current sense in h-bridge opamp
It isn't very clear why you think you want to measure the polarity of the current; if you're looking for a zero crossing, this would look like a minimum in the measured current through the sense resistor, and to detect this you don't need the polarity of the current through the load before and after the minimum.
Also, if there's so much noise on the measured current, you are unlikely to be able to improve the detection of zero-crossing simply by knowing the polarity of the current through the load.
And, if the load *is* purely resistive+inductive, the current through it won't carry on increasing when the PWM switches direction, because the phase-shift can't exceed 90 degrees; the current through the sense resistor is never more than the current being put into the load through the MOSFETs.
Therefore, if you want to limit the load current because you don't want to damage the MOSFETs, you also don't need to know the polarity of the current because if 10A is flowing through the sense resistor then 10A is flowing through two of the MOSFETs.
So, what are you trying to achieve? And, is the load a simple resitor+inductor?
HTH
barny