skip to main content


Title: Decoupled Data-Based Approach for Learning to Control Nonlinear Dynamical Systems
This paper addresses the problem of learning the optimal control policy for a nonlinear stochastic dynam- ical. This problem is subject to the ‘curse of dimension- ality’ associated with the dynamic programming method. This paper proposes a novel decoupled data-based con- trol (D2C) algorithm that addresses this problem using a decoupled, ‘open-loop - closed-loop’, approach. First, an open-loop deterministic trajectory optimization problem is solved using a black-box simulation model of the dynamical system. Then, closed-loop control is developed around this open-loop trajectory by linearization of the dynamics about this nominal trajectory. By virtue of linearization, a linear quadratic regulator based algorithm can be used for this closed-loop control. We show that the performance of D2C algorithm is approximately optimal. Moreover, simulation performance suggests a significant reduction in training time compared to other state of the art algorithms.  more » « less
Award ID(s):
1850206
NSF-PAR ID:
10315328
Author(s) / Creator(s):
; ; ; ;
Date Published:
Journal Name:
IEEE Transactions on Automatic Control
ISSN:
0018-9286
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. This paper addresses trajectory optimization for hypersonic vehicles under atmospheric and aerodynamic uncertainties using techniques from desensitized optimal control (DOC), wherein open-loop optimal controls are obtained by minimizing the sum of the standard objective function and a first-order penalty on trajectory variations due to parametric uncertainty. The proposed approach is demonstrated via numerical simulations of a minimum-final-time Earth reentry trajectory for an X-33 vehicle with an uncertain atmospheric scale height and drag coefficient. Monte Carlo simulations indicate that dispersions in the final position footprint and the final energy can be significantly reduced without closed-loop control and with little tradeoff in the performance metric set for the trajectory. 
    more » « less
  2. This paper develops a closed-loop approach for ink-jet 3D printing. The control design is based on a distributed model predictive control scheme, which can handle constraints (such as droplet volume) as well as the large-scale nature of the problem. The high resolution of ink-jet 3D printing make centralized methods extremely time-consuming, thus a distributed implementation of the controller is developed. First a graph-based height evolution model that can capture the liquid flow dynamics is proposed. Then, a scalable closed-loop control algorithm is designed based on the model using Distributed MPC, that reduces computation time significantly. The performance and efficiency of the algorithm are shown to outperform open-loop printing and closed-loop printing with existing Centralized MPC methods through simulation results. 
    more » « less
  3. We study an optimal control problem arising from a generalization of rock-paper-scissors in which the number of strategies may be selected from any positive odd number greater than 1 and in which the payoff to the winner is controlled by a control variable \begin{document}$ \gamma $\end{document}. Using the replicator dynamics as the equations of motion, we show that a quasi-linearization of the problem admits a special optimal control form in which explicit dynamics for the controller can be identified. We show that all optimal controls must satisfy a specific second order differential equation parameterized by the number of strategies in the game. We show that as the number of strategies increases, a limiting case admits a closed form for the open-loop optimal control. In performing our analysis we show necessary conditions on an optimal control problem that allow this analytic approach to function.

     
    more » « less
  4. null (Ed.)
    Visual-inertial SLAM is essential for robot navigation in GPS-denied environments, e.g. indoor, underground. Conventionally, the performance of visual-inertial SLAM is evaluated with open-loop analysis, with a focus on the drift level of SLAM systems. In this paper, we raise the question on the importance of visual estimation latency in closed-loop navigation tasks, such as accurate trajectory tracking. To understand the impact of both drift and latency on visualinertial SLAM systems, a closed-loop benchmarking simulation is conducted, where a robot is commanded to follow a desired trajectory using the feedback from visual-inertial estimation. By extensively evaluating the trajectory tracking performance of representative state-of-the-art visual-inertial SLAM systems, we reveal the importance of latency reduction in visual estimation module of these systems. The findings suggest directions of future improvements for visual-inertial SLAM. 
    more » « less
  5. The traditional locomotion paradigm of quadruped robots is to use dexterous (multi degrees of freedom) legs and dynamically optimized footholds to balance the body and achieve stable locomotion. With the introduction of a robotic tail, a new locomotion paradigm becomes possible as the balancing is achieved by the tail and the legs are only responsible for propulsion. Since the burden on the leg is reduced, leg complexity can be also reduced. This paper explores this new paradigm by tackling the dynamic locomotion control problem of a reduced complexity quadruped (RCQ) with a pendulum tail. For this specific control task, a new control strategy is proposed in a manner that the legs are planned to execute the open-loop gait motion in advance, while the tail is controlled in a closed-loop to prepare the quadruped body in the desired orientation. With these two parts working cooperatively, the quadruped achieves dynamic locomotion. Partial feedback linearization (PFL) controller is used for the closed-loop tail control. Pronking, bounding, and maneuvering are tested to evaluate the controller’s performance. The results validate the proposed controller and demonstrate the feasibility and potential of the new locomotion paradigm. 
    more » « less