Trajectory generation and optimized control of a truck-mounted arm for manipulation tasks

Industry 4.0 will be characterized by an even increasing development and employment of robotic technologies. Their ability to work without supervision, enhanced by the potential improvements of Artificial Intelligence in environmental sensing and self-learning task solving, will allow to increase efficiency and autonomy of production lines and storage warehouse. In mobile manipulators dexterity of robotic arm is combined to mobility of ground vehicle, allowing to operate autonomously in hazardous and hostile environments. Because mobile manipulators rely on batteries for power, their movements need to be planned efficiently to conserve energy and maximize operational time. In this thesis the problem of planning trajectories for a 6-DOF manipulator in presence of an obstacle is addressed as a multi-objective optimization problem; safe energy-optimal and safe peak power-optimal trajectories are sought first, then a possible trade-off between all the three cost function is explored. Trajectory planning is performed in Cartesian space with third-order B-splines, and control points optimized leveraging a Genetic Algorithm.