I have been reading about different types of underwater navigation. According to this Youtube video, these are the following navigation systems:

  1. Inertial navigation systems (use of Inertial Measurement Unit)
  2. Acoustic (e.g., long-baseline)
  3. Geophysical (use of external environment information)

This article states sonar affects whales and the use of the acoustic navigational systems should be minimized. It seems like the combination of inertial and geophysical underwater navigation is not enough. So what kind of underwater navigation system one should use in order to be environmentally friendly but also accurate?


Usually, most underwater robots are running on just inertial navigation and surfacing occasionally. This tends to work better than one would expect and is what most large scale deployments end up using. Additionally, most of these use a Doppler Velocity Log, Pressure Sensor, and Compass in addition to their IMU.

Geophysical localization is fairly precise, SLAM solutions have been developed for mobile and underwater robotics for a long time. I believe underwater SLAM solutions are in a fairly mature place and if your robot is operating in a feature-rich environment and has the ability to use LIDAR or cameras you could get a fairly good localization solution.

As an aside, usually, the sonar on robots is much less powerful and less overused than shipboard sonar. It's probably not a huge issue unless you plan to explicitly work near whales (but I am not a marine biologist).

  • $\begingroup$ What if my underwater robot is supposed to travel hundreds of kilometres without surfacing? It would only surface at the destination. I dont think IMU and camera sensors would be enough. DVL is part of an acoustic sensor so I scratched it from the available sensors in my small scenario. $\endgroup$ – Ship Jan 2 at 14:35
  • $\begingroup$ In that scenario, you'll likely have to do geographic localization using an underwater LIDAR. I see really no other option if you think dead reckoning won't be accurate enough. $\endgroup$ – Octopuscabbage Jan 5 at 20:19
  • 2
    $\begingroup$ You don't say what accuracy you need when you do surface, nor how much room or money you have for inertial sensors, nor how deep you need to run. If you've got a good compass, a knot meter, and a database of expected tides & currents then I'd be surprised if you couldn't go 100km and surface within 1km of your target. You could probably back that up with sensors for temperature & turbidity and get a long way to independent navigation, even if you couldn't get down far enough to "watch" the bottom. $\endgroup$ – TimWescott Jan 5 at 21:07

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.