Adaptive Model-Predictive Motion Planning for Navigation in Complex Environments

摘要

Outdoor mobile robot motion planning and navigation is a challenging problem in artificial intelligence. The search space density and dimensionality, system dynamics and environmental interaction complexity, and the perceptual horizon limitation all contribute to the difficultly of this problem. It is hard to generate a motion plan between arbitrary boundary states that considers sophisticated vehicle dynamics and all feasible actions for nontrivial mobile robot systems. Accomplishing these goals in real time is even more challenging because of dynamic environments and updating perception information. This thesis develops effective search spaces for mobile robot trajectory generation, motion planning, and navigation in complex environments. Complex environments are defined as worlds where locally optimal motion plans are numerous and where the sensitivity of the cost function is highly dependent on state and motion model fidelity. Examples include domains where obstacles are prevalent, terrain shape is varied, and the consideration of terramechanical effects is important.