Swarm Of Vehicles Research Articles

Machine teaching in Swarm Metaverse under different levels of autonomy.

Shepherding algorithms enable scalable swarm control via the utilization of one or a few control agents. Despite their demonstrated effectiveness in controlling swarms of point-particle agents, shepherding algorithms have been barely evaluated in controlling realistic swarms of uncrewed vehicles (UxVs). Furthermore, existing shepherding algorithms face significant challenges in dealing with complex environments such as those featuring obstacles. We address these research gaps by studying the use of human demonstrations for teaching herding behaviours to machine learning controllers. In particular, we focus on how the level of autonomy used for collecting human demonstrations affects the effectiveness of the resulting swarm controller performance. Our experimental investigation shows that demonstrations collected under a high level of autonomy result in a significantly higher success rate than those collected under a low level of autonomy. Our findings highlight that providing high-level commands for the human demonstrator is more effective even when the demonstrations is used for training a low-level controller.This article is part of the theme issue 'The road forward with swarm systems'.

A configuration maintenance and cooperative navigation method for UAV swarms under intermittent GNSS conditions

PurposeIn response to the challenges of maintaining the configuration and navigation stability of low-cost unmanned aerial vehicle swarms under intermittent global navigation satellite system (GNSS) signal conditions, this study aims to introduce a fast network topology generation algorithm and a hybrid covariance filter.Design/methodology/approachFirst, using spatially stable structures and the principle of three-sphere intersection, connectivity between nodes is rapidly generated, ultimately forming a network topology with tetrahedrons as the basic unit. This ensures the stability of the configuration. Subsequently, a problem arises from the improper distribution of internal confidence within the system when some nodes are connected to GNSS, whereas others rely solely on ranging. In response, a hybrid covariance method with independent relative and absolute covariance matrices is proposed, which can improve the overall navigation precision of the swarm.FindingsSimulation results show that the approach achieves rapid convergence of relative positioning errors to less than 0.5 m for internode distances over 100 m. When one, two and three anchor nodes are accessed, the positioning accuracy of the proposed method is improved by 31.59 %, 64.53 % and 64.48 %, respectively, compared with the existing methods.Originality/valueThe proposed method can stabilize configurations and improve overall positioning accuracy, providing support for addressing distributed navigation issues in intermittent GNSS signals.

ANALYSIS OF A COMPREHENSIVE APPROACH TO THE FORMATION OF TEST DATASETS FOR TRAINING UAV SWARMS UNDER DYNAMIC CONDITIONS

The article addresses the issue of generating test datasets for the training of swarms of unmanned aerial vehicles (UAVs) under complex and dynamic operational conditions, which are in constant change. The study emphasises the necessity of considering various factors, including the presence of obstacles, terrain features, and challenges associated with the lack of a stable GPS signal. Proper test dataset formation ensures swarm reliability and combat effectiveness by enabling training algorithms to pre-emptively account for diverse scenarios. The analysis of existing methods highlights three main directions. Firstly, clustering techniques (e.g. K-means, DBSCAN) enable the automatic grouping of numerous potential scenarios, identification of typical and rare conditions, and avoidance of data duplication that does not contribute to broader scenario coverage. Secondly, the application of genetic algorithms facilitates the search for globally optimal parameter configurations, taking into account the multidimensional nature of the problem (simultaneous changes in UAV positioning, variability of weather conditions, and various types of obstacles). This approach helps identify critical combinations of factors that are often overlooked by other methods. Thirdly, machine learning methods (including neural networks, support vector machines, and multi-agent reinforcement learning) equip swarms with the ability to adaptively 'learn' from historical data, respond to new types of threats, and predict future developments. The article proposes a comprehensive approach that integrates the advantages of clustering, genetic algorithms, and machine learning. Initially, clustering is employed to structure a broad range of scenarios, categorising them from the simplest to the most complex conditions. At the next stage, genetic algorithms analyse each cluster, identifying key scenario parameters that could reduce swarm performance. Simultaneously, machine learning methods enable the development of adaptive models capable of promptly adjusting their behaviour based on obtained results. This approach ensures a balanced test dataset that encompasses both typical and non-trivial cases, thereby facilitating more flexible and informed configuration of swarm control systems. The practical significance of this approach lies in the substantial enhancement of the combat readiness of UAV swarms. These swarms are able to learn to perform effectively under predictable conditions and to acquire the necessary skills to operate in complex scenarios with limited resources. Future research will focus on improving the process of forming adaptive and test datasets to ensure high combat readiness of UAV swarms. This approach will substantially mitigate risks during combat missions and maximise the potential of swarms in challenging and rapidly changing environments.

