ORCA: A multi-tier, physics based robotics simulator

• Main Authors: Emmanouil Hourdakis, Giorgos Chliveros, Panos Trahanias
• Other Authors: Emmanouil, Hourdaki, Giorgos, Chlivero, Panos, Trahanias
• Format: Article in Journal/Newspaper
• Language: English
• Published: Università degli studi di Bergamo 2014
• Subjects: Physics simulation; robotic; sensor modeling; graphic; distributed processing; Settore ING-INF/05 - Sistemi di Elaborazione delle Informazioni; Orca
• Online Access: http://hdl.handle.net/10446/87669; https://doi.org/10.6092/JOSER_2014_05_02_p13

Physics-based simulation has grown to become an indispensable tool in robotics research, since it greatly facilitates development and evaluation of complex robotics algorithms. Today it is considered a standard practice, prior to any deployment on hardware platforms, and an important part of the development activities in large-scale projects. To cope with this role, simulators must address additional issues related to large-scale model development, including multi-modal processing, provisions for algorithmic integration, and fast, scalable graphics. To meet those needs, we have developed ORCA, a versatile, physics-based robotics simulator that has underpinned the research efforts of numerous European and national funded projects. ORCA integrates a middleware back-end directly into the server/simulator environment, making the sharing of resources a seamless task, promoting code re-use and facilitating the integration of software modules. In the current paper, we outline the main traits of the software package, its ability to render and process large-scale environments, as well as the property to distribute computing load across available resources. To demonstrate the software's applicability, we discuss three different use-cases, in which ORCA was used at the heart of their development efforts.