Convergence of a WMR to the source of unknown signal distribution in a cluttered environment

Many mechanical systems that suffer from non-holonomic constraints have attracted a lot of attention recently due to the theorem of Brockett. One such system is a wheeled mobile robot (WMR). Equipped with a proper sensors the robot has an advantage over a human since it can easily maneuver in dangerous environments, and could be rapidly modified or improved to fulfill any specific tasks. Thereby, it can be used to detect and track radioactive, chemical, thermal or any other signal in order to localize and approach its source. This work presents an extremum seeking controller designed with the main aim to safely drive the vehicle towards the source of radiated signal with no prior knowledge about the environment and analytical structure of nonlinear signal distribution. Sliding mode controllers are employed to control WMR during the obstacle free mode, and also while maneuvering within a cluttered environment. To assure the timely convergence to the source, Kalman filter is used to predict the future trajectory evolution over sufficiently long time interval.