Speech understanding in air intercept controller training system design

This report describes the requirements for a speech recognition and understanding system to support an automated training system for Air Intercept Controllers (AIC). A combined isolated word recognition (IWR) and limited connected speech recognition (LCSR) system was developed and tested in a laboratory AIC training system model. Speech stylization constraints were minimized, resulting in particularly challenging recognition requirements. Integration of the IWR and LCSR techniques proved difficult.