堆叠覆盖环境下的机械臂避障抓取是一个重要且有挑战性的任务。针对机械臂在堆叠环境下的避障抓取任务,本文提出了一种基于图像编码器和深度强化学习(deep reinforcement learning,DRL)的机械臂避障抓取方法Ec-DSAC(encoder and crop fo...堆叠覆盖环境下的机械臂避障抓取是一个重要且有挑战性的任务。针对机械臂在堆叠环境下的避障抓取任务,本文提出了一种基于图像编码器和深度强化学习(deep reinforcement learning,DRL)的机械臂避障抓取方法Ec-DSAC(encoder and crop for discrete SAC)。首先设计结合YOLO(you only look once)v5和对比学习网络编码的图像编码器,能够编码关键特征和全局特征,实现像素信息至向量信息的降维。其次结合图像编码器和离散软演员-评价家(soft actor-critic,SAC)算法,设计离散动作空间和密集奖励函数约束并引导策略输出的学习方向,同时使用随机图像裁剪增加强化学习的样本效率。最后,提出了一种应用于深度强化学习预训练的二次行为克隆方法,增强了强化学习网络的学习能力并提高了控制策略的成功率。仿真实验中Ec-DSAC的避障抓取成功率稳定高于80.0%,验证其具有比现有方法更好的避障抓取性能。现实实验中避障抓取成功率为73.3%,验证其在现实堆叠覆盖环境下避障抓取的有效性。展开更多
目的颗粒的堆积高度反映了包装袋的填充密度,堆积高度越大,则填充密度越小;反之,填充密度越大。探究不同振动参数(振动时间、振动频率、振动幅度和振动方向)对颗粒堆积高度的影响规律,以提高包装袋的填充密度。方法基于DEM(Discrete Ele...目的颗粒的堆积高度反映了包装袋的填充密度,堆积高度越大,则填充密度越小;反之,填充密度越大。探究不同振动参数(振动时间、振动频率、振动幅度和振动方向)对颗粒堆积高度的影响规律,以提高包装袋的填充密度。方法基于DEM(Discrete Element Method),利用EDEM数值模拟软件建立PET颗粒堆积模型,并通过Matlab图像处理技术与实验相结合验证模型的准确性;在此模型基础上,仿真模拟颗粒在静止和振动状态下的堆积行为。结论竖直方向的振动更能降低颗粒的堆积高度,增大填充密度;随着振动时间的延长,颗粒的堆积高度逐渐降低,随后趋于平稳,最大可降低约17.70%;随着振动频率的增加,颗粒的堆积高度显著降低,最大可降低约16.67%;随着振动幅度的减小,颗粒的堆积高度逐渐降低,最大可降低约18.59%。结果通过改变振动皮带机的振动参数,可以有效提升包装袋的颗粒填充密度。展开更多
Aiming at the problem of small area human occlusion in gait recognition,a method based on generating adversarial image inpainting network was proposed which can generate a context consistent image for gait occlusion a...Aiming at the problem of small area human occlusion in gait recognition,a method based on generating adversarial image inpainting network was proposed which can generate a context consistent image for gait occlusion area.In order to reduce the effect of noise on feature extraction,the stacked automatic encoder with robustness was used.In order to improve the ability of gait classification,the sparse coding was used to express and classify the gait features.Experiments results showed the effectiveness of the proposed method in comparison with other state-of-the-art methods on the public databases CASIA-B and TUM-GAID for gait recognition.展开更多
文摘堆叠覆盖环境下的机械臂避障抓取是一个重要且有挑战性的任务。针对机械臂在堆叠环境下的避障抓取任务,本文提出了一种基于图像编码器和深度强化学习(deep reinforcement learning,DRL)的机械臂避障抓取方法Ec-DSAC(encoder and crop for discrete SAC)。首先设计结合YOLO(you only look once)v5和对比学习网络编码的图像编码器,能够编码关键特征和全局特征,实现像素信息至向量信息的降维。其次结合图像编码器和离散软演员-评价家(soft actor-critic,SAC)算法,设计离散动作空间和密集奖励函数约束并引导策略输出的学习方向,同时使用随机图像裁剪增加强化学习的样本效率。最后,提出了一种应用于深度强化学习预训练的二次行为克隆方法,增强了强化学习网络的学习能力并提高了控制策略的成功率。仿真实验中Ec-DSAC的避障抓取成功率稳定高于80.0%,验证其具有比现有方法更好的避障抓取性能。现实实验中避障抓取成功率为73.3%,验证其在现实堆叠覆盖环境下避障抓取的有效性。
文摘目的颗粒的堆积高度反映了包装袋的填充密度,堆积高度越大,则填充密度越小;反之,填充密度越大。探究不同振动参数(振动时间、振动频率、振动幅度和振动方向)对颗粒堆积高度的影响规律,以提高包装袋的填充密度。方法基于DEM(Discrete Element Method),利用EDEM数值模拟软件建立PET颗粒堆积模型,并通过Matlab图像处理技术与实验相结合验证模型的准确性;在此模型基础上,仿真模拟颗粒在静止和振动状态下的堆积行为。结论竖直方向的振动更能降低颗粒的堆积高度,增大填充密度;随着振动时间的延长,颗粒的堆积高度逐渐降低,随后趋于平稳,最大可降低约17.70%;随着振动频率的增加,颗粒的堆积高度显著降低,最大可降低约16.67%;随着振动幅度的减小,颗粒的堆积高度逐渐降低,最大可降低约18.59%。结果通过改变振动皮带机的振动参数,可以有效提升包装袋的颗粒填充密度。
基金Project(51678075) supported by the National Natural Science Foundation of ChinaProject(2017GK2271) supported by Hunan Provincial Science and Technology Department,China
文摘Aiming at the problem of small area human occlusion in gait recognition,a method based on generating adversarial image inpainting network was proposed which can generate a context consistent image for gait occlusion area.In order to reduce the effect of noise on feature extraction,the stacked automatic encoder with robustness was used.In order to improve the ability of gait classification,the sparse coding was used to express and classify the gait features.Experiments results showed the effectiveness of the proposed method in comparison with other state-of-the-art methods on the public databases CASIA-B and TUM-GAID for gait recognition.
