Road infrastructure has become a key enabler for achieving full autonomous driving. However, research on this topic is still awaiting answers, especially at the experimentation side. For this reason, we present AUTODRAITEC1, a novel AI-based system that is deployed on the road infrastructure to control the driving of Connected and Autonomous Vehicles (CAVs). The system deploys a hybrid machine learning approach comprised of a supervised learning classifier to characterize the behaviors of human drivers, with a deep reinforcement learning policy to provide speed recommendations for CAVs. This new architecture aims to enhance the situational awareness for autonomous driving systems, and improve the explainability of AI actions through the understanding of others human-drivers behaviors. Beside simulation evaluation, a Proof of Concept (PoC) of the system is presented. Using a 1:18 scale testbed that faithfully replicates real-world driving scenarios, we demonstrate that AUTODRAITEC consistently succeeds in avoiding accidents, enhancing safety distance and efficiency, while preserving the traffic flow rate. The presented solution is also scalable to different driving use cases.

Neural networks have demonstrated remarkable success in learning and solving complex tasks in a variety of fields including cognitive cities. Nevertheless, the rise of those networks in modern computing has been accompanied by concerns regarding their vulnerability to adversarial attacks. In this work, we propose a novel gradient-free, gray box, incremental attack that targets the training process of neural networks. The proposed attack, which implicitly poisons the intermediate data structures that retain the training instances between training epochs acquires its high-risk property from attacking data structures that are typically unobserved by professionals. Hence, the attack goes unnoticed despite the damage it can cause. Moreover, the attack can be executed without the attackers' knowledge of the neural network structure or training data, making it more dangerous. The proposed attack was tested under a sensitive application of secure cognitive cities, namely, biometric authentication. The conducted experiments showed that the proposed attack is effective and stealthy. Finally, the attack effectiveness property was concluded from the fact that it was able to flip the sign of the loss gradient in the conducted experiments to become positive, which is noisy and unstable training. Moreover, the attack was able to decrease the inference probability in the poisoned networks compared to their unpoisoned counterparts by 15.37%, 14.68%, and 24.88% for the Densenet, VGG, and Xception, respectively. Finally, the attack retained its stealthiness despite its high effectiveness. This was demonstrated by the fact that the attack did not cause a notable increase in the training time, in addition, the Fscore values only dropped by an average of 1.2%, 1.9%, and 1.5% for the poisoned Densenet, VGG, and Xception, respectively.

This paper addresses the problem of estimating the location of Internet of Things (IoT) Long Range Wide Area Networks (LoRaWAN) devices from time of arrival differences measured at gateways. An Extended Kalman Filter (EKF) based approach is considered to aggregate the measurements obtained at different time instants. Particular attention is paid to the processing of outliers. Based on experimental data obtained from field measurements conducted on a real LoRaWAN™ network an insight into the realistic localization accuracy of the considered localization approach is provided.

The combination of multiple-input multiple-output (MIMO) technology and orthogonal frequency division multiplexing (OFDM) has been considered as an attractive solution for the next generation wireless communication networks. In comparison with single-input single-output (SISO) systems, channel estimation in the MIMO scenario becomes more challenging, owing to the increased number of independent transmitter-receiver links to be estimated. In this paper, we introduce an enhanced iterative channel estimation method, which is capable of providing genetically optimized channel information for uplink multi-user MIMO OFDM receivers, approaching the optimal benchmark arrangement utilizing perfect channel knowledge. The effectiveness of the proposed scheme is demonstrated by a range of numerical results under realistic scenarios.

Despite the assistance of well-recognized hot technologies such as Multiple-Input Multiple-Output (MIMO) and Orthogonal Frequency Division Multiplexing (OFDM), cell-edge mobile users still suffer significant performance degradation in uplink, thus restricting the system coverage. In the context of the European B3G/4G WINNER system, we propose a hybrid solution which takes the advantages of both Genetic Algorithm (GA) and turbo processing, resulting in an enhanced MIMO OFDM receiver. Simulation results show that the proposed scheme is capable of achieving a good link performance, whilst improving the robustness of the multi-links and thus the cell coverage.

In this paper we present a class of structured LDPC codes over GF(q) which are suitable for efficient encoding. We derive the conditions under which the Phi matrix defined by Richardson & Urbanke in (2001) is an identity matrix over GF(q). If the parity-check matrix satisfies the derived conditions, the inversion of the Phi matrix can be removed in encoding process so that the computational complexity grows linearly with the code length. Simulation results show that efficiently encodable structured LDPC codes have no performance degradation due to the constraint on their design structure, compared with randomly constructed LDPC codes.

In this paper, we investigate the performance-complexity trade-off for two leading-edge channel coding techniques, namely duo-binary turbo-codes (DBTC) and block LDPC codes (BLDPCC). We assess their respective decoding complexity in terms power consumption and cycle delays, as a function of both the code rate and packet length, to derive a domain of suitability for both techniques. We also compare their capabilities to serve as mother codes for rate-compatible punctured codes (RCPC) and finally analyze their achievable throughput, as a function of the degree of parallelism.

In this paper, we present an analytical approach to study the power estimate of the interference in a MC-CDMA transmission scheme. The receiver is implementing a simple RAKE architecture that is not near-far resistant. The regenerated signal (RS) technique is applied on a symbol basis for power estimation of the jamming signal. An exact equation of this estimate is obtained for L uncorrelated diversity branches, and the impact of diversity order is discussed.

The study presented in this paper underlines the capacity increase resulting from the joint use of multicarrier (MC) adaptive modulation techniques and MC-CDMA receiver in the context of beyond 3G systems. The performance enhancement due to such near optimal rate and power adaptation decreases with diversity order. Moreover, results obtained through two different kinds of wireless channels (ETSI BRAN A and C) highlight convergence phenomena, independent of initial channel diversity order. Finally, complete single user detection (SUD) techniques are compared in tit is context of adaptive modulation.