Helene Vorobieva

Automatic Parallel Parking in Tiny Spots: Path Planning and Control

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.

Geometric Path Planning for Automatic Parallel Parking in Tiny Spots

This 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 retrieving the vehicle from the parking spot and reversing the obtained path to park the vehicle. Two methods for parking in tiny spaces, where parking in one trial is not possible, are proposed. In these cases, the number of needed trials to park the vehicle can be calculated from a simple formula or from an iterative algorithm. The proposed planning methods are independent of the initial position and the orientation of the vehicle. Reference trajectories are generated so that the vehicle can park by following them. Simulations are provided for both methods.

Lane keeping and lane departure avoidance by rear wheels steering

The driver active assistances have gained recently in importance and become usual features in commercialized vehicles. Nevertheless, the sharing of the vehicle control with the driver, especially for the steering assistances, is an open issue. The rear wheels steering could be an answer to this question. This study contains the development of a control law for the rear steering able to provide lane departure avoidance and lane keeping. The design of the control law is casted as a LMI (Linear Matrix Inequality) optimization problem in order to take into account constraints, as the rear steering actuator limitation. A modulation factor is added to allow a flexible activation strategy of the rear steering assistance. Preliminary simulation results are presented to explain the concept.