High Accuracy Speech Recognition in Noisy Environments

Wide acceptance of spoken language interfaces with automated equipment is contingent upon its continued accurate performance in adverse noise environments. Significant to the research is the ability to perform Natural Language Speech (NLS) recognition and synthesis in both a high and variable noise environment. The noise environment may contain varying amounts of non-stationary, stationary and impulse components or even competing speech. A practical small self-contained and portable module containing the necessary micro-processor and software plus standard connectors for speech enabling other automated computer systems wiil perform speech recognition and voice response in real-time. The goal NLS module for high noise environments would also contain a display and acoustic man-machine interface for application flexibility. The research of multiple noise abatement and speaker voice tracking technologies is implemented in the system together with state-of-the-art speech recognition, synthesis and natural language understanding. The combination of these emerging technologies are expected to be very resistant to accuracy degradation caused by both continuous and impulse noises present in military armored vehicles and on factory floors, as well as in the presence of other voices. An integrated micro-miniature system will be designed that is able to understand spoken statements, speak responses in dialog fashion, display all input and output on an internal display and exchange data with an external computer. BENEFITS: The first product is a microcomputer prototype device containing the high-accuracy natural language speech system for noise environments, suitable for OEM and computer integrators of automated controls, automobiles, construction machinery, aircraft and spacecraft. The device can control machinery, hands-free tele-communications, operation of unmanned vehicles, ground/air traffic control, maintenance, riot control. police activities simulator training and the entertainment industry.