A complex research system has been described which is to be used for R&D for controller training. Work has been described which sought to enhance the state of the art in computer voice recognition capability. The problems of real-time voice recognition were discussed, and some ideas for their solution advanced. A description of a modularized laboratory version of a GCA controller training system was given. A departure from the traditional form of adaptive training was taken. Emphasis was given to informative feedback during the course of an approach, as well as summary information at the end, to facilitate teaching the concepts. Student and instructor need feedback, but of a different sort and at different times. Capabilities were described for research on the vocabularies of other controller-type tasks. The modularity of this system requires replacement of only the simulated task display, controller model, and performance measurement system for conversion to the appropriate training system. Procurement of a voice synthesizer was noted and advantages to training of such a capability were indicated.