Capturing uncertainty in models of complex dynamical systems is crucial to designing safe controllers. Stochastic noise causes aleatoric uncertainty, whereas imprecise knowledge of model parameters leads to epistemic uncertainty. Several approaches use formal abstractions to synthesize policies that satisfy temporal specifications related to safety and reachability. However, the underlying models exclusively capture aleatoric but not epistemic uncertainty, and thus require that model parameters are known precisely. Our contribution to overcoming this restriction is a novel abstraction-based controller synthesis method for continuous-state models with stochastic noise and uncertain parameters. By sampling techniques and robust analysis, we capture both aleatoric and epistemic uncertainty, with a user-specified confidence level, in the transition probability intervals of a so-called interval Markov decision process (iMDP). We synthesize an optimal policy on this iMDP, which translates (with the specified confidence level) to a feedback controller for the continuous model with the same performance guarantees. Our experimental benchmarks confirm that accounting for epistemic uncertainty leads to controllers that are more robust against variations in parameter values.
This position paper reflects on the state-of-the-art in decision-making under uncertainty. A classical assumption is that probabilities can sufficiently capture all uncertainty in a system. In this paper, the focus is on the uncertainty that goes beyond this classical interpretation, particularly by employing a clear distinction between aleatoric and epistemic uncertainty. The paper features an overview of Markov decision processes (MDPs) and extensions to account for partial observability and adversarial behavior. These models sufficiently capture aleatoric uncertainty but fail to account for…
Decision makers involved in the management of civil assets and systems usually take actions under constraints imposed by societal regulations. Some of these constraints are related to epistemic quantities, as the probability of failure events and the corresponding risks. Sensors and inspectors can provide useful information supporting the control process (e.g. the maintenance process of an asset), and decisions about collecting this information should rely on an analysis of its cost and value. When societal regulations encode an economic perspective that is not aligned with that of the decisio…
University of Texas at Austin（德克萨斯大学奥斯汀分校）；Radboud University（拉德堡德大学）；University of California, Berkeley（加州大学伯克利分校）；Eindhoven University of Technology（埃因霍温理工大学）
Uncertain partially observable Markov decision processes (uPOMDPs) allow the probabilistic transition and observation functions of standard POMDPs to belong to a so-called uncertainty set. Such uncertainty, referred to as epistemic uncertainty, captures uncountable sets of probability distributions caused by, for instance, a lack of data available. We develop an algorithm to compute finite-memory policies for uPOMDPs that robustly satisfy specifications against any admissible distribution. In general, computing such policies is theoretically and practically intractable. We provide an efficient…
Radboud University（拉德堡德大学）；University of Oxford（牛津大学）；University of Birmingham（伯明翰大学）；University of Kentucky（肯塔基大学）
Controllers for autonomous systems that operate in safety-critical settings must account for stochastic disturbances. Such disturbances are often modelled as process noise, and common assumptions are that the underlying distributions are known and/or Gaussian. In practice, however, these assumptions may be unrealistic and can lead to poor approximations of the true noise distribution. We present a novel planning method that does not rely on any explicit representation of the noise distributions. In particular, we address the problem of computing a controller that provides probabilistic guarant…
Safe control of constrained linear systems under both epistemic and aleatory uncertainties is considered. The aleatory uncertainty characterizes random noises and is modeled by a probability distribution function (PDF) and the epistemic uncertainty characterizes the lack of knowledge on the system dynamics. Data-based probabilistic safe controllers are designed for the cases where the noise PDF is 1) zero-mean Gaussian with a known covariance, 2) zero-mean Gaussian with an uncertain covariance, and 3) zero-mean non-Gaussian with an unknown distribution. Easy-to-check model-based conditions for…
The Ohio State University（俄亥俄州立大学）；Ford Greenfield Labs（福特格林菲尔德实验室）；Ford Motor Company（福特汽车公司）
In this paper, an online evolving framework is proposed to detect and revise a controller's imperfect decision-making in advance. The framework consists of three modules: the evolving Finite State Machine (e-FSM), action-reviser, and controller modules. The e-FSM module evolves a stochastic model (e.g., Discrete-Time Markov Chain) from scratch by determining new states and identifying transition probabilities repeatedly. With the latest stochastic model and given criteria, the action-reviser module checks validity of the controller's chosen action by predicting future states. Then, if the chos…
Radboud University（拉德堡德大学）；University of Oxford（牛津大学）；University of Kentucky（肯塔基大学）；University of Twente（特文特大学）
