Seabed Sentry, Anduril unveils its networked seabed sensor

Seabed Sentry is the new AI-enabled underwater sensor network from Anduril designed to provide persistent surveillance of the sea floor, according to a 3rd April press release. The system can communicate between its nodes in real time and provide “highly accurate, persistent autonomous awareness on the seafloor.” 

The sensors can be deployed to depths of 500 metres, which is typically sufficient for most coastal areas and the bottlenecks created by terrain and maritime transit routes that most marine traffic uses. The Baltic Sea has a maximum depth of 459 metres and averages 55 metres, an article from the John Nurminen Foundation explains. 

Like most Anduril systems, Seabed Sentry is networked and controlled using the Lattice software and can be deployed from the company’s autonomous submersible vehicles. It employs Sea Spear, a lightweight deployable sonar from the US company Ultra Maritime. Sea Spear is designed to be deployed from submarines and other naval vessels, to increase the sensing range of the platform over its built-in sensors. Anduril also states that the payload of Seabed Sentry can be altered and changed as required to include optical sensors, environmental sensors and active as well as passive sonar. The system has an endurance of months and it can be recovered and recharged allowing the sensors to be reused. 

The key feature of the system is its artificial intelligence (AI) on the edge, which enables Seabed Sentry to compute and understand what it is sensing before communicating that information. Put very simply, there are two types of computing, or two locations for sensor computing to be done. Edge computing is where the sensor, say a CCTV camera, radar on a satellite, or sonar, is fitted with its own computer running the AI algorithms and software. This means the sensor data can be processed by the computer and AI as it is received, producing recommendations or clearer target indications to the operator. In the case of a video feed, this might include the identification of a tank and indication that it is a hostile vehicle, for example. 

If this is not done at the edge – where the sensor is – then the data must be passed to a computer somewhere else, either a cloud computing facility or the user’s terminal. This might be another server that is connected to the feeds of several sensors or a ground control station. One of the big differences between these two modalities is bandwidth. If the data is processed on the edge, it can be turned into smaller data packages that are easier to send over a busy radio network. Without this processing, it might mean passing the data in its raw form over the network, which can be bandwidth intensive in the case of things like video feeds. 

How does Seabed Sentry send data?

The Anduril Dive-XL autonomous underwater vehicle is shown here. It would be capable of deploying the Seabed Sentry. Credit: Anduril

All of the above is particularly important for subsea networks. Anduril states that Seabed Sentry is wireless and networked, indicating that it must send data whilst underwater, which is very difficult. Salt water soaks up radio frequency signals which forces a reliance on different means of data transfer. Some options include acoustic waves or very low frequency radio links, both of which have extremely limited bandwidth and are generally constrained to transmitting packets of data. Acoustic transmission converts electrical energy into sound waves to send data, and then converts sound waves into electrical energy to receive data using something called a transducer. 

Another option is electromagnetic field technology, such as the system advertised by the UK’s CSignum, which is able to transmit data through water, ice, soil and into the air, the company states. L3 Harris has also developed the Communicating Using Underwater Ultrasonic Wireless (CUUUWi) system, which allows subsurface platforms to communicate with above surface users, allowing low bandwidth communication at distances up to 10 km, and higher bandwidths at much shorter ranges. 

Anduril states that the Seabed Sentry has an acoustic communications relay. But does not state how it is able to share data with a command node to integrate into a wider anti-submarine warfare network. However, an image of the Seabed Sentry includes the company name, Sonardyne. Sonardyne provides a number of subsea capabilities including Modem 6, which it describes as a “reliable and cost-effective tool for simple point-to-point wireless underwater sensor data transfer.” Modem 6 provides a data rate Up to 9,000 bps, which is close to the speeds achieved by dial-up internet connections, and would effectively limit what Seabed Sentry could transmit. 

Sonardyne also offers BlueComm, an optical wireless through-water communication system providing bandwidths up to 10 mb/s, with a range up to 150 m. 10 mb/s is sufficient bandwidth for web browsing or watching a standard definition video, however, this capacity appears to depend on the nodes in the network being very close together. Anduril states that Seabed Sentry uses acoustic communications links, indicating that Modem 6 is more likely to be the solution, or a variant thereof. 

Regardless of how the data is communicated between the Seabed Sentry nodes, it should hopefully be clear now that compressing that data by edge computing is an important element of making this type of capability work. 

Calibre comment

Anti-submarine warfare (ASW) is sometimes referred to as ‘awfully slow warfare.’ It is a painstaking and resource-intensive process as vast areas must be swept and patrolled for submarines that are already hard to find. Let’s take the Baltic Sea as an example, it is 1,300 km long with a surface area of 385,000 square kilometers without the Kattegat and Danish Straits, providing a total volume of 21.7 thousand cubic kilometers. The Baltic Sea is, in the words of naturalniebaltyckie.pl, “quite small,” at least as far as seas go. And yet, it represents an enormous area to patrol with hundreds if not thousands of places for a submarine to hide. This is a problem when western navies have relatively few vessels and aircraft built for the ASW mission. However, it is hoped that networked sensors like Anduril’s Seabed Sentry, as well as autonomous sub-surface capabilities like BAE’s Herne XLAUV will be able to provide the mass that western forces need to track adversary submarines. Ensuring these assets can talk to each other and fuse what they have sensed into a single recognisable picture is absolutely key to realising this vision for the future of subsurface security.