AI On The Sea

in STEMGeeks10 days ago

It’s been 400 years since the legendary journey of Mayflower. Now a new ship named Mayflower will head to the seas. But this time, it will have no humans on board and will be controlled by an AI.


trimaran-2806625_1920.jpg

Image by willoh from Pixabay

In the year 1620, a galleon named Mayflower sailed from Plymouth into the New World. It carried about thirty crew members and about 102 passengers that later settled the East of the current United States of America.

Part of the celebrations of the four-hundredth anniversary of Mayflower should be the voyage of one of the first ships controlled by artificial intelligence which is named Mayflower 400. While the voyage should have started last year, it was sadly stopped by the coronavirus pandemic. Now, everything seems to indicate that it will finally sail this year on the 15th of May.

One big benefit – in these times especially – of a ship controlled by AI is that nobody can get sick. The ship is fully autonomous and will not need a single human to cross the Atlantic. Mayflower 400 is a trimaran that is 15 meters long and is capable of displacing 9 tons of water. The ship gets its energy from solar panels that fully cover its upper side. To navigate, the ship uses a sophisticated system with six cameras and radar. And the ship won’t just sail but also do research into ocean pollution and observe ocean mammals with a hydrophone.

The brain behind the project is Brett Phaneuf from the non-profit charity ProMare that focuses on the importance of oceans for the global climate. Rosie Lickorish – a new technology specialist at IBM and one of the partners of the project – is convinced that autonomous ships will play a major role in ocean research as a ship without a human crew can focus purely on research and attempt to sail in the toughest of conditions.

Hundreds of experts from many countries including the USA, India, Switzerland, and more are working on the Mayflower 400 and the data gathered will be freely available to anyone. Mayflower 400 will sail along a similar path as the original Mayflower – from Plymouth in Britain to Plymouth in Massachusetts and the journey should take three weeks.

The AI on board of Mayflower 400 was trained to identify obstacles it can encounter during its journey and react correctly. Thousands of images of sea objects and various scenarios were used to train the AI. Hopefully, it will survive the journey.

Sources:


*https://techxplore.com/news/2021-04-ai-captain-autonomous-ship-maiden.html


  • If you like the content I’m producing about science maybe you will like the content I produce about gaming as well! Be sure to check out my other posts!
Sort:  

There are quite some research done into the field of "autonomous ships" - and I believe there are already some in hard production, doing useful work today (though, not sail powered). I also think there are some smaller autonomous research equipment moving around in the sea. Perhaps something to do research into? :-)

The crew is usually the weakest point on a ship (particularly on a good sail boat). There are several stories where the crew have abandoned ship and been rescued due to bad weather, with the boat still floating around years later.

Another thing I've though about - there are lots of accidents happening every year due to silly human mistakes. A year ago, a Norwegian frigate with super-expensive radar equipment and navigational systems managed to crash right into a gigantic brightly lit oil tanker, despite being warned over VHF that they were on collision course - it's completely ridiculous. I've also had several accidents when I've been out sailing, i.e. crashing into land while falling asleep or having the attention somewhere else, failing to see that I've asked the autopilot to take the boat directly over a rock ... worst was when I let someone else have the rudder ... there was one rock in the sea, it was clearly marked both on the map and physically. Said person was steering manually and had a lookout both on the sea and on the navigation system and managed to crash straight into this rock. I could go on ... big ships crashing into bridges or going straight on land because the captain was drunk or fell asleep over the rudder ... yes, it happens, far too frequently.

There are of course risks with autonomous systems, all programs are buggy, there can be unexpected problems onboard that needs either human human hands and/or head(s) to be resolved, other floating objects/ships that cannot be properly understood (or maybe not even seen) by the radar ... but still ... I believe a lot of accidents could have been avoided either by autonomous controlling of the ship - but instead of replacing the crew completely, it may be an idea to fix better warning systems.

All modern navigation systems come with alarms, but they usually go off way too often, they are often turned off, and the system usually don't make any difference i.e. between a "warning" and a "critical" alarm ... and often it's possible to go crash into an island or another ship without any alarms going off at all. In the case of the frigate crashing with the tank ship, a simple AIS alarm could have prevented it. However, such alarms were turned off because there were "too many false alarms due to moored or anchored ships". I have the possibility to turn on alarms on waypoint arrival, low depth (under the bow), out of course and objects within a radar guard zone ... but putting on and adjusting all those alarms can be a major hassle, the alarms are rarely useful and too often an annoyance. And the most important alarm: "you're on collision course with a rock or land" - it's simply missing. The navigator knows where I am, and it usually knows about the land and the rocks right in front of the boat , it should be very possible to program it to give a critical loud alarm whenever the ship is in immediate danger of collision, but no ... that's not how it works.

The crew is usually the weakest point on a ship (particularly on a good sail boat). There are several stories where the crew have abandoned ship and been rescued due to bad weather, with the boat still floating around years later.

Another thing I've though about - there are lots of accidents happening every year due to silly human mistakes. A year ago, a Norwegian frigate with super-expensive radar equipment and navigational systems managed to crash right into a gigantic brightly lit oil tanker, despite being warned over VHF that they were on collision course - it's completely ridiculous. I've also had several accidents when I've been out sailing, i.e. crashing into land while falling asleep or having the attention somewhere else, failing to see that I've asked the autopilot to take the boat directly over a rock ... worst was when I let someone else have the rudder ... there was one rock in the sea, it was clearly marked both on the map and physically. Said person was steering manually and had a lookout both on the sea and on the navigation system and managed to crash straight into this rock. I could go on ... big ships crashing into bridges or going straight on land because the captain was drunk or fell asleep over the rudder ... yes, it happens, far too frequently.

There are of course risks with autonomous systems, all programs are buggy, there can be unexpected problems onboard that needs either human human hands and/or head(s) to be resolved, other floating objects/ships that cannot be properly understood (or maybe not even seen) by the radar ... but still ... I believe a lot of accidents could have been avoided either by autonomous controlling of the ship - but instead of replacing the crew completely, it may be an idea to fix better warning systems.

All modern navigation systems come with alarms, but they usually go off way too often, they are often turned off, and the system usually don't make any difference i.e. between a "warning" and a "critical" alarm ... and often it's possible to go crash into an island or another ship without any alarms going off at all. In the case of the frigate crashing with the tank ship, a simple AIS alarm could have prevented it. However, such alarms were turned off because there were "too many false alarms due to moored or anchored ships". I have the possibility to turn on alarms on waypoint arrival, low depth (under the bow), out of course and objects within a radar guard zone ... but putting on and adjusting all those alarms can be a major hassle, the alarms are rarely useful and too often an annoyance. And the most important alarm: "you're on collision course with a rock or land" - it's simply missing. The navigator knows where I am, and it usually knows about the land and the rocks right in front of the boat , it should be very possible to program it to give a critical loud alarm whenever the ship is in immediate danger of collision, but no ... that's not how it works.

Congratulations @kralizec! You have completed the following achievement on the Hive blockchain and have been rewarded with new badge(s) :

You have been a buzzy bee and published a post every day of the week

You can view your badges on your board and compare yourself to others in the Ranking
If you no longer want to receive notifications, reply to this comment with the word STOP

To support your work, I also upvoted your post!

Check out the last post from @hivebuzz:

False-Positive phishing alert reported by antivirus software
Feedback from the May 1st Hive Power Up Day
Support the HiveBuzz project. Vote for our proposal!

This is a very cool project. I imagine that autonomous sailing is easier than autonomous driving, at least until the ship tries to navigate a crowded port.