Title: Intention Recognition and Trajectory Tracking in Human-Robot Interaction Using Deep Learning and VR

 

Abstract: 

Recognizing intentions and tracking trajectories in Human-Robot Interaction (HRI) is crucial for enabling robots to anticipate and respond to human actions effectively. This study investigates the use of deep learning techniques for classifying and tracking human intentions and locations in HRI, with data gathered from Virtual Reality (VR) environments. By utilizing VR, a controlled and immersive setting is established, allowing for precise monitoring and recording of human behaviors. Ensemble deep learning models, particularly Convolutional Neural Networks (CNNs), Long Short-Term Memory (LSTM) networks, and Transformers, were trained on this rich dataset to recognize and predict human intentions and track their trajectories with high accuracy. While CNN and CNN-LSTM models delivered high accuracy, they faced challenges in correctly identifying certain intentions. In contrast, the CNN-Transformer model outperformed the others, achieving higher precision, recall, and F1-scores in classification, along with strong performance in trajectory tracking. This approach highlights the potential for enhancing HRI by equipping robots with the ability to understand and respond to human intentions in real-time, fostering more intuitive and effective human-robot collaboration. The experimental results underscore the promise of deep learning in advancing intention recognition in HRI.