This thesis deals with the path planning of car-like mobile robots. As nowadays autonomous robots and vehicles are of great significance, it is becoming essential to develop, test and optimize path planning algorithms.
One of the major tasks of my thesis is to present the increment-dimensional path planning method for car-like mobile robots. Increment-dimensional path planning algorithm is a multi-step algorithm. This method performs holonomic path planning, then nonholonomic path planning is performed based on the output of the holonomic path planner.
This thesis presents multiple versions of RRT (Rapidly exploring Random Tree) algorithm in details. The modifications that are made are also presented with the explanation of their benefits.
The question then arises about whether it is necessary to use a multi-step algorithm?
To answer this question, I have implemented a one-step path planning algorithm, which is able to calculate the global path in one step. This thesis includes mathematical background and the structure of this method.
Last, but not least a comparison is made between the increment-dimensional path planning and the one-step path planning algorithms based on several points of view. In addition to that, simulations are made with the help of LOGITECH FORMULA™ FORCE EX steering wheel, to examine the response of the wheel, to the reference steering signal, calculated by the path planning algorithms above.