Abstract:This paper studies formal synthesis of controllers for continuous-space systems with unknown dynamics to satisfy requirements expressed as linear temporal logic formulas. Formal abstraction-based synthesis schemes rely on a precise mathematical model of the system to build a finite abstract model, which is then used to design a controller. The abstraction-based schemes are not applicable when the dynamics of the system are unknown. We propose a data-driven approach that computes a growth bound of the system using a finite number of trajectories. The computed growth bound together with the sampled trajectories are then used to construct the abstraction and synthesise a controller. Our approach casts the computation of a growth bound as a robust convex optimisation program (RCP). Since the unknown dynamics appear in the optimisation, we formulate a scenario convex program (SCP) corresponding to the RCP using a finite number of sampled trajectories. We establish a sample complexity result that gives a lower bound for the number of sampled trajectories to guarantee the correctness of the growth bound computed from the SCP with a given confidence. Our sample complexity result requires knowing a possibly conservative bound on the Lipschitz constant of the system. We also provide a sample complexity result for the satisfaction of the specification on the system in closed loop with the designed controller for a given confidence. Our data-driven synthesised controller can provide guarantees on satisfaction of both finite and infinite-horizon specifications. We show that our data-driven approach can be readily used as a model-free abstraction refinement scheme by modifying the formulation of the system’s growth bounds and providing similar sample complexity results. The performance of our approach is shown on three case studies.

Deductive Controller Synthesis for Probabilistic Hyperproperties

Controller Synthesis for Hyperproperties

Probabilistic Hyperproperties of Markov Decision Processes

Permissive Controller Synthesis for Probabilistic Systems

Formal Controller Synthesis for Markov Jump Linear Systems with Uncertain Dynamics

Probabilities Are Not Enough: Formal Controller Synthesis for Stochastic Dynamical Models with Epistemic Uncertainty

Probabilistic Plan Synthesis for Coupled Multi-Agent Systems

Bayesian Statistical Model Checking for Multi-agent Systems using HyperPCTL*

Statistical Model Checking for Hyperproperties

Multi-Objective Controller Synthesis with Uncertain Human Preferences

Learning Robust Policies for Uncertain Parametric Markov Decision Processes

Correct-by-Construction Control for Stochastic and Uncertain Dynamical Models via Formal Abstractions

Formal controller synthesis for hybrid systems using genetic programming

Negotiating the Probabilistic Satisfaction of Temporal Logic Motion Specifications

Search and Explore: Symbiotic Policy Synthesis in POMDPs

Controller synthesis for linear temporal logic and steady-state specifications

Scalable control synthesis for stochastic systems via structural IMDP abstractions

Hybrid Controller Synthesis for Nonlinear Systems Subject to Reach-Avoid Constraints.

Unpredictable Planning Under Partial Observability

Data-driven abstraction-based control synthesis

Hyper parametric timed CTL