Design and development of algorithms and control strategies for automatic parking systems: advanced driver assistance system

Abstract

The relationship between customers/consumers and the automotive industry adds to the need for electric, intelligent and autonomous vehicles to be independent, even during maneuvers considered difficult, aiming at the safety of vehicle occupants and third parties. In this work, a type C algorithm was developed for the automatic parking of a vehicle in parallel, with computational validation (Software In-the Loop). The great differential of the developed C algorithm adopts the strategy of the main vehicle being misaligned in relation to the secondary (parked) vehicle. That is, it is not mandatory that the main vehicle is in the axle-to-axle position of the secondary vehicle to perform the automatic parallel parking maneuver. A minimum value was determined in relation to the position (x) and a minimum value in relation to the position (y) with the reference being the secondary (parked) vehicle. The geometric points are the positions of the 2 parked vehicles and the parking space is equal to the displacement of the free space between the 2 parked vehicles. When this displacement is greater than the total length of the vehicle, the maneuver is performed, otherwise the maneuver is canceled because there is no guarantee that it will be successful; these geometric relations give rise to the trajectory that the main vehicle must travel. The developed algorithm aims at the subsequent implantation of the Hardware In-the Loop (HIL) model, specifically the validation of the algorithm in a real prototype or vehicle, since the computational execution/validation of the algorithm was successful. The type C algorithm if programmed in an electronic microcontroller allows the control of the vehicle to perform safe maneuvers and without human intervention, so it is considered an intelligent automatic parallel parking system (embedded system) capable of identifying a parallel parking space, checking if the size of the space is adequate to park the vehicle in the reverse direction, detecting possible obstacles during the maneuver, through ultrasonic sensors; other types of measurement sensors can be used. As other proposed contributions, the tuning methods for the longitudinal controller (PID) and lateral controller (MPC) stand out, as well as a complete study on the kinematics and vehicle dynamics and the functional modeling of a vehicle with 4 wheels. The MATLAB/Simulink software helped with the various accurate computer simulations, to ensure Software In-the Loop (SIL) validation.