Hi @yvesyves, welcome to the forum! 
Thanks for transferring this over from your email - hopefully others will find it interesting/of use.
Initial Response (from email)
We don’t have additional views, and the meaning of the thruster configurations is one of the things we’re working on improving our documentation for at the moment.
Thruster Contributions/Factors
In ArduSub each thruster is considered to have contributions to roll / pitch / yaw / throttle (vertical) / forward / lateral, and for the default configurations the intent/expectation is that contributions to the translation planes should be through the vehicle’s center of mass (e.g. if you tell your ROV to go straight down, it will activate all thrusters proportionately to their ‘throttle’ factor, and will expect to go straight downwards with no rotation about any of the axes).
For the 4-thruster simplerov configuration those factors are defined here - basically thrusters 1 and 2 are expected to contribute only to forwards and yaw, so should be in the same horizontal plane as the vehicle’s center of mass, and thrusters 3 and 4 are expected to contribute only to vertical and roll, so should be in the same vertical plane as the vehicle’s center of mass.
ArduSub Motion Control
An important concept to understand is that ArduSub does its motion control in the body frame, so its control for each direction and rotation axis is done independently. This works very well and intuitively for vehicles that are intended to be controlled like that, but vehicles that are intended to be controlled with composite motion need to have those separate components specified by the pilot.
As an example, the design in your image has thrusters with independent and intuitive control of the forwards motion, and the roll and yaw axes. With the vertical thrusters where they are there’s a large pitch component but very little vertical component - moving up/down with such a design would practically always be accomplished by pitching the vehicle and then going forwards, which is also quite efficient from a drag perspective. Note that that conflicts with what the 4-thruster simplerov configuration expects - instead of the expected strong vertical component and negligible pitch component you have the opposite. It’s of course possible to move the vertical thrusters closer to the center (assuming the center of mass of your vehicle is at/near the center of volume/buoyancy) but most likely you don’t actually want independent vertical control in this case - it’s poorly suited to a long-thin vehicle because of the significant drag involved compared to moving forwards while pitched.
Setting Custom Thruster Factors
For ArduSub to have the best idea of the available motion components it would be best to make a custom configuration with the relevant pitch contribution specified, instead of using the 4-thruster simplerov configuration. Note that since there is still some vertical contribution from the vertical thrusters you can leave the vertical factors as 1.0 - the factors are normalised by the relative strength of all available contributions to a given motion type, so ‘small vertical’ maps to 1.0 if all other thrusters contribute ‘no vertical’.
Hopefully the above information is a good starting point. In addition, check out the posts tagged with ‘vehicle-design’ - there’s a bunch of info in them covering thruster positioning, and vehicle control directions and balancing
Response to changes
That’s significantly shorter than the one you’ve pictured, in which case it might make more sense to move the vertical thrusters back to the middle and just use the standard 4-thruster simplerov setup.
If you’re instead thinking you might just want to use our enclosure end-caps with a custom length of 6" tube then the above points about adding the pitch contribution still apply. Also, you might be interested in the Under Pressure pressure-vessel design software, to help determine a reasonable depth capacity for your tube, and/or which thickness is appropriate for the depth you need to go to