated a ready-made network or multistatic ZOP system based on SensorMAX. The trials had a much more specific scope and showed above all the possibility of current data analysis with PRHA and a rapid adaptation of the model to newly recognized signals.

AI as another hydroacoustics tool

Lockheed Martin develops SensorMAX from his own resources and at this stage the system remains a demonstration solution. Therefore, this year's tests do not mean its acceptance for equipment US Navy or the start of mass deployment on helicopters MH-60R.

However, they show a direction in which hydroacoustic analysis can develop. The more sensors are involved in the search for a submarine, the greater the problem becomes, not just to obtain information, but to quickly organize and use it.

However, it can help to identify where to continue the search, and in the activities of the ZOP several saved minutes may decide whether the detected underwater object can be followed further.