Intersection Crossing with Future-Focused Control Barrier Functions

— In this paper, we introduce a future-focused control barrier function (ff-CBF) approach to an unsignaled four-way intersection crossing problem for a collection of com- municating automobiles. Our novel ff-CBF encodes that vehicles take control actions that avoid predicted collisions over an arbitrarily long future time horizon, thereby defining a virtual barrier. We then propose a relaxed-virtual CBF (rv-CBF) that inherits the predictive power of the ff-CBF while allowing relaxations of the virtual barrier away from the physical barrier between vehicles. We study the efficacy of the ff-CBF and rv-CBF based controllers via a series of simulated trials of the intersection scenario and highlight how the rv-CBF based controller empirically outperforms a benchmark controller from the literature by improving intersection throughput while preserving safety and feasibility properties.