Risk-Aware Fixed-Time Stabilization of Stochastic Systems under Measurement Uncertainty
2403.20258
0
0
🏅
Abstract
This paper addresses the problem of risk-aware fixed-time stabilization of a class of uncertain, output-feedback nonlinear systems modeled via stochastic differential equations. First, novel classes of certificate functions, namely risk-aware fixed-time- and risk-aware path-integral-control Lyapunov functions, are introduced. Then, it is shown how the use of either for control design certifies that a system is both stable in probability and probabilistically fixed-time convergent (for a given probability) to a goal set. That is, the system trajectories probabilistically reach the set within a finite time, independent of the initial condition, despite the additional presence of measurement noise. These methods represent an improvement over the state-of-the-art in stochastic fixed-time stabilization, which presently offers bounds on the settling-time function in expectation only. The theoretical results are verified by an empirical study on an illustrative, stochastic, nonlinear system and the proposed controllers are evaluated against an existing method. Finally, the methods are demonstrated via a simulated fixed-wing aerial robot on a reach-avoid scenario to highlight their ability to certify the probability that a system safely reaches its goal.
Create account to get full access
Introduction
The paper discusses the importance of certifying the stability and convergence properties of complex dynamical systems, particularly when modeling the systems as stochastic differential equations (SDEs) to account for uncertain behavior. Traditional control Lyapunov functions (CLFs) have been used to guarantee asymptotic stability or convergence, but they only provide guarantees in the limit as time tends to infinity. Many real-world systems, however, operate over finite time intervals or have temporal and spatiotemporal constraints, requiring stronger notions of finite-time stability (FTS) or fixed-time stability (FxTS).
The paper makes three key contributions:
-
It introduces novel classes of risk-aware fixed-time CLFs (RA-FxT-CLFs) and risk-aware path-integral CLFs (RA-PI-CLFs) for the fixed-time stabilization of stochastic, nonlinear systems with stochastic measurement noise.
-
It proves that the use of RA-FxT-CLFs or RA-PI-CLFs for control design certifies that the associated goal set is probabilistically FxTS with a specified probability.
-
It demonstrates the efficacy of RA-FxT-CLF- and RA-PI-CLF-based control laws in a comparative study against a control law derived from stochastic FxT-CLFs on a nonlinear system.
The paper aims to address the open problem of characterizing the level of associated risk, i.e., the probability of failing to reach the equilibrium or goal set, in stochastic control design for systems with finite-time or fixed-time constraints.
Preliminaries and Problem Formulation
The paper discusses a class of nonlinear, stochastic systems described by stochastic differential equations. Several mathematical preliminaries are provided, including definitions of standard distributions, vector/matrix spaces, and the Gauss error function.
The main focus is on stabilizing the trajectories of the nonlinear stochastic system to a goal set around the origin. A state observer is introduced to estimate the system state, and an estimate-feedback controller is designed. The notions of finite-time and fixed-time stability in probability are formally defined. The problem is then stated as designing an estimate-feedback controller such that the goal set is rendered fixed-time stable with a specified probability.
Risk-Aware Fixed-Time Stabilization
This section introduces two approaches to risk-aware fixed-time stabilization: the risk-aware fixed-time control Lyapunov function (RA-FxT-CLF) and the risk-aware path integral control Lyapunov function (RA-PI-CLF). The use of either will render a goal set fixed-time stable with probability pg*.
For the RA-FxT-CLF, the paper shows that if certain conditions are met, the goal set is rendered pg-FxTS, meaning it is stable in probability and locally finite-time attractive with probability pg and uniformly bounded settling time.
For the RA-PI-CLF, the paper shows that if the conditions are met, the goal set is rendered pg*-FxTS, meaning it is stable in probability and locally finite-time attractive with probability pg* and uniformly bounded settling time.
The paper then introduces an optimization-based control law that can use either the RA-FxT-CLF or RA-PI-CLF to filter a nominal control policy in real-time. This control law seeks to minimize the deviation from the nominal input while satisfying the risk-aware stabilization conditions.
