From human to humanoid locomotion—an inverse optimal control approach

Abstract

The purpose of this paper is to present inverse optimal control as a promising approach to transfer biological motions to robots. Inverse optimal control helps (a) to understand and identify the underlying optimality criteria of biological motions based on measurements, and (b) to establish optimal control models that can be used to control robot motion. The aim of inverse optimal control problems is to determine-for a given dynamic process and an observed solution-the optimization criterion that has produced the solution. Inverse optimal control problems are difficult from a mathematical point of view, since they require to solve a parameter identification problem inside an optimal control problem. We propose a pragmatic new bilevel approach to solve inverse optimal control problems which rests on two pillars: an efficient direct multiple shooting technique to handle optimal control problems, and a state-of-the art derivative free trust region optimization technique to guarantee a match between optimal control problem solution and measurements. In this paper, we apply inverse optimal control to establish a model of human overall locomotion path generation to given target positions and orientations, based on newly collected motion capture data. It is shown how the optimal control model can be implemented on the humanoid robot HRP-2 and thus enable it to autonomously generate natural locomotion paths.

Investigating optimal technique in the presence of motor system noise: application to the double layout somersault dismount on high bar

“…If this had been the case an inverse optimal control approach could have been used to find the appropriate parameters and weightings which would produce a close match to the recorded performances (Mombaur, Truong & Laumond, 2010). When such an approach is applied to recreating human movement patterns, components of the criterion are often related to producing smooth movements, which are a feature of human movement (Mombaur et al, 2010). Components within optimisation criteria which minimise joint torque, torque change, or jerk are often required to cope with inadequate or absent representations of human muscle and its innervation.…”

Why Don't We Move Slower? The Value of Time in the Neural Control of Action

“…To get persuaded of this, it suffices to realize that a model predicting MT accurately would be meaningless if it does not predict the correct system trajectories in fixed time as well. In what follows, we shall assume that l is either known from the literature or can be identified via some inverse OC procedure performed in fixed time (e.g., Mombaur et al, 2010;Berret et al, 2011). The function g is the infinitesimal (i.e., instantaneous) CoT we are looking for, whose antiderivative is the actual CoT and will be denoted by G͑t͒ Ϫ G͑0͒ ϭ ͵ 0 t g͑s͒ds (in the sequel, we will assume that G(0) ϭ 0).…”

On Inverse Optimal Control Problems of Human Locomotion: Stability and Robustness of the Minimizers

“…From a practical point of view, it amounts to find in L a cost whose minimizing trajectories fit accurately experimental data (see Figure 2). To this purpose Mombaur et al [14] proposed a purely numerical approach based on parameter identification that is used to implement human-like motion on humanoid robots. This method furnishes a plausible solution but does not give further insight into the considered inverse optimal control problem.…”