TJBot-A Caring, Emotionally Intelligent Robot

February 1, 2019

Overview

My winter project for the Northwestern MSR program is to build/program a simple, caring emotionally intelligent robot using IBM’s TJBot and Watson Services. The target audience is intended to be all people who want a companion, entertainment, or a support system during a difficult time. I will be custom naming the TJBot “David” in honor of my favorite filmmaker David Lynch.

Motivation

My Career goal is to improve the lives of people using Social Robotics. I believe this project is a good learning tool to investigate how Social Robotics research can benefit a broad audience and improve the well-being of people in versatile ways. Currently there are challenges facing social robotics developers, such as having difficulty separating the technology from simple voice assistants. In addition, many chatbots and smart speakers that exist today have also shown to cause frustration in users, not only by erroneous responses, but also due to a lack of human emotion. Hence, I believe it is important to try to implement more natural and empathetic, conversational language in the technology to humanize the experience, make users more comfortable, and develop more of a rapport with machines.

While being inspired by chatbot therapists and companions that exist today, such as Replika and Woebot, I am motivated to pave the way for transforming this technology in the form of a Social Robot. I believe a cute robot that is physically always there, vs. on a phone screen, can make people feel even more comfort. Further, I believe designing a robotic system that implements an empathic interaction with users, and is capable of recognizing emotions in realistic, spontaneous environments has the power to potentially cure loneliness, and even save lives for suicidal individuals. Therefore I believe this project serves as an exciting gateway to utilizing more advanced research in social robotics, multimodal human emotion recognition, and affective computing to understand the needs of users, and develop more robust emotion recognition techniques.

FINAL DEMO VIDEO

The final demo video shown below demonstrates a dialogue with the TJBot, who I named David (after my favorite filmmaker David Lynch). In this dialogue, the main goal was to utilize an initial tone of sadness from the user as input to the Watson services, which then begins a dialogue interaction between David and the user based on the sad tone. In addition, TJBot software capabilities were utilized which lights up a blue LED based on sadness. David receives, detect, and responds based on the initial tone of sadness. From there,a natural, flowing dialogue for which David sounds concerned, and further asks the user the details of his sadness begins. I am simulating a scenario for which I have gone throuh a breakup. David then attempts to make me feel better and gets feedback from me on whether his response made me feel better. The resulting response from me indicates a positive emotion, and David lights his “sad blue” LED light from blue to “happy yellow” light and waves his arm three times as my happiness past a certain threshold of “happiness.” The demo also demonstrates the use of humor, chosen in inspiration of utilizing multiple methods of improving human well being. See the test in action below.

View project details on github here.

Find preliminary test files and weekly status notes for my project in this PROGRESS section.

IN WORK: System flow diagram updates and data flow description updates.