The Magdalena Ridge Observatory Interferometer (MROI) is a long baseline optical interferometer under construction in New Mexico, USA. Once complete, it will study active galactic nuclei, young stellar objects and mass transfer in dynamical systems at visible and near infrared wavelengths with sub-milliarcsecond angular resolution. The facility will gradually grow to include ten relocatable 1.4 metre diameter unit telescopes with baselines up to 347 metres. The optical beam trains will deliver light from the telescopes over hundreds of metres in vacuum pipes to a central beam combining laboratory that houses the fringe tracker and science instruments.

One major improvement of the MROI over contemporary interferometers is its aim to track fringes on targets as dim as magnitude 14 in the H band, which will unlock more of the night sky for high resolution imaging than ever before. Beam alignment stability is crucial for meeting this sensitivity goal because errors in beam angle and position would provoke fringe visibility loss in interferometric measurements. The top-level error budget for the MROI defines a tight alignment tolerance: every beam must overlap with the nominal optical axis within 15 milliarcseconds (on-sky) in angle and 1% of the pupil diameter in position at all times during observations.

An Automated Alignment System (AAS) has been proposed to meet this technically and logistically challenging goal. Its job is to initialise the ten beam trains prior to the start of science observations, then monitor and correct alignment drifts caused by overnight temperature swings. All of this will happen with minimal human intervention by utilising a set of artificial light sources and detectors distributed throughout each beamline. Start-of-night alignment procedures must be completed within one hour, while intra-night procedures must consume no more than 5 minutes every hour.

This thesis describes the development of three core components of the AAS:

1. A Shack-Hartmann detector for measuring the angle and position of collimated reference and starlight beams.
2. A dual-wavelength reference light source to be placed near each unit telescope that acts as a bright proxy for starlight.
3. A broadband reference light source that resides in the beam combining laboratory.

The objectives over the time frame of this project were as follows:

1. Establish an error budget and derived requirements for the hardware.
2. Build and verify the performance of concept prototypes in a laboratory.
3. Build and verify the performance of preliminary designs at the observatory.
4. Progress as far as possible towards a final design for these three components.

Questions relating to stability and sensitivity have been answered through results from concept prototyping, deployment at the observatory and simulations. The final design of the AAS will depend on the outcome of a small number of future studies, but this work has ensured that the remaining technical risk is low.