Soft Touch

Soft touch research focuses on understanding and replicating the human sense of touch, primarily for robotic manipulation and human-computer interaction. Current efforts concentrate on developing vision-based tactile sensing, using techniques like DeepSDF and supervised variational autoencoders to reconstruct 3D shapes and learn grasping strategies from tactile data, often integrating this with advanced AI models like large vision-language models for improved control and understanding. This work is significant for advancing robotics, particularly in delicate tasks like surgery and underwater exploration, and for creating more immersive and interactive virtual experiences.