Fog computing network is designed as an extension of the cloud due to the need for a supporting platform capable of ensuring the requirements of Internet of Thing (IoT). The growth of fog based fifth generation mobile communication (5G) system is challenged by the need for data sharing security. In fact, without properly securing access to Fog node resources in IoT network, services providers may not be able to achieve the desired performance. Indeed, fog computing obviously confront numerous security and privacy risks, due to its features, such as huge scale geolocation, heterogeneity and mobility. Thus, we propose a security model that is based on cooperation between IoT and fog. This model integrates an efficient access control process associated with a monitoring scheme to ensure secure cooperation between diverse resources and different operational parts. Indeed, a comprehensive scheduling process and resource allocation mechanism using our security model is proposed to improve the intended performance of the system. In fact, our main contribution is to introduce a distributed access control based on security resource management framework for fog-IoT networks, and proactive security scheme under ultra-trustworthiness and low-latency constraints. After evaluation based on iFogSim, we have proved that our scheme not only provides low latency with high security and privacy, but also reduces the complexity of administration and management of security and resources mechanisms.

Human Activity Recognition (HAR) is one of the critical subjects of research in health and human machine interaction fields in recent years. Algorithms such as Support Vector Machine (SVM), K-Nearest Neighbors (K-NN), Decision Tree (DT) and many other algorithms were previously implemented to serve this common goal but most of the traditional Machine learning proposed solutions were not satisfying in term of accuracy and real time testing process. For that, a human activities analysis and recognition system with an embedded trained ANN model on Raspberry PI for an online testing process is proposed in this work. This paper includes a comparative study between the Artificial Neural Network (ANN) and the Recurrent Neural Network (RNN), using signals produced by the accelerometer and gyroscope, embedded within the BlueNRG-Tile sensor. After evaluate algorithms performance in terms of accuracy and precision which reached an accuracy of 82% for ANN and 99% for RNN, obtained ANN model was implemented in a Raspberry PI for real-time predictions. Results show that the system provides a real-time human activity recognition with an accuracy of 86%.

VANETs (Vehicular ad hoc networks) are an important communication to offer safety information and other infotainment applications to vehicles. To secure these networks, it is important to ensure the authentication and confidentiality of communicationV2V (vehicle to vehicle) and V2I (Vehicles to infrastructure). However, in VANET environments, vehicles are moving with a very high speed on the roads, so a strong security can degrade the performance of the protocol. Among the research that presents some schemes to secure VANET communications, we mention the Dual Authentication and Key Management Techniques that suffers from several lack of security. In this paper, we present an enhancement to this solution to obtain a deliberate compromise between quality of service and security. Security analysis proves that our scheme is secure and the result of performance validation is promising.

Cloud computing is a distributed and dynamic computing model that still has many difficulties. Intrusion prevention based on access control is an essential research challenge to better ensure the security of the cloud network. The flexibility and the scalability of the application of access control are the essential issues in a cloud environment. This paper proposes the use of risk and role based dynamic access control for preventing intrusions targeting the cloud computing, where the main enhancement of the proposed procedure is that access decision is mainly based on the resources available in the cloud services providers. The proposed scheme get the access decision based on a mixture of eXtensible Access Control Markup Language (XACML) decisions, risk analysis, and vulnerabilities assessment. Experimental results based on a comparison with related work are also presented to show the efficiency of the proposed approach.

With the emergence of Internet of Things (IoT) and the growth of cloud computing technology, Electronic Health systems are integrated as a significant and active domain, which enables the development of medical practices with an affordable cost. Thus, the questions about preserving the security and privacy of the sensitive user's information are raised. Hence, in order to convince users to move their sensitive health records to cloud networks, a strong and secure access control schemes should be employed to protect their data. In fact, access control schemes play a key role in the protection of e-Health records, by which several policies will coexist to grant security and privacy in a network. That is why, one of the essential targets in this work is to retain a secure access to health cloud services. Thus, this paper offers a certificate-based access control strategy, where we introduce an access method for e-health cloud that is mainly based on trust assessment. The main features of the proposed scheme are the integration of trust concept with the monitoring process in order to ensure more security in access control. We more explained the methodology of the proposed approach through appropriate evaluation results, which improves system security and performance by minimizing the time spent to have permissions to access services and optimizing the overall system resource utilization.