I want to implement a precise positioning for my ROV. From what I have seen during my researches, the DVL and the USBL are the main sensors for this task.
But the DVL isn’t cheaper, the position calculated from it will drift and it doesn’t work well if you are too far away from the seabed.
On the other hand, the precision of the USBL seems more than acceptable (more or less few centimeters in the best cases) and its frequency of new positioning data seems enough for the models that I have seen.
So why using a DVL ?
Hi!
An USBL give you world positioning all the time, but it jumps some meters at distance.
A DVL give you very precise movements, but not world positioning.
That means that it drifts by distance/time, but also if you deploy at great depth where you have to use surface positioning as starting reference.
On work class systems you use both working together with advance calman filters.
You use USBL for estmates and averages, DVL for movements.
This gives good positioning, but complicated and expensive.
If you’d like to learn more about the differences between positioning systems like USBL, and position aiding systems like DVL, this wiki page from our website is a great start: Nortek | A Complete Guide to Underwater Navigation
One key thing to remember with USBL is that it requires continuous interaction between the ROV and a surface transponder, which can be unreliable in real world conditions, whereas DVL/INS solutions are fully self contained.
Traditional DVLs can only be used for relative velocities and must pair with an INS to give real world position, however newer models such as the Nucleus actually contain their own AHRS. In the specific case of the Nucleus, this can now be used to continuously output position, in addition to attitude, depth, altitude, water temperature and measurements of current speed.
Thanks @nicolettep, and great to be having this discussion as I’m trying to get my head around pros/cons of various. On USBL: My understanding is that USBL transceivers have a sensing ‘cone’ that the ROV must remain within. This becomes a major limitation in shallow water applications. Correct?
Hi @PeterM ! The effective sensing area seems to vary between different systems. As @k.deboer mentioned, some systems have an area you should avoid, while other systems only mention line of sight.
A more common issue with USBLs in shallow waters has to do with multipath as sound can bounce off the bottom or surface which can confuse the sensor. It won’t know what path was taken, so there’s some additional uncertainty about the measurement.
