7 Conclusions

 Our simulations of co-phasing algorithms, using a triplet hierarchy, show convergence for both the cases of unresolved and resolved objects. With 3 different wavelengths, zero path difference has been achieved with initial errors of the order of $10~\mu$m. Simulations with limited photon rates have shown that resolved images of bright stars are in principle obtainable with apertures of 10 cm, using adaptive phasing at visible wavelengths. Arrays of large adaptive telescopes should extend the limiting magnitudes towards m_v =20 on Earth and m_v =35 in space. Possible refinements of these algorithms remain to be explored, but the simulations demonstrate the effectiveness of the triplet algorithm, used with image sharpness criteria, for phasing a multi-aperture array. This establishes the practical feasibility of obtaining high-resolution snapshot images on compact objects, with future interferometers spanning up to thousands of kilometers. In particular, it supports current projects towards obtaining resolved images of exo-planets with sufficient quality to see clouds, continents and vegetated areas (Boccaletti et al. 1999).

Acknowledgements

One of us (E.P.) would like to express his gratitude to Alastair Macgregor for his supervision in the early stages of this project. He would also like to thank Neal Jackson for his supervision and fruitful discussions during the continuation of this project at Jodrell Bank.

We are very much indebted to Laurent Koechlin and Luc Arnold for their critical reading of the manuscript and useful suggestions. This work started at the University of Northumbria as a final year project and continued at the Observatoire of Haute Provence and NRAL, Jodrell Bank.