Title: On the synthesis of control policies from noisy example datasets: a probabilistic approach

Abstract: In this note we consider the problem of synthesizing optimal control policies for a system from noisy datasets. We present a novel algorithm that takes as input the available dataset and, based on these inputs, computes an optimal policy for possibly stochastic and nonlinear systems that also satisfies actuation constraints. The algorithm relies on solid theoretical foundations, which have their key roots into a probabilistic interpretation of dynamical systems. The effectiveness of our approach is illustrated by considering an autonomous car use case. For such use case, we make use of our algorithm to synthesize a control policy from noisy data allowing the car to merge onto an intersection, while satisfying additional constraints on the variance of the car speed