Performance of Smart Personal Assistant Applications Based on Speech Recognition Technology using IoT-based Voice Commands

Abstract

The aim of this research is to compare the performance of three Smart Personal Assistant (SPA) applications: Google Assistant, Amazon Alexa, and Apple Siri. Experiments designed using smartphones, smart speakers, and control system devices show that the pronunciation performance has the highest success rate with Google Assistant (95%) for responding to voice commands and the lowest with Apple Siri (80%). Performance tests of the SPA voice command features show that Google Assistant has the fastest response time (0.62 seconds) for answering questions on common tasks and Apple Siri has the slowest (2.58 seconds). Performance tests of the system devices show that Google Assistant has the fastest response time for controlling home appliances using IoT-based voice commands (1.03 seconds) and Apple Siri has the slowest (5.96 seconds). Therefore, the Google Assistant application is the best of the three SPA voice commands. We expect this research will be useful for helping elderly and disabled people in their interactions with the IoT environment, as they will not need to move to turn on and turn off electrical appliances. They can remotely control these easily using smartphones, without having to physically touch the appliances.