NIT Rourkela’s human-like AI robot can understand everyday language, recognise emotions
New Delhi, Dec 11 (IANS/WISHAVWARTA) Researchers at the National Institute of Technology (NIT) Rourkela have developed a robotic system designed to interact with people in the most human-like manner.
Developed using Artificial Intelligence (AI) and Large Language Models (LLMs), the social robot can understand everyday language, follow spoken commands, answer questions, and engage in real-time conversations beyond pre-coded responses.
It is also capable of recognising users’ emotions by interpreting facial expressions, such as happy, neutral, or sad. It can greet users, provide instructions, and reply using natural-sounding speech.
The system, which can be used in homes, classrooms, offices, hospitals, and community spaces, can also detect simple gestures such as waving and raising a hand, and react appropriately.
“The indigenously developed robotic system significantly aligns with the Government of India’s Make in India and Atmanirbhar Bharat initiatives. In contrast to existing systems available globally, NIT Rourkela’s social robot introduces a uniquely integrated framework that combines gesture, emotion, speech, and LLM-based conversational abilities within a cost-effective platform suitable for Indian contexts,” said Dr. Anup Nandy, Associate Professor, Department of Computer Science and Engineering, NIT Rourkela.
The Institute has also secured a patent for the developed innovation and system.
The robot uses the Raspberry Pi system to process user inputs such as spoken commands or text-based queries.
The LLM analyses the input, determines the context, and produces an appropriate, human-like response. This output is then delivered through the robot’s speech system using Google Text-to-Speech.
The robot is equipped with a wheel-based platform and navigation system for easy movement. The built-in distance-sensing module allows it to avoid obstacles and ensure safe movement in a crowded setup.
The developed robot, detailed in the paper published in the Computers and Electrical Engineering journal, is expected to cost between Rs 80,000 and Rs 90,000 based on manufacturing scale and component optimisation, said the researchers.
—
