A Cross-Platform Tactile Capabilities Interface for Humanoid Robots

Abstract

This article presents the core elements of a cross-platform tactile capabilities interface (TCI) for humanoid arms. The aim of the interface is to reduce the cost of developing humanoid robot capabilities by supporting reuse through cross-platform deployment. The article presents a comparative analysis of existing robot middleware frameworks as well as the technical details of the TCI framework that builds on the existing YARP platform. Currently, the TCI framework includes robot arm actuators with robot skin sensors. It presents such hardware in a platform-independent manner, making it possible to write robot control software that can be executed on different robots through the TCI frameworks. The TCI framework supports multiple humanoid platforms, and this article also presents a case study of a cross-platform implementation of a set of tactile protective withdrawal reflexes that have been realized on both the NAO and iCub humanoid robot platforms using the same high-level source code.

Cognitive robotics software development aspects based on experiments of future software engineers

“…2. The layered architecture of TCI [20] Another useful cognitive robot model is the layered architecture of TCI (Tactile Capabilities Interface). As middleware layer, TCI [20] uses YARP (Yet Another Robot Platform).…”

“…The layered architecture of TCI [20] Another useful cognitive robot model is the layered architecture of TCI (Tactile Capabilities Interface). As middleware layer, TCI [20] uses YARP (Yet Another Robot Platform). YARP implements a variant of the observer pattern, which is a type of distributed publish-subscribe.…”