V Numerical Case Studies
This section demonstrates the use of the proposed risk-aware fixed-time control Lyapunov function (RA-FxT-CLF) and risk-aware prescribed input control Lyapunov function (RA-PI-CLF) controllers through two numerical examples.
The first example is an empirical study on a 2D nonlinear system. The system was simulated under the proposed RA-FxT-CLF and RA-PI-CLF controllers, as well as a stochastic FxT-CLF (S-FxT-CLF) controller. The empirical results show that the proposed controllers satisfy their theoretical probabilistic fixed-time stability bounds in all trials. Decreasing the tolerable risk of failure generally leads to faster convergence, suggesting the controllers with stronger risk aversion seek their goals more aggressively. Performance also decreases in the presence of imperfect measurements, though the proposed methods still outperform the S-FxT-CLF controller on average.
The second example illustrates the application of the RA-FxT-CLF controller to a fixed-wing UAV kinematic model, where the objective is to reach a goal set while avoiding unsafe ellipsoid obstacles. The results show the controller successfully steers the UAV to the goal while avoiding the obstacles.
Overall, the section demonstrates the efficacy of the proposed RA-FxT-CLF and RA-PI-CLF controllers through these numerical studies, highlighting their ability to provide risk-aware control with probabilistic guarantees.
Conclusion
This paper proposes two Lyapunov-based methods for risk-aware fixed-time stabilization of a class of stochastic, nonlinear systems with measurement uncertainty. The methods are shown to render a goal set probabilistically fixed-time stable with a specified probability. An empirical study on a nonlinear system and a fixed-wing UAV example validate the proposed approaches in simulation. Future work aims to reduce conservatism, combine the methods with risk-aware barrier functions, and investigate their use for a priori system verification.
Appendix A Proof of Lemma 1
The proof follows from a previous result in Section 3 of the referenced paper [27]. The cumulative distribution function (CDF) of the supremum of a scalar process {x_t: t ∈ [0,∞)} over an interval [0,T] is given by:
F_x(a,T) = P(sup_{0≤t≤T} x(t) < a) = ∫_{T}^∞ q_a(θ|x_0) dθ
where q_a is the first passage-time probability density function with respect to level a > 0.
For a standard Wiener process {w_t: t ∈ [0,∞)}, the CDF of the supremum is:
F_w(a,T) = sqrt(2/π) ∫_{0}^{a/sqrt(T)} e^{-s^2/2} ds = erf(a/sqrt(2T))
This provides the closed-form expression for the CDF of the supremum of a standard Wiener process over the interval [0,T].
Appendix B Proof of Lemma 2
The provided text demonstrates steps to derive an expression for the expected value of the square of the integral of a vector-valued stochastic process over a time interval. It starts by invoking a result inspired by a previous lemma, using the vector dot product. It then shows that the expected value of the square of the integral can be expressed as the sum of the expected values of the squares of the integrals of the individual components of the vector process. This result follows from the independence of the increments of the Wiener processes corresponding to the different vector components, and the Ito isometry property. The final expression obtained is the expected value of the sum of the integrals of the squares of the vector components over the time interval.
This summary was produced with help from an AI and may contain inaccuracies - check out the links to read the original source documents!
Related Papers
A Stability-Based Abstraction Framework for Reach-Avoid Control of Stochastic Dynamical Systems with Unknown Noise Distributions
Thom Badings, Licio Romao, Alessandro Abate, Nils Jansen
0
0
Finite-state abstractions are widely studied for the automated synthesis of correct-by-construction controllers for stochastic dynamical systems. However, existing abstraction methods often lead to prohibitively large finite-state models. To address this issue, we propose a novel abstraction scheme for stochastic linear systems that exploits the system's stability to obtain significantly smaller abstract models. As a unique feature, we first stabilize the open-loop dynamics using a linear feedback gain. We then use a model-based approach to abstract a known part of the stabilized dynamics while using a data-driven method to account for the stochastic uncertainty. We formalize abstractions as Markov decision processes (MDPs) with intervals of transition probabilities. By stabilizing the dynamics, we can further constrain the control input modeled in the abstraction, which leads to smaller abstract models while retaining the correctness of controllers. Moreover, when the stabilizing feedback controller is aligned with the property of interest, then a good trade-off is achieved between the reduction in the abstraction size and the performance loss. The experiments show that our approach can reduce the size of the graph of abstractions by up to 90% with negligible performance loss.
