Hi Umer27, I read some of your other thread too.
I designed a commercial high-efficiency switchmode solar MPPT a few years back, and if you really want to deliver maximum power into a 12v battery the ONLY factor you need to optimise is the CURRENT delivered into the battery itself. So you need to think of it as a "maximum current point tracker". This in effect gives you the greatest power into the battery itself, no need to complicate things trying to measure or control "power".
Of course if you want to measure voltage too, then your AVR can do a multiply calc and show "power" on the display to entertain the user.
Re the mechanics, in your other post you mentioned trying to make it as efficient as possible, I would suggest doing the sensing optically to determine solar position using 2 light sensors and one lens. Then close-loop the error output through the AVR to control your actuator. Your AVR can maximise efficiency by running the motor and leadscrew infrequently and leaving the entire actuator system in a zero power state for 90+% of the time, leadscrews are good for that because they are resistant to being backdriven by the wind.
For cloud conditions etc the AVR can guess solar posiiton based on previous days and its 24hour clock, an average of where the sun was at that time on previous days.
I think on the PIClist years ago someone had a mechanical tracker system where the PIC actually was able to realign its own 24hour clock based on solar zenith (noon) averaging, so it kept flawless time all year round and used that time to control the mechanical position of the solar panels. One of the benefits was it only needed one light sensor (noon detector) and dirt problems etc of the sensor were much more easily addressed than dual sensors (like when a bird craps on ONE of the 2 sensors).