RT Journal Article
T1 Efficient local navigation approach for autonomous driving vehicles
A1 López Fernández, Joaquín
A1 Sánchez Vilariño, Pablo
A1 Sanz Domínguez, Rafael
A1 Paz Domonte, Enrique
K1 1203.04 Inteligencia Artificial
K1 304.12 Dispositivos de Control
K1 3317 Tecnología de Vehículos de Motor
AB This paper presents an efficient and practical approach for a car navigation system (CVM-Car) based on the velocity space optimization paradigm. The method calculates the velocity control commands to keep the car in the lane while avoiding the obstacles detected by the proximity sensors. The car has to follow a road path consisting of a sequence of lanelets. This approach is a lower-level reactive control that combines the pure pursuit method to obtain a reference curvature and a reactive control algorithm that keeps the vehicle in the center of the lane’ s free space while avoiding obstacles that can partially block it. CVM-Car formulates local obstacle avoidance as a constrained  optimization problem in the velocity space of the car. In addition to the vehicle dynamics and obstacles constraints included by the curvature method, car-shape and non-holonomic restrictions are considered in the CVM-Car velocity space. The method has been applied to an autonomous vehicle prototype.
PB IEEE Access
SN 21693536
YR 2021
FD 2021-05-28
LK http://hdl.handle.net/11093/2457
UL http://hdl.handle.net/11093/2457
LA eng
NO IEEE Access, 9: 79776-79792 (2021)
NO Ministerio de Economía y Competitividad | Ref. TRA2015-70501-C2-2-R
DS Investigo
RD 23-sep-2023