Detection of object's characteristics for mobile robot SLAM in dynamic environments

For mobile robots to be successful, they have to navigate safely in populated and dynamic environments. While recent research has provided a variety of localization methods that can track robots well in static environments, there is still a need to develop methods that can robustly localize mobile robots in dynamic environments where objects might block the robot's sensors for extensive/short periods of time as objects move in the environment with varying velocities. For a mobile agent to be able to execute its task in these scenarios, it has to be able to detect and register any changes in the environment and modify its internal map accordingly so it can be used to localize itself. This thesis proposes a simple but effective sensory system aiming at providing autonomous agents with the capability to readily classify moving or stationary obstacles in the environment. By using the proposed sensory device, the robot can monitor their surrounding area with multiple sensors and readily analyze sensory data looking for changes in the environment. Every detected object is associated with a “signature” that helps the robot to identify obstacles that might change their position. Thus, helping the robot to determine its position within the given changing terrain.