Multimodal smart robotic assistant

Abstract

Robotic assistants can help human beings in reducing the manual efforts and the risk factor in hazardous situations. The robotic assistants can be made smart by facilitating their control of operation using multiple modalities. In this paper, we develop a smart robotic assistant (SRA) whose body is controlled using tilt-gesture of a smart-phone, and its arm and claws using human voice commands. The voice command signal is processed using a speech recognition module that converts the speech signal received into BCD logic control commands. The SRA is capable of performing different movements such as turns, start/stop operations, and picking up an object from one location and placing it at another location. The effectiveness of the tilt-gesture and voice commands is examined through various experiments, with encouraging results. Potential applications of SRA are also discussed.