Non-monotonic Mapping for Mobile Robots in Dynamic Environments

Aim: To develop effective methods for mobile robots to build maps in environments containing moving objects. Significance: For many applications, a mobile robot must be able to develop and maintain a map of its environment autonomously. However, existing map building techniques assume that the environment surrounding the robot is static. This assumption cannot be justified in the real world where people, other robots and furniture all change position over time. Outcomes: This project will develop mapping methods for objects that change position occasionally (eg furniture), methods for detecting and mapping dynamic objects (eg people) and methods for simultaneously estimating the robot position and mapping the environment.