驾驶意图识别能有效提高自车对其他交通参与者的轨迹预测能力,是实现智能车自主决策和规划的基础。然而动态复杂交通环境下周围车辆的交互是实现准确可靠驾驶意图识别亟待解决的挑战之一。为提高在动态复杂交通场景下驾驶意图识别的准确...驾驶意图识别能有效提高自车对其他交通参与者的轨迹预测能力,是实现智能车自主决策和规划的基础。然而动态复杂交通环境下周围车辆的交互是实现准确可靠驾驶意图识别亟待解决的挑战之一。为提高在动态复杂交通场景下驾驶意图识别的准确率,提出基于双向图长短时记忆网络(Bidirectional graph long short term memory,Bi-GLSTM)网络的驾驶意图时序识别模型。首先基于局部加权回归散点平滑法对原始数据集中的位置、速度和加速度进行平滑处理,并联合纵横向运动参数为数据标注驾驶意图;然后建立图注意力神经网络,分析和提取周围车辆与目标车辆之间的交互特征,嵌入注意力机制,分析周围车辆对目标驾驶意图的重要性,增强模型对相关性较大的车辆运动状态关注程度;融合周围车辆交互特征和目标车辆历史运动特征,为提高模型在动态复杂交通环境下的鲁棒性和可靠性,基于双向长短时记忆网络提取特征之间的时序特征;最后在公开数据集HighD上训练并验证模型的有效性,结果表明相比于图神经网络、循环神经网络等模型,识别准确率分别提高了11.33%、55.31%;通过可视化注意力权重,说明所提出的模型也一定程度上解决了可解释性问题。展开更多
The uncertainty influences may result in performance deterioration and instability to the steer by wire(SBW) system. Thus, it must make the control system keep robust stability from uncertainty, and have good robustne...The uncertainty influences may result in performance deterioration and instability to the steer by wire(SBW) system. Thus, it must make the control system keep robust stability from uncertainty, and have good robustness. In order to effectively restrain the interference and improve steering stability, this paper presents a μ synthesis robust controller based on SBW system, which considers the effect of model uncertainty and external disturbance on the system dynamics. Taking the ideal yaw rate tracking, interference suppression and excellent robustness as the control objectives, the μ synthesis robust controller is designed using linear fractional transformation theory to deal with the uncertainty. Then, it is testified through time domain and robustness simulation analysis. Simulation results show that the proposed controller can not only ensure robustness and robust stability of the system quite well, but improve handling stability of the vehicle effectively. The results of this study provide certain theoretical basis for the research and application of SBW system.展开更多
文摘驾驶意图识别能有效提高自车对其他交通参与者的轨迹预测能力,是实现智能车自主决策和规划的基础。然而动态复杂交通环境下周围车辆的交互是实现准确可靠驾驶意图识别亟待解决的挑战之一。为提高在动态复杂交通场景下驾驶意图识别的准确率,提出基于双向图长短时记忆网络(Bidirectional graph long short term memory,Bi-GLSTM)网络的驾驶意图时序识别模型。首先基于局部加权回归散点平滑法对原始数据集中的位置、速度和加速度进行平滑处理,并联合纵横向运动参数为数据标注驾驶意图;然后建立图注意力神经网络,分析和提取周围车辆与目标车辆之间的交互特征,嵌入注意力机制,分析周围车辆对目标驾驶意图的重要性,增强模型对相关性较大的车辆运动状态关注程度;融合周围车辆交互特征和目标车辆历史运动特征,为提高模型在动态复杂交通环境下的鲁棒性和可靠性,基于双向长短时记忆网络提取特征之间的时序特征;最后在公开数据集HighD上训练并验证模型的有效性,结果表明相比于图神经网络、循环神经网络等模型,识别准确率分别提高了11.33%、55.31%;通过可视化注意力权重,说明所提出的模型也一定程度上解决了可解释性问题。
基金Projects supported by Zhejiang Provincial Natural Science Foundation of China(No.LR16H300001)National Natural Science Foundation of China(No.31670008)
文摘棘白菌素类抗真菌药物米卡芬净和阿尼芬净的合成工艺中包括一个关键步骤:水解除去FR901379分子和棘白菌素B(echinocandin B,ECB)分子的脂肪酸侧链,形成环状的6元多肽核心。FR901379和ECB的水解可以分别被FR901379酰基酶和阿库来菌素A酰基酶(aculeacin A acylase,AAC)催化,因此,FR901379酰基酶和AAC的发掘、表征和生产在米卡芬净和阿尼芬净的工业生产上具有重要的应用价值。本研究首先筛选到了犹他游动放线菌SW1311,发现该菌株发酵液具有酰基酶活性,并摸索了不同发酵条件对酰基酶活性的影响。然后将犹他游动放线菌SW1311中的AAC基因克隆到改造过的质粒载体p IJ8660中,并将该质粒转化到天蓝色链霉菌(Streptomyces coelicolor)A3(2)中对AAC基因进行高表达,得到重组菌株sSCO-AAC。最后将sSCO-AAC生产的AAC活性和所需培养时间与犹他游动放线菌SW1311进行比较,表征了sSCO-AAC的发酵液水解FR901379的反应。结果表明,犹他游动放线菌SW1311中的酰基酶具有水解FR901379和青霉素V的酰基活性。用重组菌株sSCO-AAC生产的AAC活性比犹他游动放线菌SW1311的高4.6倍,且该重组菌株所需的培养时间比犹他游动放线菌SW1311缩短了30%。该结果不仅将ACC的应用范围从阿尼芬净合成拓宽到了米卡芬净合成,而且还揭示出天蓝色链霉菌A3(2)可以作为一个良好的AAC表达菌株。本研究对阿尼芬净和米卡芬净的工业化生产具有潜在的应用价值。
基金supported by the Visiting Scholar Foundation of the State Key Lab of Mechanical Transmission in Chongqing University(Grant Nos.SKLMT-KFKT-2014010&SKLMT-KFKT-201507)the National Natural Science Foundation of China(Grant Nos.51375007&51605219)+1 种基金the Fundamental Research Funds for the Central Universities(Grant No.NE2016002)the Natural Science Foundation of Jiangsu Province(Grant No.SBK2015022352)
文摘The uncertainty influences may result in performance deterioration and instability to the steer by wire(SBW) system. Thus, it must make the control system keep robust stability from uncertainty, and have good robustness. In order to effectively restrain the interference and improve steering stability, this paper presents a μ synthesis robust controller based on SBW system, which considers the effect of model uncertainty and external disturbance on the system dynamics. Taking the ideal yaw rate tracking, interference suppression and excellent robustness as the control objectives, the μ synthesis robust controller is designed using linear fractional transformation theory to deal with the uncertainty. Then, it is testified through time domain and robustness simulation analysis. Simulation results show that the proposed controller can not only ensure robustness and robust stability of the system quite well, but improve handling stability of the vehicle effectively. The results of this study provide certain theoretical basis for the research and application of SBW system.
