With increasing congestion in urban airspace, centralized Unmanned Aircraft System Traffic Management (UTM) face scalability challenges and risks of single-point failures, driving the need for robust decentralized solutions to ensure safe autonomous UAV navigation. This paper introduces the Graph-Attentive Deep Collision Avoidance (GADCA) framework, a decentralized Multi-Agent Reinforcement Learning (MARL) approach for collision avoidance in urban airspace. GADCA extends the Multi-Agent Deep Deterministic Policy Gradient (MADDPG) by integrating Graph Attention Networks (GAT) to model dynamic inter-agent relationships, enabling UAVs to navigate effectively using local observations and shared situational awareness. Addressing limitations in prior work, GADCA incorporates 3D dynamics, continuous velocity action spaces, and scalable fleet sizes for robust performance. Unlike centralized UTM methods, GADCA operates in a decentralized manner during execution, where each UAV independently makes collision-avoidance decisions based on local and shared information, without relying on a central controller. It combines decentralized MARL execution with graph-based relational reasoning to enhance collision avoidance. Experimental results show that GADCA outperforms MADDPG across fleets of 5 to 10 UAVs, achieving 45.8% to 67.7% improvements in learning performance, 12.0% to 54.5% enhancements in task completion, up to 23% reduction in conflicts, and 58% to 100% reduction in collision risk, with no physical collisions recorded. The framework’s dynamic adjacency matrix and collective reward mechanism effectively balance navigation efficiency and safety in complex 3D urban environments, using a highly realistic Gauss-Markov mobility model.

International audience

In this study, we propose an advanced Orthogonal Genetic Algorithm (OGA) specifically developed to tackle the Traveling Salesman Problem with Drones (TSP-D), a multifaceted optimization challenge that necessitates precise synchronization between a truck and a drone for effective delivery tasks. The OGA integrates Orthogonal Crossover and Region-Based Mutation strategies, thereby enhancing the algorithm's proficiency in optimizing drone routing in a range of TSP-D scenarios. This novel approach significantly augments the algorithm's adaptability and exploratory capabilities within the intricate search space. Our comprehensive experimental analysis rigorously evaluates the performance of the proposed OGA against established algorithms in a variety of TSP-D instances. The results from these evaluations reveal that our approach substantially surpasses conventional algorithms in terms of both convergence speed and solution quality. This enhanced performance underscores the OGA's efficacy and robustness in optimizing complex paths in TSP-D scenarios.

Despite being a cornerstone of scientific progress, reproducibility remains a key challenge in deep reinforcement learning (DRL). This paper presents a reproducibility study of the RELIANCE method, a DRL-based approach for drone collision avoidance originally proposed by Ouahouah et al. We identify critical flaws in RELIANCE’s exploration strategy and reward function, which impede learning efficiency and task completion. By refining the exploration strategy and introducing a distance-based reward shaping mechanism, we achieve notable improvements in collision avoidance performance, reward accumulation, and success rates. Our modifications yield a 17-fold increase in success rates and a 15.56% reduction in collisions, alongside higher cumulative rewards. These findings highlight the vital role of early exploration and carefully crafted reward functions in DRL applications.

The anticipated proliferation of Unmanned Aerial Vehicles (UAVs) in the airspace in the coming years has raised concerns about how to manage their flights to avoid collisions and crashes at various stages of flight. To this end, many Unmanned Aircraft Traffic Management systems (UTM) have been designed. These systems use various methods for managing UAV conflicts. Several surveys have reviewed conflict resolution methods for UAVs. However, to the best of our knowledge, there is no survey specifically addressing conflict detection and resolution methods in UTM, particularly those using AI-based methods. Therefore, this article serves as a comprehensive survey of all UAVs conflicts detection and resolution methods proposed in the literature and their use in the UTM systems. This survey classifies the methods into two categories: classical (non-learning) methods and learning-based methods. Classical methods typically rely on pre-defined algorithms or rules for UAVs to avoid collisions, whereas Artificial Intelligence-based methods, including Machine Learning (ML) and especially Reinforcement Learning (RL), enable UAVs to adapt to their environment, autonomously resolve conflicts, and exhibit intelligent behavior based on their experiences. It also presents their application in the conflict resolution service for UTMs. Additionally, the challenges and issues associated with each type of methods are discussed. This article can serve as a foundational resource for researchers in guiding their selection of methods for conflict resolution, particularly those relevant to UTM systems.

