Hi,
I feel sorry that I am not being able to support this thread as I should and could not complete the conversion of my SMD design to Hobbyist version. (I wish though) Here are some quick tips that may help you babalola:

Synchronous Rectification Deals with Switching on the upper FETs on a per cycle,phase basis so that the rectifier efficiency is best while charging, where the Body Diode is superseded by the RDSon of the MOSFET. For the upper one this may sound simple if you are sensing the input AC and know the Transformer connected as ( Primary Start/Finish vs Secondary Start/Finish, My Pro Design detects it) but for the lower one it gets tricky.

So lets forget Synchronous for now and let the upper MOSFETs behave just like diodes.

Apply PWM to both the Lower MOSFETs simultaneously but only in the middle of a single cycle, with upper limited duty cycle.
To experiment what is the limit with your transformer and profile entire system.
Have a load () connected to the Bridge +DC- and ensure Caps in order of 10000uF
Run a slow loop to increase the PWM (till you fry your MOS )
Idea is to figure out if you are actually achieving any boost

And Tahmid thanks for the great post. I wish I could actually spend sometime giving back to the community for which I am whatever I am today. But I personally would not buy the idea of, a set of different lookup tables for voltage regulation. You can have a 8x8 unsigned multiply within 54uS as stated by Darrel somewhere else and an unrolled Bit tested, rolling multiply takes about 42-44uS to call from PBP. For a proportional control and tricky retention of portion's of older error (I call it pseudo I, cause a software automatically introduces the integration Time) at least a PI system can be designed.

What I have learned from experience is that sometimes it is the Hardware that is crap than my own code (Loads of Crap ) which is to be blamed. Post a schema if possible.

Something to read : http://www.researchgate.net/publicat...fe3d4e9275.pdf

Thanks to all
Regards
Sougata