An Intent Recognition Method for Aerial Swarm Based on Attention Pooling Mechanism

In the realm of unmanned aerial vehicle (UAV) swarm intent recognition, conventional approaches predominantly focus on the attributes derived from singular targets at discrete instances. This trend leads to a significant limitation: the inability to effectively harness and capture the collective feature information of the entire swarm over temporal sequences. To address this gap, this study introduces a comprehensive end-to-end UAV swarm intent recognition approach. Initially, this method utilizes the distance threat coefficient and angular threat coefficient between UAVs to construct the graphical structural representation of the UAV swarm. Subsequently, an innovative deep learning framework, designated as attention-pool based on graph attention network and long short-term memory, which integrates a graph attention network, a novel graph pooling strategy, and a long short-term memory, is developed. This architecture can process the graphically structured data derived from the swarm modeling and accurately deduce the collective intent. Through experimental validation and analyses against existing methodologies, as well as ablation studies, it is evidenced that the model outperforms state-of-the-art methods in terms of accuracy of intent recognition.

Advanced Path Planning for UAV Swarms in Smart City Disaster Scenarios Using Hybrid Metaheuristic Algorithms

In disaster-stricken areas, rapid restoration of communication infrastructure is critical to ensuring effective emergency response and recovery. Swarm UAVs, operating as mobile aerial base stations (MABS), offer a transformative solution for bridging connectivity gaps in environments where the traditional infrastructure has been compromised. This paper presents a novel hybrid path planning approach combining affinity propagation clustering (APC) with genetic algorithms (GA), aimed at maximizing coverage, and ensuring quality of service (QoS) compliance across diverse environmental conditions. Comprehensive simulations conducted in suburban, urban, dense urban, and high-rise urban environments demonstrated the efficacy of the APC-GA approach. The proposed method achieved up to 100% coverage in suburban settings with only eight unmanned aerial vehicle (UAV) swarms, and maintained superior performance in dense and high-rise urban environments, achieving 97% and 93% coverage, respectively, with 10 UAV swarms. The QoS compliance reached 98%, outperforming benchmarks such as GA (94%), PSO (90%), and ACO (88%). The solution exhibited significant stability, maintaining consistently high performance, highlighting its robustness under dynamic disaster scenarios. Mobility model analysis further underscores the adaptability of the proposed approach. The reference point group mobility (RPGM) model consistently achieved higher coverage rates (95%) than the random waypoint model (RWPM) (90%), thereby demonstrating the importance of group-based mobility patterns in enhancing UAV deployment efficiency. The findings reveal that the APC-GA adaptive clustering and path planning mechanisms effectively navigate propagation challenges, interference, and non-line-of-sight (NLOS) conditions, ensuring reliable connectivity in the most demanding environments. This research establishes the APC-GA hybrid as a scalable and QoS-compliant solution for UAV deployment in disaster response scenarios. By dynamically adapting to environmental complexities and user mobility patterns, it advances state-of-the-art emergency communication systems, offering a robust framework for real-world applications in disaster resilience and recovery.

Gradient-tracking-based distributed Nesterov accelerated algorithms for multiple cluster games over time-varying unbalanced digraphs

This paper studies the design of accelerated distributed algorithms for seeking the Nash Equilibrium (NE) of multiple cluster games over time-varying unbalanced digraphs, where players in the same cluster cooperatively minimize the summation of the local cost function and do not concern the interest of other clusters. In this game, the players have limited access to others’ decisions, while they can communicate with others over the inter-cluster and intra-cluster topologies. Accelerated distributed algorithms are proposed based on the decision estimation, Nesterov acceleration, and pseudo gradient estimation to seek the NE of multiple cluster games. We prove that the proposed algorithm linearly converges to the NE using the multistep contraction and linear systems of inequalities. Moreover, three variants of the proposed algorithm are also given for dealing with cases where only partial communication topologies are time-varying and gossip-type. Lastly, the effectiveness of proposed algorithms and the acceleration effect are verified by solving the intrusion–interception confrontation problem of Unmanned Vehicle (UV) swarms in simulations.

