A task allocation problem for the heterogeneous unmanned aerial vehicle (UAV) swarm in unknown environments is studied in this paper.Considering that the actual mission environment information may be unknown,the UAV s...A task allocation problem for the heterogeneous unmanned aerial vehicle (UAV) swarm in unknown environments is studied in this paper.Considering that the actual mission environment information may be unknown,the UAV swarm needs to detect the environment first and then attack the detected targets.The heterogeneity of UAVs,multiple types of tasks,and the dynamic nature of task environment lead to uneven load and time sequence problems.This paper proposes an improved contract net protocol (CNP) based task allocation scheme,which effectively balances the load of UAVs and improves the task efficiency.Firstly,two types of task models are established,including regional reconnaissance tasks and target attack tasks.Secondly,for regional reconnaissance tasks,an improved CNP algorithm using the uncertain contract is developed.Through uncertain contracts,the area size of the regional reconnaissance task is determined adaptively after this task assignment,which can improve reconnaissance efficiency and resource utilization.Thirdly,for target attack tasks,an improved CNP algorithm using the fuzzy integrated evaluation and the double-layer negotiation is presented to enhance collaborative attack efficiency through adjusting the assignment sequence adaptively and multi-layer allocation.Finally,the effectiveness and advantages of the improved method are verified through comparison simulations.展开更多
As a crucial process in the coordinated strikes of unmanned aerial vehicles(UAVs), weapon-target assignment is vital for optimizing the allocation of available weapons and effectively exploiting the capabilities of UA...As a crucial process in the coordinated strikes of unmanned aerial vehicles(UAVs), weapon-target assignment is vital for optimizing the allocation of available weapons and effectively exploiting the capabilities of UAVs. Existing weapon-target assignment methods primarily focus on macro cluster constraints while neglecting individual strategy updates. This paper proposes a novel weapon-target assignment method for UAVs based on the multi-strategy threshold public goods game(PGG). By analyzing the concept mapping between weapon-target assignment for UAVs and multi-strategy threshold PGG, a weapon-target assignment model for UAVs based on the multi-strategy threshold PGG is established, which is adaptively complemented by the diverse cooperation-defection strategy library and the utility function based on the threshold mechanism. Additionally, a multi-chain Markov is formulated to quantitatively describe the stochastic evolutionary dynamics, whose evolutionary stable distribution is theoretically derived through the development of a strategy update rule based on preference-based aspiration dynamic. Numerical simulation results validate the feasibility and effectiveness of the proposed method, and the impacts of selection intensity, preference degree and threshold on the evolutionary stable distribution are analyzed. Comparative simulations show that the proposed method outperforms GWO, DE, and NSGA-II, achieving 17.18% higher expected utility than NSGA-II and reducing evolutionary stable times by 25% in large-scale scenario.展开更多
This paper presents a path planning approach for rotary unmanned aerial vehicles(R-UAVs)in a known static rough terrain environment.This approach aims to find collision-free and feasible paths with minimum altitude,le...This paper presents a path planning approach for rotary unmanned aerial vehicles(R-UAVs)in a known static rough terrain environment.This approach aims to find collision-free and feasible paths with minimum altitude,length and angle variable rate.First,a three-dimensional(3D)modeling method is proposed to reduce the computation burden of the dynamic models of R-UAVs.Considering the length,height and tuning angle of a path,the path planning of R-UAVs is described as a tri-objective optimization problem.Then,an improved multi-objective particle swarm optimization algorithm is developed.To render the algorithm more effective in dealing with this problem,a vibration function is introduced into the collided solutions to improve the algorithm efficiency.Meanwhile,the selection of the global best position is taken into account by the reference point method.Finally,the experimental environment is built with the help of the Google map and the 3D terrain generator World Machine.Experimental results under two different rough terrains from Guilin and Lanzhou of China demonstrate the capabilities of the proposed algorithm in finding Pareto optimal paths.展开更多
