Can tf/tf2 resolve "hidden transforms" ?

edit

Howdy folks,

I have question relating to the following transform tree:

[/nav_frame]->[/base_link]->[/sensor_platform] -> [/IMU_AHRS]

The sensor platform is literally a platform that has sensors mounted on it (GPS antenna, IMU-AHRS, etc.) and is attached to the robotic platform (/base_link).

The nav_frame is a local ENU frame in which I do my navigation computations: it's placed on the WGS84 ellipsoid at the same location as the base_link, and it always points north.

I got an IMU-AHRS system (with magnetometer and world magnetic map) mounted on the sensor_platform that I will use to resolve the transformation between nav_frame and base_link.

So, I have the following situation:

[/nav_frame]->[/base_link] UNKNOWN, this is what I need to resolve

[/base_link] ->[/sensor_platform] is a known and fixed transform

[/sensor_platform] -> [/IMU_AHRS] is a known and fixed transform

My Question: Can tf (or tf2) somehow automatically (internally) resolve the hidden transform [/nav_frame]->[base_link] when it's given the measured transform [/nav_frame] -> [/IMU_AHRS] as produced by the IMU-AHRS system, or do I (as the programmer) have to explicitly compute this transform (which is easy enough using the IMU-AHRS measurements and the fixed transforms relating the IMU_AHRS frame w.r.t. the base_link) and explicitly publish this as the [/nave_frame] -> [/base_link] transform?

if tf/tf2 can do this "internally", then how?

Thanks in advance

Best regards,

Galto