An Embedded Multi-Modal Fingertip Sensor for Dexterous Anthropomorphic Hands

The growing need for intelligent and adaptive robotic manipulation has intensified research into multi-modal sensing technologies that can emulate human fingertip perception capabilities. This paper presents the development of a novel biomimetic multi-modal sensorized fingertip that offers advantages in sensor architecture, sensing performance, and system integration. The fingertip module integrates optoelectronic tactile sensing, a 9-axis IMU, and long-range ToF proximity detection in an ultra-compact form factor. A built-in microcontroller enables LED control via PWM and provides standard wireless interfaces, ensuring compatibility with future battery-powered setups. The sensor is designed for seamless integration with anthropomorphic hands and provides tactile, proximity, and inertial data simultaneously, enabling adaptive grasping and contact-aware manipulation. Experiments validate its performance in contact detection, orientation estimation, and distance measurement. This multimodal, biomimetic fingertip offers a promising solution for next-generation robotic and prosthetic hands, enabling delicate object handling and advanced human-robot interaction.