ProxySKIN: a Multi-Modal Robotic Skin for Human-Robot Interaction

This dissertation presents the design, implementation and testing of a new sensing technology from the hardware up to the cognitive perception and control levels, based on networks of proximity and tactile sensors providing a unified proxi-tactile perception of the environment for collaborative robots. This technology, called ProxySKIN, is meant to cover large areas of the robot body and aims to support a safe and efficient collaboration between industrial manipulators and humans. In particular, this skin-like sensory system aims to provide a complete and seamless proxy-tactile perception of the robot surroundings, where proximity and tactile information can be intended as a new extended sensing mode. This idea opens new challenges for the development of robots with high cognitive capabilities and motivates: (I) the study of perception and control methods for autonomous and human-driven robot operations; (II) the implementation of new types of robot tasks exploiting joint proxy-tactile feedback; (III) the development of techniques to enforce safety in human-robot interaction (HRI); (IV) The development of techniques for self-localization of robots and models for robot-centric environment representation. This research activity is at the core of the European project Sestosenso and is structured into three main sections that follow the logical thread of the project. After introducing the general context and providing some insights on the project objectives, the second chapter of this dissertation addresses the design and implementation of the sensing architecture from a hardware and software perspective, focusing on the design choices and requirements that led to the final prototype release. The third chapter focuses on the communication infrastructure designed for large-area multi-modal sensor networks, addressing critical application-oriented challenges, such as bandwidth optimization, cable routing efficiency, network scalability, and fault tolerance. The fourth chapter describes the task priority control architecture implemented to concurrently manage tactile and proximity sensor feedback in two relevant application scenarios, thus demonstrating the effectiveness of the proposed technology.