Optimization-based speed planning for mobile robots and industrial manipulators

Abstract

This thesis deals with the generation of the reference speed profile to be followed by a mobile robot or an industrial manipulator in order to complete an assigned task in minimum-time. The speed planning is a non trivial problem. Indeed, depending on the kind of robot considered and the purpose of the application, the sought speed profile has to take into account a series of limitations and constraints. The efficiency of the speed planning algorithms is a mandatory feature when the robot has to operate in dynamic environments in which the motion replanning can be frequently requested. This thesis proposes algorithms which exploit the structure of the finite dimensional formulations of the speed planning problem in order to get the optimal solution with the best possible time complexity and trying as much as possible to avoid using external solvers. Numerical tests show that the proposed algorithms, which in some cases have the best possible computational complexity, are significantly faster than algorithms already existing in literature.