Selected problems of automatic obstacle avoiding

International Automotive Conference 2018

Gidlewski, M.^1,3 ; Jemiol, L.^3 ; ardecki, D.^1,2

Automotive Industry Institute, ul. Jagielloska 55, Warszawa

03-301, Poland^1

Military University of Technology, ul. Gen. Witolda Urbanowicza 2, Warszawa

00-908, Poland^2

University of Technology and Humanities, ul. Malczewskiego 29, Radom

26-600, Poland^3

关键词: Computational tests;    Controlled objects;    Controlling process;    Mechatronic systems;    Obstacle-avoiding;    Reference signals;    Regulating systems;    Vehicle control system;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/421/3/032009/pdf
DOI  :  10.1088/1757-899X/421/3/032009

来源: IOP

【 摘 要 】

The paper presents selected problems concerning the concept of automatic controlling of motor vehicle motion in case of sudden appearance of an obstacle. If the vehicle cannot be brought to a halt before hitting the obstacle and the adjacent lane is clear, circumvention of the obstacle is the only way to avoid a collision. In the concept of the vehicle control system, elements of general control theory and simplified models of longitudinal and lateral vehicle dynamics have been used. The vehicle controlling process consists of two stages: braking and then changing the lane at a speed determined by the braking stage. The first stage is implemented with using standard mechatronic systems that automatize the braking process, while the second stage requires the use of a special controller coupled with the steering system. The controller generates a "bang-bang" reference signal and corrects it by means of regulating systems that minimize the deviations of the actual (measured) signals describing vehicle's trajectory from the reference signal. The control algorithm developed has been based on very simplified and, thanks to that, efficient reference models. The algorithm was validated by simulation tests, where a model of a medium-size truck was used as the virtual controlled object. In the model (3D, nonlinear, MBS type, experimentally verified), a very detailed description of vehicle motion, even in boundary conditions, can be taken into account. The paper presents fragments of the research work, including unpublished results of computational tests.