Superior Speech Recognition
For 40+ years, UFA has led the aviation industry in speech recognition technology. The fundamentally superior functionality that supports all UFA products goes well beyond competitors to achieve the greatest possible accuracy.
Built specifically for UFA products by the award-winning researchers at Raytheon BBN, the speech recognition component of each of our solutions is a proprietary technology that provides voice recognition control for all of UFA’s products, including ATCoach,® ATTower,® ATWorld, ATLIVE, ATVehicle and ATSpeak. Trained on air traffic management language, our speech recognition provides up to 98% accuracy and infuses all of our technology with the capability to recognize language with a wide range of accents, audio qualities, and speed.
According to the FAA, automatic speech recognition techniques play a "crucial role" in "enhancing safety for air traffic control (ATC) in terminal environments." UFA's standout capabilities set us apart and set a higher standard for safety across the aviation industry.
Speech Recognition and Real-Time Response
UFA's voice application both recognizes speech and generates appropriate feedback/responses using a Text-To-Speech system. Our speech recognition component accepts a user’s spoken commands, provides core recognition functions, coordinates and applies valid commands to the simulation exercise, and provides appropriate pilot voice responses.
The system's responses can be customized to meet any scenario. Responses are generated in real time based on data from the response files and the simulation application. Multiple female and male pilot voices with distinct accents are available to increase training realism, and the user can control the speech rate, volume, and pronunciation of each voice.
Our speech recognition component can be configured for use with English ICAO, FAA 7110.65, or custom phraseology. It is currently supports bilingual recognition for English and German.
Grammar-Based Recognition for ATC Training
Purpose-built for air traffic control language and grammar, UFA's voice recognition tool includes grammar files and tools that define the allowable voice commands or phraseology recognized by the system. UFA optimizes each grammar to further increase recognition performance. In ATC applications, the grammar is typically defined in accordance with FAA, ICAO or other radio-telephony requirements. Grammar files also include airspace-specific information such as aircraft callsigns, facility names, and navigational aids. Our speech recognition application provides a mechanism, ATVgen, for dynamically adding this information to grammar files.
The system's sensitivity to specific air traffic management language enables the technology to increase voice recognition for non-native speakers of US English. UFA has developed several international Voice Models to support a variety of non-native speakers of US English accents and dialects.
