The main focus of this thesis is on the motion planning and control of mobile robots in dynamic unstructured environments in which the primary challenge is to formulate and deal with uncertainty. This thesis contributes to the motion planning problem in three distinct yet related aspects that can together present a model predictive approach to enhance autonomy of mobile robots in dynamic unknown environments. The first contribution of this thesis is to introduce a robust yet probing control algorithm. The proposed algorithm is based on the output-feedback tube-based model predictive control (MPC). The performance of the algorithm has been enhanced using the partially-closed loop strategy. The tube-based approach requires uncertainties to be modeled in the set-theoretic framework, whereas the partially closed-loop strategy is modeled in the probabilistic framework. A key component of the algorithm is related to proposing the relationship between these two different paradigms. The proposed framework utilizes the uncertainty fusion in the probabilistic framework and collision avoidance in the set-theoretic framework. The efficiency of the proposed algorithm is verified using thorough numerical simulations and experiments. The second contribution of this thesis is in regards with linearization of stochastic nonlinear systems. A statistical linearization method, unscented transform, is proposed to replace the analytical linearization method in MPC. The advantage and disadvantage of such replacement has been examined through extensive numerical simulations. The numerical simulation indicates that statistical linearization has two important advantages. First, the proposed approach is derivative free that is it can be applied to complex systems for which no analytical model exists. Second, it is more accurate so that it enhances performance of the planning algorithm. However, the tradeoff is that the analytical linearization is computationally less expensive. The third contribution of this thesis is related to the formulation of the robust tube-based MPC scheme for incremental smoothing and mapping known as active iSAM problem in the literature. In addition to utilizing a robust MPC scheme, the active iSAM utilizes the optimization-based method, iSAM, to solve the simultaneous localization and mapping SLAM problem. Extensive numerical simulations have been conducted to verify the performance of the algorithm. ; Applied Science, Faculty of ; Engineering, School of (Okanagan) ; Graduate
MPC- based mobile robots motion planning and control in uncertain dynamic environment
2015-01-01
Theses
Electronic Resource
English
DDC: | 629 |
Motion planning of mobile robots
IEEE | 2004
|Motion Planning for Mobile Robots in a Time-Varying Environment
British Library Online Contents | 1996
|Trajectory Planning for Mobile Robots in a Dynamic Environment
British Library Conference Proceedings | 1993
|Dynamic Motion Planning for Mobile Robots Using Potential Field Method
British Library Online Contents | 2002
|