ETRW on Speech Processing in Adverse Conditions Cannes-Mandelieu, France November 10-13, 1992

Integrating speech recognition into a helicopter

D. R. Starks (1), M. J. Morgan (2)

(1) Avionics Division, Canadian Marconi Company, Kanata, Ontario, Canada
(2) Flight Research Laboratory, Institute for Aerospace Research, National Research Council of CaOttawa, Ontario, Canada Ottawa, Ontario, Canada

In advanced cockpits the pilot's hands and eyes are kept busy. Speech recognition could provide an effective alternative input medium which would decrease the workload thereby increasing the mission's chance of success. Previous studies of speech recognition in avionics have indicated that for speech to be accepted in the cockpit, recognition performance had to increase and become consistent across speakers and acoustic conditions. A Fly-By-Voice (FBV) program was developed at the Flight Research Laboratory (FRL) of the National Research Council (NRC) which was intended to demonstrate and advance speech recognition technology for use in adverse environments such as the helicopter cockpit. The Speech Research Centre of the NRC, with the participation of Canadian Marconi Company (CMC) and the Neil Squire Foundation (NSF) developed a robust speech recognition system that achieves high accuracy in simulated noise and stress. The system has been realized in a prototype hardware platform. This paper describes the processes involved with the integration of the hardware recogniser into a helicopter and concludes the technology is robust to the adverse environments encountered and that speech recognition is a viable and useful addition to the advanced helicopter cockpit.

Full Paper

Bibliographic reference.  Starks, D. R. / Morgan, M. J. (1992): "Integrating speech recognition into a helicopter", In SPAC-1992, 195-198.