الفهرس 
INTRODUCTIONOnly 14 pages are availabe for public view
 |  | from | 110 |  |  |
| | | from | 110 | | |
Abstract
Currently, Automotive Driving Assistance Systems (ADAS) technologies are developing rapidly. Many OEMs are giving huge intentions to invent new ADAS features, and that will lead to developing more safe and reliable auto-pilot systems. Eventually, Artificial Intelligence (AI) plays an important role in developing algorithms that are widely used to develop ADAS main systems, such as, environment perception, path planning, and decision-making, and motion control. Sensors’ data fusion sub-systems are used to build features information related to the surrounding environment which provides information about the road, objects, edges, contours, etc. A number of advanced driver assistance systems, for instance, Lane Keep Assist (LKA), is considered as an important assist ADAS feature that is used in the different autonomous level. Further details related to the control of LKA using adaptive Model Predictive Control (MPC) will be discussed in this work.The rationale behind using MPC technology is that it has been progressed steadily over more than 30 years ago. MPC applications are currently showing a stable architecture in many industries like Chemicals, Aerospace, Defense, and Food processing. Recently, the work area of MPC’s technologies is rapidly changing, so it is difficult to maintain track of the accelerating development in academic research and industrial applications. MPC is currently showing large potential for usage in automotive applications as automobile subsystems are progressively coordinated to enhance fuel economy and safety. Accordingly, innovative chances for MPC are arising, which includes coordination of braking and powertrain in torque vectoring, management of engine and transmission to improve fuel economy and responsiveness, and control of complex engines. This study discusses an approximate Neural Network (NN) model which is proposed to imitate the behavior of adaptive MPC. The proposed model can be used for nonlinear systems with nonlinear constraints related to automotive applications (i.e. LKA). The obtained result can guarantee stability, robustness, and constraint satisfaction for the learned MPC. In this study, it is illustrated that the Approximated Adaptive Model Predictive Control (AAMPC) can reduce computational complexity and be a better solution for embedded systems. The overall simulation model has been developed using MATLAB®, Simulink®, Model Predictive Control ToolboxTM, and Automated Driving System ToolboxTM. In particular, scenarios were generated using the Scenario Designer application with the Auto-mated Driving System Toolbox that allows us to develop training data and test ADAS (LKA feature).