Design And Construction Of Robot Head System Using Speech Recognition To Express Emotions

Liew, Quek Shiung (2019) Design And Construction Of Robot Head System Using Speech Recognition To Express Emotions. Project Report. UTeM, Melaka, Malaysia. (Submitted)

Text (24 Pages) Design And Construction Of Robot Head System Using Speech Recognition To Express Emotions.pdf - Submitted Version Download (297kB)

Abstract

Humanoid robot is a robot that has the characteristics of a human. It is designed to improve the quality of human life. Humanoid robot is able to have human-like behaviours and engage in effective robot-human synergy. Facial expressions are very important in communication between human. The objectives of this project are to develop a robot head system that is able to express five basic expressions which are happiness, sadness, anger, surprise and fear, to control the robot’s expressions using voice commands and to analyse the performance of the robot head system in terms of position, the recognition rate of each facial expression and the speech recognition rate. The control points of the robot head are proposed by combining the action units. The robot head that is able to express various emotions have been provided with 5 DOFs. A survey has been conducted on the design after finished drawing using SolidWorks. Evaluation on the conceptual design of the robot head is done, later, to be implemented on the final product of the robot head. Hardware experiment has been conducted and Arduino Mega is used as the controller for the system. Besides, a microphone with voice recognition module is designed to control the expressions of the robot using speech recognition. This project also focused on the robot head system’s ability to function in terms of position of the actuators, recognition rate for facial expressions and speech recognition. The result of the relative position error for the chosen actuator is less than 5%. This increases the ability of the robot head to express the emotions more effectively. Besides, a survey related to the recognition rate for the five facial expressions has been done on 30 subjects. The result obtained for most of the recognition rate is more than 50% except for fear which is 23.34%. The speech recognition rate is repeated for 30 times. The recognition rate for most of the emotions are above 90%. The objectives of this project have been achieved and a robot head system which is able to express emotions using speech recognition has been successfully designed and constructed. However, in the future work, the facial expression recognition rate is expected to be improved by adding more DOFs or improve the hardware integration. The robot head system can also be integrated with other robotic bodies and addition of sensors.

Actions (login required)

View Item