Environment Classification Approach for Mobile Robots

The type of environment plays an essential role in mobile robotics. Autonomous robots usually work in static and single type environments. Unfortunately, in the real world, there are many situations when the type of the environment changes. This paper deals with the design of the system for environment recognition working in real time on a mobile robot. Once the environment is recognized and classified, the parameters of the robot can be adapted. Consequently, the robot can handle operations in multiple environments. The paper contains information about previous-related work in the environment classification, description of the proposed system, and experiments. Experiments are focused on change detection and environment classification using visual and non-visual sensors. Moreover, two non-visual change detection approaches are proposed in the experiment section.