A collaborative surface target detection and localization method for an unmanned surface vehicle swarm

A single unmanned surface vehicle (USV) designed for marine missions suffers from limited payload, low efficiency and weak intelligence, while a swarm of USVs shows significant advantages in mission flexibility, diverse payload and task efficiency. One of the key issues for an USV swarm is how to achieve highly efficient collaborative perception. To address this issue, a method framework of collaborative surface target detection and localization based on multiple sensors for a swarm including 4 USVs is designed. First, perception systems are constructed, a joint calibration method for different sensors is proposed, and a lightweight target detection method improved with attention mechanism and lightweight adaptive spatial feature fusion is designed. Second, a specialized fusion method using sensor principles based on an extended Kalman filter (EKF) is proposed for a single USV to obtain a target state model. Third, the obtained target models from different USVs are registered with fuzzy matching and integrated into the complete model in a geographic coordinate system. The proposed method is applied to the collaborative perception system on our developed 4 USV swarm and verified in real marine environment and simulation. Experimental results show that our proposed method framework significantly improves the accuracy, efficiency, and reliability of the target detection and localization. The proposed LAF-YOLOv8-s reduces the model size by 5.1M, while the mean average precision (mAP) reaches 68.7%, which is significantly superior to other methods. The average collaborative localization error is reduced by 2.9m. The dataset is available at https://github.com/maochenyu1/WSLight.

SYNERGISTIC CONFLICT-FREE UAV SWARM MOVEMENT MODELING USING METAHEURISTIC APPROACHES

The paper considers a new synergistic conflict-free movement model for unmanned aerial vehicle (UAV) swarm on the battlefield, based on the metaheuristic approach of particle swarm optimization (PSO). The paper presents an improved algorithm that ensures effective swarm coordination, traffic safety and efficient communication between UAVs. The model is designed to be used in various combat conditions and demonstrates the ability of the swarm to adapt to dynamic changes on the battlefield and perform tasks with high efficiency. Improvements to the PSO algorithm include the addition of collision avoidance force vectors, which allows each UAV take into account the position of its neighbors and avoid conflict situations. This ensures a more stable and smoother UAV swarm movement, reducing the risk of collisions and increasing the overall effectiveness of combat missions. The model also provides for the ability to adapt to changing conditions on the battlefield, which allows the UAV swarm to respond effectively to new threats and challenges.The simulation results show that the proposed metaheuristic approach based on the improved PSO algorithm is capable of calculating suboptimal trajectories for UAV swarm, minimizing the risk of collisions and improving the overall performance of combat missions. A comparative analysis with the classical PSO algorithm has revealed the advantages of the proposed model in the context of the efficiency of coordination and safety of UAV swarm movement. These results confirm the prospects of using the developed approach to control UAV swarms in combat conditions.The proposed algorithm allows each UAV in the swarm to take into account the other UAV position and speed, which allows maintaining the optimal distance between them, reducing collision probability. This is achieved by introducing an avoidance force vector that is directed away from other UAVs. This approach allows a swarm of UAVs to act as a single organized structure, which significantly increases the efficiency of performing tasks in complex and dynamic combat conditions. In addition, the model takes into account various combat scenarios, including obstacle avoidance, target acquisition, and retreat to a safe distance. This makes the algorithm a versatile tool for managing UAV swarms in real-world combat conditions, where the speed of reaction and accuracy of task execution are crucial.

Distributed Task Allocation for Multiple UAVs Based on Swarm Benefit Optimization

The auction mechanism stands as a pivotal distributed solution approach for addressing the task allocation problem in unmanned aerial vehicle (UAV) swarms, with its rapid solution capability well-suited to meet the real-time requirements of aerial mission planning for UAV swarms. Building upon the auction mechanism, this paper proposes a distributed task allocation method for multi-UAV grounded in swarm benefits optimization. The method introduces individual benefit variation to quantify the effect of a task on the benefit of a single UAV, thereby enabling direct optimization of swarm benefit through these individual benefit variations. Within the formulated individual benefit calculation, both the spatial distance between tasks and UAVs and the initial task value along with its temporal decay are taken into account, ensuring a thorough and accurate assessment. Additionally, the method incorporates real-time updates of individual benefits for each UAV, reflecting the dynamic state of task benefit fluctuations within the swarm. Monte Carlo simulation experiments demonstrate that, for a swarm size of 16 UAVs and 80 tasks, the proposed method achieves an average swarm benefit improvement of approximately 2% and 4% over the Consensus-Based Bundle Algorithm (CBBA) and Performance Impact (PI) methods, respectively, thus validating its effectiveness.

