A framework for migrating control strategies and algorithms developed in virtual robots to physical systems

Abstract

Virtual robotic environments are commonly used to develop, prototype and simulate control strategies and algorithms for single or multi-robot systems. Once developed, the control strategies and algorithms can then be implemented on physical systems. However, the transition from virtual to physical systems can be rather tedious due to different programming languages, application programming interfaces, and incompatibilities between the virtual sensor and robot models, and the physical system. Transitioning the implemented control strategies and algorithms to other robot platforms may also require cumbersome recoding of the initial implementation. This paper introduces the use of Mobile-R, a mobile agent-based robot control and deployment platform for virtual and physical systems, for developing and migrating virtual control strategies and algorithms to physical systems. Mobile-R provides a common set of programming and control abstractions, and middleware for both virtual and physical robots alleviating the tedious conversion process from virtual to physical system implementation. A real-world validation experiment involving a virtual and a physical Pioneer 3DX mobile robot is presented.