FLEXIBLE MECHANOSENSORS FOR HUMANOID ROBOTS WITH ENHANCED TACTILE PERCEPTION

This thesis addresses key limitations hindering the advancement of humanoid robots, specifically in achieving dexterous manipulation, ensuring safe and intuitive human-robot interaction, and developing sophisticated artificial skins. It highlights the critical need for enhanced tactile sensing beyond simple binary contact, enabling robots to perceive object proximity and exert nuanced force control during manipulation. The work reviews existing robotic hand designs and control systems, underscoring the challenges in replicating human-like dexterity and adaptability. It also examines human-robot collaboration, emphasizing the potential of integrated multisensory systems, particularly tactile sensing, for safer interaction. Furthermore, the thesis explores the current state of artificial skins, identifying limitations in robustness, sensitivity, and integration. To address these challenges, this research presents the development of a human-like robotic arm, employing a bio-inspired design leveraging the mechanical properties of elastomeric materials. The thesis details the comprehensive mechanical modeling and characterization of these elastomers using constitutive models and experimental data from various strain tests, providing crucial parameters for arm design and control. Finally, the integration and testing of sensors within the developed robotic arm system are included. The concluding sections will discuss the findings and propose future research directions focused on enhancing manipulation, collaboration, and sensory capabilities in humanoid robots.