软件定义网络(software-defined networks,SDN)流量调度提升网络性能和资源利用率、实现节能和负载均衡至关重要.传统的多目标优化算法在高流量和网络动态性增加的情况下显著影响算法的收敛速度,难以满足复杂网络环境的多样化需求.针对...软件定义网络(software-defined networks,SDN)流量调度提升网络性能和资源利用率、实现节能和负载均衡至关重要.传统的多目标优化算法在高流量和网络动态性增加的情况下显著影响算法的收敛速度,难以满足复杂网络环境的多样化需求.针对此问题,提出了一种基于深度强化学习的流量预测在线路由算法——OTPR-DRL:根据流量特征预测关键流和普通流,结合网络状态和流量信息建立线性规划问题获得关键流路由的最优解.为满足普通流不同服务质量(quality of service,QoS)需求,引入通用效用函数实现多目标优化,通过多智能体和优先级经验回放机制为普通流选择路由.实验结果表明,在高流量强度下,OTPR-DRL与现有的算法相比,提高了收敛速度,至少降低了10.26%的网络传输时延,3.09%的丢包率,提高了1.70%的吞吐率.展开更多
A method for moving object recognition and tracking in the intelligent traffic monitoring system is presented. For the shortcomings and deficiencies of the frame-subtraction method, a redundant discrete wavelet transf...A method for moving object recognition and tracking in the intelligent traffic monitoring system is presented. For the shortcomings and deficiencies of the frame-subtraction method, a redundant discrete wavelet transform （RDWT） based moving object recognition algorithm is put forward, which directly detects moving objects in the redundant discrete wavelet transform domain. An improved adaptive mean-shift algorithm is used to track the moving object in the follow up frames. Experimental results show that the algorithm can effectively extract the moving object, even though the object is similar to the background, and the results are better than the traditional frame-subtraction method. The object tracking is accurate without the impact of changes in the size of the object. Therefore the algorithm has a certain practical value and prospect.展开更多
Aiming at solving the problem of missed detection and low accuracy in detecting traffic signs in the wild, an improved method of YOLOv8 is proposed. Firstly, combined with the characteristics of small target objects i...Aiming at solving the problem of missed detection and low accuracy in detecting traffic signs in the wild, an improved method of YOLOv8 is proposed. Firstly, combined with the characteristics of small target objects in the actual scene, this paper further adds blur and noise operation. Then, the asymptotic feature pyramid network (AFPN) is introduced to highlight the influence of key layer features after feature fusion, and simultaneously solve the direct interaction of non-adjacent layers. Experimental results on the TT100K dataset show that compared with the YOLOv8, the detection accuracy and recall are higher. .展开更多
为缓解逆向可变车道交叉口左转流压力导致的通行效率低、运行不稳定及能耗问题,本文考虑不同自动化水平下自动驾驶车辆(Autonomous Vehicle,AV)与网联自动驾驶车辆(Connected and Autonomous Vehicle,CAV)的混行趋势,构建一种可变车道...为缓解逆向可变车道交叉口左转流压力导致的通行效率低、运行不稳定及能耗问题,本文考虑不同自动化水平下自动驾驶车辆(Autonomous Vehicle,AV)与网联自动驾驶车辆(Connected and Autonomous Vehicle,CAV)的混行趋势,构建一种可变车道背景下左转车辆的轨迹控制方法。在控制时域内建立以通行效率最大化与能耗最小化为目标的多目标轨迹优化模型,引入AV与CAV在信息感知和协同能力上的差异,实现动态调控车队加减速和变道行为,并采用分支定界算法求解变道信息与行驶轨迹。进一步设计持续型、集中型与均衡型3类典型交通情境,结合不同CAV渗透率进行数值模拟分析。结果表明:轨迹优化控制模型能够有效改善车队的纵向时空分布,降低停车次数与队列波动;在持续型、集中型与均衡型情境下,分别在CAV渗透率为20%、50%与100%时效果最佳,相较于能耗优先与效率优先方法,持续型场景综合指标分别下降约51.5%与38.9%,集中型分别下降16.2%与40.8%,均衡型情境分别下降10.7%与33.2%;同时,轨迹控制模型在车头时距与可变车道长度变化下表现出较强的鲁棒性。展开更多
针对复杂道路场景下现有算法对道路上交通标志检测精度不高以及漏检误检问题,文中提出了一种改进YOLOv5(You Only Look Once version 5)检测算法。在主干网络使用轻量化网络RepVGG(Re-parameterized Visual Geometry Goup)增强模型特征...针对复杂道路场景下现有算法对道路上交通标志检测精度不高以及漏检误检问题,文中提出了一种改进YOLOv5(You Only Look Once version 5)检测算法。在主干网络使用轻量化网络RepVGG(Re-parameterized Visual Geometry Goup)增强模型特征提取能力,通过增加残差分支结构获得更多特征细节信息。将坐标注意力机制(Coordinate Attention,CA)融入RepVGG模块来提升模型对于小目标的识别与定位能力。在特征融合方面,引入加权双向特征金字塔结构(Bi-directional Feature Pyramid Network,BiFPN)重构颈部网络,充分利用不同尺度特征信息来增强网络特征融合能力。将原CIoU(Complete Intersection over Union)损失函数替换为EIoU(Efficient IoU)损失函数,以此提高回归框的稳定性,加速收敛。实验结果表明,改进算法在CCTSDB(Chinese Traffic Sign Detection Bench mark)数据集的均值平均精度为98.9%。相较于原YOLOv5算法,所提算法的平均精度提升了2.5百分点,召回率提升了4.5百分点,减少了漏检和误检发生概率,同时满足实时检测要求。展开更多
In this paper, a traffic signal control method based on fuzzy logic (FL), fuzzy-neuro (FN) and multi-objective genetic algorithms (MOGA) for an isolated four-approach intersection with through and left-turning movemen...In this paper, a traffic signal control method based on fuzzy logic (FL), fuzzy-neuro (FN) and multi-objective genetic algorithms (MOGA) for an isolated four-approach intersection with through and left-turning movements is presented. This method has an adaptive signal timing ability, and can make adjustments to signal timing in response to observed changes.The 'urgency degree' term, which can describe the different user's demand for green time is used in decision-making by which strategy of signal timing can be