Storage Availability Prediction in Unmanned Aerial Vehicle Swarms Using Agent‐Based Simulation

ABSTRACTIn practice, when an unmanned aerial vehicle (UAV) swarm is not executing a mission, its UAVs will be stored as inventory. To ensure that the UAV swarm can be quickly deployed when needed, it is necessary to assess and predict its storage state. Due to the flexible configuration of UAV swarms and the complex factors that affect them during storage, existing storage state indicators and prediction methods cannot meet the requirements of UAV swarm storage. In order to address these issues, a UAV swarm storage availability prediction method based on agent‐based simulation (ABS) is proposed. Considering the degradation of health status, maintenance, support, and other factors during the storage period of UAVs, a UAV swarm storage state measurement metric that covers the storage cycle is proposed. Based on this metric, a UAV swarm storage availability model is established. Then, considering the dynamic adaptability and internal complex interactions of UAV swarms, the ABS is used to realize the modeling and prediction of UAV swarm storage availability. Finally, a UAV swarm rescue case is used to illustrate its scientific validity and accuracy. Therefore, this study offers a scientific and efficient method for measuring the availability of UAV swarms, providing valuable insights for rapid response and decision‐making during the transition from storage to deployment. It also presents a viable approach for availability modeling and prediction in complex, emergent swarm systems.

A Dynamic Task Allocation Method for Unmanned Aerial Vehicle Swarm Based on Wolf Pack Labor Division Model

A Dynamic Task Allocation Method for Unmanned Aerial Vehicle Swarm Based on Wolf Pack Labor Division Model

Extending Conflict-Based Search for Optimal and Efficient Quadrotor Swarm Motion Planning

Multi-agent pathfinding has been extensively studied by the robotics and artificial intelligence communities. The classical algorithm, conflict-based search (CBS), is widely used in various real-world applications due to its ability to solve large-scale conflict-free paths. However, classical CBS assumes discrete time–space planning and overlooks physical constraints in actual scenarios, making it unsuitable for direct application in unmanned aerial vehicle (UAV) swarm. Inspired by the decentralized planning and centralized conflict resolution ideas of CBS, we propose, for the first time, an optimal and efficient UAV swarm motion planner that integrates state lattice with CBS without any underlying assumption, named SL-CBS. SL-CBS is a two-layer search algorithm: (1) The low-level search utilizes an improved state lattice. We design emergency stop motion primitives to ensure complete UAV dynamics and handle spatio-temporal constraints from high-level conflicts. (2) The high-level algorithm defines comprehensive conflict types and proposes a motion primitive conflict detection method with linear time complexity based on Sturm’s theory. Additionally, our modified independence detection (ID) technique is applied to enable parallel conflict processing. We validate the planning capabilities of SL-CBS in classical scenarios and compare these with the latest state-of-the-art (SOTA) algorithms, showing great improvements in success rate, computation time, and flight time. Finally, we conduct large-scale tests to analyze the performance boundaries of SL-CBS+ID.

Task allocation for UAV swarms under communication attacks: An approach based on game theory and negotiation mechanism

This research delves into the method of task allocation for Unmanned Aerial Vehicle (UAV) swarms utilizing game theory and negotiation mechanisms in the face of communication network disruptions. Initially, a UAV swarm topology network characterized by a scale-free structure is established. Within this network, specific nodes undergo a single, deliberate batch attack. To address this, an edge-compensation strategy is formulated for network restoration, aiming to reinstate network connectivity to an optimal level. Subsequently, an algorithm is developed, integrating a bi-directional selection negotiation mechanism and coalition game theory, to determine the initial coalition for UAV task allocation. Following this, a traversal algorithm is employed to optimize the initial coalition, ultimately yielding the final task allocation coalition result. This methodology proves effective in repairing the topology network of UAV swarms during attack conditions, leading to the derivation of an optimal UAV allocation scheme that enhances the success rate of task allocations.