4/3/2024
Distributionally Robust Policy and Lyapunov-Certificate Learning
Kehan Long, Jorge Cortes, Nikolay Atanasov
0
0
This article presents novel methods for synthesizing distributionally robust stabilizing neural controllers and certificates for control systems under model uncertainty. A key challenge in designing controllers with stability guarantees for uncertain systems is the accurate determination of and adaptation to shifts in model parametric uncertainty during online deployment. We tackle this with a novel distributionally robust formulation of the Lyapunov derivative chance constraint ensuring a monotonic decrease of the Lyapunov certificate. To avoid the computational complexity involved in dealing with the space of probability measures, we identify a sufficient condition in the form of deterministic convex constraints that ensures the Lyapunov derivative constraint is satisfied. We integrate this condition into a loss function for training a neural network-based controller and show that, for the resulting closed-loop system, the global asymptotic stability of its equilibrium can be certified with high confidence, even with Out-of-Distribution (OoD) model uncertainties. To demonstrate the efficacy and efficiency of the proposed methodology, we compare it with an uncertainty-agnostic baseline approach and several reinforcement learning approaches in two control problems in simulation.
4/8/2024
🔗
Myopically Verifiable Probabilistic Certificates for Safe Control and Learning
Zhuoyuan Wang, Haoming Jing, Christian Kurniawan, Albert Chern, Yorie Nakahira
0
0
This paper addresses the design of safety certificates for stochastic systems, with a focus on ensuring long-term safety through fast real-time control. In stochastic environments, set invariance-based methods that restrict the probability of risk events in infinitesimal time intervals may exhibit significant long-term risks due to cumulative uncertainties/risks. On the other hand, reachability-based approaches that account for the long-term future may require prohibitive computation in real-time decision making. To overcome this challenge involving stringent long-term safety vs. computation tradeoffs, we first introduce a novel technique termed `probabilistic invariance'. This technique characterizes the invariance conditions of the probability of interest. When the target probability is defined using long-term trajectories, this technique can be used to design myopic conditions/controllers with assured long-term safe probability. Then, we integrate this technique into safe control and learning. The proposed control methods efficiently assure long-term safety using neural networks or model predictive controllers with short outlook horizons. The proposed learning methods can be used to guarantee long-term safety during and after training. Finally, we demonstrate the performance of the proposed techniques in numerical simulations.
4/29/2024
🚀
Learning-Based Optimal Control with Performance Guarantees for Unknown Systems with Latent States
Robert Lefringhausen, Supitsana Srithasan, Armin Lederer, Sandra Hirche
0
0
As control engineering methods are applied to increasingly complex systems, data-driven approaches for system identification appear as a promising alternative to physics-based modeling. While the Bayesian approaches prevalent for safety-critical applications usually rely on the availability of state measurements, the states of a complex system are often not directly measurable. It may then be necessary to jointly estimate the dynamics and the latent state, making the quantification of uncertainties and the design of controllers with formal performance guarantees considerably more challenging. This paper proposes a novel method for the computation of an optimal input trajectory for unknown nonlinear systems with latent states based on a combination of particle Markov chain Monte Carlo methods and scenario theory. Probabilistic performance guarantees are derived for the resulting input trajectory, and an approach to validate the performance of arbitrary control laws is presented. The effectiveness of the proposed method is demonstrated in a numerical simulation.
4/17/2024