We need to know the relative position of the optical modules within 20 cm. This corresponds to the precision one can expect for the relative timing of photo-multipliers with a few photoelectrons. The angular resolution of the detector depends on the timing accuracy.
This can be achieved in principle with a triangular sonar base and acoustic detectors. Such a system is under preparation.
We plan to take advantage of the know-how of different groups working in the field. There may be alternative solutions to internal alignment. For example, a pulsed light source could in principle achieve the same goal. It remains to be proven and compared. One also has to get an absolute positioning of the whole detector to be able to point to individual sources. Solutions exist with acoustic devices connected to surface detectors coupled to differential GPS.