Not yet reviewed by Pith; the record is open.

This paper extends the gap-based navigation technique in Potential Gap by guaranteeing safety for nonholonomic robots for all tiers of the local planner hierarchy, so called Safer Gap. The first tier generates a Bezier-based collision-free path through gaps. A subset of navigable free-space from the robot through a gap, called the keyhole, is defined to be the union of the largest collision-free disc centered on the robot and a trapezoidal region directed through the gap. It is encoded by a shallow neural network zeroing barrier function (ZBF). Nonlinear model predictive control (NMPC), with Keyhole ZBF constraints and output tracking of the Bezier path, synthesizes a safe kinematically-feasible trajectory. Low-level use of the Keyhole ZBF within a point-wise optimization-based safe control synthesis module serves as a final safety layer. Simulation and experimental validation of Safer Gap confirm its collision-free navigation properties.