For the land we will have self-driving cars. And for the seas?

Drones Self-driving cars. Flying robotaxi. If you believe the headlines of the last few years, in the future, ground-based transport will replace robotic stage and devices, which practically do not require human intervention, well, except perhaps to download the application. But what about the remaining 70% of the surface of our planet - the part that is covered with water? Of course, there are already underwater drones that can record video in 4K for the next BBC documentary. Remote-controlled vehicles (ROVs) are capable of going down thousands of meters to explore ocean openings or repair industrial infrastructure.

However, most robots areToday, on or under water, is still heavily dependent on the human factor. And this is understandable, given the unstructured environment of the seas and weak communication capabilities for everything that moves below the waves. Autonomous underwater vehicles (AUV) are probably the closest to smart ocean machines today, but more often they follow pre-programmed instructions.

The new generation of waterfowl robots conqueringthe seas — and involving artificial intelligence, machine vision, advanced sensors, and other technologies — are ready to dive into the depths. Let's see what the engineers have prepared for us.

Marine transformer

Nick Radford, Technical Director of HoustonMechatronics Inc. (HMI), hesitates to say the word "autonomy", talking about the best creation of its startup - Aquanaut. He prefers the term "total control".

Aquanaut seems to come down from the movie screen"Transformers". The underwater robot begins each mission in the form of a submarine capable of autonomously traveling up to 200 kilometers on a battery, depending on the destination.

</ p>

When Aquanaut reaches its destination -the oil and gas industry is first in the queue of those that the HMI wants to change - its four specially designed and built linear actuators are set to work. Aquanaut folds into a robot with its head, upper torso and two manipulators, while remaining afloat.

The moment of illumination, when the engineer understood how to turn a submarine into a robot, arose on the day when the table was laid out in the office. The answer suddenly seemed obvious.

“We’re just going to build a big, giant, underwater folding table,” says Radford.

The hardware was not the only problema team of veteran NASA robotics like Redford that they should have solved. To get rid of expensive support vessels and large teams of people needed to manage traditional ROVs, Aquanaut had to be able to feel its environment in the smallest detail and transmit this information back to headquarters using an underwater communication system that existed telephone line.

To solve the problem of low bandwidthability, HMI equipped Aquanaut with a computer vision system consisting of acoustic, optical and laser sensors. All this dense data is compressed using proprietary HMI technology and transmitted to a single human operator, who controls Aquanaut with a conventional computer mouse. Even a joystick is required.

“I don’t know anyone who has achieved this level of autonomy in dealing with the environment,” says Radford.

Earlier this year, HMI received 20 milliondollars from Transocean, one of the world's largest offshore drilling contractors. That should be enough to complete the prototype Aquanaut, which, according to Redford, is 99.8% ready. Loud demonstrations are already scheduled for the beginning of the next year, and commercial deployment is planned for 2020.

“What gives us an incredible advantage is that we were born and raised to create robotic systems for remote locations,” says Radford. "This is my life, I made a bet on it."

On cruise control

Meanwhile, a startup from Boston is trying to solvethe problem of ensuring the autonomy of ships at sea. Sea Machines received about $ 12.5 million in investment, and Toyota AI was on the list of investors.

Sea Machines seeks inspiration in the industry.autonomous cars, developing the so-called “ship intelligence” systems, which can modify existing commercial ships or freshly built workers ships.

</ p>

At the beginning of this year, a startup announced the conclusiondeals with Maersk, the world's largest container shipping company, to deploy an artificial intelligence, computer vision system and LiDAR on a new Danish company container ship. The technology works in the same way as advanced driver assistance systems that are installed in cars to avoid hazards. Proof of concept will lay the foundation for a future autonomous collision avoidance system.

Not only startups are trying to create autonomousnavigation systems. Radford noted that Rolls Royce leads the development of autonomous ships. His Intelligence Awareness system uses almost all types of high-profile modern technologies: neural networks, augmented reality, virtual reality and LiDAR.

In augmented reality mode, for example, liveThe video stream from the ship’s sensors could simultaneously detect static and moving objects, superimposing on the scene information about the types of vessels in the area, as well as distance from them, heading and other data.

While safety is primarythe motivation for ship automation — more than 1,100 ships have been lost in the last ten years — these new technologies can make ships more efficient and less expensive to operate.

Sea hunting meets science

Marine robots can play an important role insaving the seas from environmental threats. In particular, from such a threat as a lionfish. According to the Ocean Support Fund, the population of these poisonous endemic fish is growing rapidly as the fish finds new habitats and eats thirty times more than it enters the stomach. It is rapidly reducing the population of young reef fish. Scientists and engineers are developing autonomous robots to destroy an invasive predator.

In Worcester Polytechnic Institute, for example,Students build robots with spears that use machine learning and computer vision to distinguish the lionfish from other aquatic species. Students have taught algorithms on thousands of different images of lionfish. The result was a killing machine for killing, which boasts an accuracy of more than 95%.

A new wave of smart, independent robots dives,swims and travels through the ocean and its deepest depths. These autonomous systems are not necessarily designed to replace people — rather, so that we can safely go where we could not before, or to increase safety at sea. Perhaps these innovations will inspire robots that will ever plunge into the depths of aquatic planets away from Earth.

Believe it? Tell in our chat in Telegram.