Search for: All records

Self-driving vehicles are very susceptible to cyber attacks. This paper aims to utilize a machine learning approach in combating cyber attacks on self-driving vehicles. We focus on detecting incorrect data that are injected into the data bus of vehicles. We will utilize the extreme gradient boosting approach, as a promising example of machine learning, to classify such incorrect information. We will discuss in details the research methodology, which includes acquiring the driving data, preprocessing it, artificially inserting incorrect information, and finally classifying it. Our results show that the considered algorithm achieve accuracy of up to 92% in detecting the abnormal behavior on the car data bus. 

As 5G systems are starting to be deployed and becoming part of many daily life applications, there is an increasing interest on the security of the overall system as 5G network architecture is significantly different than LTE systems. For instance, through application specific virtual network slices, one can trigger additional security measures depending on the sensitivity of the running application. Drones utilizing 5G could be a perfect example as they pose several safety threats if they are compromised. To this end, we propose a stronger authentication mechanism inspired from the idea of second-factor authentication in IT systems. Specifically, once the primary 5G authentication is executed, a specific slice can be tasked to trigger a second-factor authentication utilizing different factors from the primary one. This trigger mechanism utilizes the re-authentication procedure as specified in the 3GPP 5G standards for easy integration. Our second-factor authentication uses a special challenge-response protocol, which relies on unique drone digital ID as well as a seed and nonce generated from the slice to enable freshness. We implemented the proposed protocol in ns-3 that supports mmWave-based communication in 5G. We demonstrate that the proposed protocol is lightweight and can scale while enabling stronger security for the drones. 

Unmanned Aerial Vehicles (UAVs), or drones, are increasingly being utilized for public safety circumstances including post-disaster recovery of destroyed communication infrastructure. For instance, drones are temporarily positioned within an affected area to create a wireless mesh network among public safety personnel. To serve the need for high-rate video-based damage assessment, drone-assisted communication can utilize high- bandwidth millimeter wave (mmWave) technologies such as IEEE 802.11ad. However, short-range mmWave communication makes it hard for optimally- positioned drones to be authenticated with a centralized network control center. Therefore and assuming that there are potential imposters, we propose two lightweight and fast authentication mechanisms that take into account the physical limitations of mmWave communication. First, we propose a drone-to-drone authentication mechanism, which is based on proxy signatures from a control center. Accordingly, any newly joining drone can authenticate itself to an exist one rather than attempting to authenticate to the outof-reach control center. Second, we propose a drone-to- ground authentication mechanism, to enable each drone to authenticate itself to its associated ground users. Such authentication approach is based on challenge-response broadcast type, and it is still utilizing fast proxy signature approach. The evaluation of the proposed authentication mechanisms, conducted using NS-3 implementation of IEEE 802.11ad protocol, show their efficiency and practicality. 

Unmanned aerial vehicle (UAV) plays prominent role in enhancing backhaul connectivity and providing extended coverage areas due to its mobility and flexible deployment. To realize these objectives simultaneously, we present a new framework for positioning the UAV to maximize the small-cells backhaul network connectivity characterized by its Fiedler value, the second smallest eigenvalue of the Laplacian matrix representing the network graph, while maintaining particular signal-to-noise ratio constraint for each individual user equipment. Moreover, we show that the localization problem can be approximated by a low complexity convex semi-definite programming optimization problem. Finally, our extensive simulations verify the approximation validity and demonstrate the potential gain of UAV deployment. 

In this work, we utilizie unmanned aerial vehicles (UAVs) to enhance the connectivity of the backhaul of small-cells (SCs) network and achieve better coverage for those user equipment (UE) in deep fade and not attached to any SCs. Deploying UAVs as a backbone for communication systems has gained an increasing prominence, especially in emergency network breakdown as an efficient and reliable alternative to restore the network connectivity. Moreover, to meet the ever increasing demand on high date rate applications, larger bandwidth is needed which is realizable in mm-wave frequencies range from 20 to 60 GHz. In mm-wave frequencies range, transmitted signals suffer from deleterious path loss mainly due to direct path blockages and significant penetrations losses. We exploit the mobility of UAVs and optimize its position to cope with such impairments. To the best of our knowledge, this work is the first to optimize the UAVs’ locations to jointly increase the connectivity between SCs and improve coverage by overcoming mm-wave harmful pathloss.