Safety applications are the cornerstone of the envisioned Vehicle Ad Hoc Network (VANET). Early transmission of alert messages following car accidents can avoid further potential crashes and save lives. However, the limited terrestrial coverage on highways, particularly in rural areas with low traffic density, hinders the deployment of this service. One promising solution involves integrating Unmanned Aerial Vehicles (UAVs), commonly known as drones, into VANETs to serve as flying relays. These UAVs can re-broadcast alert or warning messages between vehicles, bridging communication gaps. In this paper, we propose a unified UAVs-VANET architecture where UAVs relay messages among vehicles on rural highways. To evaluate our approach, we consider a case study involving a road accident on a Tunisian highway (Tunis - Bou Salem 75 km). We use the SUMO simulator in conjunction with NS3. The obtained results showed that 100% of vehicles are alerted within 3.85 seconds. The study also investigates how the number of deployed UAVs impacts the number of alerted vehicles.

The routing challenges in Flying Ad Hoc Networks (FANETs), characterized by high-speed Unmanned Aerial Vehicles (UAVs), limited UAV battery life, intermittent links, network partitioning, and dynamic topologies, have led to the development of specialized routing protocols based on Reinforcement Learning (RL). In this context, the Q-Learning algorithm is the most commonly used RL algorithm. It relies on two primary hyperparameters: the learning rate and discount factor. The protocol's efficiency hinges on the selection of these parameters. To tackle this challenge, numerous adaptive Q-Learning routing protocols introduce novel functions to dynamically adjust the learning parameters. Therefore, this paper delves into an examination of these parameters and introduces a novel taxonomy categorizing them into three distinct classes: linear function-based adjustment, exponential function-based adjustment, and grid search-based adjustment. This paper highlights that the prevailing adjustment function for the learning rate follows a decreasing exponential pattern, while the discount factor adheres to a linear function. This equilibrium facilitates swift adaptation to changes while ensuring a stable transition between short-term and long-term rewards. Such balance is essential for efficient and effective routing in FANETs.

The increasing use of Unmanned Aerial Vehicles (UAVs), especially in the civil sector, has brought to the fore-front the challenge of managing and integrating them within national airspace alongside manned aircraft. Unmanned Traffic Management (UTM) is a specialized system designed to oversee unmanned aircraft operations. A UTM system provides an array of services and can adopt either a centralized or decentralized architecture. The decentralized architecture serves as a possible solution to the single point of failure problem; however, it may introduce security issues in the areas of identification, flight plan management and communication links. These issues invariably erode the performance and efficiency of the UTM. Renowned for its robust security capabilities, blockchain technology presents an effective means to enhance the security framework of a UTM. In this article, we delineate the security challenges that could impact a UTM, and we explore the avenues through which blockchain technology has been leveraged to address these challenges.

The Information-Centric Networking (ICN) paradigm is shaping the foreseen future Internet architecture by focusing on the data itself rather than its hosting location. It is a shift from a host-centric communication model to a content-centric model supporting among others unique and location-independent content names, in-network caching and name-based routing. By leveraging the easy data access, and reducing both the retrieval delay and the load on the data producer, the ICN can be a viable framework to support the Internet of Things (IoT), interconnecting billions of heterogeneous constrained objects. Among several ICN architectures, the Named Data Networking (NDN) is considered as a suitable ICN architecture for IoT systems. However, its default caching approach lacks a data freshness mechanism, while IoT data are transient and frequently updated by the producer which imposes stringent requirements in terms of information freshness. Furthermore, IoT devices are usually resource-constrained with harsh limitations on energy, memory and processing power. We propose in this paper a caching strategy and a novel cache freshness mechanism to monitor the validity of cached contents in an IoT environment while minimizing the caching process cost. We compared our solution to several relevant schemes using the ccnSim simulator. Our solution exhibits the best system performances in terms of hop reduction ratio, server hit reduction ratio and response latency, yet it provides the lowest cache cost and significantly improves the content validity.

This article describes how the named data networking (NDN) has recently received a lot of attention as a potential information-centric networking (ICN) architecture for the future Internet. The NDN paradigm has a great potential to efficiently address and solve the current seminal IP-based IoT architecture issues and requirements. NDN can be used with different sets of caching algorithms and caching replacement policies. The authors investigate the most suitable combination of these two features to be implemented in an IoT environment. For this purpose, the authors first reviewed the current research and development progress in ICN, then they conduct a qualitative comparative study of the relevant ICN proposals and discuss the suitability of the NDN as a promising architecture for IoT. Finally, they evaluate the performance of NDN in an IoT environment with different caching algorithms and replacement policies. The obtained results show that the consumer-cache caching algorithm used with the Random Replacement (RR) policy significantly improve NDN content validity in an IoT environment.

Conflict detection is an important research issue in Unmanned Aerial Vehicles to ensure safety and collision free flights. In this paper, we first propose a comprehensive analytical framework for a three dimensional conflict detection. Then, we propose SLIDE a new Straight LIne conflict DEtection and alerting algorithm for a set of UAVs to safely share a common airspace. SLIDE is fully distributed and requires a limited state information exchange between UAVs. The assumptions of precise state information and packet-loss free communications are relaxed so as to guarantee the applicability and efficiency of the algorithm in real world situations. A thorough discussion is also presented to deal with appropriate tuning of the different parameters of the collision detection framework. Extensive simulations based on OMNeT++ are used to validate SLIDE and evaluate its performance. Simulation results indicate that SLIDE guarantees a reduced number of false and missed alarms even in high density traffic scenarios and communication perturbed environments, yet it leaves adequate time to accomplish the required maneuver actions.