Adaptive communication power control for enhancing attack resilience in UAV networks

A swarm of Unmanned Aerial Vehicles (UAVs) comprises multiple UAVs that are capable of completing tasks beyond the capabilities of a single UAV. Due to the unique challenges of UAV missions, these vehicles often operate far from base stations, making network connectivity crucial for the successful completion of UAV swarm missions. However, existing methods do not account for strategies to maintain the overall connectivity of the UAV network when nodes are under attack. To address this issue, we propose a method named Adaptive Communication Power Control (ACPC) that dynamically adjusts UAV communication power to mitigate potential connectivity losses caused by node failures or malicious attacks. This adjustment ensures that the network can maintain information exchange among the remaining UAVs even in the event of disruptions. Additionally, we introduce a novel evaluation method to assess the overall connectivity of the network and validate ACPC through simulations of UAV swarm missions where some UAVs experience failures. Using this evaluation method, we measured the network’s state before and after attacks and recovery and calculated the additional energy consumption required by the UAVs. The results indicate that our method can increase the resilience of the UAV network by up to 5.36 times, while only raising the total communication energy consumption to 1.53 times.

Distributed Decision‐Making of General Linear Systems in Multi‐Coalition Games and Its Application to USV Swarm Confrontation

ABSTRACTThis article explores a decision‐making problem with partial information in a multi‐coalition game that involves both cooperation and competition. Considering heterogeneous players with general linear systems, we propose a distributed algorithm that employs average consensus mechanisms to estimate the gradient of the coalition function among players within each coalition and utilizes leader‐following protocols to estimate the actions of all players across multiple coalitions. Furthermore, to reduce communication costs, a dynamic event‐triggered mechanism (ETM) is introduced into the average consensus and leader‐following protocols. The ETM is asynchronous, eliminating the requirement for a global clock and allowing players to transmit information only when the triggering condition is satisfied. By the Lyapunov analysis, all actions asymptotically converge to the Nash equilibrium by the proposed algorithms. Moreover, the unmanned surface vehicle (USV) swarm confrontation is formulated as a coalition game. In the scenario of safeguarding territorial integrity, we devise specific tasks for both the invading and defending USVs, encompassing elements such as formation, intrusion, defense, and so forth. The effectiveness of the proposed algorithms is validated through comprehensive simulations.

End-to-End Latency Optimization for Resilient Distributed Convolutional Neural Network Inference in Resource-Constrained Unmanned Aerial Vehicle Swarms

An unmanned aerial vehicle (UAV) swarm has emerged as a powerful tool for mission execution in a variety of applications supported by deep neural networks (DNNs). In the context of UAV swarms, conventional methods for efficient data processing involve transmitting data to cloud and edge servers. However, these methods often face limitations in adapting to real-time applications due to the low latency of cloud-based approaches and weak mobility of edge-based approaches. In this paper, a new system called deep reinforcement learning-based resilient layer distribution (DRL-RLD) for distributed inference is designed to minimize end-to-end latency in UAV swarm, considering the resource constraints of UAVs. The proposed system dynamically allocates CNN layers based on UAV-to-UAV and UAV-to-ground communication links to minimize end-to-end latency. It can also enhance resilience to maintain mission continuity by reallocating layers when inoperable UAVs occur. The performance of the proposed system was verified through simulations in terms of latency compared to the comparison baselines, and its robustness was demonstrated in the presence of inoperable UAVs.