Controllers for dynamical systems that operate in safety-critical settings must account for stochastic disturbances. Such disturbances are often modeled as process noise in a dynamical system, and common assumptions are that the underlying distributions are known and/or Gaussian. In practice, however, these assumptions may be unrealistic and can lead to poor approximations of the true noise distribution. We present a novel controller synthesis method that does not rely on any explicit representation of the noise distributions. In particular, we address the problem of computing a controller tha…
In this paper, a proactive dynamic spectrum sharing scheme between 4G and 5G systems is proposed. In particular, a controller decides on the resource split between NR and LTE every subframe while accounting for future network states such as high interference subframes and multimedia broadcast single frequency network (MBSFN) subframes. To solve this problem, a deep reinforcement learning (RL) algorithm based on Monte Carlo Tree Search (MCTS) is proposed. The introduced deep RL architecture is trained offline whereby the controller predicts a sequence of future states of the wireless access net…
In this work, we address the problem of formal safety verification for stochastic cyber-physical systems (CPS) equipped with ReLU neural network (NN) controllers. Our goal is to find the set of initial states from where, with a predetermined confidence, the system will not reach an unsafe configuration within a specified time horizon. Specifically, we consider discrete-time LTI systems with Gaussian noise, which we abstract by a suitable graph. Then, we formulate a Satisfiability Modulo Convex (SMC) problem to estimate upper bounds on the transition probabilities between nodes in the graph. Us…
Learning to Adapt Control Barrier Functions Under Epistemic and Aleatoric Uncertainty
Department of Robotics, University of Michigan（机器人学系，密歇根大学）；Charles Stark Draper Laboratory（查尔斯·斯泰克·德拉珀实验室）；Department of Aerospace Engineering, University of Michigan（航空航天工程系，密歇根大学）
In this paper, we propose a compositional framework for the synthesis of safety controllers for networks of partially-observed discrete-time stochastic control systems (a.k.a. continuous-space POMDPs). Given an estimator, we utilize a discretization-free approach to synthesize controllers ensuring safety specifications over finite-time horizons. The proposed framework is based on a notion of so-called local control barrier functions computed for subsystems in two different ways. In the first scheme, no prior knowledge of estimation accuracy is needed. The second framework utilizes a probabilit…
Learning Surrogate LPV State-Space Models with Uncertainty Quantification
Eindhoven University of Technology（埃因霍温理工大学）；European Space Agency, ESTEC（欧洲航天局ESTEC）；HUN-REN Institute for Computer Science and Control（HUN-REN计算机科学与控制研究所）
We present a case study applying learning-based distributionally robust model predictive control to highway motion planning under stochastic uncertainty of the lane change behavior of surrounding road users. The dynamics of road users are modelled using Markov jump systems, in which the switching variable describes the desired lane of the vehicle under consideration and the continuous state describes the pose and velocity of the vehicles. We assume the switching probabilities of the underlying Markov chain to be unknown. As the vehicle is observed and thus, samples from the Markov chain are dr…
A Framework for Adaptive Stabilisation of Nonlinear Stochastic Systems
Leibniz University Hannover, Institute of Automatic Control（汉诺威莱布尼茨大学自动控制研究所）；University of Melbourne, Department of Electrical and Electronic Engineering（墨尔本大学电气与电子工程系）
We present SySCoRe, a MATLAB toolbox that synthesizes controllers for stochastic continuous-state systems to satisfy temporal logic specifications. Starting from a system description and a co-safe temporal logic specification, SySCoRe provides all necessary functions for synthesizing a robust controller and quantifying the associated formal robustness guarantees. It distinguishes itself from other available tools by supporting nonlinear dynamics, complex co-safe temporal logic specifications over infinite horizons and model-order reduction. To achieve this, SySCoRe generates a finite-state abs…
Ludwig Maximilian University of Munich（慕尼黑大学）；University of Colorado Boulder（科罗拉多大学博尔德分校）；Newcastle University（纽卡斯尔大学）
A novel reinforcement learning scheme to synthesize policies for continuous-space Markov decision processes (MDPs) is proposed. This scheme enables one to apply model-free, off-the-shelf reinforcement learning algorithms for finite MDPs to compute optimal strategies for the corresponding continuous-space MDPs without explicitly constructing the finite-state abstraction. The proposed approach is based on abstracting the system with a finite MDP (without constructing it explicitly) with unknown transition probabilities, synthesizing strategies over the abstract MDP, and then mapping the results …
This paper studies the synthesis of controllers for discrete-time, continuous state stochastic systems subject to omega-regular specifications using finite-state abstractions. We present a synthesis algorithm for minimizing or maximizing the probability that a discrete-time stochastic system with finite number of modes satisfies an omega-regular property. Our approach uses a finite-state abstraction of the underlying dynamics in the form of a Bounded-parameter Markov Decision Process (BMDP) arising from a finite partition of the system's domain. Such abstractions allow for a range of transitio…