In this paper,based on a bidirectional parallel multi-branch feature pyramid network(BPMFPN),a novel one-stage object detector called BPMFPN Det is proposed for real-time detection of ground multi-scale targets by swa...In this paper,based on a bidirectional parallel multi-branch feature pyramid network(BPMFPN),a novel one-stage object detector called BPMFPN Det is proposed for real-time detection of ground multi-scale targets by swarm unmanned aerial vehicles(UAVs).First,the bidirectional parallel multi-branch convolution modules are used to construct the feature pyramid to enhance the feature expression abilities of different scale feature layers.Next,the feature pyramid is integrated into the single-stage object detection framework to ensure real-time performance.In order to validate the effectiveness of the proposed algorithm,experiments are conducted on four datasets.For the PASCAL VOC dataset,the proposed algorithm achieves the mean average precision(mAP)of 85.4 on the VOC 2007 test set.With regard to the detection in optical remote sensing(DIOR)dataset,the proposed algorithm achieves 73.9 mAP.For vehicle detection in aerial imagery(VEDAI)dataset,the detection accuracy of small land vehicle(slv)targets reaches 97.4 mAP.For unmanned aerial vehicle detection and tracking(UAVDT)dataset,the proposed BPMFPN Det achieves the mAP of 48.75.Compared with the previous state-of-the-art methods,the results obtained by the proposed algorithm are more competitive.The experimental results demonstrate that the proposed algorithm can effectively solve the problem of real-time detection of ground multi-scale targets in aerial images of swarm UAVs.展开更多
在林业管理中,及时发现火灾并识别其规模对于安全防护和治理火灾至关重要。针对现有火灾检测算法存在的精度低、漏检误检和实时性不足等问题,提出一种无人机航拍图像下火灾实时检测算法——MDSYOLOv8。以YOLOv8为基线算法,将骨干网络第...在林业管理中,及时发现火灾并识别其规模对于安全防护和治理火灾至关重要。针对现有火灾检测算法存在的精度低、漏检误检和实时性不足等问题,提出一种无人机航拍图像下火灾实时检测算法——MDSYOLOv8。以YOLOv8为基线算法,将骨干网络第7层卷积模块和颈部网络卷积模块替换成动态蛇形卷积(DSConv),提高算法的特征提取性能,并强化算法对微小特征的学习能力;然后在颈部与检测头之间添加多维协作注意力机制(MCA),加强颈部特征融合,增强算法对小目标的检测能力,并抑制无关背景信息;最后使用SIoU损失函数替换原YOLOv8中的CIoU损失函数,加快算法的收敛速度和回归精度。实验结果表明,MDSYOLOv8在公开数据集KMU上对烟雾目标的检测精度mAP达到95.89%,相较于基线YOLOv8提高了3.33个百分点,具有卓越的检测性能。此外,本研究采集互联网上的无人机航拍火灾图像制作UFF(UAV field fire)数据集,主要对象为火焰和烟雾,包含森林和城市等火灾隐患可能发生场景。在自制数据集UFF上进行深度实验分析,MDSYOLOv8的检测精度达到93.98%,检测速度为54帧/s,并且能同时识别烟雾和火焰两种火灾场景中的主要目标,与主流目标检测方法相比,在检测精度和效率方面均展现出明显优势,更加契合航拍场景下的火灾检测应用。展开更多
针对动态不确定战场环境下多无人机对多区域、多目标的协同察打任务规划过程中存在的信息不确定、任务多约束及航迹强耦合的多目标优化与决策问题,结合Dubins航迹规划算法,提出了一种融合多种改进策略的灰狼优化算法(grey wolf optimiza...针对动态不确定战场环境下多无人机对多区域、多目标的协同察打任务规划过程中存在的信息不确定、任务多约束及航迹强耦合的多目标优化与决策问题,结合Dubins航迹规划算法,提出了一种融合多种改进策略的灰狼优化算法(grey wolf optimization algorithm incorporating multiple improvement strategies,IMISGWO).首先,针对动态环境带来的无人机巡航速度及察打任务消失时间的不确定性,基于可信性理论建立了以最大化任务收益为指标的任务规划数学模型;其次,为实现该问题的快速求解,设计了初始解均匀分布、个体通信机制调整、动态权重更新和跳出局部最优等策略,提升算法解搜索能力;最后,构建了多无人机察打一体典型任务仿真场景,通过数字仿真以及虚实结合半实物仿真试验验证了算法的可行性和有效性.仿真结果表明:算法在求解不确定环境下耦合航迹的多无人机察打一体任务规划问题时,能够生成多机高效的任务执行序列和满足无人机飞行性能约束的飞行轨迹,且能够适用于无人机数量增加导致问题复杂度增加情形下此类问题的求解.展开更多
基金National Natural Science Foundation of China (12202293)Sichuan Science and Technology Program (2023NSFSC0393,2022NSFSC1952)。
文摘A task allocation problem for the heterogeneous unmanned aerial vehicle (UAV) swarm in unknown environments is studied in this paper.Considering that the actual mission environment information may be unknown,the UAV swarm needs to detect the environment first and then attack the detected targets.The heterogeneity of UAVs,multiple types of tasks,and the dynamic nature of task environment lead to uneven load and time sequence problems.This paper proposes an improved contract net protocol (CNP) based task allocation scheme,which effectively balances the load of UAVs and improves the task efficiency.Firstly,two types of task models are established,including regional reconnaissance tasks and target attack tasks.Secondly,for regional reconnaissance tasks,an improved CNP algorithm using the uncertain contract is developed.Through uncertain contracts,the area size of the regional reconnaissance task is determined adaptively after this task assignment,which can improve reconnaissance efficiency and resource utilization.Thirdly,for target attack tasks,an improved CNP algorithm using the fuzzy integrated evaluation and the double-layer negotiation is presented to enhance collaborative attack efficiency through adjusting the assignment sequence adaptively and multi-layer allocation.Finally,the effectiveness and advantages of the improved method are verified through comparison simulations.
