The topic of this MSc thesis is the path planning and trajectory control of autonomous car-like mobile robots. The goal is to examine the existing roadmap-based path planning methods and implement a solution for trajectory control. An autonomous robot must be able to find out its position and orientation in the workspace, distinguish obstacles from free workspace, and reach the goal configuration along a collision-free path, without human interaction.
As far as the path planning is concerned the common kinematic model of car-like mobile robots is described alongside the continuous curvature paths that the car-like robots can follow. Next some of the existing roadmap-based multi-query global planners are briefly investigated. An implementation of a Probabilistic Roadmap and a simple local planner for continuous curvature paths is shown. The path planning is illustrated by simulation results.
In regards of the trajectory control the implementation of the necessary components of the closed loop control are described. First a control algorithm is implemented and simulated. Fort the state estimation of the robot a camera has been used. The camera is positioned above the workspace. The segmentation of the free region and the obstacles is done by processing the top view image of the workspace. The description of the used equipment is followed by the tests of the trajectory control, and the thesis ends with a summary of further development possibilities.