Human-Robot Interaction in Assisting Individuals with Disabilities using Deep Learning

Human-Robot Interaction in Assisting Individuals with Disabilities using Deep Learning

This thesis explores the enhancement of human-robot interaction (HRI) to assist individuals with disabilities using deep learning techniques, with a particular focus on humanoid robots such as Pepper. The primary goal is to enable robots to understand and respond more naturally and intuitively to individuals' specific needs, integrating natural language processing, gestures, and emotions. Current challenges lie in the robots' ability to dynamically adapt to complex scenarios, which limits their effectiveness. Through the collection and annotation of interaction data, deep learning algorithms, such as convolutional neural networks for gesture recognition and Transformer models for language processing, will be utilized. The multimodal integration of these visual and auditory data will help develop more personalized and responsive interactions. The implementation of this framework will be tested on assistance tasks such as mobility and emotional support, with the ultimate goal of improving autonomy and the quality of life for individuals with disabilities.

Objectives:

• Develop accurate gesture and emotion recognition for more intuitive robot interactions with individuals with disabilities.
• Improve natural language understanding to enhance context-aware robot responses to user needs.
• Create and test a multimodal interaction framework for personalized assistance, focusing on mobility and emotional support.

Human-Robot Interaction (HRI), Deep Learning, Gesture Recognition, Natural Language Processing (NLP), Humanoid Robots.

The candidate should hold a Master's degree in Robotics, Computer Science, or a related field, with a solid foundation in machine learning and artificial intelligence. They should have experience with deep learning techniques, including convolutional neural networks (CNNs) and Transformer models, and be familiar with human-robot interaction (HRI), particularly in the context of assistive technologies. Proficiency in programming languages ​​like Python, along with hands-on experience using AI/ML frameworks such as TensorFlow or PyTorch, is essential. Additionally, the candidate should have an interest in disability studies or assistive technology applications, with a strong motivation to improve autonomy and quality of life through robotics.

Prof. Badr ELKARI

b.elkari@ueuromed.org and cedoc.admission@ueuromed.org

04/10/2024