Minimizing attention in a motion control context

In this paper we motivate the need for an algorithmic solution to the problem of segmenting and representing vector-valued functions of time with the goal of identifying a set of standard parts which can be used as the basis for a efficient description of an entire family. We motivate the use of such an approach in a feedback control context where one can use such efficient descriptions together with concatenation and scaling operations to generate quite general trajectories. Our results include a procedure for jointly specifying feedback gains and open-loop controls leading to the definition of efficient motion control languages.