Advancing Noise Robust Automatic Speech Recognition for Command and Control Applications

Abstract

Abstract : This is a technical assessment paper intended for use by engineers and research scientist working on the development and integration of Automatic Speech Recognition (ASR), it will cover the state of speech and recognition technologies with emphasis on noise robust command and control (C2) application. The reliable elimination of the keyboard and mouse in mounted and un-mounted C2 systems has been a desire of systems developers and requirements writers since the development of PC-based ASR systems in the early 1990's. However, current research and commercial quality ASR applications never had the noise robustness to support a truly tactical C2 application. As ASR achieved limited operational success in noisy environments around the 2002 timeframe, the C2 requirements evolved to include the emerging system of systems approach and multilingual operational environments in support of the Global War On Terrorism (GWOT) in such environment's, the system must understand not just words as commands (ASR), but to understand phrases and sentences (semantic and syntactic) and reply in a conversational manner (speech and natural language generation). If the keyboard and mouse are to be truly eliminated, a system now needs to conduct a natural conversation with an operator and possibly others in the operational environment. This paper will cover the advances, limitations, and reasonable expectations from several levels: Research Scientist and Engineers, Program Executive Office (PEO), Program Manager (PM), and requirements office. I will also discuss the major technical challenges that remain as well as some risk assessment to help decision makers align expectations with reasonable availability dates based on current and future research efforts.