Learning Stabilizable Dynamical Systems via Control Contraction Metrics

@article{Singh2019LearningSD,
  title={Learning Stabilizable Dynamical Systems via Control Contraction Metrics},
  author={Sumeet Singh and Vikas Sindhwani and Jean-Jacques E. Slotine and Marco Pavone},
  journal={ArXiv},
  year={2019},
  volume={abs/1808.00113}
}
We propose a novel framework for learning stabilizable nonlinear dynamical systems for continuous control tasks in robotics. [...] Key Method By leveraging tools from contraction theory, statistical learning, and convex optimization, we provide a general and tractable semi-supervised algorithm to learn stabilizable dynamics, which can be applied to complex underactuated systems. We validated the proposed algorithm on a simulated planar quadrotor system and observed notably improved trajectory generation and…Expand
Learning Stabilizable Dynamical Systems via Control Contraction Metrics
TLDR
The key idea is to develop a new control-theoretic regularizer for dynamics fitting rooted in the notion of stabilizability, which guarantees that the learned system can be accompanied by a robust controller capable of stabilizing any open-loop trajectory that the system may generate. Expand
Learning stabilizable nonlinear dynamics with contraction-based regularization
TLDR
The results support the conjecture that the use of stabilizability constraints as a form of regularization can help prune the hypothesis space in a manner that is tailored to the downstream task of trajectory generation and feedback control, resulting in models that are not only dramatically better conditioned, but also data efficient. Expand
Learning-based Robust Motion Planning With Guaranteed Stability: A Contraction Theory Approach
This letter presents Learning-based Autonomous Guidance with RObustness and Stability guarantees (LAG-ROS), which provides machine learning-based nonlinear motion planners with formal robustness andExpand
Universal Adaptive Control for Uncertain Nonlinear Systems
TLDR
This work presents an adaptive control framework for nonlinear systems with unmatched uncertainties that addresses several of the limitations of existing methods through two key innovations, leveraging contraction theory and a new type of contraction metric that is able to track feasible trajectories generated by an adapting reference model. Expand
Contraction Metrics in Adaptive Nonlinear Control
TLDR
This work uses contraction metrics to derive an adaptive controller for stabilizable nonlinear systems by constructing a distance-like function differentially rather than explicitly, and can be applied to a larger class of non linear systems as a result. Expand
Neural Stochastic Contraction Metrics for Learning-based Robust Control and Estimation
TLDR
The NSCM framework allows autonomous agents to approximate optimal stable control and estimation policies in real-time, and outperforms existing nonlinear control and estimating techniques including the state-dependent Riccati equation, iterative LQR, EKF, and the deterministic neural contraction metric, as illustrated in simulation results. Expand
Adaptive Nonlinear Control With Contraction Metrics
TLDR
This letter derives direct adaptive control algorithms for nonlinear systems nominally contracting in closed-loop, but subject to structured parametric uncertainty from methods based on feedback linearization or backstepping. Expand
Fitting a Linear Control Policy to Demonstrations with a Kalman Constraint
TLDR
A heuristic method, based on the alternating direction method of multipliers (ADMM), is proposed, to approximately solve the problem of learning a linear control policy for a linear dynamical system from demonstrations of an expert regulating the system. Expand
Learning Dynamical Systems with Side Information
TLDR
The overall learning methodology combines ideas from convex optimization, real algebra, dynamical systems, and functional approximation theory, and can potentially lead to new synergies between these areas. Expand
...
1
2
...

References

SHOWING 1-10 OF 42 REFERENCES
Learning Stabilizable Dynamical Systems via Control Contraction Metrics
TLDR
The key idea is to develop a new control-theoretic regularizer for dynamics fitting rooted in the notion of stabilizability, which guarantees that the learned system can be accompanied by a robust controller capable of stabilizing any open-loop trajectory that the system may generate. Expand
Safe Model-based Reinforcement Learning with Stability Guarantees
TLDR
This paper presents a learning algorithm that explicitly considers safety, defined in terms of stability guarantees, and extends control-theoretic results on Lyapunov stability verification and shows how to use statistical models of the dynamics to obtain high-performance control policies with provable stability certificates. Expand
Improved Learning of Dynamics Models for Control
TLDR
This paper presents an extension to Data As Demonstrator for handling controlled dynamics in order to improve the multiple-step prediction capabilities of the learned dynamics models. Expand
Gaussian Networks for Direct Adaptive Control
A direct adaptive tracking control architecture is proposed and evaluated for a class of continuous-time nonlinear dynamic systems for which an explicit linear parameterization of the uncertainty inExpand
Robust online motion planning via contraction theory and convex optimization
TLDR
This work presents a framework for online generation of robust motion plans for robotic systems with nonlinear dynamics subject to bounded disturbances, control constraints, and online state constraints such as obstacles and demonstrates the approach through simulations of a 6-state planar quadrotor navigating cluttered environments in the presence of a cross-wind. Expand
Control Contraction Metrics: Convex and Intrinsic Criteria for Nonlinear Feedback Design
TLDR
It is shown that sufficient conditions for exponential stabilizability of all trajectories of a nonlinear control system are necessary and sufficient for feedback linearizable systems and also derive novel convex criteria for exponential stabilization of anonlinear submanifold of state space. Expand
Learning potential functions from human demonstrations with encapsulated dynamic and compliant behaviors
TLDR
The proposed controller can be realized as a unification approach between “realtime motion generation” and “variable impedance control”, with the advantages that it has guaranteed stability as well as does not rely on following a reference trajectory. Expand
Neural learning of vector fields for encoding stable dynamical systems
TLDR
A novel learning scheme that represents vector fields by means of neural networks, where asymptotic stability of the induced dynamics is explicitly enforced through utilizing knowledge from Lyapunov stability theory, in a predefined workspace leads to robust motion generation. Expand
Robust Control Contraction Metrics: A Convex Approach to Nonlinear State-Feedback ${H}^\infty$ Control
TLDR
RCCM is a Riemannian metric that verifies differential L^{2} -gain bounds in closed-loop, and guarantees robust stability of arbitrary trajectories via small gain arguments, and can be transformed to a convex optimization problem. Expand
Learning Stable Nonlinear Dynamical Systems With Gaussian Mixture Models
TLDR
A learning method is proposed, which is called Stable Estimator of Dynamical Systems (SEDS), to learn the parameters of the DS to ensure that all motions closely follow the demonstrations while ultimately reaching and stopping at the target. Expand
...
1
2
3
4
5
...