Jean-Philippe Georges

The deployment of autonomous vehicles has the potential to significantly lessen the variety of current harmful externalities, (such as accidents, traffic congestion, security, and environmental degradation), making autonomous vehicles an emerging topic of research. In this paper, a literature review of autonomous vehicle development has been conducted with a notable finding that autonomous vehicles will inevitably become an indispensable future greener solution. Subsequently, 5 different deep learning models, YOLOv5s, EfficientNet-B7, Xception, MobilenetV3, and InceptionV4, have been built and analyzed for 2-D object recognition in the navigation system. While testing on the BDD100K dataset, YOLOv5s and EfficientNet-B7 appear to be the two best models. Finally, this study has proposed Hessian, Laplacian, and Hessian-based Ridge Detection filtering techniques to optimize the performance of those 2 models. The results demonstrate that these filters could increase the mean average precision by up to 11.81%, and reduce detection time by up to 43.98% when applied to YOLOv5s and EfficientNet-B7 models. Overall, all the experiment results are promising and could be extended to other domains for semantic understanding of the environment. Additionally, various filtering algorithms for multiple object detection and classification could be applied to other areas. Different recommendations and future work have been clearly defined in this study.

Software-defined networking was recently introduced and proposed to separate the control from the data plane. This architecture introduces new challenges, particularly with regard to security and safety. To address the safety challenges, it is necessary to set up a multi controller architecture to provide redundancy. In addition, the second controller can have a security benefit because it can be used to validate the decisions taken by the first controller. However, communication between the controllers is necessary in these architectures, which may be exploited by an attacker to spread across the controllers, resulting in a security issue. This study aims to develop a multi controller architecture without communication between controllers. The control is executed by the nominal controller, which performs the data plane computation, whereas the second controller is in charge of verifying the consistency of the controller’s decisions, i.e., the management traffic. We first formulated the activity of the command and then provided conditions to determine a consistent control. These conditions include a time boundary, which corresponds to the tolerance for a delay in the response time of the controller, and structural properties to verify the consistency of the path setup. Moreover, we proposed a detection algorithm that is divided into two parts: first, a learning phase that aims to learn the consistent path set up by the controller, and second, a running phase which aims to verify that the controller sets up paths that are similar to the learned path. This algorithm was evaluated in terms of its reactivity, precision, and recall. To evaluate this, we considered three use cases: a distributed denial of service (DDOS) attack, an attack to send malicious packets on the network, and a failure of the controller.