I would think that it would be far more reliable to position the panel in a similar fashion that telescopes are postioned. Point it to the position where the object (sun) is known to be and then track it with a simple time based drive. After all, the sun does not wander about the sky aimlessly. A table of data would be stored in EEPROM which hold the daily sun's offset from the equator.
This way the solar storage batteries are not run down trying to locate the sun on a very cloudy day.

Hal