In a move that strengthens our position in the race for artificial intelligence and the Fourth Industrial Revolution, Misr University for Information Technology has unveiled the latest electronic program for interactive intelligent autonomous driving.
Prof. Dr. Hoda Mokhtar, Dean of the Faculty of Computer Science and Information at Egypt University of Informatics, emphasised that autonomous driving technologies pose a technical challenge due to their direct impact on traffic safety, transport efficiency, and the availability of autonomous driving services for people with special needs. Despite significant advances in the fields of computer vision, sensor integration, and machine learning, current autonomous driving systems still suffer from two major problems:
1. Limited generalisation and adaptability due to the difficulty of dealing with unfamiliar scenarios, leading to wrong decisions that may increase the risk of accidents.
2. Opacity of decision-making (the black box problem): Most current autonomous driving systems lack transparency, which reduces public confidence in them and thus hinders the regulatory approval of these systems, which are still banned in many countries around the world.
She said that Egypt University of Informatics, through a team of fourth-year students from the Faculty of Computer and Information Sciences, has successfully developed the DriveFusion model, a multimodal model based on the Qwen2.5-VL model. It integrates inputs from the camera, GPS coordinates, speed data, and user text queries into a unified input platform, enabling the model to produce low-level control commands (such as steering and acceleration) and high-level linguistic interpretations that support transparent and understandable decision-making.
Dr. Huda confirmed the team’s success in developing a comprehensive autonomous driving model that addresses the fundamental challenges of limited adaptability and lack of transparency in current technologies. We have successfully combined the capabilities of large multimodal language models with specialised automotive components. The new model thus offers a highly efficient solution capable of explaining decisions. It contributes significantly to the field of explainable artificial intelligence in safety-critical applications and paves the way for a future in which autonomous vehicles are safer, more reliable and more trustworthy.
She said that this project represents the best form of cooperation between universities and industry professionals. In addition to my personal supervision of the student team developing the programme, which consists of Omar Samir, Ibrahim Ahmed, Youssef Walid, Ahmed Walid, Sajid Samer and Mahmoud Khaled, all fourth-year students at the faculty, I am also supervised by Dr. Ibrahim Sobh, a senior expert in the field of artificial intelligence at Valeo, one of the world’s largest companies in the field of embedded programming for cars, and Engineer Taqi Muhammad, a teaching assistant at the Faculty of Computer and Information Sciences.
A team of computer science and information technology students at Misr University for Information Technology designs a program for autonomous driving
