Voice directed communications system archetecture

In combination

• a plurality of voice-directed communications system platforms each associated with a respective public communications network node, each of said platforms being operative for establishing telephone connections via their associated node in response to voice directions received from respective subscribers via the associated network node, for receiving from the associated node speech signals characterizing respective identifiers associated with respective ones of said subscribers and for modeling said received speech signals in accordance with a predetermined modeling scheme,
• at least one voice verification unit, said at least one voice verification unit being associated with a respective group of said identifiers and associated with a respective database having stored therein voice templates characterizing respective ones of the associated group of identifiers, and wherein said at least one voice verification unit is operative for verifying the identities of respective ones of said subscribers of the associated group using the associated ones of said stored voice templates and corresponding models of said speech signals received from respective ones of said platforms, and
• a packet network for interconnecting each of said voice directed communications system platforms with said centralized voice verification system, said verification system receiving said models of said speech signals via said packet network and returning via said network the results of such verification to the originators of said models.

A voice directed communications system architecture comprising:

• a plurality of voice directed communications system platforms each associated with a telecommunications network node,
• a centralized voice verification system, and
• a data network for interconnecting each of said voice directed communications system platforms with said centralized voice verification system, and wherein said centralized voice verification system comprises a plurality of voice verification units, each of said voice verification units being associated with a respective range of subscriber identifiers and including
• a database for storing a plurality of voice templates characterizing respective ones of the associated range of subscriber identifiers, and
• at least one digital signal processor operative for verifying a subscriber's identity based on the subscriber's associated stored template and associated speech signals received from a particular one of said platforms via said network and for returning the results of said verification to said one of said platforms via said network.