Signal Temporal Logic Control Synthesis among Uncontrollable Dynamic Agents with Conformal Prediction
Abstract
The control of dynamical systems under temporal logic specifications among uncontrollable dynamic agents is challenging due to the agents’ a-priori unknown behavior. Existing works have considered the problem where either all agents are controllable, the agent models are deterministic and known, or no safety guarantees are provided. We propose a predictive control synthesis framework that guarantees, with high probability, the satisfaction of signal temporal logic (STL) tasks that are defined over the system and uncontrollable stochastic agents. We use trajectory predictors and conformal prediction to construct probabilistic prediction regions for each uncontrollable agent that are valid over multiple future time steps. Specifically, we reduce conservatism and increase data efficiency compared to existing works by constructing a normalized prediction region over all agents and time steps. We then formulate a worst-case mixed integer program (MIP) that accounts for all agent realizations within the prediction region to obtain control inputs that provably guarantee task satisfaction with high probability. To efficiently solve this MIP, we propose an equivalent MIP program based on KKT conditions of the original one. We illustrate our control synthesis framework on two case studies.
Yiqi Zhao
PhD Student, Computer Science
Yiqi (Nick) Zhao is a PhD Student for Computer Science at the University of Southern California.