Design of a speech recognition interface for a personal emergency response and health monitoring system

Personal emergency response systems are employed by seniors to help contact assistance in case of a fall or injury. An automated interface and planning module for a personal emergency response and health monitoring system has been developed using speech recognition to communicate with the senior over a microphone array and speaker. This interface would reduce the volume of false alarms plaguing call centres, allow the senior to decide when and whom to contact if help is needed, and facilitate communication from a health monitoring module. The effectiveness of the microphone array, speech recognition engine, and the planning module were investigated in three pilot tests. These demonstrated that a microphone array was able to improve the recognition rate; a yes/no vocabulary was able to achieve high accuracy of detection, and the prototype was able to conduct a dialog with subjects and make an appropriate decision.