Algorithmization of constrained monotonic maneuvers for an advanced driver assistant system in the intelligent urban buses

Precise Docking at Charging Stations for Large-Capacity Vehicles: An Advanced Driver-Assistance System for Drivers of Electric Urban Buses

“…At the beginning of a docking task, on the driver's command and after the perception subsystem recognizes the charging station, the path planner computes a reference path joining the current configuration of the bus with a target configuration determined by an initial relative location of the charger with respect to the vehicle frame. The planner looks for a path that satisfies the following constraints and optimization criteria [9] (in order of importance): 1) a collision-free path (with respect to known static obstacles and taking into account bus body dimensions) 2) a curvature-limited path (the maximal absolute curvature of the path cannot violate the maximal admissible steering angle max 0 b of a bus) 3) a sufficiently smooth path (the absolute curvature rate cannot exceed imposed steering rate limitations) 4) a shortest-length path with a minimal number of segments (two-segment paths are preferred, constituting the splines up to a seventh degree). Thanks to specific properties of the docking maneuver, most of the planning computations are performed using analytical (geometrical) relations, which helps speed up the planning process.…”

“…Thanks to specific properties of the docking maneuver, most of the planning computations are performed using analytical (geometrical) relations, which helps speed up the planning process. In more complicated cases, the planner numerically searches for the path using the state lattices approach [9], [14], which casts the path-planning problem to a graph-search problem after building an implicit multidimensional grid in the vehicle's configuration space by recursively expanding a predefined set of path primitives. The path computed by the planner is provided in a nonparameterized form by via the so-called level curve approach (see, e.g., [9]), where the equation ( , )…”

“…This article presents a prototype ADAS for drivers of electric urban buses, which is being cooperatively developed by Poznań University of Technology and European manufacturer Solaris Bus & Coach (SBC). The proposed system builds on and integrates components previously created by the authors in the context of vision-based detection [8] as well as motion planning and feedback control algorithms [9], [10]. The main contribution to the field of intelligent transportation comes from the new concept of the complete ADAS, which adheres to the requirements of bus manufacturers and operators by relying only on onboard sensors and computational systems while being affordable and scalable to various city bus models.…”

A Self-Supervised Learning Approach to Rapid Path Planning for Car-Like Vehicles Maneuvering in Urban Environment

“…denote the first and second derivative of the path in the i-th gluing point, respectively. All values describing i-th gluing point are expressed in the local coordinate system associated with (i − 1)-th gluing point in order to enable the path P to be represented in the control-law as the level-curve [16]. As a result, segments are described by the 5-th degree polynomials in the local coordinate system associated with the (i − 1)-th gluing point, defined as…”