AN AUTONOMOUS SIMULATION BASED SYSTEM FOR ROBOTIC SERVICES IN PARTIALLY KNOWN ENVIRONMENTS

Abstract

This paper is focused on presenting the architecture and the implementation of a semi autonomous simulation based system which is able to navigate into a partially known environment. Most of robotic agent does not support mobility according the requirements of applications. Our implementation provides a semi autonomous system which is used by an operator performing several services with better quality and low costs. The design is a module based architecture which supports the robot navigation using simulation offline software and on line at the moment when the agency percepts obstacles. The robot changes states of its mobility in real time building a new strategy to achieve the normal path received from a simulator that execute and communicate the path calculated by a simple navigational algorithm in the virtual static environment. Using a communication protocol between robotic unit and simulator software, it is possible to correct data and to improve the system behavior through a wireless communication. The physical system is tested in a laboratory environment. It is situated in a environment which is the same designed in the simulator. The robot updates its coordinates in the virtual environment and the simulation runs exactly according the navigation algorithms until the sensor module transmits to simulation software new data of obstacle presence. We evaluate the performance of the system and the results confirm an improved behavior of robotic agent in extreme situations of dynamic environment.