We present shrinking horizon model predictive control for discrete-time linear systems under stochastic disturbances with constraints encoded as Signal Temporal Logic (STL) specification. The control objective is to satisfy a given STL specification with high probability against stochastic uncertainties while maximizing the robust satisfaction of an STL specification with minimum control effort. We formulate a general solution, which does not require precise knowledge of probability distributions of (possibly dependent) stochastic disturbances; only the bounded support of the density functions and moment intervals are used. For the specific case of disturbances that are normally distributed, we optimize the controllers by utilizing knowledge of the probability distribution of the disturbance. We show that in both cases, the control law can be obtained by solving optimization problems with linear constraints at each step. We experimentally demonstrate effectiveness of this approach by synthesizing a controller for an HVAC system.
24 September 2018 in IEEE Transactions on Automatic Control
Controller synthesis techniques for continuous systems with respect to temporal logic specifications typically use a finite-state symbolic abstraction of the system model. Constructing this abstraction for the entire system is computationally expensive, and does not exploit natural decompositions of many systems into interacting components. We describe a methodology for compositional symbolic abstraction to help scale controller synthesis for temporal logic to larger systems. We introduce disturbance bisimulation, which strengthens the standard approximate alternating bisimulation relation used in control. It extends naturally to systems which are composed of weakly interconnected sub-components, possibly connected in feedback, and models the coupling signals as disturbances. We show how networks of incrementally input-to-state stable, nonlinear, continuous-time control systems can be abstracted compositionally, so that all local abstractions are simultaneously disturbance bisimilar to their continuous counterparts. Furthermore, our construction ensures that the final composed abstraction is disturbance bisimilar to the original system. Finally, we discuss how
Systematic testing of autonomous vehicles operating in complex real-world scenarios is a difficult and expensive problem. We present Paracosm, a framework for writing systematic test scenarios for autonomous driving simulations. Paracosm allows users to programmatically describe complex driving situations with specific features, e.g., road layouts and environmental conditions, as well as reactive temporal behaviors of other cars and pedestrians. A systematic exploration of the state space, both for visual features and for reactive interactions with the environment is made possible. We define a notion of test coverage for parameter configurations based on combinatorial testing and low dispersion sequences. Using fuzzing on parameter configurations, our automatic test generator can maximize coverage of various behaviors and find problematic cases. Through empirical evaluations, we demonstrate the capabilities of Paracosm in programmatically modeling parameterized test environments, and in finding problematic scenarios.
7 January 2025 in Proceedings of the ACM on Programming Languages
Deciding termination is a fundamental problem in the analysis of probabilistic imperative programs. We consider the qualitative and quantitative probabilistic termination problems for an imperative programming model with discrete probabilistic choice and demonic bounded nondeterminism. The qualitative question asks if the program terminates almost-surely, no matter how nondeterminism is resolved. The quantitative question asks for a bound on the probability of termination. Despite a long and rich literature on the topic, no sound and relatively complete proof systems were known for these problems. In this paper, we provide sound and relatively complete proof rules for proving qualitative and quantitative termination in the assertion language of arithmetic. Our rules use variant functions as measures of distances to termination as well as supermartingales that restrain the growth of these variants almost-surely. Our completeness result shows how to construct suitable supermartingale and variant functions from an almost-surely terminating program. We also show that proofs of termination in many existing proof systems can be transformed to proofs in our system, pointing to its applicability in practice. As an application of our proof rule, we show an explicit proof of almost-sure termination for the two-dimensional random walker.
28 December 2019 in Discrete Event Dynamic Systems
Reactive synthesis and supervisory control theory both provide a design methodology for the automatic and algorithmic design of digital systems from declarative specifications. The reactive synthesis approach originates in computer science, and seeks to synthesise a system that interacts with its environment over time and that, doing so, satisfies a prescribed specification. Here, the distinguishing feature when compared to other synthesis problems in computer science is that the interaction is temporal in that it explicitly refers to a sequence of computation cycles. Supervisory control originates in control theory and seeks to synthesise a controller that – in closed-loop configuration with a plant – enforces a prescribed specification over time. The distinguishing feature compared to other branches of control is that all dynamics are driven by discrete events as opposed to continuous signals. While both methods apparently are closely related, the technical details differ significantly. We provide a formal comparison which allows us to identify conditions under which one can solve one synthesis problem using methods from the other one; we also discuss how the resulting solutions compare. To facilitate this comparison, we give a unified introduction to reactive synthesis and supervisory control and derive formal problem statements and a characterisation of their solutions in terms of ω-languages. Recent contributions to the two fields address different aspects of the respective problem, and we expect the formal relationship identified in this paper to be useful in that it allows the application of algorithmic techniques from one field in the other.