Autonomous Vehicles (AVs) rely extensively on GPS signals for navigation, exposing them to a wide range of GPS spoofing attacks, from simplistic signal manipulation to sophisticated, coordinated falsification. Existing detection and mitigation solutions, both conventional and AI-based, face several critical limitations: they struggle to adapt to novel or evolving GPS spoofing strategies, rely on shallow or handcrafted features that fail to capture the semantic complexity of signal distortions, and often lack scalability, as they are typically designed for isolated scenarios and cannot be readily extended to heterogeneous AV fleets. In this study, we introduce a novel framework that integrates multiple Lightweight Language Models (LightLMs), including BERT, RoBERTa, DistilBERT, and TinyBERT, with Reinforcement Learning (RL) algorithms such as Q-Learning, Deep Q-Network, Advantage Actor-Critic, and Proximal Policy Optimization, to improve detection and mitigation of GPS spoofing. The LightLMs are used to convert structured GPS-related features into enriched state embeddings, which serve as input to the RL agent. These embeddings provide semantically meaningful representations that help the agent recognize complex spoofing behaviors and apply mitigation strategies. To train and evaluate the proposed models, we build a Python-based simulation environment that emulates multiple spoofing scenarios and integrates LightLM-driven state inputs. Experimental results across the datasets used show that RL models enhanced with LightLM-generated state representations significantly outperform their baseline counterparts in detection accuracy, mitigation efficiency, and response time. The results also demonstrate the proposed approach’s scalability, generalizability, and operational reliability for secure AV navigation.

Recently, Unmanned Aerial Systems (UAS) have become heavily reliant on communication, navigation, and other critical components such as sensors and actuators, which are essential for operations in both civilian and defense applications. However, this increasing reliance makes UAS more vulnerable to attacks and faults, posing rising threats. While there have been many advances in UAS security, a significant number of studies have proposed artificial intelligence (AI)-enhanced solutions to address these challenges. Yet, no research has explored the potential of generative AI (GenAI) in this domain. GenAI stands out due to its ability to detect and prevent cyberattacks by continuously learning and adapting to new threats and vulnerabilities. In this paper, we propose DroneDefGANt, a GenAI-based approach that combines the capabilities of generative adversarial networks (GAN) and transformer models. The DroneDefGANt is designed to detect both external UAS attacks like GPS spoofing and jamming, and internal attacks such as actuator faults. Through evaluations using synthetic datasets, DroneDefGANt surpassed various conventional AI models, demonstrating superior accuracy and robustness, particularly in the presence of Gaussian noise.

Networked Autonomous Systems (AS) represent a networking paradigm where nodes autonomously make decisions, operating independently or collaboratively to achieve network goals. Unlike traditional networks, networked AS emphasizes decentralized control, enabling nodes to adapt dynamically without relying heavily on centralized controllers. This autonomy promotes agility, scalability, and resilience, making networked AS applicable to various domains like IoT and wireless communications. However, this architecture is vulnerable to various threats, including Advanced Persistent Threat (APT), which exploits vulnerabilities leading to resource depletion and hindering legitimate network operations. APTs, particularly prevalent in IoT-enabled networked AS, pose a significant risk by overwhelming network resources and disrupting services. To address these challenges, this paper introduces a new neural network-based framework tailored to detecting APT in networked AS environments. The proposed framework leverages generative adversarial networks and Deep Learning (DL) to accurately identify malicious activities in networked AS. We evaluated our framework using the newly released dataset TON_IoT, which contains a comprehensive variety of APT and addresses gaps in existing datasets in the context of AS. We obtained a significant improvement in APT detection by comparing DL techniques such as Conventional Neural Networks, Gated Recurrent Units, Long Short-Term Memory, Multilayer Perceptron (MLP), and combining MLP with Auto-Encoder (MLP+AE). The proposed framework demonstrates important accuracy and robustness by analyzing the selected DL techniques’ performance as influenced by variations in the training set size.

Unmanned aerial systems (UAS) and unmanned aerial vehicles (UAVs) have been utilized in numerous civilian and military applications. However, with the increasing deployment of UAS, several cybersecurity challenges have emerged, particularly vulnerabilities in GPS transmissions. GPS jamming attacks pose a significant threat to UAS, resulting in performance degradation and capability loss. To address this challenge, we propose a novel approach for detecting jamming attacks on GPS transmissions in UAS. The proposed approach combines Naive Bayes and Multi-Layer Perceptron algorithms to offer efficient and precise detection of GPS jamming attacks. We evaluated the proposed approach’s performance against various machine learning (ML) models using a synthetic dataset. The experimental results demonstrate the effectiveness of the proposed approach in detecting jamming attacks in GPS transmissions, achieving better accuracy than other ML models under several jammers. Index Terms---AI, ML, UAS/UAV, GPS Jamming Attacks.

Implantable medical devices (IMDs) necessitate secure and energy-efficient communication systems to safeguard patient data and maintain device functionality. In this paper, our primary focus is on examining the impact of varying channel models on the success of Physical Layer Key Generation (PLKG) for Implantable Medical Devices (IMDs), whilst also introducing an energy-aware approach. We utilize the unique properties of double$\kappa-\mu$Shadowed fading channels, a generalized model allowing us to cover a range of special cases, including Nakagami-m, Rician, and Rayleigh fading. We apply a moving average filter to improve the stability of received signal strength (RSS) measurements and employ an adaptive threshold for quantization. To reconcile the generated keys, we adapt the Cascade protocol, ensuring its lightweight operations. Energy awareness is a paramount feature in our PLKG system, achieved by limiting reconciliation attempts to prevent rapid battery depletion. Through rigorous simulations, we present a comprehensive analysis of the performance of our energy-aware PLKG system under various fading conditions

Prediction of power consumption in smart grid and microgrid systems has become a major issue, it represents one of the most important factors in energy management systems (EMS). Recently, several models based on artificial intelligence techniques have been proposed to predict electricity consumption and production, mainly for household energy efficiency. In this paper, we evaluate different algorithms to predict the daily power consumption in a university campus microgrid context. We investigate the implementation of different prediction models in three different real datasets, considering four performance indicators to analyze their accuracy, such as Mean Squared Error, Mean Absolute Error, Mean Absolute Percentage Error, and R-square. Different approaches using time series: ARIMA, SARIMA, machine learning: SVM, XGBOOST, and deep learning: RNN, LSTM, and LSTM-RNN hybrid model were evaluated. Results prove that deep learning approaches achieve better results than time series and machine learning forecasting models. In this work, we prove that the RNN-LSTM hybrid model is the most appropriate model for university campus microgrid case with an accuracy between 83% and 93%.

In order to conserve energy in wireless sensor networks (WSNs), sensor nodes are partitioned into clusters. Clustering algorithm provides an effective way to extend the network lifetime of WSNs. The operation of clustering algorithm is divided into cluster heads (CHs) selection phase and cluster formation phase. However, most of the previous researches have focused on CHs selection, and have not considered the cluster formation phase, which is important problem in WSNs and can drastically affect the network lifetime in WSNs. In this paper, cluster formation using fuzzy logic (CFFL) approach has been proposed to prolong network lifetime and reduce energy consumption in WSNs. this approach uses fuzzy logic in the formation cluster phase, two fuzzy parameters are used. These parameters are residual energy which is energy level of each CH and closeness to base station (BS) which is the distance between the CH and the BS. Simulation results show that the proposed approach consumes less energy and prolongs the network lifetime compared with Low Energy Adaptive Clustering Hierarchy (LEACH) protocol.