Abstract:We investigate the problem of synthesizing distributionally robust control policies for stochastic systems under safety and reach-avoid specifications. Using a game-theoretical framework, we consider the setting where the probability distribution of the disturbance at each time step is selected from an ambiguity set defined by the Wasserstein distance. The goal is to synthesize a distributionally robust control policy that ensures the satisfaction probability exceeds a specified threshold under any distribution within the ambiguity set. First, for both safety and reach-avoid specifications, we establish the existence of optimal policies by leveraging the dynamic programming principles. Then we demonstrate how the associated optimization problem can be efficiently solved using the dual representation of Wasserstein distributionally robust optimization. Furthermore, for safety specifications in particular, we introduce a novel concept of distributionally robust control barrier certificates and show how these enable the efficient synthesis of controllers through sum-of-squares programming techniques. Finally, our experimental results reveal that incorporating distributional robustness during the synthesis phase significantly improves the satisfaction probability during online execution, even with limited statistical knowledge of the disturbance distribution.

What problem does this paper attempt to address?

### What problem does this paper attempt to solve? This paper aims to solve the problem of how to synthesize distributionally robust control strategies for stochastic systems in uncertain environments to ensure that the safety and reach - avoid specifications of the system are met. Specifically, the researchers focus on: 1. **Controller design in uncertain environments**: - The system will be affected by external disturbances or adversarial inputs during operation. - The distribution of these disturbances may be unknown or partially known, so a robust method is required to deal with these uncertainties. 2. **Safety and reach - avoid specifications**: - **Safety specification**: It requires the system to stay within a certain safety area throughout the operation. - **Reach - avoid specification**: It requires the system to reach the target area while avoiding obstacles. 3. **Distributionally robust optimization framework**: - Using the ambiguity set defined by the Wasserstein distance, consider that the disturbance distribution at each time step may come from any distribution in this ambiguity set. - The goal is to design a distributionally robust control strategy such that the probability of satisfying the specification exceeds a given threshold under any distribution in the ambiguity set. ### Main contributions of the paper 1. **Existence of optimal control strategies**: - Through the dynamic programming principle, it is proved that under the Wasserstein ambiguity set, there exists an optimal control strategy that can satisfy the safety and reach - avoid specifications. 2. **Computational feasibility**: - A method based on dual representation is proposed, which transforms the original infinite - dimensional optimization problem into a finite - dimensional problem, thus effectively solving the optimization problem in the dynamic programming equation. 3. **Efficient synthesis of safety specifications**: - The concept of distributionally robust control barrier certificates is introduced to approximate the lower bound of the satisfaction probability and transform it into a more easily solvable sum - of - squares program. ### Summary By introducing the distributionally robust optimization framework, this paper solves the problem of stochastic systems satisfying safety and reach - avoid specifications in uncertain environments. The research results not only extend the existing theoretical framework but also provide an effective computational method, which can improve the robustness and reliability of the system in practical applications.

Distributionally Robust Control Synthesis for Stochastic Systems with Safety and Reach-Avoid Specifications

Efficient Strategy Synthesis for Switched Stochastic Systems with Distributional Uncertainty

Safe and Stable Control Synthesis for Uncertain System Models via Distributionally Robust Optimization

Data-Driven Distributionally Robust System Level Synthesis

Wasserstein Distributionally Robust Control of Partially Observable Linear Stochastic Systems

Distributionally Robust Infinite-horizon Control: from a pool of samples to the design of dependable controllers

Distributionally Robust Density Control with Wasserstein Ambiguity Sets

Stochastic Lyapunov-Barrier Functions for Robust Probabilistic Reach-Avoid-Stay Specifications

Distributionally Robust Policy and Lyapunov-Certificate Learning

Distributionally robust stochastic optimal control

A Distributionally Robust Optimization based Method for Stochastic Model Predictive Control

Statistical Learning of Distributionally Robust Stochastic Control in Continuous State Spaces

Adaptive dynamic programming and distributionally robust optimal control of linear stochastic system using the Wasserstein metric

Robust Control for Dynamical Systems with Non-Gaussian Noise via Formal Abstractions

Robust Data-Driven Safe Control using Density Functions

Feasibility Analysis and Regularity Characterization of Distributionally Robust Safe Stabilizing Controllers

Data-drivenDistributionallyRobustOptimal Stochastic ControlUsing theWassersteinMetric

Robust affine control of linear stochastic systems

Distributionally Robust Control for Chance-Constrained Signal Temporal Logic Specifications

Wasserstein Distributionally Robust Control and State Estimation for Partially Observable Linear Systems

Distributionally robust uncertainty quantification via data-driven stochastic optimal control