Cognito: A Multipurpose Robotic Assistant for Higher Education Environment

摘要

The automation of redundant and mundane tasks has become an integral aim of the robotics field. Hence, in this research, the system developed for the university students and instructors aids in taking attendance through facial/fingerprint recognition, transcribing the lectures, correcting exams, and displaying facial expressions. The proposed system, termed Cognito, incorporates various hardware and software technologies to achieve these tasks. Firstly, in a facial recognition task, a Histogram of Oriented Gradients (HOG) and Support Vector Machines (SVM) are used to perform face detection and match the detected faces to the trained images. Secondly, in the fingerprint recognition task, the distinct points on the finger are scanned and compared to the existing fingerprints. Both tasks are used for making the attendance process in an automated manner, thus reducing the time and effort required from the instructor. Thirdly, the transcription feature is implemented through the use of Google speech-to-text API which takes speech as input and stores the speech converted to text as a file. Fourthly, the correction of MCQ-based exams uses OpenCV, Optical Mark Recognition (OMR) and Canny Edge algorithm to compare a student’s answer sheet with a predefined correct answer sheet and also display the final grade. Fifthly, in the emotion identification task, the system interacts with users to detect human emotion (using CNN with Haar features), and based on that emotion, the corresponding message appears on the face of the robot. These functionalities of Cognito contribute to the learning process by saving time and automating repetitive tasks at a higher education setup.