For extra terrestrial systems, there are proper directional tracking systems with a definite purpose. The solar panels are to be eventually directed to receive the solar beam.
I think that it can be done. you monitor the voltage and current for each PV array/string. The one with the highest output is the one facing the sun.Also, if I am not mistaken there are test cells available for such cases. Instead of monitoring the entire array, you just monitor the test cell.
You have an arrangement rather similar to, though not as complicated as, the eye of an insect - an ommatidium. You can find the sun direction by post processing the outputs of the cells using some sort of spatial filter. This will also remove all ambiguity due to indirect lighting and haze - providing there is enough direct light. But why go to all this trouble when as Stelios says, there are well known algorithms for locating the directions of the sun and stars in space?