基金supported by the National Natural Science Foundation of China (No. 62073267)。
文摘As a crucial process in the coordinated strikes of unmanned aerial vehicles(UAVs), weapon-target assignment is vital for optimizing the allocation of available weapons and effectively exploiting the capabilities of UAVs. Existing weapon-target assignment methods primarily focus on macro cluster constraints while neglecting individual strategy updates. This paper proposes a novel weapon-target assignment method for UAVs based on the multi-strategy threshold public goods game(PGG). By analyzing the concept mapping between weapon-target assignment for UAVs and multi-strategy threshold PGG, a weapon-target assignment model for UAVs based on the multi-strategy threshold PGG is established, which is adaptively complemented by the diverse cooperation-defection strategy library and the utility function based on the threshold mechanism. Additionally, a multi-chain Markov is formulated to quantitatively describe the stochastic evolutionary dynamics, whose evolutionary stable distribution is theoretically derived through the development of a strategy update rule based on preference-based aspiration dynamic. Numerical simulation results validate the feasibility and effectiveness of the proposed method, and the impacts of selection intensity, preference degree and threshold on the evolutionary stable distribution are analyzed. Comparative simulations show that the proposed method outperforms GWO, DE, and NSGA-II, achieving 17.18% higher expected utility than NSGA-II and reducing evolutionary stable times by 25% in large-scale scenario.
基金supported by the National Natural Science Foundation of China(6167321461673217+2 种基金61673219)the Natural Science Foundation of the Jiangsu Higher Education Institutions of China(18KJB120011)the Postgraduate Research and Practice Innovation Program of Jiangsu Province(KYCX19_0299)
文摘This paper presents a path planning approach for rotary unmanned aerial vehicles(R-UAVs)in a known static rough terrain environment.This approach aims to find collision-free and feasible paths with minimum altitude,length and angle variable rate.First,a three-dimensional(3D)modeling method is proposed to reduce the computation burden of the dynamic models of R-UAVs.Considering the length,height and tuning angle of a path,the path planning of R-UAVs is described as a tri-objective optimization problem.Then,an improved multi-objective particle swarm optimization algorithm is developed.To render the algorithm more effective in dealing with this problem,a vibration function is introduced into the collided solutions to improve the algorithm efficiency.Meanwhile,the selection of the global best position is taken into account by the reference point method.Finally,the experimental environment is built with the help of the Google map and the 3D terrain generator World Machine.Experimental results under two different rough terrains from Guilin and Lanzhou of China demonstrate the capabilities of the proposed algorithm in finding Pareto optimal paths.
文摘In this paper,based on a bidirectional parallel multi-branch feature pyramid network(BPMFPN),a novel one-stage object detector called BPMFPN Det is proposed for real-time detection of ground multi-scale targets by swarm unmanned aerial vehicles(UAVs).First,the bidirectional parallel multi-branch convolution modules are used to construct the feature pyramid to enhance the feature expression abilities of different scale feature layers.Next,the feature pyramid is integrated into the single-stage object detection framework to ensure real-time performance.In order to validate the effectiveness of the proposed algorithm,experiments are conducted on four datasets.For the PASCAL VOC dataset,the proposed algorithm achieves the mean average precision(mAP)of 85.4 on the VOC 2007 test set.With regard to the detection in optical remote sensing(DIOR)dataset,the proposed algorithm achieves 73.9 mAP.For vehicle detection in aerial imagery(VEDAI)dataset,the detection accuracy of small land vehicle(slv)targets reaches 97.4 mAP.For unmanned aerial vehicle detection and tracking(UAVDT)dataset,the proposed BPMFPN Det achieves the mAP of 48.75.Compared with the previous state-of-the-art methods,the results obtained by the proposed algorithm are more competitive.The experimental results demonstrate that the proposed algorithm can effectively solve the problem of real-time detection of ground multi-scale targets in aerial images of swarm UAVs.
文摘在林业管理中,及时发现火灾并识别其规模对于安全防护和治理火灾至关重要。针对现有火灾检测算法存在的精度低、漏检误检和实时性不足等问题,提出一种无人机航拍图像下火灾实时检测算法——MDSYOLOv8。以YOLOv8为基线算法,将骨干网络第7层卷积模块和颈部网络卷积模块替换成动态蛇形卷积(DSConv),提高算法的特征提取性能,并强化算法对微小特征的学习能力;然后在颈部与检测头之间添加多维协作注意力机制(MCA),加强颈部特征融合,增强算法对小目标的检测能力,并抑制无关背景信息;最后使用SIoU损失函数替换原YOLOv8中的CIoU损失函数,加快算法的收敛速度和回归精度。实验结果表明,MDSYOLOv8在公开数据集KMU上对烟雾目标的检测精度mAP达到95.89%,相较于基线YOLOv8提高了3.33个百分点,具有卓越的检测性能。此外,本研究采集互联网上的无人机航拍火灾图像制作UFF(UAV field fire)数据集,主要对象为火焰和烟雾,包含森林和城市等火灾隐患可能发生场景。在自制数据集UFF上进行深度实验分析,MDSYOLOv8的检测精度达到93.98%,检测速度为54帧/s,并且能同时识别烟雾和火焰两种火灾场景中的主要目标,与主流目标检测方法相比,在检测精度和效率方面均展现出明显优势,更加契合航拍场景下的火灾检测应用。
