3 ms·
You could maybe pull that off on a small boat, but collision avoidance on large vessels can mean planning a long time ahead. You pretty much need a human. The
by chmod775 5y ago
You could maybe pull that off on a small boat, but collision avoidance on large vessels can mean planning a long time ahead. You pretty much need a human.
The largest vessels need 30 minutes to come to a full stop. Just turning is obviously easier, but it's still... slow and a huge arc.
- goodpoint 5y agoThe higher the inertia and momentum - the easier it is to predict routing conflict. A system could warn at -30 min, alarm at -25 and steer + slow down if nobody acknowledges the alarm.
- Gwypaas 5y agoThe issue is alarm fatigue. Couple this with the regular result of a collision avoidance situation being to go at most ~0.5 NM behind. You also have issue with corners where the vectors will intersect. There are projects to start trying to share waypoints through something similar to AIS which may help reducing the number of alarms to something workable. Look here for a nice example of chaos (or anywhere around Europe for that matter): https://www.marinetraffic.com/en/ais/home/centerx:2.9/centery:51.9/zoom:8 https://www.marinetraffic.com/en/ais/home/centerx:2.9/center...
- goodpoint 5y agoThat looks crowded indeed, and I was thinking about the use case of absent crew (as on the yacht) in relatively uncongested seas. Still, if ships where to broadcast their planned route every minute they could even handshake/agree a conflict avoidance strategy, like planes do. It still feel like an easier problem that implementing a self-driving car.