Information-Centric Networking (ICN) is an emerging network paradigm based on name-identified data objects and in-network caching. Therefore, ICN contents are distributed in a scalable and cost-efficient manner. With the rapid growth of IoT traffic, ICN is intended to be a suitable architecture to support IoT networks. In fact, ICN provides unique persistent naming, in-network caching and multicast communications which reduce the data producer load and the response latency. Using ICN in an IoT environment requires a study of caching policies in terms of cache placement strategies and cache replacement policies. To this end, we address, in this paper, caching challenges with the aim to identify which caching policies are suitable for IoT networks. Simulation findings show that the combination of the consumer-cache caching strategy and the RR cache replacement policy is the most convenient in IoT environments in terms of hop reduction ratio, server hit reduction and response latency.

In Vehicular Ad-hoc Networks (VANETs), any vehicle uses a location service to get an up-to-date data of all vehicles' positions in its vicinity. Many location services rely on infrastructures to perform location update and location query tasks. However, the infrastructure penetration rate may differ from one region to another. We propose in this paper a new location service called Vehicle-Aided Location Service in urban environment (VALS) which resolves the problem of the shortage of infrastructure in some city districts. VALS replaces the missing Road-Side Units (RSU) with vehicles offering RSUs functionalities. It is based on a hierarchical and cluster-based design. Simulation results show that VALS achieves a high success ratio (86%), get most vehicles' data (82%) and a good position information level accuracy (less than 20 m) with a low RSU penetration rate (40%).

This paper deals with conflict detection for Unmanned Aerial Vehicles (UAVs). We provide an exhaustive mathematical framework for 3D conflict detection and propose a novel distributed conflict detection algorithm. The suggested algorithm requires minimal communication. Indeed, only the state information (i.e., position and velocity vectors) is periodically exchanged between the UAVs. We carry out simulations, using the OMNET++ simulator, to evaluate its performance. Simulation results indicate that the proposed algorithm performs well in terms of false alarms, missed alarms and scalability to increased number of UAVs.

Abstract Ad hoc wireless communications rely on beaconing to manage and maintain several network operations and to share relevant network parameters among network nodes. Beacon frames are sent at the start of each beacon interval. The frequency of beaconing depends on whether beacon intervals are fixed size or may be adapted and regulated according to the perceived network and workload conditions. On the other hand, current mobile nodes usually embed several heterogeneous wireless interfaces that urge the design of an adaptive beaconing strategy. In this paper, we propose an autonomous and adaptive beaconing strategy for multi‐interfaced mobile wireless nodes that is capable of regulating the beacon interval size dynamically according to the perceived network conditions. The proposed strategy is based on a joint dynamic estimation of both the announcement traffic indication message (ATIM) window and the beyond‐ATIM window sizes for each beacon interval. Extensive simulations were conducted using OMNeT++ to ascertain the improvements achieved by autonomously regulating the entire beacon interval to take into account the network and workload conditions. Obtained results showed that the two proposed approaches improve significantly the efficiency of the network in terms of throughput, end‐to‐end delay, and power consumption. The proposed fixed beacon interval (fixed‐BI) approach stands as an enhanced version of the power‐saving multi‐channel MAC protocol (PSM‐MMAC). The proposed variable beacon interval (variable‐BI) approach, which regulates dynamically both of the ATIM and the beyond‐ATIM windows, outperforms both the fixed‐BI and the PSM‐MMAC. In particular, under nominal traffic loads, the end‐to‐end delay of the variable‐BI is much lower than those provided by the fixed‐BI and PSM‐MMAC. Copyright © 2015 John Wiley & Sons, Ltd.

Information-Centric Networking (ICN) is a new paradigm proposing a shift in the main Internet architecture from a host-centric communication model to a content-centric model. ICN architectures target to meet user demands for accessing the information regardless of its location. A major building block of ICNs concerns caching strategies. Concomitantly, Machine-to-Machine (M2M) technologies are considered the main pattern for the Internet of Things (IoT). Unifying M2M and ICN into a single framework raises the challenge of cache coherence. In this paper, we propose a novel cache coherence mechanism to check the validity of cache contents. We also propose a caching strategy suitable to M2M environment. Extensive experimentations are conducted to evaluate the performance of our proposals. They show that the combination of our two proposed schemes results in a notable improvement in content validity at the expenses of a certain degradation in both server hit and hop reduction ratios.

