Artificial neural networks-based simulation of obstacle detection with a mobile robot in a virtual environment

Abstract

Mobile robot navigation is primarily a task that occurs in a real environment. However, simulating obstacles and robot movements in a virtual environment can provide significant advantages and yield good results, as demonstrated in this paper. By employing artificial neural networks (ANNs), it is possible to develop a trained system in a virtual environment that can detect obstacles using data collected from various sensors. In this study, infrared (IR) sensors and a camera were utilized to gather information from the virtual environment. The MatLab Simulink software package was used as a tool to train the artificial neural networks. Detection and avoidance of obstacles were simulated in the RobotinoSIM virtual environment.