Learning conversation with a mobile robot

Abstract

This study proposes to design a mobile robot to assist language learning. This should motivate learners to practice speaking and listening in order to improve their English skill. This project provides a low-cost robot for all students who are willing to spend time learning English inside and outside the classroom. This research proposes to deliver a robot with wheels, cameras, microphones and speakers. Raspberry Pi 3 and Arduino provide the computation power to interact with and control the robot. The robot converses with English learners to accomplish specific tasks. In one task, a learner asks the robot to travel on a street map with simple roads and locations such as a library, a park, and an elementary school. The utterances of the learner are recognized with a Google cloud service, and the voice response of the robot is generated with text-to-speech technology. When the robot walks on a street map, it does image processing to follow lanes and recognize intersections and a number of prescribed locations. An empirical experiment was run to study the learning process of learners. Some interesting results were found.