Browse Prior Art Database

Speech-Driven System for Identifying Materials

IP.com Disclosure Number: IPCOM000116987D
Original Publication Date: 1995-Dec-01
Included in the Prior Art Database: 2005-Mar-31
Document File: 4 page(s) / 117K

Publishing Venue

IBM

Related People

Kruse, D: AUTHOR [+4]

Abstract

Disclosed is a system using speech recognition for the identification of a material from a four-digit materials identifier required to be affixed to the container of any hazardous material or industrial chemical in the United States. This type of system is particularly valuable in an emergency situation, involving, for example, an overturned truck or a chemical splashed on an individual in a laboratory.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 46% of the total text.

Speech-Driven System for Identifying Materials

      Disclosed is a system using speech recognition for the
identification of a material from a four-digit materials identifier
required to be affixed to the container of any hazardous material or
industrial chemical in the United States.  This type of system is
particularly valuable in an emergency situation, involving, for
example, an overturned truck or a chemical splashed on an individual
in a laboratory.

      A standard hazardous materials manual has thousands of
materials, in a format requiring the technician to first look up the
hazardous material number and then to look up an action code listed
according to the hazardous material number.  In an emergency
situation, this manual may be used by someone wearing heavy gloves
and holding a flashlight under poor lighting.  Furthermore, a
conventional hazardous materials manual does not have a cross
referencing system for checking to avoid using the wrong action code
for a material.

      As a solution for this time-consuming process, a custom
speech-recognition system allows a user to read the four-digit
materials identifier into a computing system, which provides a visual
response on a display screen and verbal feedback, indicating the
material being identified and instructions for its appropriate
handling.  The process may be started by telling the system to
identify an unknown substance.  For example, say "hazardous
materials" or just "hazmat."  Alternately, the system may be notified
using a pointing device or the button of a press-to-talk microphone,
or using such a button together with speaking a word or phrase.  The
user then reads the hazardous material identifier, such as a
four-digit code, and the system responds, for example, by displaying
the name of the hazardous material with actions to be taken on the
screen.  As audio feedback, the system either plays a pre-recorded
audio message or uses a text-to-speech capability to emphasize the
most important emergency response or actions required of the user.

      The system preferably has additional features providing
interaction with the user.  For example, the system may confirm, on
screen and through audio, that it has understood the input of the
user.  For example, the hazardous materials identifier can be
displayed prominently, allowing the user to make corrections if
needed.  Preferably, the system also includes a provision allowing
the user to interrupt the playback of a message, through audio or
through mechanical means, such as the depression of a button, to stop
the system if he realizes the materials identifier was not heard
correctly, or to ask a specific question about the material.
Preferably, the speech recognition system includes code for handling
specific types of common questions.

      The system may be able to use additional alpha-numeric
identifiers, such as the Universal Product Code (UPC) of a household
chemical.  Other input, s...