The postgraduate doctoral student (Shahed Sebeh Ghintab) from the Control and Systems Engineering Department received a doctorate degree in (Control Engineering) with a grade of (very good) for her thesis entitled (Deep Learning Based Autonomous Driving in Urban Areas). The discussion was held on Monday, March 4, 2024, in the discussion Hall (Hall No. 9) in the department,

and the discussion committee consisted of:

1- Prof. Dr. Hazem Ibrahim Ali / Chairman

2- Prof. Dr. Hadeel Nusrat Abdullah / Member

3- Asst. Prof. Dr. Firas Abdul Razzaq Rahim / Member

4- Asst. Prof. Dr. Osama Ali Awad / Member

5- Asst. Prof. Dr. Bushra Kadhum Oleiwi / Member

6- Prof. Dr. Muhammad Youssef Hassan / Member and Supervisor

Autonomous vehicles have emerged as an interesting field in engineering, offering a potential solution to reduce accidents caused by human error and maximize the use of parking in cities. This work investigates the significant impact of autonomous vehicle technology on transportation infrastructure, emphasizing the necessity of a comprehensive framework that integrates perception, localization, planning, and control to ensure reliable autonomous navigation. The thesis proposes an adaptive integrated derivative Type-2 fuzzy logic controller. Advanced technology has been employed in localization, where convolutional neural networks were used and trained on images of red, blue, and green types after modifying one of the networks known as AlexNet. The A* algorithm was used for path planning. For environmental perception, a network called You Only Look Once (YOLO) version 3 was utilized, providing accurate localization results for autonomous driving within the city despite various weather conditions. Only color images were relied upon, without resorting to expensive sensors, such as radar and LiDAR technology. The algorithm for merging regular images with depth images was adopted to obtain good results when examining the network trained using the MATLAB program. Subsequently, the trained network was rebuilt using the Python language so that it could be linked with the simulation program for autonomous driving cars called CARLA. The simulation results with the trained network and simulation showed remarkable and excellent performance in perceiving the surrounding environment of the vehicles. The achieved results from localization, perception, and path planning were used alongside the proposed controller, whose variables were tuned using the fly optimization algorithm, which demonstrated the accuracy and robustness of the proposed algorithm through the improvement ratio of the mean squared error for the steering angle compared to the reference results. In particular, the proposed algorithm achieved an improvement in the mean squared error for the steering angle by 89.68% compared to the results of reference [77]. As for speed, there was an improvement in the square error rate compared to the same reference with an improvement rate of 98.52%. The proposed solution effectively adjusts the parameters of the control unit accurately, while staying flexible and adaptable to changing conditions. It can be implemented in real-world scenarios to achieve safe and successful navigation. The defense was attended by the Assistant Head of the Department for Scientific Affairs and Postgraduate Studies (Prof. Dr. Abbas Hussein Issa) along with a group of professors in the department. On this occasion, we congratulate the student (Shahed Sebeh Ghintab) and wish her success and prosperity.