This paper presents three novel model reference adaptive control (MRAC) systems for nonlinear plants in which the matched, nonparametric uncertainty is only known to reside in a vector-valued reproducing kernel Hilbert space (RKHS). The first MRAC system is based on an extension of the classical projection operator for robust MRAC systems and assures ultimate boundedness of the trajectory tracking error. This MRAC system allows the user to design part of the control input, known as the compensator. The second MRAC method modifies the previous one and, based on an estimate of the largest admissible uncertainty in the RKHS norm, employs a particular choice of the compensator to assure asymptotic convergence of the tracking error to zero. The last MRAC method uses the previous one and, leveraging the error bounding method framework, employs an additional learning law to assure asymptotic convergence of the tracking error without any information on the largest admissible uncertainty. While the conventional approaches to adaptive error bounding methods usually leave it to an analyst to derive the required error bounds on a case-by-case basis, the proposed approaches work for any functional uncertainty that is contained in any native space of vector-valued functions defined in terms of an operator kernel.
This paper describes the formulation and experimental testing of a novel method for the estimation and approximation of submanifold models of animal motion. It is assumed that the animal motion is supported on a configuration manifold $Q$ that is a smooth, connected, regularly embedded Riemannian submanifold of Euclidean space $X\approx \mathbb{R}^d$ for some $d>0$, and that the manifold $Q$ is homeomorphic to a known smooth, Riemannian manifold $S$. Estimation of the manifold is achieved by finding an unknown mapping $\gamma:S\rightarrow Q\subset X$ that maps the manifold $S$ into $Q$. The overall problem is cast as a distribution-free learning problem over the manifold of measurements $\mathbb{Z}=S\times X$. That is, it is assumed that experiments generate a finite sets $\{(s_i,x_i)\}_{i=1}^m\subset \mathbb{Z}^m$ of samples that are generated according to an unknown probability density $\mu$ on $\mathbb{Z}$. This paper derives approximations $\gamma_{n,m}$ of $\gamma$ that are based on the $m$ samples and are contained in an $N(n)$ dimensional space of approximants. The paper defines sufficient conditions that shows that the rates of convergence in $L^2_\mu(S)$ correspond to those known for classical distribution-free learning theory over Euclidean space. Specifically, the paper derives sufficient conditions that guarantee rates of convergence that have the form $$\mathbb{E} \left (\|\gamma_\mu^j-\gamma_{n,m}^j\|_{L^2_\mu(S)}^2\right )\leq C_1 N(n)^{-r} + C_2 \frac{N(n)\log(N(n))}{m}$$for constants $C_1,C_2$ with $\gamma_\mu:=\{\gamma^1_\mu,\ldots,\gamma^d_\mu\}$ the regressor function $\gamma_\mu:S\rightarrow Q\subset X$ and $\gamma_{n,m}:=\{\gamma^1_{n,j},\ldots,\gamma^d_{n,m}\}$.
31 October 2018 in Journal of Dynamic Systems, Measurement, and Control
Bioinspired design of robotic systems can offer many potential advantages in comparison to traditional architectures including improved adaptability, maneuverability, or efficiency. Substantial progress has been made in the design and fabrication of bioinspired systems. While many of these systems are bioinspired at a system architecture level, the design of linkage connections often assumes that motion is well approximated by ideal joints subject to designer-specified box constraints. However, such constraints can allow a robot to achieve unnatural and potentially unstable configurations. In contrast, this paper develops a methodology, which identifies the set of admissible configurations from experimental observations and optimizes a compliant structure around the joint such that motions evolve on or close to the observed configuration set. This approach formulates an analytical-empirical (AE) potential energy field, which “pushes” system trajectories toward the set of observations. Then, the strain energy of a compliant structure is optimized to approximate this energy field. While our approach requires that kinematics of a joint be specified by a designer, the optimized compliant structure enforces constraints on joint motion without requiring an explicit definition of box-constraints. To validate our approach, we construct a single degree-of-freedom elbow joint, which closely matches the AE and optimal potential energy functions and admissible motions remain within the observation set.