| This project studies the problem of motion planning for mobile robots. The aim is to define an obstacle-free path for a mobile robot which will consider its geometry with the goal to give it a greater autonomy in the planning of movements on long distances in a partially known environment. Starting from a first base path defined by the original long distance planner of the robot and a polygonal representation of the environment based on a 3D irregular triangular mesh, created from the information given by a three-dimensional laser scanner placed on the robot, a new obstacle free path is generated considering the robot geometry and eliminating as much as possible the saw tooth present in the base path. An algorithm based on a cost function makes possible to define this path, minimizing the collision risk, the distance to be traveled and the strong changes of orientation of the robot.; Different cases are studied in order to establish the performance of different approaches. |