The number of vehicles on roads keeps increasing continuously, making the management of traffic flow, especially in big cities more and more challenging. One of the key enablers for having smooth traffic flows and better mobility is to rely on real-time traffic monitoring systems. These systems allow road operators to implement intelligent traffic management strategies such as the dynamic adjustment of timing and phasing of traffic lights and the adaptive road congestion charging. Moreover, better informed travelers will plan smartly their journeys and hence potentially contribute in reducing traffic jams. Traditional real-time traffic monitoring usually get real-time data from GPS-equipped fleets and fixed sensors installed in specific locations. In this paper, a new real-time traffic monitoring based on emerging vehicular communication systems is proposed. The system enables traffic monitoring with higher reliability, accuracy, and granularity. The cluster-based V2X traffic data collection mechanism is able to gather more than 99% of the available data and reduce the overhead to one quarter when compared to other approaches.

Sensing coverage in wireless sensor networks is one of the most fundamental issues, which have been extensively addressed in the literature. It is viewed as one of the critical performance measures in large-scale sensor networks. In this context, researchers have designed several coverage protocols. Throughout the variety of research works in this topic, most interests focused purely on the coverage problem under the restrictive assumption that each deployed node is equipped with a GPS receiver that provides a sensor node with its accurate location. However, in some wireless sensor network applications GPS service may be inaccessible, unpractical and very expensive. Faced to this challenge, several GPS-less localization algorithms for wireless sensor networks have been proposed. Such localization algorithms enable sensor nodes to locate themselves with some degree of accuracy. In this paper, we address the issue of maintaining coverage from the perspective of a GPS-less localization. We particularly integrate two well-known coverage and GPS-less localization solutions, namely CCP and AT-Dist. This integration yields key insights for handling coverage and GPS-less localization in a unified framework in contrast to several existing approaches that address the two issues in isolation. To the best of our knowledge, this will be the first work that integrates those two solutions. Extensive simulations show the effectiveness of this integrated framework to provide guaranteed coverage and localization.

The HIP-Based M2M Overlay Network (HBMON) is a virtual, self-organized and secure M2M network built on the top of Internet, composed of scattered mobile devices. A fundamental requirements of this overlay network is to ensure session survivability upon end-host movement. The Host Identity Protocol (HIP) provides a regular mobility support in our M2M Overlay network. However, HIP is not able to handle the simultaneous mobility case, where both communicating end-points simultaneously acquires a new topologically correct IP address. We propose in this paper a novel solution to manage the simultaneous mobility (also known as double jump) of M2M devices within our overlay. For this purpose we enhance the HIP rendez-vous server in order to fully manage the double jump case. We analytically evaluate the signaling cost of our solution. Then, we implement our double jump solution within the OMNeT++ network simulator. Finally, we evaluate the application recovery time of an M2M device experiencing a double jump situation.

Wireless sensor networks (WSN) have received significant attention over the last few years as they afford a growing number of applications in various fields. At the same time, these networks provide numerous challenges due to their constraints, primarily related to energy scarcity. To overcome energy waste caused by collisions and contention based algorithm, the channel assignment mechanisms, like TDMA1, seem to be an effective way for scheduling node transmissions. To solve channel assignment problems, graph coloring theory has been exploited in many research works, primarily in order to assure collision-free communications. In this paper, we present a novel distributed coloring algorithm for WSNs taking into account the constraints of a real WSN environment. Our collision aware coloring algorithm assures a 2 hop nodes coloring, in a deterministic time execution, without requiring a neighborhood discovering phase. Performance evaluation results have shown the effectivness of our algorithm in terms of exchanged control packets per node as well as the chromatic number.

This paper introduces CrossWalk, a novel integrated cross-layer medium access control/routing protocol based on a receiver-oriented contention resolution mechanism for data dissemination in wireless sensor networks. Traditional approaches for data dissemination such as flooding or simple random walks suffer respectively from a large message forwarding overhead and an excessively long cover time. In this scope, CrossWalk, based on biased random walks, aims to achieve a convenient cover time at the cost of a reasonable overhead. In particular, we demonstrate by an analytical study that the proposed biasing strategy based on a decreasing truncated geometric distribution over a fixed contention window favors a data packet to progress-in a unicast fashion-towards nodes in less explored vicinity by making them more likely to win the contention for the medium access. We therefore show by extensive simulations that CrossWalk outperforms common random walks in terms of partial coverage and efficient network resources consumption by making less redundant message transmissions.

Multihoming and mobility in IPv6 are usually considered as two disjoints concepts. This has lead to the development of the two protocols family separately. However, the new IPv6 terminals are most of the time mobile and are equipped with multiple interfaces. Thus, we need to adopt a new vision toward multihoming and mobility and propose new solutions to manage both of them. In this paper, we focus on the SHIM6 protocol - a new multihoming protocol proposed by the IETF - and we investigate to what extent it can achieve mobility.