AWOL:
Could you use a parallax method to get an indication of range?
Two pairs of LDRs mounted on servos turn to a light source so that when turned, illumination on each pair of LDRs is equal.
Simple trig gives a rough idea of range.
I would want a detector with wide view and another with very narrow view turn together until the narrow view detector caught the source.
The baseline of the triangle and angle-measure methods limits the accuracy and range possible. When the baseline is human eyes width, parallax ends about 20 feet out. When it is as wide as Earth orbit we can get parallax on close stars. Some reptiles move their heads side to side just to increase their baseline, maybe there's a robot that does that already.
Also consider that nothing says that the baseline has to be horizontal.