determined. Using a fuzzy logic controller, we can determine whether to extend or terminate the current signal phase and select the sequences of phases. In this paper, a method based on fuzzy-neuro can be used to predict traffic parameters used in fuzzy logic controller. The feasibility of using a multi-objective genetic algorithm ( MOGA) to find a group of optimizing sets of parameters for fuzzy logic controller depending on different objects is also demonstrated. Simulation results show that the proposed methed is effecfive to adjust the signal timing in response to changing traffic conditions on a real-time basis, and the controller can produce lower vehicle delays and percentage of stopped vehicles than a traffic-actuated controller.展开更多
文摘软件定义网络(software-defined networks,SDN)流量调度提升网络性能和资源利用率、实现节能和负载均衡至关重要.传统的多目标优化算法在高流量和网络动态性增加的情况下显著影响算法的收敛速度,难以满足复杂网络环境的多样化需求.针对此问题,提出了一种基于深度强化学习的流量预测在线路由算法——OTPR-DRL:根据流量特征预测关键流和普通流,结合网络状态和流量信息建立线性规划问题获得关键流路由的最优解.为满足普通流不同服务质量(quality of service,QoS)需求,引入通用效用函数实现多目标优化,通过多智能体和优先级经验回放机制为普通流选择路由.实验结果表明,在高流量强度下,OTPR-DRL与现有的算法相比,提高了收敛速度,至少降低了10.26%的网络传输时延,3.09%的丢包率,提高了1.70%的吞吐率.
文摘A method for moving object recognition and tracking in the intelligent traffic monitoring system is presented. For the shortcomings and deficiencies of the frame-subtraction method, a redundant discrete wavelet transform （RDWT） based moving object recognition algorithm is put forward, which directly detects moving objects in the redundant discrete wavelet transform domain. An improved adaptive mean-shift algorithm is used to track the moving object in the follow up frames. Experimental results show that the algorithm can effectively extract the moving object, even though the object is similar to the background, and the results are better than the traditional frame-subtraction method. The object tracking is accurate without the impact of changes in the size of the object. Therefore the algorithm has a certain practical value and prospect.
文摘Aiming at solving the problem of missed detection and low accuracy in detecting traffic signs in the wild, an improved method of YOLOv8 is proposed. Firstly, combined with the characteristics of small target objects in the actual scene, this paper further adds blur and noise operation. Then, the asymptotic feature pyramid network (AFPN) is introduced to highlight the influence of key layer features after feature fusion, and simultaneously solve the direct interaction of non-adjacent layers. Experimental results on the TT100K dataset show that compared with the YOLOv8, the detection accuracy and recall are higher. .
文摘为缓解逆向可变车道交叉口左转流压力导致的通行效率低、运行不稳定及能耗问题,本文考虑不同自动化水平下自动驾驶车辆(Autonomous Vehicle,AV)与网联自动驾驶车辆(Connected and Autonomous Vehicle,CAV)的混行趋势,构建一种可变车道背景下左转车辆的轨迹控制方法。在控制时域内建立以通行效率最大化与能耗最小化为目标的多目标轨迹优化模型,引入AV与CAV在信息感知和协同能力上的差异,实现动态调控车队加减速和变道行为,并采用分支定界算法求解变道信息与行驶轨迹。进一步设计持续型、集中型与均衡型3类典型交通情境,结合不同CAV渗透率进行数值模拟分析。结果表明:轨迹优化控制模型能够有效改善车队的纵向时空分布,降低停车次数与队列波动;在持续型、集中型与均衡型情境下,分别在CAV渗透率为20%、50%与100%时效果最佳,相较于能耗优先与效率优先方法,持续型场景综合指标分别下降约51.5%与38.9%,集中型分别下降16.2%与40.8%,均衡型情境分别下降10.7%与33.2%;同时,轨迹控制模型在车头时距与可变车道长度变化下表现出较强的鲁棒性。
基金This project was supported by China Postdoctoral Science Foundation: "Research on Traffic Signal Control Method for Urban Intersection Based on Intelligent Techniques, 2001" .
文摘In this paper, a traffic signal control method based on fuzzy logic (FL), fuzzy-neuro (FN) and multi-objective genetic algorithms (MOGA) for an isolated four-approach intersection with through and left-turning movements is presented. This method has an adaptive signal timing ability, and can make adjustments to signal timing in response to observed changes.The 'urgency degree' term, which can describe the different user's demand for green time is used in decision-making by which strategy of signal timing can be determined. Using a fuzzy logic controller, we can determine whether to extend or terminate the current signal phase and select the sequences of phases. In this paper, a method based on fuzzy-neuro can be used to predict traffic parameters used in fuzzy logic controller. The feasibility of using a multi-objective genetic algorithm ( MOGA) to find a group of optimizing sets of parameters for fuzzy logic controller depending on different objects is also demonstrated. Simulation results show that the proposed methed is effecfive to adjust the signal timing in response to changing traffic conditions on a real-time basis, and the controller can produce lower vehicle delays and percentage of stopped vehicles than a traffic-actuated controller.
