Questions about Hydrodynamics parameters (Inconsistencies)

[lblunschi]

Hi everyone, hi @evan-palmer,

First, thank you for creating and maintaining this repository and the related projects, they’ve been extremely helpful for my work so far.

I’m currently working on a model-based controller for the BlueROV2 and I am looking for hydrodynamic parameters that would allow me to create an accurate motion model. While looking at the Gazebo simulation setup, I noticed that the added mass and linear damping terms are set to zero:

blue/blue_description/description/bluerov2/gazebo.xacro

Lines 84 to 109 in 812fea6

	<gazebo>
	<plugin filename="gz-sim-hydrodynamics-system" name="gz::sim::systems::Hydrodynamics">
	<link_name>${prefix}base_footprint</link_name>
	<!-- Added mass -->
	<xDotU>0</xDotU>
	<yDotV>0</yDotV>
	<zDotW>0</zDotW>
	<kDotP>0</kDotP>
	<mDotQ>0</mDotQ>
	<nDotR>0</nDotR>
	<!-- First order stability derivative -->
	<xU>0</xU>
	<yV>0</yV>
	<zW>0</zW>
	<kP>0</kP>
	<mQ>0</mQ>
	<nR>0</nR>
	<!-- Second order stability derivative -->
	<xUabsU>-33.8</xUabsU>
	<yVabsV>-54.269</yVabsV>
	<zWabsW>-73.371</zWabsW>
	<kPabsP>-4.0</kPabsP>
	<mQabsQ>-4.0</mQabsQ>
	<nRabsR>-4.0</nRabsR>
	</plugin>
	</gazebo>

I was wondering if this was intentional, and if so, what the reason for this choice was.

In contrast, the controller configuration in the blue_demos repository uses nonzero values for both added mass and linear damping (also the quadratic damping factors are inconsistent with the ones used in the simulator):

blue/blue_demos/control_integration/config/bluerov2_controllers.yaml

Lines 40 to 49 in 812fea6

	hydrodynamics:
	mass: 10.0
	weight: 98.1
	buoyancy: 100.06
	moments_of_inertia: [0.16, 0.16, 0.16]
	added_mass: [-5.50, -12.70, -14.60, -0.12, -0.12, -0.12]
	center_of_buoyancy: [0.0, 0.0, 0.0]
	center_of_gravity: [0.0, 0.0, 0.0]
	linear_damping: [-4.03, -6.22, -5.18, -0.07, -0.07, -0.07]
	quadratic_damping: [-18.18, -21.66, -36.99, -1.55, -1.55, -1.55]

Additionally, the parameters used for the heavy configuration differ from those in the auv_controllers repository (both chained and individual controllers which are designed for the BlueRov2 Heavy, given the 8 thruster configuration):

blue/blue_demos/control_integration/config/bluerov2_heavy_controllers.yaml

Lines 46 to 55 in 812fea6

	hydrodynamics:
	mass: 13.5
	weight: 114.80
	buoyancy: 112.80
	moments_of_inertia: [0.16, 0.16, 0.16]
	added_mass: [-5.50, -12.70, -14.60, -0.12, -0.12, -0.12]
	center_of_buoyancy: [0.0, 0.0, 0.0]
	center_of_gravity: [0.0, 0.0, 0.0]
	linear_damping: [-4.03, -6.22, -5.18, -0.07, -0.07, -0.07]
	quadratic_damping: [-18.18, -21.66, -36.99, -1.55, -1.55, -1.55]

vs
https://github.com/Robotic-Decision-Making-Lab/auv_controllers/blob/20f768da4fd283a48aa506648b4a9f9168584332/auv_control_demos/chained_controllers/xacro/chained.model.xacro#L3-L33

I may be misunderstanding how these hydrodynamics parameters are used, so I wanted to ask:

• Are the added mass and linear damping terms in the Gazebo model intentionally set to zero?
• Is there a recommended source for validated parameters for the BlueROV2?

Any clarification would be appreciated.

[evan-palmer]

Hi @lblunschi

Are the added mass and linear damping terms in the Gazebo model intentionally set to zero?

Yes, the Gazebo parameters differ because I didn't find the simulated hydrodynamic behavior matched what I was observing during my hardware tests. I adjusted those to get something "reasonable", and wouldn't recommend using those on hardware.

Is there a recommended source for validated parameters for the BlueROV2?

The parameters that you linked here are based on the parameters that Wil Marais used in some of his experiments for this paper. I believe that he got these parameters from 6-DoF Modelling and Control of a Remotely Operated Vehicle - Chu-Jou Wu, which uses a least squares method to obtain the parameters. These are also the parameters that I am currently using on hardware. A related simulation paper that I have seen is An Open-Source Benchmark Simulator: Control of a BlueROV2 Underwater Robot, but I haven't tried these parameters on hardware yet (I'm skeptical of the drag coefficients).

I'll note that my simulation testing with Gazebo is primarily intended to perform integration testing and to get a general sense of the controller gains that I will need. The gains and dynamics can change depending on the operating conditions (e.g., a river vs the ocean), which is what motivated our transition to data-driven modeling and adaptive control.

[lblunschi]

Thank you for the detailed response and for sharing the source of the parameters that are used. :)

[evan-palmer]

Of course. Please feel free to reach out if you have any additional questions.