Automated vehicle control for critical situations

 
Design of automated vehicle control systems for critical situations, taking into account the state of the wheel-ground contact
1 Mar 2022– 28 Feb 2024
External identifier
2019-2.1.11-TÉT-2020-00125
Cost
1 990 000 HUF
SZTAKI cost
1 990 000 HUF
Founded %
100%
 

The aim of the research project

When controlling automated vehicles, it is essential that the control system has information about the state of the ego-vehicle and its environment. The knowledge of the tyre-road friction characteristics is a particularly important segment of this information.

For conventional driver assistance systems, an approximate knowledge of the aforementioned characteristics is sufficient, as it is assumed that the driver has detailed environmental information. Automated vehicle control, however, needs to have full information about the tyre-road friction to maintain stable vehicle motion in critical situations (e.g. tyre slip and/or high speed and collision avoidance manoeuvres).

The aim of this research is to develop control design and estimation methods for critical cases associated with vehicles that have automated functions considering the tyre-road friction characteristic. The research will be implemented through the design of two functions of high importance. These functions are the detection of critical situations for the automated ego-vehicle and the design of fault-tolerant control schemes based on real-time tyre-pressure estimation using low-cost sensors.

The tyre-road friction estimation tasks are solved by fusing machine learning-based data and model-based methods, while the control tasks are solved by designing fault-tolerant, robust and nonlinear reconfigurable control.

The research will result in new estimation and control design methods that can be used to guarantee safe motion of automated vehicles under varying tyre-road friction characteristics and to meet the requirements for vehicle dynamic quality requirements with higher efficiency compared to model-based control methods used in driver assistance systems.

The effectiveness of the developed methodological results will be demonstrated partly through simulations and partly through the evaluation of measurements on real test vehicles (e.g. on the ZalaZone test track and during road measurements).

 

 

Participants
  • Faculty of Mechanical Engineering and Naval Architecture, University of Zagreb,