This paper presents automatic parallel parking for a passenger vehicle, with highlights on a path-planning method and on experimental results. The path-planning method consists of two parts. First, the kinematic model of the vehicle, with corresponding geometry, is used to create a path to park the vehicle in one or more maneuvers if the spot is very narrow. This path is constituted of circle arcs. Second, this path is transformed into a continuous-curvature path using clothoid curves. To execute the generated path, control inputs for steering angle and longitudinal velocity depending on the traveled distance are generated. Therefore, the traveled distance and the vehicle pose during a parking maneuver are estimated. Finally, the parking performance is tested on a prototype vehicle.

International audienceThis paper deals with path planning for car-like vehicle in parallel parking problems. Our path planning method uses simple geometry of the vehicle kinematic model. The presented strategy consists in two parts: create a simple geometric path for the parallel parking in one or more maneuvers, formed by circle arcs and then transform it to a continuous-curvature path with the use of clothoids. Accordingly, a car-like vehicle can follow such path without reorienting its front wheels at stop. The proposed planning method is independent of the initial position and of the orientation of the vehicle. Control inputs for steering angle and longitudinal velocity are generated so that the vehicle can park by following them

No abstract

No abstract

This paper compares 3 ASR 1 control laws, a PI controller, a linearizing feedback and a flatness based control. First, the controllers are designed based on a two state equations vehicle model. Then, the controllers are experimentally validated on an hybrid vehicle (VELROUE) with rear electric wheels and front axle ICE 2 traction. Finally, the nonlinear control laws are compared to the PI controller that is classically used in the automotive industry.