Motion planning is a challenging task in mobile robot applications. The main technical difficulties include: 1) the need for handling nonlinear system dynamics and planning trajectories in a short period; 2) the requirement of allowing the mobile robot to handle all the complex constraints during operations; and 3) the need for high robustness to deal with the environment uncertainty.

In this thesis, we investigate a safety-preserving motion planning problem in a noisy environment. Since most existing results consider safety in deterministic systems, this thesis proposes a probabilistic approach to solve the mobile robot motion planning problem subject to safety constraints with Gaussian noises. We present an optimization-based control framework using model predictive control (MPC) and...[ Read more ]

Motion planning is a challenging task in mobile robot applications. The main technical difficulties include: 1) the need for handling nonlinear system dynamics and planning trajectories in a short period; 2) the requirement of allowing the mobile robot to handle all the complex constraints during operations; and 3) the need for high robustness to deal with the environment uncertainty.

In this thesis, we investigate a safety-preserving motion planning problem in a noisy environment. Since most existing results consider safety in deterministic systems, this thesis proposes a probabilistic approach to solve the mobile robot motion planning problem subject to safety constraints with Gaussian noises. We present an optimization-based control framework using model predictive control (MPC) and develop an algorithm that converges and guarantees the safety of the motion planning with on a prescribed probability threshold. Our approach is evaluated in simulations and a physical platform. Experiment results show that our control framework can preserve safety constraints and achieve optimality of the system simultaneously.