Prioritized optimal control: A hierarchical differential dynamic programming approach

摘要: This paper deals with the generation of motion for complex dynamical systems (such as humanoid robots) to achieve several concurrent objectives. Hierarchy tasks and optimal control are two frameworks commonly used this aim. The first one specifies objectives a number quadratic functions be minimized under strict priorities. second minimizes an arbitrary user-defined function future state system, thus considering its evolution in time. Our recent work on prioritized merges advantages both these methods. reformulates original algorithm precise goal improving computational speed. We extend dynamic programming method hierarchy tasks. compared our approach simulation previous classical control. measured improvement represents another step towards application online model predictive robots. believe that could key unlock (so far unexploited) capabilities mechanical systems.