As the use of autonomous robotic vehicles is increasing in everyday applications, issues regarding the performance of these vehicles become important research problems. The main performance concerns are optimization issues, such as travel time, length and energy expenditure.
This thesis studied the time optimal navigation problem of an autonomous mobile robot, navigating in a planar environment containing a single polygonal obstacle. The robot is subjected to state-dependent safety constraints that prevent an inevitable collision with the obstacle during navigation. The mobile robot is also subject to a constraint on its minimal turning radius. While being subjected to these constraints, the goal of the mobile robot is to navigate from a given start point and heading to a target point, which can possibly be located on the obstacle boundary (a docking problem). The thesis obtained analytical and numerical results concerning the safe time optimal path. Specifically, the safe time optimal path consists of three types of path primitives: minimum radius of turn circular arcs, full linear acceleration, and generalized cycloid curves which represent navigation at speeds on the verge of inevitable collision with the obstacle. Extensive numerical examples will demonstrate the results.
