Giving smart devices a body

• Main Author: Ng, Jing Hang
• Other Authors: Dino Accoto
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2020
• Subjects: Engineering::Mechanical engineering::Mechatronics; Engineering::Mechanical engineering::Robots
• Online Access: https://hdl.handle.net/10356/141821
• Institution: Nanyang Technological University

The field of robotics is advancing rapidly. Newer and smarter technologies are being developed for industrial applications, making advanced robotics a pillar of Industry 4.0. However, adoption rate of such technology for domestic applications is still low. An aspect of technology that has had a huge penetration in the populace is smartphones, where people use for work, play, and stuff. The technology in smartphones have advanced by leaps and bounds over the years and are now able to execute computations that even exceeds laptops from a decade ago. Application developers have also been able to take advantage of the capabilities of these smartphones by developing applications empowered with machine intelligence such as face filters, facial recognition, speech-to-text, image classifier, text classifier and other developments with machine learning platforms such as TensorFlow. With the intention of making use of the smart capabilities of a smartphone and integrating it with a robotic platform to conceptualize a smartphone enabled robot to improve the population penetration of robots in domestic situations, this project was established. The goal is to create two Android applications. The first will use the object detection and tracking capabilities of the TensorFlow Lite API to instruct a robotic platform. The second uses the TensorFlow Lite API’s speech recognition capabilities to move on the robotic platform on a user’s command. The project involved Android programming, interfacing with the TensorFlow Lite API, creating appropriate instructions and interfacing with the robot through the Universal Serial Bus (USB). The application was not implemented successfully, due to time and resource constraint that led to a lack of Android application debugging, so stability is not measurable. The author was able to identify the main issues with the application and architecture of the setup. However, this report intends to document the progress of the project as well as the insights gained from the project so that a better version of the application can be developed.