Development of Interactive Visual Recognition Assistant Bot

摘要

The rapid progress in the field of robotics have enabled robots to communicate with humans like friends and provide a personal touch in human lives. Assistive robots are beneficial for people with disabilities and help elderly people in their daily chores and improves their day-to-day living. The interaction between robots and humans over voice commands through voice assistant systems like Alexa, Siri, Cortana, Bixby, Google Assistant etc., have called for research in the field of Human Robot Interaction via Language Processing. With the use of NLP (Natural Language Processing), NLU (Natural Language Understanding) and NLG (Natural Language Generation) and Computer Vision, robots can understand human language by breaking down the spoken sentence into 3 parts; Wake word, Invocation name, Utterance and find the intent using the keywords extracted. Voice control is an effective way to control the robot and helps in performing task in areas that pose high risk for humans to enter. AI powered chatbots for communication with isolated patients, Assistive robots to deliver food, medication and other necessities etc. have proved to be quite beneficial in today's times. This paper illustrates the development of the software architecture of an Assistive Robot that can substitute humans in areas that are not advisable for humans especially in hospitals where services rendered to the patients directly through human contact can cause spread of infections. The robot will feature voice control system which can be operated using voice commands as well as have one-to-one personalized communication with the patients about general topics which makes the patients feel comfortable and also provide an interface for doctors to communicate with the patients. The robot can also provide answers to questions asked by the patients by searching for the answers over the internet or its own database. The robot is capable of interacting with the patient using their names which it recognizes through face recognition techniques.