【标题】Kodiak Robotics CEO on long-haul trucking's autonomous future
【参考中译】科迪亚克机器人公司首席执行官谈长途卡车运输的自主未来
【类型】 期刊
【作者】 Jacob Moreton
【摘要】 A recent fundraising round demonstrates the attraction of autonomous trucking over passenger car applications. Will self-driving cars ever become a reality? Perhaps, but recent months have proven that the path will be anything but smooth. In April 2021, Waymo Chief Executive John Krafcik quit, later followed by its Chief Financial Officer and head of automotive partnerships. US media outlets speculated that executives at the company had grown frustrated with the slow pace of progress. Meanwhile several autonomous developers have faced serious operational controversies. Most recently, Parisian taxi firm G7 removed Tesla Model 3 vehicles from its fleet in December 2021 after one was involved in a fatal collision, although the cause remains unclear.
【参考中译】 最近的一轮融资显示了自动卡车运输相对于乘用车应用的吸引力。自动驾驶汽车会成为现实吗？也许吧，但最近几个月的事实证明，这条道路绝不会一帆风顺。2021年4月，Waymo首席执行官约翰·克拉夫西克辞职，随后该公司首席财务官兼汽车合作伙伴关系负责人辞职。美国媒体猜测，该公司高管对进展缓慢感到失望。与此同时，几家自主开发公司面临着严重的运营争议。最近，巴黎出租车公司G7于2021年12月将一辆特斯拉Model 3汽车从其车队中移除，此前一辆特斯拉Model 3汽车发生了致命碰撞，尽管原因尚不清楚。
【来源】 Automotive World Magazine 2022
【入库时间】 2023/3/29
【标题】Digital twin-driven deep reinforcement learning for adaptive task allocation in robotic construction
【参考中译】数字双驱动深度强化学习在机器人施工中的自适应任务分配
【关键词】 Digital twin; Proximal policy optimization (PPO); Deep reinforcement learning (DRL); Autonomous robot; Adaptive task allocation
【参考中译】 数字双胞胎；近邻策略优化；深度强化学习；自主机器人；自适应任务分配
【作者】 Dongmin Lee; SangHyun Lee; Neda Masoud; M. S. Krishnan; Victor C. Li
【摘要】 In order to accomplish diverse tasks successfully in a dynamic (i.e., changing over time) construction environment, robots should be able to prioritize assigned tasks to optimize their performance in a given state. Recently, a deep reinforcement learning (DRL) approach has shown potential for addressing such adaptive task allocation. It remains unanswered, however, whether or not DRL can address adaptive task allocation problems in dynamic robotic construction environments. In this paper, we developed and tested a digital twin-driven DRL learning method to explore the potential of DRL for adaptive task allocation in robotic construction environments. Specifically, the digital twin synthesizes sensory data from physical assets and is used to simulate a variety of dynamic robotic construction site conditions within which a DRL agent can interact. As a result, the agent can learn an adaptive task allocation strategy that increases project performance. We tested this method with a case project in which a virtual robotic construction project (i.e., interlocking concrete bricks are delivered and assembled by robots) was digitally twinned for DRL training and testing. Results indicated that the DRL model's task allocation approach reduced construction time by 36% in three dynamic testing environments when compared to a rule-based imperative model. The proposed DRL learning method promises to be an effective tool for adaptive task allocation in dynamic robotic construction environments. Such an adaptive task allocation method can help construction robots cope with uncertainties and can ultimately improve construction project performance by efficiently prioritizing assigned tasks.
【参考中译】 为了在动态(即随时间变化)的施工环境中成功完成不同的任务，机器人应该能够对分配的任务进行优先排序，以优化其在给定状态下的性能。最近，深度强化学习(DRL)方法已经显示出解决这种自适应任务分配的潜力。然而，DRL是否能够解决动态机器人施工环境中的自适应任务分配问题仍然没有答案。在本文中，我们开发并测试了一种数字双驱动DRL学习方法，以探索DRL在机器人施工环境中自适应任务分配的潜力。具体地说，数字双胞胎从实物资产合成感官数据，并用于模拟各种动态的机器人建筑工地条件，在这些条件下，DRL代理可以进行交互。因此，代理可以学习提高项目绩效的自适应任务分配策略。我们通过一个案例项目对该方法进行了测试，在该案例项目中，一个虚拟机器人施工项目(即，由机器人运送和组装联锁混凝土砖)被数字化地孪生，用于DRL培训和测试。结果表明，与基于规则的命令式模型相比，DRL模型的任务分配方法在三个动态测试环境中的施工时间减少了36%，为动态机器人施工环境中的自适应任务分配提供了一种有效的工具。这种自适应任务分配方法可以帮助施工机器人应对不确定性，并通过有效地对分配的任务进行优先排序来最终提高施工项目的性能。
【来源】 Advanced engineering informatics 2022, vol.53
【标题】A framework and method for Human-Robot cooperative safe control based on digital twin
【参考中译】基于数字孪生的人-机器人协同安全控制框架与方法
【关键词】 Human-robot collaboration; Digital twin; Safety control; Machine vision; Convolutional neural network
【参考中译】 人机协作；数字孪生；安全控制；机器视觉；卷积神经网络
【作者】 Hao Li; Wenfeng Ma; Haoqi Wang; Gen Liu; Xiaoyu Wen; Yuyan Zhang; Miying Yang; Guofu Luo; Guizhong Xie; Chunya Sun
【摘要】 Human-robot collaboration (HRC) combines the robot's mechanical properties and predictability with human experience, logical thinking, and strain capabilities to alleviate production efficiency. However, ensuring the safety of the HRC process in-real time has become an urgent issue. Digital twin extends functions of virtual models in the design phase of the physical counterpart in the production phase through virtual-real interactive feedback, data fusion analysis, advanced computational features, etc. This paper proposes an HRC safety control framework and corresponding method based on the digital twin. In the design phase, virtual simulation and virtual reality technology are integrated to construct virtual twins of various HRC scenarios for testing and analyzing potential safety hazards. In the production phase, the safety distance between humans and robots of the HRC scene is monitored and calculated by an iterative algorithm according to machine vision and a convolutional neural network. Finally, the virtual twin is driven based on real-scene data, real-time online visual monitoring, and optimization of the HRC's overall process. A case study using ABB-IRB1600 is presented to verify the feasibility of the proposed approach.
【参考中译】 人-机器人协作(HRC)将机器人的机械特性和可预测性与人类的经验、逻辑思维和应变能力相结合，以降低生产效率。然而，确保人权委员会进程的实时安全已成为一个紧迫的问题。数字孪生模型通过虚实交互反馈、数据融合分析、先进的计算特性等扩展了虚拟模型在生产阶段物理模型设计阶段的功能，提出了一种基于数字孪生模型的HRC安全控制框架和方法。在设计阶段，将虚拟仿真和虚拟现实技术相结合，构建了各种HRC场景的虚拟双胞胎，用于测试和分析潜在的安全隐患。在生产阶段，根据机器视觉和卷积神经网络的迭代算法，对HRC场景中人与机器人之间的安全距离进行监测和计算。最后，基于实时场景数据、实时在线视觉监控和人权委员会整体流程的优化来驱动虚拟孪生兄弟。以ABB-IRB1600为例，验证了该方法的可行性。