Comparative Reliability Analysis of Unmanned Aerial Vehicle Swarm Based on Mathematical Models of Binary-State and Multi-State Systems

A key aspect in evaluating the performance of a UAV or its swarm is reliability. The reliability is calculated based on various mathematical models. Traditionally, Binary-State System (BSS) models, which assess two states—operational and faulty—are employed. However, some studies suggest using a Multi-State System (MSS) model, which allows for a detailed analysis by considering multiple states beyond just operational and faulty. Both mathematical models allow for the evaluation of Unmanned Aerial Vehicle (UAV) swarms based on availability, which is considered as a probability of swarm mission implementation. There is one more similar assessment computed based on MSS, which is named the probabilities of the performance level. There are not any recommendations for applications of these mathematical models and assessments for reliability analyses of UAV swarms. This paper introduces a comparative study on the availability of UAV swarms using both BSS and MSS models and the probability of performance levels of UAV swarms. This study provides quantitative and qualitative recommendations to exploit these mathematical models and assessments for UAV swarms according to computational complexity and informativeness. The comparative analysis shows that the evaluation of UAV swarm failure should be based on BSS, and the analysis of operation states should be implemented based on probabilities’ performance levels instead of swarm availability. These results are confirmed by quantitative and statistical examinations of UAV swarms of different types based on both BSS and MSS. The number of UAVs is changed from 2 to 20 in these examinations.

Research on Target Allocation for Hard-Kill Swarm Anti-Unmanned Aerial Vehicle Swarm Systems

In response to the saturated attacks by low, slow, and small UAV swarms, there is currently a lack of effective countermeasures. Counter-UAV swarm technology is an important issue that urgently requires breakthroughs. This paper conducts research on a mid–short-range hard-kill counter-swarm scenario where fewer swarms confront multiple swarms and stronger swarms confront weaker swarms. The requirement is for counter-swarm UAVs to quickly penetrate the swarm at mid–short range and collide with as many incoming UAVs as possible to destroy them. To address the sparse solution space problem, an improved genetic algorithm that integrates multiple strategies is adopted to calculate the spatial density distribution of the incoming swarm. A baseline is identified through gradient descent that maximizes the density integral in a straight-line direction. Based on this baseline, the solution space for single strikes on the swarm is filtered. During the solution process, an elite strategy is introduced to prevent the overall degradation of the population performance. Additionally, the feasibility of the flight trajectory needs to be assessed. A piecewise cubic spline interpolation method is used to optimize the flight trajectory, minimizing the maximum curvature. Ultimately, multiple counter-swarm UAV targets within the swarm and their corresponding trajectories are obtained.

Construction of kill webs with heterogeneous UAV swarms in dynamic contested environments

With the concept of "mosaic warfare," a novel combat style that involves constructing "kill webs" with unmanned aerial vehicle (UAV) swarms has emerged. However, little research has focused on this specific task scenario, particularly concerning the self-organization and adaptive collaboration of heterogeneous combat units in dynamic contested environments. Considering the scales and highly dynamic natures of such swarms, an adaptive communication network mechanism is developed based on the Molloy-Reed criterion. In contrast with common offline/noncombat task scenarios, the self-organization process is refined through agent-based modeling, and a combat effectiveness evaluation is introduced to provide enhanced task execution incentives. The proposed dynamic consensus-based coalition algorithm (DCBCA) addresses UAV intelligence defects such as "confusion," "forgetfulness," and "recklessness" during the dynamic target selection process, enabling effective bottom-up kill webs construction. Extensive simulation results demonstrate that the algorithmic system outlined in this paper can support the efficient and resilient operations of large-scale heterogeneous UAV swarms. The DCBCA outperforms the dynamically improved consensus-based grouping algorithm (CBGA) and the consensus-based timetable algorithm (CBTA) in terms of performance and convergence speed.

A Survey on Security of UAV Swarm Networks: Attacks and Countermeasures

The increasing popularity of Unmanned Aerial Vehicle (UAV) swarms is attributed to their ability to generate substantial returns for various industries at a low cost. Additionally, in the future landscape of wireless networks, UAV swarms can serve as airborne base stations, alleviating the scarcity of communication resources. However, UAV swarm networks are vulnerable to various security threats that attackers can exploit with unpredictable consequences. Against this background, this paper provides a comprehensive review on security of UAV swarm networks. We begin by briefly introducing the dominant UAV swarm technologies, followed by their civilian and military applications. We then present and categorize various potential attacks that UAV swarm networks may encounter, such as denial-of-service attacks, man-in-the-middle attacks and attacks against Machine Learning (ML) models. After that, we introduce security technologies that can be utilized to address these attacks, including cryptography, physical layer security techniques, blockchain, ML, and intrusion detection. Additionally, we investigate and summarize mitigation strategies addressing different security threats in UAV swarm networks. Finally, some research directions and challenges are discussed.