Hello Radanipour,
how should I say this in a polite way? A discussion forum is a great place to get hints an recommendations, but its not a good place to let other people do your own work ;-) Even if I would know, how your transducer behaves over temperature (which I still don't do), I would leave the analysis to you.
If you execute this analysis and get stuck at specific questions, I'd propose that you present these specific questions in this forum and look for hints and recommendations. And in the end, only the real world matters: the circuit will behave as it behaves, no matter what we discussed here before ;-)
As general comments I'd say:
1) I'd worry more about the temperature behaviour of the (still unkown) transducer itself (like changing efficiency) than about the temperature behaviour of the excitation current.
2) I've never put an inductive load in the feedback path of a current feedback amplifier. My first thought looking at your schematic was, that this is somehow strange, but my second thought was, that it might nevertheless work. Now it's time to find out, so go ahead.
3) If you're worried about accuracy of your current source, then you may consider to split the accuracy part from the power part. In the datasheet of the ADA4870 there is even a chapter about this topic, proposing a composite amplifier. (A precision opamp takes care for accuracy, while the ADA4870 takes care for output power.) The circuit won't work exaclty like that for your current source, but it may be a base for your further circuit ideas.
best regards
Achim
