Real-life speech-enabled system to enhance interaction with rfid networks in noisy environments

Ontology-based pattern generator and root semantic analyser for spoken dialogue systems

“…The dialog interpreter used in this application, like our previous work [7,8,6], relies on a pattern matching language to understand user's utterances. More specifically, the Artificial Intelligence Markup Language (AIML), an XML compliant language that is designed to create chat bots that can fool users into thinking that they are chatting with a person.…”

“…Here speech utterances are captured in a buffer and passed either directly to the automatic speech recognition (ASR) system or to the noisy speech enhancer according to the user preferences. Noisy speech is enhanced by an online signal subspace enhancement technique based on the Variance of the Reconstruction Error of the Karhunen-Loeve Transform (VRE-KLT) [6], whose main advantage is that it is a wave-in-wave-out method and does not require the separation of noise information. The N-best ASR recognition outputs are then fed to a dialog manager, which parses them and returns either a user-friendly message to the user to indicate that it did not understand the meaning of the utterance or the most likely command with its parameters to the business logic end.…”