Nalezení optimální kinematické struktury robotického manipulátoru pro danou úlohu

Abstract

This dissertation focuses on the development of methods for the design and optimization of the kinematic structure of robotic manipulators tailored to the specific requirements of a given task. The main objective is to establish a comprehensive methodology that enables the generation of a suitable manipulator configuration ensuring reachability of the target poses while respecting mechanical constraints. The solution also includes optimization of the base position within the workspace. The proposed approach integrates kinematic structure design with motion planning in a unified optimization framework. Candidate configurations are generated using a genetic algorithm and evaluated for their ability to reach the defined poses in an environment containing obstacles. The evaluation considers not only reachability but also possible collisions between robot links and external objects. The manipulator geometry is modeled using a capsule representation, which allows for efficient collision detection. Path planning between the initial and target configurations is performed using the Rapidlyexploring Random Tree (RRT) algorithm. The resulting discrete path is subsequently simplified and used as the basis for generating a continuous trajectory. Trajectory planning is formulated as a multi-objective optimization problem that considers both total motion time and motion smoothness, expressed by minimizing jerk. The result is a time-parametrized profile of joint variables that satisfies predefined limits on position, velocity, and acceleration. The time profile is generated using quintic polynomials with boundary conditions applied at all waypoints. The proposed methodology links structural and motion planning aspects into a single coherent process. It is validated through a set of experimental scenarios involving motion between two defined poses in an environment with fixed obstacles. The results demonstrate that the approach provides feasible and efficient solutions without the need for manual intervention in the design or planning stages.