Irfan Khan
Combined lateral and longitudinal control for autonomous driving based on Model Predictive Control.
Rel. Nicola Amati, Andrea Tonoli, Angelo Bonfitto. Politecnico di Torino, Corso di laurea magistrale in Automotive Engineering (Ingegneria Dell'Autoveicolo), 2019
|
PDF (Tesi_di_laurea)
- Tesi
Licenza: Creative Commons Attribution Non-commercial No Derivatives. Download (7MB) | Preview |
Abstract: |
Autonomous ground vehicles, as an important part of intelligent transportation system, are attracting more attention than ever before. Their control system usually consists of three modules: environment perception, planning and decision-making, and vehicle control. Vehicle control is one of the most critical part of the whole architecture, as it is responsible for the vehicle guidance considering both safety and comfort. In general, control can be divided into lateral control and longitudinal velocity control. Their inter coordination leads to autonomous vehicle motion. This thesis is focused on the development of a combined lateral and longitudinal controller for autonomous driving based on Model Predictive Control (MPC). The proposed strategy utilizes an adaptive MPC to perform lateral guidance and speed regulation by acting on the steering angle and acceleration/deceleration to minimize the vehicle’s lateral deviation and relative yaw angle with respect to the reference trajectory, while driving the vehicle at the maximum acceptable longitudinal speed. The technique exploits a stereo camera that utilizes the synthetic data coming from the simulated scenario for lane detection and reference trajectory generation i.e. center line of the lane, to perform the lateral guidance. Longitudinal control strategy is realized with a reference speed generator, which calculates the maximum speed by previewing the path ahead of the vehicle and stability of the vehicle at the same time. The proposed controller is tested with three different scenarios: highway, interurban and urban driving to check the performance at different speeds and varying environment. Dynamics of the vehicle is modeled using a 3 degree of freedom rigid vehicle model, while the internal plant model for MPC is modeled using a linear bicycle model. The overall system has been developed using MATLAB®, Simulink®, Model Predictive Control Toolbox™ and Automated Driving System Toolbox™. In particular, scenarios were generated using the Scenario Designer application with the Automated Driving System Toolbox that allows to develop and test Advanced Driver Assistance Systems (ADAS) and autonomous driving systems providing computer vision algorithms and generating synthetic data. |
---|---|
Relators: | Nicola Amati, Andrea Tonoli, Angelo Bonfitto |
Academic year: | 2018/19 |
Publication type: | Electronic |
Number of Pages: | 83 |
Subjects: | |
Corso di laurea: | Corso di laurea magistrale in Automotive Engineering (Ingegneria Dell'Autoveicolo) |
Classe di laurea: | New organization > Master science > LM-33 - MECHANICAL ENGINEERING |
Aziende collaboratrici: | UNSPECIFIED |
URI: | http://webthesis.biblio.polito.it/id/eprint/10667 |
Modify record (reserved for operators) |